Energy & Fuels 1997, 11, 909-914
909
Determination of Sulfides and Thiols in Petroleum Distillates Using Solid-Phase Extraction and Derivatization with Pentafluorobenzoyl Chloride Jane S. Thomson,* John B. Green, and Ted B. McWilliams BDM-Petroleum Technologies, P.O. Box 2543, Bartlesville, Oklahoma 74005 Received September 23, 1996X
An improved method for differentiating sulfides, thiophenes, and thiols in petroleum distillates is presented. Thiophenes are separated from sulfides and thiols via solid-phase extraction. Reaction with pentafluorobenzoyl chloride is employed to form thiol esters, while sulfides do not react. Thiol elution times increase sufficiently after derivatization to provide separation from sulfides during gas chromatographic analysis. In addition, electron impact mass spectra of derivatized thiols exhibit an intense, characteristic 195 fragment ion, which aids in their identification.
Introduction Sulfur compounds have been undesirable components in petroleum since the early days of the industry.1 In recent years, allowable sulfur levels in transportation fuels have been drastically lowered by government regulations to combat air pollution. Industrial processes to reduce sulfur to allowable levels proceed via thermal cracking and/or catalytic desulfurization. A knowledge of desulfurization reactants and products is useful for the optimization of sulfur removal processes. The main classes of sulfur compounds found in virgin petroleum are thiophenes and their benzologs, cyclic sulfides, and thiols. Since thiophenic compounds are typically more resistant to removal by desulfurization processes, they generally constitute the most prevalent class of sulfur compounds found in gasoline and other distillate products. This has led to an extensive body of research on separation and speciation of benzologs of thiophenes, recently reviewed by Andersson and Schmidt.2 There is less information available on the separation and speciation of sulfides and thiols, however. Owing to their structural diversity and low concentration, direct speciation of sulfides and thiols in whole distillates is quite difficult. For this reason, analyses for these compounds have typically been limited to techniques that rely on elaborate preseparation procedures prior to identification.3-8 However, low molecular weight sulfides and thiols have been determined directly in the presence of hydrocarbons using sulfur-specific GC deAbstract published in Advance ACS Abstracts, June 1, 1997. (1) Rall, H. T.; Thompson, C. J.; Coleman, H. J.; Hopkins, R. L. Sulfur Compounds in Crude Oil. Bull.sU.S., Bur. Mines 1972, 659, 1-187. (2) Andersson, J. T.; Schmid, B. J. Chromatogr. 1995, 693, 325338. (3) Snyder, L. R. Anal. Chem. 1961, 33, 1538-1543. (4) Vogh, J. W.; Dooley, J. E. Anal. Chem. 1975, 47, 816-821. (5) Nishioka, M.; Campbell, R. M.; Lee, M. L.; Castle, R. M. Fuel 1986, 65, 270-273. (6) White, C. M.; Douglas, L. J.; Perry, M. B.; Schmidt, C. E. Energy Fuels 1987, 1, 222-226. (7) Payzant, J. D.; Mojelsky, T. W.; Strausz, O. P. Energy Fuels 1989, 3, 449-454. X
S0887-0624(96)00165-X CCC: $14.00
tection methods such as atomic emission,2,9,10 flame photometry,11 or chemiluminescence.12 An alternative approach is to determine the total sulfide and/or thiol content. For example, the sulfide content has been estimated in whole samples by oxidation of the sulfides to sulfoxides followed by quantitation using the sulfoxide band in the IR.13 Another common approach is titration of thiols according to UOP Method 163-89.14 The major drawback to these group type methods is the lack of structural information obtained. This report describes a method for separation of sulfides and thiols from hydrocarbons, thiophenes, and other weakly adsorbed compound types via rapid solidphase extraction (SPE). Pentafluorobenzoyl chloride (PFBC) is used to derivatize thiols, allowing their differentiation from sulfides, which do not react, during subsequent GC/MS analysis. An optional step in the method enables isolation of compound types strongly held by the SPE column. In most cases, this fraction will be composed largely of phenolic compounds. Experimental Section The overall separation scheme is outlined in Figure 1. Concentrates were prepared from distillates by SPE on alumina (5.0 g, Woelm no. 02087, ICN Biomedical, Aurora, OH), activated to Brockman activity IV (400 °C for 4 h) before use. Short glass columns (1.25 cm i.d. × 10 cm, with a 3 cm × 4 cm solvent reservoir) were plugged with silanized glass (8) Green, J. A.; Green, J. B.; Grigsby, R. D.; Pearson, C. D.; Reynolds, J. W.; Shay, J. Y.; Sturm, G., Jr.; Thomson, J. S.; Vogh, J. W.; Vrana, R. P.; Yu, S. K.-T.; Diehl, B. H.; Grizzle, P. L.; Hirsch, D. E.; Hornung, K. W.; Tang, S.-Y.; Carbognani, L.; Hazos, M.; Sanchez, V. Analysis of Heavy Oils: Method Development and Application to Cerro Negro Heavy Petroleum; Topical Report NIPER-452 (2), NTIS Report No. DE9000201; NTIS: Springfield, VA, 1989; Chapter 5B. (9) Albro, T. G.; Dreifuss, P. A.; Wormsbecher, R. F. J. High Resolut. Chromatogr. 1993, 16, 13-17. (10) McCormack, A. J.; Sudhakar, C.; Levine, S. A.; Patel, M. S.; McCann, J. M. LC-GC 1994, 12 (1), 31-35. (11) Bradley, C.; Schiller, D. J. Anal. Chem. 1986, 58, 3017-3021. (12) Di Sanzo, F. P.; Bray, W.; Chawia, B. J. High Resolut. Chromatogr. 1994, 17, 255-258. (13) Green, J. B.; Yu, S. K.-T.; Pearson, C. D.; Reynolds, J. W. Energy Fuels 1993, 7, 119-126. (14) UOP Method 163-89. UOP Laboratory Test Methods for Petroleum and Its Products; UOP Inc.: Des Plaines, IL, 1991; Vol. 1.
© 1997 American Chemical Society
910 Energy & Fuels, Vol. 11, No. 4, 1997
Thomson et al. Table 1. Response Factors of Sulfides and Derivatized Thiols Relative to Diphenylsulfide RT (min)
Figure 1. Solid-phase extraction (SPE) scheme used to separate thiophenes, sulfides, thiols, and other polar compounds in fuel distillates boiling below 650 °F. wool and filled with 2.5:97.5% (v/v) dichloromethane:pentane, and alumina was added with gentle tapping. Next, a layer of anhydrous, granular Na2SO4 (about 0.5 cm) was added to hold the alumina in place during subsequent steps. The eluant was drained to the top of the Na2SO4 layer, and 10 mL of distillate (or a sufficient quantity to contain up to 10 mg of thiols) was charged to the top of the column. Samples were then sequentially eluted under argon at a slight positive pressure using 20 mL portions of the following solvents: for fraction I, 2.5: 97.5% (v/v) dichloromethane:pentane (hydrocarbons, aromatics, and alkylthiophenes); for fraction II, 100% methyl tertbutyl ether (alkyl-, cycloalkyl-, and cyclic sulfides and alkylthiols); for fraction III, 10:90% (v/v) methanol:methyl tertbutyl ether (arylthiols and phenols). Solvents were removed via evaporation under argon. Sulfur contents were determined by microcoulometric titration using ASTM Method D 3120. Derivatizations of fraction II were performed in 5 mL heavy wall glass reaction vials (Supelco, Bellefonte, PA, no. 3-3299) fitted with Teflon caps (Supelco, no. 3-3303). Blends of either six to eight pure compounds (0.3 mg/compound) or petroleum concentrates (5-10 mg thiols) in dichloromethane or pentane were placed in vials and concentrated to about 0.5 mL. Diphenylsulfide (0.3 mg) was added as an internal standard. PFBC (0.5 mL, PCR, Inc., Gainesville, FL, no. 13670-5) and 1.0 mL cyclopentane were added to the vials, which were capped and held at room temperature for 20 min with periodic shaking. Samples were then extracted 3 times with 1 N NaOH to remove unreacted PFBC. The resulting organic layer was passed through anhydrous, granular Na2SO4 and evaporated under argon to a final volume of 0.3 mL. GC/MS analyses were performed using a Kratos (Ramsey, NJ) MS-80 mass spectrometer coupled to a Hewlett-Packard (Avondale, PA) Model 5890 GC equipped with a cool-on-column inlet and capillary direct interface. A Data General Nova 4-based DS-55 data system was used for data acquisition and analysis. Samples were injected onto a capillary column (Restek Corp., Bellefonte, PA, Rtx-1 fused silica, 105 m, 0.25 mm i.d., 0.5 µm film), held for 2 min at 30 °C, programmed at 20 °C/min to 70 °C, then programmed at 2 °C/min to 270 °C, and held for 10 min. Other instrument conditions were the following: GC/MS interface, 290 °C; helium column flow, 1 mL/min; column head pressure, 3.0 kg/cm2; 70 eV ionizing voltage; 1,000 resolution; scan rate, 0.5 s/decade; source pressure, 10-5 Torr; source temperature, 300 °C. In cases where analysis of fraction III (phenols, other polar types) was desired, it was performed using a previously described procedure for GC/MS analysis of phenols as their trifluoroacetate esters.15 The GC/MS conditions for this analysis were the same as specified above. (15) Green, J. B.; Yu, S. K.-T.; Vrana, R. P. J. High Resolut. Chromatogr. 1994, 17, 439-451.
avg RFa
PFBC Derivatives 1-methyl-1-propanethiol 2-methyl-3-pentanethiol 1-pentanethiol 3-heptanethiol 1-heptanethiol 3-methylbenzenethiol 2,4-dimethylbenzenethiol 1-octanethiol 1-nonanethiol 1-decanethiol
57.45 67.78 68.05 74.43 80.53 79.64 86.04 86.28 91.72 96.90
1.63 1.74 1.67 2.34 1.81 1.15 1.12 b 2.07 1.89
Other Compounds cyclohexanone 2,5-dimethylthiacyclopentane, trans 2,5-dimethylthiacyclopentane, cis 4-methylthiacyclohexane 2-ethyl-2-methylthiacyclopentane diphenylsulfide
23.60 23.80 24.05 28.15 31.72 74.05
0.57 0.58 0.58 0.58 0.55 1.00
a
5.1% average variation between duplicates. b Not determined.
Relative response factors (RRF) for sulfur compounds, both derivatized and unreacted, were obtained using diphenylsulfide as an internal standard. These were calculated as shown in eq 1 below:
RRF ) (AX/AIS)(WIS/WX)
(1)
where A ) area percent based on the GC/MS total ion current, W ) weight, X ) sulfide or thiol, and IS ) diphenyl sulfide internal standard. Typical RRF values are shown in Table 1. Concentrations of sulfur compounds (on a ppm distillate sulfur basis) were then calculated as shown in eq 2, below:
ppm sulfur ) (AX/AIS)(WIS/WD)(1/RRF)(WF S)(106) (2) where A ) area percent based on the GC/MS total ion current, W ) weight, X ) sulfide or thiol, IS ) diphenylsulfide internal standard, D ) distillate, RRF ) relative response factor, and WF S ) weight fraction of sulfur in each compound. The separation was developed using a 250-375 °F crude oil distillate known to contain thiols (198 ppm by titration using UOP Method 163-89) and cyclic sulfides. The distillate was produced from an Isthmus crude oil distilled using ASTM Method D 2892. The whole crude contained 1.33 wt % sulfur and had an as-received API gravity of 34.2. The 250-375 °F distillate contained 0.0498 wt % sulfur and had an API gravity of 52.1. The procedure was applied to a 650 °F) Merey (Venezuela) residue. To obtain the neutral fraction, the whole residue was passed through LC columns containing macroporous anion and cation resins to remove acidic and basic compounds.17
Results and Discussion Development of SPE Separation. The first sorbent evaluated, a silica gel packed in commercially available SPE cartridges, was rejected because of insufficient thiol retention. Other sorbents investigated included cesium-loaded silica and ion-exchange resins, (16) Green, J. B.; Zagula, E. J.; Reynolds, J. W.; Wandke, H. H.; Young, L. L.; Chew, H. Energy Fuels 1994, 8, 856-867. (17) Green, J. B.; Hoff, R. J.; Woodward, P. W.; Stevens, L. L. Fuel 1984, 63, 1290-1301.
Sulfides and Thiols
both of which were too retentive. Alumina proved to be the optimum sorbent, although when long residence times were encountered (i.e., with a long, thin column), some evidence of thiol oxidation to disulfides was observed. When thiols were separated using short columns, short residence times, and slight positive argon pressure, negligible disulfide formation occurred. The solvent chosen for elution of fraction I (2.5% dichloromethane in pentane) was sufficient to elute all saturated hydrocarbons, alkylbenzenes, and thiophenes. However, multiple-ring compounds such as naphthalene and benzothiophene were partially retained on the column. Although increasing the eluant strength just slightly (5.0% dichloromethane in pentane) was sufficient for complete elution of naphthalene, it also caused partial elution of sulfides. Given this tradeoff in separation behavior, the option of maximizing sulfide retention was chosen because the interference from low levels of multiple-ring hydrocarbons and thiophenes in the analysis of fraction II was minimal in actual practice. Following initial trials with dichloromethane, methyl tert-butyl ether was later substituted for elution of fraction II. Although cyclic sulfides were completely eluted using either solvent, alkylthiol elution was incomplete using dichloromethane. Although the nominal strength of either solvent should be comparable, the superior performance of methyl tert-butyl ether may be attributed to phenomona such as the secondary solvent effects described by Snyder.18 The solvent selected to elute fraction III, 10% methanol in methyl-tert-butyl ether, eluted arylthiols and phenols both in one fraction. Attempts to include both alkyl- and arylthiols in fraction II were unsuccessful, owing to the significantly greater retention of arylthiols on alumina. Similarly, the comparable retention of arylthiols and phenols prevented their separation on alumina. Since commingling of these types presents no special difficulty in GC/MS analysis, no further attempts were made to separate them. Optimization of Derivatization Conditions. Although thiol derivatives can be formed using trifluoroacetic anhydride,15 N-methyl-N-tert-butyldimethylsilyltrifluoroacetamide,19 and 4-fluorobenzoyl chloride,20 PFBC was selected for reasons stated below. Trifluoroacetylated thiols coeluted with sulfides and exhibited noncharacteristic mass spectra. The GC characteristics of 4-fluorobenzoyl derivatives were noticeably inferior to those of PFBC derivatives. PFBC derivatives of thiols were well separated from sulfides and their identification was aided by the formation of intense m/z 195 fragment ions in the EI/MS spectra. Figure 2 illustrates the GC group separation of sulfides and thiols for the derivatized Isthmus 250-375 °F fraction II concentrate. Naphthalene is the only aromatic hydrocarbon carried into fraction II in this boiling range. Prominent thiol homologues are identified in the figure caption. The derivatization procedure was optimized using a 321-338 °F thiol concentrate isolated from Wasson (Texas) crude oil during American Petroleum Institute (18) Snyder, L. R. Principles of Adsorption Chromatography; Marcel Decker: New York, 1968; pp 216-229. (19) Landrum, D. C.; Mawhinney, T. P. J. Chromatogr. 1989, 483, 21-32. (20) Spratt, M. P.; Dorn, H. C. Anal. Chem. 1984, 56, 2038-2043.
Energy & Fuels, Vol. 11, No. 4, 1997 911
Figure 2. Total ion current GC/MS chromatogram of fraction II from SPE separation of a derivatized Isthmus 230-375 °F crude oil distillate. Peaks identified or assigned to groups include (1) naphthalene, (2) 2-methylpropanethiol, (3) C5 alkylthiol, (4) C6 alkylthiols, (5) diphenylsulfide internal standard, (6) C7 alkylthiols, (7) cyclohexanethiol, (8) C7 cycloalkylthiol + C8 alkylthiol, (9) C7 cycloalkylthiols, and (10) C8 alkylthiol.
Figure 3. Total ion current GC/MS chromatograms (same instrumental conditions) of a highly purified Wasson crude oil 320-338 °F thiol concentrate before and after derivatization with PFBC.
Project 48. The isolation of this concentrate has been described elsewhere.1 This highly purified concentrate, which contained only C7 and C8 thiols, was a difficult sample to completely react. Reagent optimization, however, allowed essentially complete conversion of the Wasson concentrate at room temperature, as shown in Figure 3. Mass Spectral Fragmentation Patterns. This section will focus on the features of spectra of PFBC derivatives of thiols, since mass spectral information about underivatized sulfides, thiaindans, and thiols is available in the general literature. These features are summarized in Figure 4. The pentafluorobenzoyl fragment (m/z 195) is always the most intense ion present. Other fragment ions that occur in all spectra include [M-195]+, pentafluorophenyl (m/z 167), and [C5F3]+ (m/z 117). Assignments of the fluorinated fragment ions listed in Figure 4 were confirmed by high-resolution MS. Molecular ions are prevalent in spectra of arylthiol derivatives but absent or weak in those of cycloalkyland alkylthiols. Conversely, [M-60]•+ ions, corresponding to the loss of -COS, are found in the spectra of derivatives of n-alkylthiols and sometimes weakly in those of cycloalkylthiols but do not occur in arylthiol spectra. Low-intensity ions found in spectra of derivatives of cycloalkyl- and alkylthiols, but generally absent in those
912 Energy & Fuels, Vol. 11, No. 4, 1997
Thomson et al.
Figure 4. Relative intensities of characteristic mass spectral ions of thiols derivatized with PFBC.
of arylthiols, are the [M-228]•+ ions, corresponding to formation of an olefin following loss of -HSCOC6F5. This fragment is quite useful in the differentiation of cyclic and acyclic thiol derivatives, since [M-228]•+ fragments found in the spectra of cycloalkylthiol derivatives have masses 2 units less than those of the corresponding acyclic compounds. Examples of typical spectra of aryl-, cycloalkyl-, and alkylthiol derivatives are shown in Figure 5. As noted above, m/z 195 is the base fragment in all three spectra. The m/z 117 nominal mass ion in the spectrum in Figure 5c is comprised of both the C5F3+ and [M-195]+ fragments mentioned previously. Applications: Straight Run Petroleum Distillate. The compounds (or compound types) tentatively identified in fraction II from the Isthmus 250-375 °F distillate described above and their estimated concentrations are listed in Table 2. These values were calculated by using eq 2. As shown in the table, cyclic sulfides containing 5-10 carbon atoms/molecule are present in this distillate, with C8 thiacycloalkanes accounting for both the largest number of isomers (24) and the largest concentration (85 ppm as sulfide sulfur). cis- and trans-2,5-dimethyl- and 2-ethylthiacyclopentanes were identified by comparison of their spectra with authentic spectra from the NIST library. Although spectra of the other dimethylthiacyclopentanes were not available for comparison, the other six C6 thiacycloalkane isomers present are probably dimethylthiacyclopentanes rather than methylthiacyclohexanes, since authentic spectra of the 2- and 3-methylthiacyclohexanes were not a good match, and the presence of 4-methylthiacyclohexane was ruled out by the absence of an m/z 116 base peak in any of the fraction II spectra. It is interesting to note that although 2,5-di-nalkylthiacyclopentanes and 2,6-di-n-alkylthiacyclohexanes have been reported by other authors21,22 as the major cyclic sulfide isomers present in bitumens, the cis- and trans-2,5-dimethylthiacyclopentanes in this (21) Barakat, A. O.; Rullko¨tter, J. Energy Fuels 1994, 8, 1168-1174.
Figure 5. Mass spectra of PFBC derivatives of (A) 3-methylbenzenethiol, (B) trans-2-methylcyclopentanethiol, and (C) 1-hexanethiol.
distillate comprised only 25% of the total C6 thiacycloalkanes present. The portion of sulfides listed as “unidentified” comprise the broad hump under the individual peaks evident in Figure 2. No acyclic alkyl sulfides were detected in this distillate. In addition to cycloalkyl sulfides, alkylthiols containing four to eight carbon atoms are also present, with C7 alkylthiols comprising both the largest number of individual isomers (8) and the highest total concentration (28 ppm as thiol sulfur). The identities of 2-methyl1-propanethiol, 1-butanethiol, and 1-pentanethiol were confirmed by a comparison of their spectra with those of authentic compounds analyzed under similar conditions. C6-C8 cycloalkylthiols are also present, with the (22) Sinninghe Damste´, J. S.; Rupstra, W. I. C.; Kock-Van Dalen, A. C.; De Leeuw, J. W.; Schenck, P. S. Geochim. Cosmochim. Acta 1989, 53, 1343-1355.
Sulfides and Thiols
Energy & Fuels, Vol. 11, No. 4, 1997 913
Table 2. Concentration of Sulfides and Thiols in Fraction II from an Isthmus 250-375 °F Distillate concentrationa (ppm S)
no. of isomers
Sulfides total C5 thiacycloalkanes 2-methylthiacyclopentane total C6 thiacycloalkanes trans-2,5-dimethylthiacyclopentane cis-2,5-dimethylthiacyclopentane 2-ethylthiacyclopentane total C7 thiacycloalkanes total C8 thiacycloalkanes total C9 thiacycloalkanes total C10 thiacycloalkanes total unidentified sulfides total sulfides
1.6 1.6 12.2 2.0 2.0 3.0 70.7 85.2 20.0 0.4 61.9 252.0
1
Thiols total C4 alkylthiol isomers 2-methyl-1-propanethiol 1-butanethiol total C5 alkylthiol isomers 1-pentanethiol total C6 alkylthiol isomers total C7 alkylthiol isomers total C8 alkylthiols total C6 cycloalkylthiol isomers cyclohexanethiol total C7 cycloalkylthiol isomers total C8 cycloalkylthiols total unidentified thiols total thiols
5.1 3.5 0.6 11.5 1.8 18.9 27.8 4.1 19.1 19.1 23.0 4.3 45.0 159.6
a
9
13 24 17 1
3 3 3 8 4 1 6 5
ppm S (distillate basis)
a
liquid product fraction I (thiophenes) fraction II (1-thiaindans) 1-thiaindan C1 1-isomers (6) C2 1-isomers (7) C3 1-isomers (9) C4 1-isomers (2) fraction II (non-thiaindan sulfides + thiols) a
Table 3. Comparison of GC/MS and Conventional Sulfur Data for Isthmus 250-375 °F Whole Distillate and Fractions I-II GC/MS
conventional
431a
530b 17.6b
252 160
ppm S (liquid product basis) GC/MS
Whole distillate basis.
whole distillate fraction I (thiophenes) fraction II (sulfides) fraction II (thiols)
Table 4. Comparison of GC/MS and Conventional Sulfur Data for a Merey Catalytic Cracked Liquid Product and Fractions I-II
198c b
Total includes thiophenes by microcoulometry. Microcoulometric titration. c UOP Method 163-89.
cyclohexanethiol derivative comprising the single largest thiol component (19 ppm). Its identity was also confirmed via comparison with the spectrum of the authentic compound. As shown in Table 3, the sulfur content for the whole distillate is estimated by adding the fraction II (sulfides and thiols) values obtained via the SPE/GC/MS procedure to the fraction I (thiophenes) values obtained by conventional means. Approximately 83% of the whole distillate sulfur (431 ppm) is accounted for by using this procedure, compared to the value obtained by conventional analysis (530 ppm). This value is acceptable considering the probable cumulative error from SPE, solvent evaporation, peak integration, sulfur analyses, and other factors. Also, it was necessary to use estimated response factors for calculation of a portion of the data listed in Tables 2 and 3, since many of the sulfur compound types were unavailable for measurement of response. FCC Liquid Product. The Merey (Venezuela) FCC liquid product was selected for further analysis in this work because of the high levels of sulfur in the feed (2.63 wt %) as well as the liquid product (0.87 wt %). As shown in Table 4, essentially all of the sulfur in this
conventional 8700a 8920a
232 22.0 84.2 65.2 54.4 6.5 ndb
Microcoulometric titration. b nd ) not detected.
product was thiophenic (102 wt % of the sulfur present in the total liquid product). The boiling point composition of the liquid product was gasoline (56 wt %), light cycle oil (25 wt %), and heavy cycle oil (19 wt %) on a gas-corrected basis. A breakdown of sulfur over these boiling ranges was not obtained, but prior work has shown that the bulk of the sulfur in FCC products is present in higher boiling ranges as benzothiophene, dibenzothiophene, and more aromatic thiophenic forms.23-25 A series of alkyl-1-thiaindans (about 2.5 wt % of the total sulfur in the liquid product) was tentatively identified in fraction II. Figure 6 shows the portion of the GC/MS total ion current chromatographic trace containing these compounds. The first compound in the series, 1-thiaindan, was identified by comparison of its GC/MS total ion current spectrum with an authentic spectrum from the NIST library. As shown in Table 4, alkyl-1-thiaindans with alkyl substituents of up to four carbons were found, with methyl-1-thiaindan isomers present in the largest concentration (84.1 ppm S) and C3 1-thiaindans exhibiting the largest number of isomers (9). The concentration of the series was estimated assuming a response similar to that of the diphenylsulfide internal standard. It is interesting to speculate on the origin of the alkyl1-thiaindans. Since the catalytic cracking process is a
Figure 6. Total ion current GC/MS chromatogram of fraction II from SPE separation of a derivatized liquid product from a bench scale catalytic cracker unit (Merey, Venezuela >650 °F neutrals feed). (23) Campagna, R. J.; Krishna, A. S.; Yanik, S. J. Oil Gas J. 1983, October 31, 128-134. (24) Huling, G. P.; McKinney, J. D.; Readal, T. C. Oil Gas J. 1975, May 19, 73-79. (25) Wollaston, E. G.; Forsythe, W. L.; Vasalos, I. A. Oil Gas J. 1971, August 2, 64-69.
914 Energy & Fuels, Vol. 11, No. 4, 1997
net producer of hydrogen,26 hydrogenation of benzothiophenes is unlikely as a source of these compounds. Their formation probably occurs via capture of H2S by alkylbenzenes, and during subsequent dehydrogenation steps, they could serve as an additional source of benzothiophenes. Although recent publications have reported the presence of low molecular weight thiols in FCC streams,9,12 it was expected that thiols boiling below butanethiol would not be detected in this sample. Gas/liquid separation in the in-house FCC unit is accomplished via two liquid traps.16 Owing to their limited efficiency, the majority of material in the C5 and C6 boiling range is carried over into the gas collection system. Although the overall liquid yields are corrected for this carryover, the liquid products actually collected are deficient in lower boiling components. For this reason, any C3 and C4 alkylthiols would likely be carried over into the gas sample and thus escape detection in a subsequent analysis of the liquid sample. It is interesting to note, however, that although the method can detect higher molecular weight thiols in the ppm range, none were observed in this sample. (26) Green, J. B.; Zagula, E. J.; Reynolds, J. W.; Young, L. L.; McWilliams, T. B.; Green, J. A. Energy Fuels 1997, 11, 46-60.
Thomson et al.
Limitations of the Method. A number of higher molecular weight polyaromatic hydrocarbons (from fluorene through alkylpyrenes) and dibenzothiophenes were present in fraction II from the higher boiling FCC product described above. Since alumina is well-known for its ability to retain higher ring-number aromatic compounds, overlap of g2-ring aromatic hydrocarbons and thiophenes with thiaindans and other sulfides plus thiols will occur during the SPE separation conditions described in this method. This overlap is noticeable in analyses of distillates boiling above 450 °F and will become severe for those boiling above 650 °F. For this reason, the method is most useful for the analysis of sulfur compound types in distillates boiling below 650 °F. Acknowledgment. The authors thank Suzie Hamilton-Gassett, Eddie Suggs, and Judy A. Green for sulfur analyses performed in conjunction with this work, and Dr. Gene P. Sturm, Jr. for technical review of the manuscript. Funding provided by the Department of Energy under Contract No. DE-AC22-94PC91008 was essential to the completion of this work. EF960165M