Capillary gas chromatographic method for determining the C3-C12

with the largest retention timein each group is taken as the standard, the relative retention times, within each group at a constant temperature are a...
2 downloads 0 Views 977KB Size
However, if each amine is classified into a group depending on the number of carbon atoms in the molecule, and the amine with the largest retention time in each group is taken as the standard, the relative retention times, within each group at a constant temperature are almost independent of the loading of PEI (Table 111), showing that Porapak Q is playing the major role. The importance of the polymer in the separation of the amines is further emphasized by comparing the results from the two polymers, Porapak Q and PAR-1 (Table IV) which show that the retention times, within each group of amines containing the same number of carbon atoms, remain independent not only of the loading of base, but also of the nature of the polystyrene. This suggests that the chromatographic separation depends o n the solubility of each amine in the polymer, which depends in turn on the proportion of carbon to nitrogen in the molecule. It is possible that any surface modification of the polymer with a polar material, such as PEI, will affect the solubility of amines in the polymer, so that the solubility, and hence retention times, of amines with the same carbon to nitrogen ratio will be changed to the same extent. Amines with a different carbon to nitrogen ratio will have their retention times altered by different amounts. Although such a scheme would explain the observed results, further work is required to produce confirmatory evidence.

The retention times of amines on PAR-1 are shorter than those using Porapak Q under the same conditions (Table 11). Moreover, the isodromos point is at a lower temperature on the boiling point axis for PAR-1 (calculated from data in Table II), and this results in the retention times of tertiary amines being shortened relative to primary and secondary. Thus, there appear to be several mixtures which can be separated by one polymer but not by the other. In general, however, by varying the loading of PEI, TEPA, or potassium hydroxide, it is possible to resolve the same mixture on both polymers (Table 11). This difference in behavior of the two polymers presumably arises from their different structural properties (for example, cross-linking, surface area, and permeability). ACKNOWLEDGMENT

The authors are grateful to Mrs. M. V. Walker and Miss M. A. Warriss for much helpful technical assistance and to Dr. E. McIntyre, I.C.I. Fibres Limited, for the chemical analysis on the polymer. RECEIVED for review August 14, 1967. Accepted November 14,1967.

Capillary Gas Chromatographic Method for Determining the C,-C Hydrocarbons in Full-Range Motor Gasolines W. N. Sanders and J. B. Maynard Shell Oil Co., Research Laboratory, Wood Riuer, Ill. 62095 A capillary gas-liquid chromatographic method has been developed to determine the individual C3-C12 hydrocarbons in full-range motor gasolines. The analyses are conducted on a 200-foot squalane capillary column in less than 2 hours. Approximately 240 chromatographic peaks are observed in the analysis of an average gasoline; 180 of them (amounting to 9699% wt of the sample) have been specifically identified. The column temperature and column inlet pressure are both programmed to obtain resolution of close-boiling hydrocarbons. Standard deviations ahd the 95% confidence limits are given for the quantitative repeatability and reproducibility of the method. Chromatographic peak identifications and the detailed quantitative composition of typical premium and regular-grade gasolines are presented, as well as the detailed composition of the American Petroleum Institute Prototype Fuel No. 1-Premium Leaded Reference. Changes that can be made in the temperature and pressure programs to obtain resolution of specific groups of hydrocarbons are discussed.

THENEED for a method to determine the detailed hydrocarbon composition of full-range motor gasolines has been evident for some time. Previously, such information was either unobtainable or required an exorbitant amount of time and effort. Generally, the data were compiled from a series of analyses such as fluorescent indicator absorption (FIA) analyses, ASTM distillations, packed column gas-liquid chromatography, and mass spectrometry. However, with

the advent of capillary columns and ultrasensitive detectors (particularly the hydrogen flame ionization and electron capture detectors), it is now possible to obtain, by gas chromatography, detailed compositional data on complex gasoline range mixtures. Schwartz, Mathews, and Brasseaux have described capillary gas-liquid chromatographic (GLC) methods for the determination of saturated hydrocarbons in the 28" to 114" C portion of petroleum, as well as the individual components in the 80" to 180" C aromatic portion of petroleum (I, 2). Martin and Winters used 500- to 800-foot capillary columns to determine hydrocarbons through Cl0 in crude oil fractions; however, long analysis times of 4 hours were reported (3). Durrett et al. employed capillary GLC for the analysis of individual refinery streams including light reformer feed, reformate, and alkylate (4, 5). McEwen (6) has employed his exhaust gas analysis (1) R. D. Schwartz and D. J. Brasseaux, ANAL.CHEM., 35, 137482 (1963). (2) R. D. Schwartz, R. G. Mathews, and D. J. Brasseaux, J . Gas Chromarog., 5 (3,251-53 (1967). (3) R. L. Martin and J. C. Winters, ANAL.CHEM.,35, 1930-33 (1963). (4) L. R. Durrett, M. C . Simmons, and I. Dvoretsky, Preprints, Division of Petroleum Chemistry, 139th National Meeting, ACS, St. Louis, Mo., March 1961, pp 63-71. ( 5 ) L. R. Durrett, L. M. Taylor, C . F. Wantland, and I. Dvoretsky, ANAL.CHEM.,35, 637-41 (1963). (6) D. J. McEwen, Ibid.,38, 1047-53 (1966). VOL. 40, NO. 3, MARCH 1968

527

Table I. Chromatographic Procedure for Determination of Ca through Clz Hydrocarbons in Full-Range Motor Gasolines Instrument Locally constructed GLC unit Detector Hydrogen-flame-ionization Locally constructed heating oven Heating device 300" C Vaporizer temperature Stream splitter temperature 270" C Detector-temperature 240' C Detector voltage 300 1-2 kliters Sample size 200/ 1 Head split ratio 200-foot X 0.010-inch id., 8% wt Column (stainless steel) squalane in rz-hexane Semiprogrammed-1.3 ml/minute obColumn flow tained at 24 psi and - 5' C to start the analysis, and held at 24 psi until the elution of o-xylene; then increased to 3.5 ml/minute obtained at 40 psi and 90" C; and finally increased to 5.0 ml/minute, obtained at 60 psi and 105" C, after the elution of n-undecane Semiprogrammed-Held at - 5" C Temperature program from start until after elution of isopentane, then heated as rapidly as possible to 25" C , held at that temperature until after elution of benzene, then heated to 90" C at 2" C/minute, held there until the elution of n-decane, then heatcd at 3" C/minute to the final temperature of 105" C

procedure t o analyze gasoline samples ; however, the detail of the method was not considered sufficient to allow adequate comparison of the individual hydrocarbon composition of a gasoline with its engine performance data and octane ratings. These comparisons can then be correlated with present and future gasoline blending schemes. The GLC method described in this article offers a relatively rapid routine analysis of full-range motor gasolines. Approximately 240 chromatographic peaks are observed in a n average gasoline, 180 of which have been specifically identified (accounting for 233 hydrocarbons). EXPERIMENTAL

Apparatus. Two G L C units constructed in this laboratory have been used for the gasoline analyses. Each unit was equipped with a conventional hydrogen-flame ionization detector (HFID), an F and M Model 240 temperature programmer, a Model VTE-1 Victoreen electrometer, and a Brown recording potentiometer (1 mV full-scale, 1 second full-scale response). In one unit, a Glas-Col 0-576 (2-liter) heating mantle in conjunction with the F and M temperature programmer functioned as the column oven, whereas the other unit was equipped with a locally constructed air-circulating oven. The air-circulating oven provided considerably better temperature control than the heating mantle, and was used in most of the analyses. For quantitative results, the gas chromatographs were monitored with an Infotronics (Model CRS-11 HSB/42) digital integrator, equipped with tape printout of retention times and peak areas. A dual-channel gas chromatograph constructed in this laboratory was used for selected analyses. The chromatographic operation of this instrument was the same as employed on the single-channel GLC units, except that the effluent from the capillary column is split equally. One half of the capillary effluent passes directly t o a conventional HFID and the other half passes through a separately thermo528

ANALYTICAL CHEMISTRY

stated mercuric perchlorate adsorption column to remove olefins and aromatics prior t o an identical HFID. The preparation and use of mercuric perchlorate adsorption columns for the removal of olefins and aromatics have been previously described (6, 7). Our adsorption columns contained a 50/50 by weight mixture of mercuric perchlorate/ firebrick, and were operated at 80" C. At this temperature, no measurable amounts of c3-C~ paraffins were removed by the adsorbent. I n addition, a Bendix Model 14-101 time-of-flight mass spectrometer (TOF-MS) was coupled to the gas chromatograph for some qualitative identifications. The coupling system used between the gas chromatograph and TOF-MS has been described (8), and consists of a splitter arrangement, whereby 90% by volume of the capillary column effluent goes to the flame detector and 10% by volume enters the mass spectrometer ionization region. Usable mass spectra were obtained from as little as lo-* of a gram of components eluted from the capillary column. Reagents and Standards. American Petroleum Institute (API) hydrocarbon standards were used whenever they were available. Additional hydrocarbon standards were obtained from Chemical Samples Co. of Columbus, Ohio, and Phillips Petroleum Co. The squalane, used as the stationary phase in the capillary column, was obtained from Eastman Chemicals. Column Preparation. The tubing used for the capillary column was Chromat I.D. obtained from Handy and Harman Tube Co. of Norristown, Pa. The stainless steel capillary column used in the analyses was coated by the dynamic method described by Ettre (9). The column was first cleaned with 20 ml of methanol followed by 10 ml of n-hexane. The solvents were forced through the column with 150 psi of nitrogen. The inside of the column was then allowed to dry under a steady stream of nitrogen prior to the coating operation. The coating was accomplished by forcing 10 ml of a solution of 8 % wt squalane in n-hexane through the column with 200 psi of nitrogen. After coating was completed, the column was preconditioned by heating to 100" C at 50 psi of helium in the chromatograph until a stable recorder baseline was obtained on the 1 X 10-lo ampere electrometer range. Chromatographic Procedure. The chromatographic procedure used in the capillary G L C analysis of the full-range finished gasolines is listed in Table I. The capillary column was cooled t o the starting temperature of - 5 " C by placing a polyethylene container filled with dry ice inside the column coil and insulating the column from ambient conditions with a two-liter heating mantel. Hydrocarbon Identification Procedures. The most common method used t o identify unknown hydrocarbons was by spiking the gasoline with API hydrocarbon standards. Generally, six or more hydrocarbon standards (with at least 5 " C difference in boiling points) were blended together a t known concentrations, The standard hydrocarbon blend was then analyzed by using the same chromatographic procedure as was used in the full-range gasoline analyses. Finally, a previously analyzed gasoline was spiked with the standard hydrocarbon blend (about 2 x vol), and the spiked gasoline reanalyzed for hydrocarbon identifications. Parallel monitoring of the G L C effluent with the Bendix TOF-MS was very helpful in establishing the carbon number and the hydrocarbon type of a particular G L C peak. In

(7) H. Mayrsohn and Q. O'Neal, Division of Water and Waste Chemistry, 148th Meeting, ACS, Chicago, Ill., September 1964. (8) J. B. Maynard, C . E. Legate, and L. B. Graiff, Combust. Flame, 11 (2), 155-164 (1967). (9) L. S . Ettre, "Open Tubular Columns in Gas Chromatography," Plenum Press, New York, 1965.

P

u

w r

U

I

15

0

MINUTES

Y

* P

1

-

m

P

rr)

1

30 lo P

0

I

I

45

I

75

60

MINUTES

3

A

I

I 90

MINUTES

I 11'5

IO5

Figure 1. Chromatogram of a typical premium-grade gasoline

some cases, it was possible to identify the hydrocarbon component from its mass spectrum. The dual-channel chromatograph was used mainly in detecting hydrocarbon types and in identification of unresolved pairs, where one component was a saturate and the other either an olefin or an aromatic. The olefin-saturate differentiations were made in the C4-C7range, in which the mercuric perchlorate adsorbent does not remove measurable amounts of paraffinic hydrocarbons, including those with reactive tertiary hydrogen atoms, such as 2-methylpentane. Aromatic-saturate differentiations were made in the C9-C12 range. Aromatics in this range are quantitatively removed by mercuric perchlorate, and it has been found that as much as 2 5 x wt of some C9-Cll saturates are removed by the adsorbent. However, we have only applied this technique qualitatively in the C9-C12carbon number range to determine if any C9-Cll saturate peaks were being masked by the large amounts of C8-G aromatics present in most gasoline samples. Sampling Procedure. Before capillary GLC analyses, gasoline samples are stored in capped 2-dram vials at 0" F. When a sample is to be analyzed, the vial is cooled in dry ice, the cap removed, a 1-2 pliter liquid sample withdrawn (using a 5-pliter Hamilton syringe), and immediately injected into the chromatograph. The sample is split 200/1 before entering the capillary column. When sampling was carried out in this manner, excellent analysis repeatability was obtained, and with no loss of light ends. Effect of Sample Evaporative Losses on GLC Results. Because of the extreme volatility of gasoline range materials, samples stored at ambient temperatures show considerable

light end loss (weathering). Several gasoline samples were stored in capped vials at ambient temperature, and analyzed periodically to determine the magnitude of weathering effects on capillary GLC results. Large differences in quantitative compositions were observed in all samples that were stored in the laboratory, with as much as 75 wt of the C4 and C5 hydrocarbons lost from some samples. Therefore, to obtain reliable quantitative results, the gasoline samples must be stored in sealed containers and/or refrigerated until analyzed. Quantitative Recovery from GLC Column. To be able to report each hydrocarbon component as per cent weight of the total hydrocarbons in the gasoline, it was necessary to demonstrate that no significant amounts of heavy hydrocarbons were lost during the GLC analysis. This was accomplished with an internal standard technique. Approximately 1OSOZ wt n-C11 was weighed into a previously analyzed gasoline, and the resulting sample analyzed in duplicate o n a 200-foot squalane column. Values of 10.54 and 10.48Z wt n-C11were obtained (corrected for n-Cll found in the original analysis of the gasoline). These values are within the estimated repeatability of the method. Previous evidence has shown that the response of an H F I D to most of the individual hydrocarbons can be assumed to be identical on a weight basis (IO). We have calibrated our detectors with the paraffins, olefins, and aromatics that are predominant in gasolines, and found the responses of the detectors to be essentially the same for all the hydrocarbons examined, Thus, no response factors were employed in our analyses. (10) W. A. Dietz, J. Gas Chromatog., 5 (2), 68-71 (1967). VOL. 40, NO. 3, MARCH 1968

529

I

15

0

I

45

30

MINUTES

I

MINUTES

I 75

60

ln ln

I 90

MINUTES

I

105

Figure 2. Chromatogram of a typical regular-grade gasoline

RESULTS AND DISCUSSION

Analyses of Full-Range Finished Gasoline. Capillary GLC analyses of full-range finished gasolines, utilizing the method described in this article, are completed in somewhat less than 2 hours, with 96% wt to 99% wt of the hydrocarbons in the gasolines identified. Analyses of premium gasolines resulted in a greater percentage of hydrocarbons identified than analyses of regular gasolines because of the differences in blending schemes. Regular gasolines contain a higher percentage of unidentified C8 and C 9 olefins and cycloparaffins (mainly from catalytically cracked gasoline), Figures 1 and 2 are reproductions of chromatograms of typical premium- and regular-grade gasolines, respectively. The identities of 233 hydrocarbon components found in the gasolines are given by peak number in Table 11. Actual hydrocarbon compositions of typical premium- and regulargrade gasoline blends in per cent weight are presented in Table I11 where 98.30% wt of the premium-grade gasoline and 96.98% wt of the regular-grade gasoline are identified, Included in Table I11 is the hydrocarbon composition of API Prototype Fuel No. 1-Premium Leaded Reference (11).

(11) W. F. Biller, “Hydrocarbon Emissions and Motor Fuel

Volatility-Progress Report,” paper presented at the America Petroleum Institute’s Division of Refining, Session on Fuels and Emissions, Los Angela, Calif., May 16, 1967.

530

ANALYTICAL CHEMISTRY

This fuel is a representative blend of the fuels manufactured in the United States during 1965, both in terms of its volatility characteristics and also in terms of its composition, and as such, was designed to represent the national average fuel. This fuel should be generally available for the standardization of chromatographic results. Repeatability of the capillary GLC method was checked by performing duplicate analyses on a premium-grade gasoline. These analyses were made in succession by one operator, utilizing the same equipment and gas-chromatographic procedure. In addition, the reproducibility of the method was demonstrated by duplicate analyses of another premiumgrade gasoline. These analyses were made by different operators, 1 month apart, using different sets of equipment (except for the capillary column), and following the same chromatographic procedure. Table IV lists the calculated average deviations and standard deviations, as well as the 95% confidence limits, for the repeatability and reproducibility data. Because of the wide range of per cent weight values encountered in a gasoline analysis (0.01% wt to 15.00% wt), the data were divided into three subgroups. Because both temperature and pressure programs are employed in the analysis procedure, it is not practical to identify peaks in a routine sample by absolute retention times, as the actual elution time of a given peak may vary somewhat from run to run, However, if the temperature and pressure programs (Table I) are closely followed, the relative elution order

Table 11. Hydrocarbon Identification of Chromatographic Peaks Shown in Figures 1 and 2 Peak Component number 1 Propane 2 Isobutane 3 Isobutylene + butene-1 4 rz-Butane 5 trans-2-Butene 6 Neopentane 7 cis-2-Butene 8 3-Methyl-1-butene 9 Isopentane 10 Pentene-1 11 2-Methyl-1-butene 12 2-Methyl-l,3-butadiene 13 rz-Pentane 14 trans-2-Pentene 15 cis-2-Pentene 16 2-Methyl-2-butene 17 3,3-Dimethyl-l-butene 18 2,2-Dimethylbutane 19 Cyclopentene 20 3-Methyl-1-pentene 4-methyl-1-pentene 21 4-Met hyl-cis-2-pentene 22 2,3-Dimethyl-1-butene 23 Cyclopentane 24 2,3-Dimethylbutane

+

25 26 27 28 29 30 31 32 33 34 35 36 37 38

+ (dcmethyl-tra,?s-2-pentene)a

2-Methylpentane 2-Methyl- 1-pentene 3-Methylpentane + (hexene-l) (2-ethyl-1-butene) cis-3-Hexene trans-3-Hexene 3-Methylcyclopentene 2-Methyl-2-pentene 3-Methyl-cis-Zpentene rz-Hexane + (4,drdimethyl-1-pentene) tru,is-2-Hexene cis-2-Hexene

+

3-Methyl-rraris-2-pentene

51

4,4-Dimethyl-truns-2-pentene Methylcyclopentane + 3,3-dimethyl-l-pentene 2,2-Dimethylpentane + 2,3-dimethyl-2-butene (2,3,3-trimethyl-l-butene) Benzene 2,4-Dimethylpentane 4,4-Dimet hyl-cis-2-pentene 2,2,3-Trimethylbutane 2,4-Dimethyl-l-pentene 1-M ethylcyclopentene + 2-methyl-cis-3-hexene 2,4Dimet hyl-2-pentene + 3-ethyl- 1-pentene + 3-methyl-1-hexene 2,3-Dimethyl-l-pentene 2-Methyl-rrans-3-hexene + 5-methyl-1-hexene 3,3-Dimethylpentane Cyclohexane + (4-methyl-cis-2-hexene) 4- Methyl- 1- hexene

52 53

3-Methyl-2-ethyl- 1-butene 5-Met hyl-trans-2-hexene

39

+

40

41 42 43 44 45 46 47 48 49 50

+ 4-methyl-traris-2-hexene

Boiling point,

c (12)

-42.07 -11.73 -6.90 -6.26

Peak Component number 54 Cyclohexene 2-Methylhexane 55 (5-methyl-cis-2-hexene) 56 2,3-Dimethylpentane

+ + (1,I-dimethylcyclopentane) + (3,4-dimethyl-cis-2-pentene)

-0.50

0.88 9.50 3.72 20.06 27.85 29.97 31.16 34.07 36.07 36.35 36.94 38,57 41.24 49.74 44.24 54.14 53.88 56.30 55.67 49.26 57.99 58.55 60.27 60.72 63.28 63.49 64.66 66.47 67.08 65.0 67.29 67.70 68.74 72.49 67.87 68.84 70.44 76.75 71.81 77.57 79.20 73.21 77.87 80.10 80.50 80.42 80.88 81.64 75.8 86 83.26 84.11 84 84.28 86 85.31 86.06 80.74 87.31 86.73 87.56 86.1 88.11

57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100

3-Methylhexane 1-cis-3-Dimethylcyclopentane 2-methyl-1-hexene

+ + 3,4-dimethyl-rraiis-2-pentene 1-traru-3-Dimethylcyclopentane + 1-heptene + 2-ethyl-1-pentene 3-Ethylpentane + 3-methyl-trans-2-hexene

1-rrans-2-Dimethylcyclopentane 2,2,4-Trimet hylpentane (tram-3-heptene) cis-3-Heptene 3-Methyl-cis-3-hexene 2-methyl-2-hexene

+

+ + 3-methyl-tra,zs-3-hexene

3-Ethyl-2-pentene trans-2-Heptene n-Heptane (3-methyl-cis-2-hexene) 2,3-Dimethyl-2-pentene cis-2-heptene

+ +

1-cis-2-Dimethylcyclopentane

Methylcyclo hexane 2,2-dimethylhexane 1,1,3-trimethylcyclopentane 2,5-Dimethylhexane ethylcyclopentane 2,4-Dimethylhexane 2,2,3-Trimethylpentane l-truns-2-cis-4-Trimethylcyclopentane Toluene 3,3-Dimethylhexane l-trans-2-Cis-3-trimethy1cyclopentane 2,3,4-TrimethyIpentane 2,3,3-Trimethylpentane 1,1,2-Trimethylcyclopentane 2,3-Dimethylhexane 2-methyl-3-ethylpentane 2-Methylheptane 4Methylheptane 3,4-Dimethylhexane ( 1-cis-2-trans-4-trimeth ylcyclopentane) 3-Ethylhexane 3-Methylheptane

+ + +

+ +

+ (3-methyl-3-ethylpentane)

1,1,3-trarzs-4-Tetramethylcyclopentane 2,2,5-TrimethyIhexane (l-cis-2-cis-4trimethylcyclopentane) 1,l-Dimethylcyclohexane 4- 1-trans-4dimethylcyclohexane 1-Cis-3-dimethylcyclohexane I -Methyl-truns-3-ethylcyclopentane 2,2,4-Trimethylhexane 1-Methyl-tram-2-ethylcyclopentane 1-methyl-cis-3-ethylcyclopentane Cycloheptane 1-Methyl-I-ethylcyclopentane l-tru~zs-2-Dimethylcyclo hexane

+

+

+ l-cis-2-cis-3-trimethylcyclopentane

n-Octane

1-cis4Dimethylcyclohexane I-tra,zs-3-Dimethylcyclohexane

2,4,4Trimethylhexane

(12) S.W. Ferris, “Handbook of Hydrocarbons,” Academic Press, Inc., New York, 1955.

Boiling point,

c (12)

82.98 90.05 89.5 89.78 87.85 87.9 91.85 91.73 91.95 90.5 90.77 93.64 94 93.48 94 91.87 99.24 95.67 95.75 95.33 95.44 93.53 96.01 97.95 98.43 94 97.40 98.5 99.57 100.93 106.84 104.89 109.10 103,47 109.43 109.84 109.29 110.63 111.97 110.2 113.47 114.76 113.73 115.61 115.65 117.65 117.71 117.73 116.73 118.53 118.93 118.26 121.6 124.08 118 119.54 119.35 120.09 120.8 126.54 121.2 121.4 118.79 121.52 123.42 123.0 125.67 124.32 124.45 130.65 Contin ued

VOL. 40, NO. 3, MARCH 1968

531

Table 11. Continued Peak number 101 102 103 104 105

Boiling point,

Component Iso-propylcyclopentane 2,3,5-Trimethylhexane 2,2-Dimethylheptane 1-Methyl-cis-2-ethylcyclopentane 2,CDimethylheptane 2,2,3-trimethylhexane 106 2,2-Dimethyl-3-ethylpentane 2-methyl-Cethylhexane 107 2,6-Dimethylheptane (1-cis-2-dimethylcyclohexane) 108 n-Propylcyclopentane 109 Ethylcyclohexane 110 2,5-Dimethylheptane 3,5-dimethylheptane 111 Ethylbenzene 112 2,4-Dimethyl-3-ethylpentane 113 3,3-Dimethylheptane 114 1,1,3-Trimethylcyclohexane 115 2,3,3-Trimethylhexane 116 l-cis-3-cis-5-Trimethylcyclohexane 117 2-Methyl-3-ethylhexane 118 pXylene 119 rn-Xylene (3,3,4trimethylhexane) 120 2,3-Dimethylheptane 121 3,CDimethylheptane 122 CMethyloctane 123 2-Methyloctane 124 3-Ethylheptane 125 3-Methyloctane 126 o-Xylene

+ + +

+

+

127 128 129 130 131 132 133 134 135 136 137 138

+ (2,2,4,5-tetramethylhexane)

2,2,CTrimethylheptane 2,2,5-Trimethylheptane 2,2,6-trimethylheptane 2,5,5-Trimethylheptane 2,4,Ctrimethylheptane Isopropylbenzene n-Nonane 3,3,5-Trimethylheptane 2,4,5-Trimethylheptane 2,3,5-trimethylheptane n-Propylbenzene 2,2,3,3-Tetramethylhexane 2,6-dimethyloctane 1-Methyl-3-ethylbenzene 1-Methyl4ethylbenzene 3,3,CTrimethylheptane 3,4,Ctrimethylheptane 3,4,5-trimethylheptane

+ +

+ +

+ +

c (12)

126.42 131.34 132.69 128.05 133.5 133.6 133.83 133.8 135.21 129.73 130.95 131.78 136.0 136.0 136.19 136.73 137.3 136.63 137.68 138.41 138.0 138.35 139.10 140.46 140.5 140.6 142.48 143.26 143.0 144.18 144.41 147.8 147.88 148 148 152.80 153 152.39 150.80 155.68 157 157 159.22 160.31 158.54 161.31 161.99 164 164 164

of the hydrocarbons does not change, even when the analysis is begun a t 0" C instead of - 5 O C. Hydrocarbons are eluted from the squalane column in essentially boiling point order, permitting easier qualitative identifications. Thus, when analyzing a new gasoline sample, the peaks are identified by their elution order from the squalane capillary column. In the analyses of over 50 gasoline samples, no difficulties in peak identifications were encountered, as all gasolines were found t o contain essentially the same hydrocarbons, differing only in concentration. In many cases during development of the method, peak identifications were checked with other techniques such as relative retention times and the dualchannel GLC unit. In each case, the peak identifications obtained by the elution order technique were found t o be accurate.

532

ANALYTICAL CHEMISTRY

Peak number Component 139 1-Methyl-2-ethylbenzene 5-methylnonane 140 4-Methylnonane 141 1,3,5-Trimethylbenzene 142 2-Methylnonane 143 tert-Butylbenzene 144 Unidentified Cloalkylate peak 145 3-Methylnonane 146 1,2,CTrimethylbenzene 147 sec-Butylbenzene 148 Isobutylbenzene 149 1-Methyl-3-isopropylbenzene 150 n-Decane 151 1,2,3-Trimethylbenzene

+

152 153 154 155 156 157

+ + indane

1-rnethyl4isopropylbenzene 1-Methyl-2-isopropylbenzene

1,3-Diethylbenzene Unidentified C,I alkylate peak 1-Methyl-3-n-propylbenzene

n-Butylbenzene 1,2-Diethylbenzene 1,Cdiethylbenzene 1-methyl-Cn-propylbenzene

+ +

158 159 160 161 162 163 164

1,3-Dimethyl-5-ethylbenzene Unidentified CI1alkylate peak 2-Methylindane 1,4-Dimethyl-2-ethylbenzene 1-Methylindane

165 166

unidentified Cl, alkylate peak 1,3-Dimethyl-4-ethylbenzen.e 1,3-Dimethyl-2-ethylbenzene 1,2-dimethyl-4-ethylbenzene

167 168 169 170 171 172 173 174 175 176 177 178 179 180

1-Methyl-2-n-propylbenzene

1-Methyl-3-terr-butylbenzene

+

+

1-Methyl-4-rert-butylbenzene

1,2-Dimethyl-3-ethylbenzene n-Undecane 1,2,4,5-Tetramethylbenzene 1,2,3,5-Tetramethylbenzene Isopentylbenzene 5-Methylindane CMethylindane n-Pentylbenzene 1,2,3,CTetramethylbenzene Tetralin Naphthalene 1,3-Dimethyl-5-terr-butylbenzene n-Dodecane a ( ) Designates minor component.

Boi 1ing point, O

c (12)

165.15 165.1 165.7 164.72 166.8 169.12

...

167.8 169.35 173.31 172.76 175.14 174.12 176.08 177.10 178.15 177 181.10

...

181.80 183.27 183.42 183.30 183.75 184.80 183.75 ... 184 186.91 186.5 189.26

...

188.41 190.01 189.75 192.76 193.91 195.89 196.8 198.0 198.9 199 203 205.46 205.4 205.57 217.96 205.1 216.28

In the original identification work, it was found that many of the 180 identified peaks in the chromatograms contained two or more components; however, almost all of these have been resolved by changing the chromatographic procedure. Squalane is considered to be essentially nonpolar, but with a decrease in temperature, the retention times of olefins will increase and aromatics decrease, relative to saturates. An example is the relative elution of benzene and 2,4-dimethylpentane. At 25 O C, benzene will be eluted after 2,4-dimethylpentane, However, as the temperature of the column is lowered, benzene will be retarded less than the saturate, and at about 10" C, benzene elutes prior to 2,4-dimethylpentane. Between 15" and 20" C, the two compounds are eluted as an unresolved pair. Another example is the relative separation of the C4-C6saturates and olefins. In the C4-Cs region, the

Table 111. Hydrocarbon Compositions of Typical Premium- and Regular-Grade Gasolines and the API Prototype Fuel No. 1-Premium-Leaded Reference All values in per cent weight; trace components less than 0.01

Regular- Premium- API (11) fuel grade grade

Component

0.14 0.30 0.02 3.93 0.16 0.02 0.13

Propane Isobutane Isobutylene butene-1 !]-Butane trans-2-Butene Neopentane cis-2-Butene 3-Methyl-1-butene Isopentane Pentene-1 2- Methyl- 1-butene rz-Pentane trans-2-Pentene cis-2-Pentene 2-Methyl-2-butene 2,2-Dimethylbutane Cyclopentene 3-Methyl-1-pentene Cmethyl1-pen tene 4-Methyl-cis-2-pentene 2,3-Dimethyl-l-butene Cyclopentane 2,3-Dimethylbutane

+

0.08

7.88 0.34 0.35 7.27 0.52 0.43 1.09 0.17 0.13

+

0.16 0.05 0.10

4-Methyl-rrutzs-2-pentene

2-Methylpentane 2-Methyl-1-pentene 3-Methylpentane (hexene-l)a (2-ethyl-1-butene) cis-3-Hexene tra/zs-3-Hexene 3-Methylcyclopentene 2- Methyl-2-pentene 3-Methyl-cis-2-pentene n-Hexane (4,4-dimethyl-lpentene) trans-2-Hexene cis-2-Hexene

+

+

+

3-Methyl-trans-2-pentene 4,4-Dimethyl-trans-2-pentene

+

Methylcyclopentane 3,3-dimethyl-l-pentene 2,2-Dimethylpentane + 2,3-di2,3,3-trimethyl-2-butene methylbutene Benzene 2,4-Dimethylpentane 2,2,3-Trimethylbutane

+

4,4-Dimethyl-cis-2-pentene

2,4-Dimethyl- 1-pentene 1-Methylcyclopentene + 2-methyl-cis-3-hexene 2,4-Dimethyl-2-pentene 3-ethyl-1-pentene + 3-methyl-1-hexene 2,3-Dimethyl- 1-pentene 2-Methyl-trans-3-hexene 5-methyl-1- hexene 3,3-Dimethylpentane 4-methyl-cis-2Cyclohexane hexene 4- Methyl- 1-hexene 4-methyltrans-Zhexene 3-Methyl-2-ethyl-1-butene 5-methyl-frans-2-hexene Cyclohexene (5-methyl-cis2-Methylhexane 2-hexene) 2,3-DimethyIpentane (1,ldimethylcyclopentane) +

+

+

+

+

+

+

+

(3,4-dimethyI-cis-2-pentene)

0.01 0.37 0.04 4.29 0.20 0.04 0.17 0.12 10.17 0.45 0.22 5.75 0.90 0.67 0.96 0.46 0.18 0.18 0.04

0.01 0.12 0.33 4.70 0.16 0.05

0.14 0.08 6.07 0.33 0.66 10.92 0.70 0.63 1.28 0.84 0.12 0.12 0.04 0.08 0.19 1.07 0.08 2.91 0.20

0.58

0.08 0.51

0.59 0.30 3.85 0.22

1.55 0.18 3.76 0.22

2.72 0.13 0.15 0.08 0.32 0.45

2.23 0.11 0.12 0.04 0.27 0.37

0.15 0.03

3.50 0.36 0.24 0.44 Trace

1.51 0.18 0.15 0.34 Trace

0.24 0.20 0.18 0.32 Trace

1.80 0.11 0.31 0.35

1.50

0.62

0.31

0.20 1.35 0.32 Trace 0.02 Trace

0.14 0.81 1.71 0.04 Trace 0.03

0.08

0.12 0.23 0.01 Trace 0.02

0.37

0.32

0.28

0.05

0.05

0.05

0.01

0.02

0.02

0.05

Trace

0.04 0.02

0.07 0.03

0.36

0.17

0.02

0.08

0.09

0.08

0.03 0.03

0.04 0.03

0.05

1.25

1.48

0.36

0.47

4.17

0.32

0.03

wt

Regular- Premium- API (121 grade fuel grade

Component

1.77

0.30

0.27

0.21

0.27

0.17

0.16 0.16

0.05

4.58 0.16

0.99 0.14

0.31 0.04 0.06

0.29 0.03 0.09

1.96

0.31

0.12 0.09

0.10 0.07

0.31 0.42 0.18 0.23 0.04 12.30

0.16 0.52 0.21 0.82 0.09 0.03 21.80

0.06 2.26 2.28 0.09

0.01 2.80 1.59 0.06

0.60 0.48 0.22

0.58

0.16 0.01

0.37 Trace

0.63

0.77

0.74

5.89

0.17 0.06 0.11

0.24 0.06 0.18

0.13

0.07

0.04

0.11

0.05

0.08

0.18

0.12

0.09

1.43 0.12 0.04 0.02

0.42

0.36

0.08

0.05

0.02 0.01 0.15 0.01

0.16 0.01 1.09 0.03

3-Methylhexane 1.41 1-cis-3-Dimethylcyclopentane 2-methyl-1-hexene 3,4-dimethyl-trans-2-pentene 0.41 1-trans-3-Dimethylcyclopentane 1-heptene 2-ethyl-lpentene 0.40 3-Ethylpentane 3-methyl-trans2-hexene 0.25 1-trans-2-Dimethylcyclopentane 0.20 2,2,4Trimethylpentane (trans3-heptene) 0.32 cis-3-Heptene 0.17 3-Methyl-cis-3-hexene 2-methyl-2-hexene + 3-methyl-trans-3-hexene 0.35 0.04 3-Ethyl-2-pentene trans-2-Heptene 0.10 n-Heptane (3-rnethyl-cis-2hexene) 1.92 2,3-Dimethyl-2-pentene + cis-2heptene 0.14 1-cis-2-Dimethylcyclopentane 0.13 Methylcyclohexane 2,2-dimethylhexane 1,1,3-trimethylcyclopentane 0.61 2,5-Dimethylhexane 0.24 Ethylcyclopentane 0.14 2,CDimethylhexane 0.34 Trace 2.2.3-Trimethvl~entane l~t~uns-2-cis-~~rimethvlcvclo~entane 0.16 Toluene (3,3-Dimethyihexane) 5.92 l-/rans-2-cis-3-Trimethy1cyclopentane 0.25 2,3,CTrimethyIpentane 0.11 2,3,3-Trimethylpentane 0.05 1,1,2-Trimethylcyclopentane 0.11 2,3-Dimethylhexane 2-methyl3-ethylpentane 0.39 2-Methylheptane 1.05 4Methylheptane 0.52 (l-cis-23,4-Dimethylhexane trans-4-trimethylcyclopentane) 0.20 3-Ethylhexane Trace (3-methyl-33-Methylheptane ethylpentane) 1.54 (l-cis-22,2,5-Trimethylhexane cis-4trimethylcyclopentane) 0.17 1,I-Dimethylcyclohexane 1-trans-4dimethylcyclohexane 1-cis-3-dimethylcyclohexane 0.27 1-Methyl-rrarts-3-ethylcyclopentane 0.12 2,2,CTrimethylhexane 0.18

+

+

+

+ +

+

+

+

+

+

+

+ +

+

+

+

+

0.50

0.06

1.76 0.28

1-Methyl-trans-2-ethylcyclopen-

+

tane 1-methyl-cis-3-ethylcyclopentane Cycloheptane + l-methyl-lethylcyclopentane 1-trans-2-Dimethylcyclohexane l-cis-2-cis-3-trimethylcyclo-

+

pentane n-Octane (1-cis-4-dimethylcyclohexane) I-frans-3-Dimethylcyclohexane 2,4,4Trimethylhexane Isopropylcyclopentane 2,3,5-Trimethylhexane 2,ZDimethylheptane

+

0.05 0.08

Continued

VOL. 40, NO. 3, MARCH 1968

533

Table 111, (Continued ) Component 1-Methyl-cis-Zethylcyclopentane 2,CDimethylheptane 2,2,3-trimethylhexane 2,2-Dimethyl-3-ethyIpentane 2-methyl-4-ethylhexane (l-cis-22,QDimethylheptane dimethylcyclohexane) n-Propylcyclopentane Ethylcyclohexane 2,5-Dimethylheptane 3,5-DimethyIheptane Ethylbenzene

+

Regular- Premium- API (ZI) grade fuel grade 0.11 0.07 0.06

+

+

+

2,4Dimethyl-3-ethylpentane

3,3-Dimethylheptane 2,3,3-Trimethylhexane 2-Methyl-3-ethylhexane p-X ylene m-Xylene (3,3,4-trimethylhexane) 2,3-Dimethylheptane 3,CDimethylheptane 4-Methyloctane 2-Methyloctane 3-Ethylheptane 3-Methyloctane o-Xylene (2,2,4,5-tetramethylhexane) 2,2,4-Trimethylheptane 2,2,5-Trimethylheptane 2,2,&trimethylheptane 2,5,5-Trimethylheptane 2,4,4trimethylheptane Isopropyl benzene n-Nonane 3,3,5-Trimethylheptane 2,4,5-Trimethylheptane 2,3,5-trimethylheptane n-Propylbenzene 2,6-Dimethyloctane (2,2,3,3-tetramethylhexane) 1-Methyl-3-ethylbenzene 1-Methyl-4-ethylbenzene 3,3&TrimethyIheptane 3,4,4-trimethylheptane 3,4,5-trimethylheptane 1-Methyl-2-ethylbenzene 5-methylnonane CMethylnonane

+

0.24

0.08

0.26

0.09

0.02

0.03

0.20 0.06 0.36

0.07 0.01 0.17

0.23 0.04 0.42

0.14 2.70 0.05 0.08 0.12 0.13 1.54

0.16 1.70 0.03 0.04

0.21 0.36 0.07 0.01 0.06 0.07 0.77

3.87 0.39 0.33

0.60

2.05 0.12

1.94 0.17

1.12 1.70

0.07

0.27

1.31

0.06 0.23 0.83 0.05

0.21 0.10 0.14 0.02

0.92 0.07 0.07 0.06

0.07 0.72

0.17 0.24

0.56 0.08

0.12 1.84 1.00

0.06

0.83 0.42

0.09 0.31 0.18

0.08

0.35

0.53

0.90

0.34 0.04

0.13

0.62 0.16 0.85

+ +

+

+

+ +

+

0.04 1.58

1.77 0.51 0.17 0.22 0.16 0.04 0.34

0.55

+

0.05

0.26

3.83 0.13 0.07 0.11 0.14 0.02

Table IV.

0.05

Component 1,3,5-Trimethylbenzene 2-Methylnonane 3-Methylnonane 1,2,4Trimethylbenzene sec-Butylbenzene Isobutylbenzene 1-Methyl-34sopropylbenzene

n-Decane 1,2,3-Trimethylbenzene 1-methyl-4isopropylbenzene

+

Regular- Premium- API (11) grade grade fuel 0.76 0.39 0.13 0.41 0.06 0.08 0.32 0.06 0.07 2.83 1.61 0.66 0.13 0.01 0.01 0.06 0.01 0.01 0.12 0.03 0.01 0.50 0.08 0.04

+

1-Methyl-2-isopropylbenzene

indane 1,3-Diethylbenzene 1-Methyl-3-n-propylbenzene

n-Butylbenzene 1,2-Diethylbenzene 1,4-diethylbenzene 1-methyl-4-n-propylbenzene 1- Methyl-2-n-propylbenzene 1,3-Dimethyl-5-ethylbenzene 2-Methylindane 1,CDimethy1-2-ethylbenzene 1-Methylindane

+ +

1-Methyl-3-rert-butylbenzene 1,3-Dimethyl-4-ethylbenzene

+

1,3-Dimethyl-2-ethyIbenzene 1,2-dimethyl-4-ethylbenzene 1- Methyl-4-rerr-butylbenzene 1,2-Dimethyl-3-ethylbenzene n-Undecane 1,2,4,5-TetramethyIbenzene 1,2,3,5-Tetramethylbenzene Isopentylbenzene 5-Methylindane CMethylindane n-Pentylbenzene 1,2,3,4-Tetramethylbenzene Tetralin Naphthalene 1,3-Dimethyl-5-tert-butylbenzene n-Dodecane Total saturates identified Total olefins identified Total aromatics identified Total components unidentified Total a

(

0.68

0.32

0.10

0.35 0.25 0.48 0.25

0.15 0.08 0.16

0.05

0.08 0.04

0.05

0.44 0.16 0.42 0.10 0.36 0.17 0.11 0.27

0.09

0.50

0.19 0.04 0.03 0.07 0.10 0.17 0.07 0.11 0.03 0.03 0.03

0.13 0.09 0.22 0.21 0.42 0.17 0.30 0.16 0.14 0.19 0.14 0.24 0.16 0.09 56.38 7.69 32.91 3.02 100.00

0.11

0.16 0.01 0.11 0.07

0.05

0.18 0.02 0.09 0.07 0.03 0.13

0.02 0.10

0.02 0.05

62.30 7.50 28.50 1.70 100.00

0.05

0.04 0.09 0.03 0.08 0.07 0.02 0.05 0.05

0.14 0.14 0.09 0.01 0.01

0.02 0.01 0.09 0.03 0.04 59.24 7.85 29.23 3.68 100.00

) Designates minor component.

Repeatability Calculations (13)

Duplicate runs on a typical premium-grade gasoline. (Same GLC unit, same operator) Range of

zwt values

No. of peaks

Deviations min-max

Average deviation

Standard deviation

65 74 21

0.00-0.02 0 . 0 0 4 . 19 0.00-0 .23

0.0057 0.012 0,065

0.0061 0.019

0.014.09 0.10-0.99 1.00-20.00

0,060

95 z7 Confidence limits f0.002 f0.004 3~0.027

Total 160 Reproducibility Calculations (13) Duplicate runs on a typical premium-grade gasoline. (Different GLC unit, different operator) Range of % wt values 0.01-0.09 0.1O-O.99 1.00-20.00

534

No. of peaks

62 79 23 Total164

ANALYTICAL CHEMISTRY

Deviations min-max 0.00-0.06

0.004.29 0.00-0.60

Average deviation 0.014 0.034 0,111

Standard deviation 0.014 0.037 0.120

z

95 Confidence limits +0.004 f0.008 rt0.052

olefins elute immediately after large saturate peaks; thus, separation from the saturates is favored by a low column temperature, which allows greater retardation of olefins relative to saturates. In the case of the C Bhydrocarbons, the opposite is true, since most of the olefins in this range elute immediately preceding saturate peaks. Thus, to obtain optimum resolution in both C4-C5 and C 6 ranges, the column temperature is held at - 5 ” C until the elution of isopentane, and then the column is warmed as rapidly as possible to 25” C. This is accomplished by activating the oven blower as soon as the dry ice is removed from the column coil. When significant deviations from the temperature program are allowed to occur, similar peak elution shifts are observed in several areas of the

gasoline chromatograms. Longer squalane capillary columns were tried, and as expected, an increase in resolution was observed, However, with a 300-foot X 0.010-inch i.d. 10% wt squalane column, the analysis time was over 4 hours. Obviously, the chromatographic procedure used for routine GLC analyses of gasolines is a compromise which gives maximum information in a reasonable time period. The capillary GLC method described in this article is also applicable to the analysis of other hydrocarbon mixtures such as catalytically cracked gasoline, catalytic reformate, alkylate, and light naphtha feedstocks. With modifications of present sampling procedures, the method could easily be applied in analysis of the individual hydrocarbons found in automotive exhaust gases.

(13) W. J. Youden, “Statistical Methods for Chemists,” Wiley, New York, 1951.

RECEIVED for review October 16, 1967. Accepted January 2, 1968.

Apparatus Combining Gas Chromatography with Spectrophotofluorometry by Means of a Flowing Liquid Interface Malcolm C. Bowman and Morton Beroza Entomology Research Division, Agricultural Research Service, U.S . Department of Agriculture, Tifton, Ga. 31 794 and B e l t s d e , M d . 20705 An apparatus has been devised to combine the high separative powers of the gas chromatograph with the high sensitivity and selective response of the spectrophotofluorometer. The solute in the effluent of the gas chromatograph is picked up by a slowly flowing stream of alcohol, and the alcohol solution is monitpred in a flow cell at the desired excitation and emission wavelengths. The combination has been used to analyze pesticides, air pollutants, and methylenedioxyphenyl compounds. Analyses in the nanogram range were possible, and sensitivity frequently exceeded that of the flame ionization detector. The device appears to have general applicability to the analysis of fluorescing substances that can be determined by gas chromatography and captured by the solvent that travels to the flow cell. The eluate of the flow cell may be collected for retention of fractions or for spectral or other analyses.

GASCHROMATOGRAPHY has been joined with a wide variety of instrumentation-e.g., the mass spectrometer ( I , 2 F t o obtain apparatus with enhanced analytical capabilities. Such combinations have improved the speed and sensitivity of analyses, circumvented the need t o trap out minute amounts of pure substance, eliminated chemical change of substances unstable subsequent to elution, provided identification more certain than possible by retention time only, and elevated the selectivity of response to allow analyses of samples with minimum cleanup. The combination of gas chromatography with spectrophotofluorometry (SPF) appeared to possess most of these virtues. SPF is widely used and well understood ( 3 ) ; the spectra of a (1) R. Ryhage, J. Lipid Res., 5, 245 (1964). (2) J. T. Watson and K. Biemann, ANAL.CHEM., 37,844 (1965). (3) S. Udenfriend, “Fluorescence Assay in Biology and Medicine,” Academic Press, New York, 1962.

great variety of compounds have been cataloged. SPF is also highly sensitive and its response can be made very selective by appropriate choice of excitation and emission wavelengths. Our own special interest in the union of the two techniques was for the analysis of such environmental contaminants as pesticides and air pollutants, many of which fluoresce strongly ; however, the combination appears to have general applicability t o fluorescent compounds that can be determined by gas chromatography. The present paper describes a device that proved to be a suitable link between the two instruments. In essence, the fluorescent substances separate in the gas chromatograph and, upon emergence, are absorbed by a flowing stream of ethanol as the carrier gas escapes; the ethanol solution then passes through 8 flow cell that is monitored by a recording spectrophotofluorometer. The combination appears t o have good sensitivity. The compounds checked could be determined in the nanogram range and some weredetectable at levels below 1 ng. EXPERIMENTAL

Gas Chromatograph (GLC). An F & M Model 700 gas chromatograph ( F & M Scientific Corp., Avondale, Pa,) was operated isothermally at oven temperatures between 130’ and 220” C. Two columns were used; both were 4-mm i.d., 6-mm 0.d. glass columns, 125 cm long. One contained 5 x w/w QF-1 on 80/100 mesh Gas-Chrom Q (Applied w/w DC Science Labs., State College, Pa.), and the other 200 o n the same support. The flow rate of the nitrogen carrier gas was 100 ml/minute. Columns were conditioned overnight at 240” C before use. The injection port was maintained 20” C above oven temperature. A ’/*-inch 0.d. stainless steel line conducted the carrier gas from the column exit t o the mixing vessel (Figure 1); it terminated in a male Luer

5z

VOL. 40, NO. 3, MARCH 1968

e

535