Aromatics

Jun 21, 2011 - Division of Mass Spectrometry Research, Korea Basic Science Institute, Ochang, 863-883, Korea ... double-bond equivalence (DBE) values ...
7 downloads 0 Views 2MB Size
ARTICLE pubs.acs.org/ac

Planar Limit-Assisted Structural Interpretation of Saturates/ Aromatics/Resins/Asphaltenes Fractionated Crude Oil Compounds Observed by Fourier Transform Ion Cyclotron Resonance Mass Spectrometry Yunju Cho,† Young Hwan Kim,‡,^ and Sunghwan Kim*,†,§ †

Kyungpook National University, Department of Chemistry, Daegu, 702-701, Korea Division of Mass Spectrometry Research, Korea Basic Science Institute, Ochang, 863-883, Korea ^ Graduate School of Analytical Science and Technology, Chungnam National University, Daejeon, 305-764, Korea § Green-Nano Materials Research Center, Daegu, 702-701, Korea ‡

ABSTRACT: Planar limits, defined as lines generated by connecting maximum double-bond equivalence (DBE) values at given carbon numbers, are proposed as a means of predicting and understanding the molecular structure of compounds in crude oil. The slopes and y-intercepts of the lines are determined by the DBE/carbon number ratios of functional groups defining the planar limits. For example, the planar limit generated by a serial addition of saturated cyclic rings has a slope of 0.25. The planar limit formed by the linear and nonlinear addition of benzene rings yields lines with slopes of 0.75 and 1, respectively. The y-intercepts of these lines were determined by additional functional groups added within a series of molecules. Plots of DBE versus carbon number for S1 class compounds observed by Fourier transform ion cyclotron resonance mass spectrometry (FTICR MS) showed that saturates/aromatics/resins/asphaltenes (SARA) fractions exhibited unique slopes and y-intercepts. The slope of the planar limit observed from a saturates fraction matched well with the slope of a planar limit generated by the serial addition of saturated cyclic rings. The slopes of planar limits of aromatics and resins fractions were very similar to that obtained from the linear addition of benzene rings. Finally, the slope of the asphaltenes fraction was almost identical to the slope obtained from the nonlinear addition of benzene rings. Simulated and experimental data show that SARA fractions exhibit different molecular structure characteristics. On the basis of the slope and y-intercept of the planar limit, the structures of molecules in SARA fractions were predicted and suggested. The use of planar limits for structural interpretation is not limited to crude oil compounds but can also be used to study other organic mixtures such as humic substances or metabolites.

A

s the world’s remaining deposits of crude oil become heavier, understanding the various chemical compositions of heavy crude oil is ever more important. In particular, structural elucidation of heavy components in crude oil can help us understand and predict the behavior of petroleum during processing.1,2 For example, such knowledge would be helpful in designing catalysts for oil processing and troubleshooting. However, structural elucidation remains difficult due to the limitations of current analytical techniques. Various techniques including NMR and X-ray diffraction (XRD) have been used to determine the molecular structure of compounds.35 Those techniques, however, provide only an average molecular structure, and studying the molecular structures of the myriad of individual compounds existing in crude oil is very difficult. Currently, Fourier transform ion cyclotron resonance mass spectrometry (FTICR MS) is one of the powerful techniques that can cope with the complexity of crude oils. Because of its high mass accuracy and ultrahigh resolution, the elemental composition of chemical compounds comprising crude oil can be calculated.7 FTICR MS has contributed greatly to our knowledge r 2011 American Chemical Society

concerning the heavy components of crude oil.2,816 Previous studies showed that DBE distributions observed by FTICR MS spectra can be used to elucidate chemical structures of crude oil and humic compounds.17,18 In another report, a “planar aromatic limit” was used to study structural variations in asphaltene samples.14,19 The planar limit was defined as the line observed in a plot of DBE versus carbon number obtained from FTICR MS data. The limit is determined by the maximum DBE values with a given number of carbon atoms. In this report, the concept of “planar limit” was expanded to include saturated cyclic and aromatic compounds, and the theoretical background for the planar limit was investigated. The slopes and intercepts of planar limits were found to vary depending on the structural features of crude oil compounds, and the planar limit could be used to understand and predict the structures of Received: May 6, 2011 Accepted: June 21, 2011 Published: June 21, 2011 6068

dx.doi.org/10.1021/ac2011685 | Anal. Chem. 2011, 83, 6068–6073

Analytical Chemistry

ARTICLE

these compounds. The concept was demonstrated with conventional saturates/aromatics/resins/asphaltenes (SARA) fractionated crude oil. The planar limits existed in each fraction and the structural features of each fraction were predicted based on the slopes and y-intercepts of the planar limits.

’ EXPERIMENTAL SECTION SARA Fractionation of Crude Oil. Arabian medium heavy oil was used in this study. The crude oils were not processed prior to fractionation. SARA fractionation was performed according to the procedure described in a previous study.20 Briefly, 0.5 g of crude oil was dissolved in 20 mL n-heptane (Burdick & Jackson ACS/HPLC) and the resulting solution was stirred for 1 h and stored overnight in the dark. The solution was filtered using a 0.7 μm pore size glass filter (GFF, Whatman, Maidstone, Kent, U.K.). The precipitate was washed with n-heptane until the solvent was colorless. The heptane-insoluble solids were isolated as the asphaltenes fraction. The asphaltenes constituted 9.32 ( 1.82% of the original crude. An airstream was directed onto the heptane-soluble fraction (maltene) to evaporate the heptane. About 0.06 g of the maltene was adsorbed onto 3 g of activated alumina (80-200 mesh, Fisher Scientific, Fairlawn, NJ) and loaded onto 25 g of neutral alumina in a 22 mm  400 mm column. The alumina was baked at 450 C for 3 h before use. The saturates fraction (52.19 ( 2.44%) was obtained by elution with 300 mL of hexane, the aromatics fraction (33.94 ( 3.28%) was eluted with 150 mL of toluene, and the resins fraction (16.14 ( 6.14%) was isolated by elution with 150 mL of 80:20 (v/v) toluenemethanol. Mass Spectrometry. Samples were prepared by diluting the oil samples to a concentration of 0.5 mg/mL in HPLC-grade toluene (Merck, Gibbstown, NJ). The prepared samples were directly injected with a syringe pump (Harvard, Holliston, MA) at a flow rate of 500 μL/h for atmospheric pressure photoionization (APPI) positive-mode analyses. Analyses were carried out with a 15 T FTICR mass spectrometer at the Korean Basic Science Institute (KBSI, Ochang-eup, Korea). APPI sources were purchased from Bruker Daltonics (Billerica, MA). Nitrogen was used as the drying and nebulizing gas. For the saturates, aromatics, and resins fractions, a nebulizing temperature of 350 C was used at a flow rate of 2.0 L/min. Higher temperatures of up to 400 C were required to efficiently ionize the asphaltenes fraction. The drying gas temperature was 210 C at a flow rate of 2.3 L/min and a spray voltage of 25003500 V; the skimmer voltage was set to 15.0 V to minimize in-source fragmentation. Each sample was analyzed three times to check reproducibility. The data were reproducible within a few percent of error. Then 4  106 data points were obtained for each spectrum to achieve high resolution and mass accuracy. Spectral Interpretation. Spectral interpretation was performed with Statistical Tool for Organic Mixtures’ Spectra (STORMS 1.0) software with an automated peak-picking algorithm for more reliable and faster results.21 The peak list obtained was first calibrated using known m/z values for the ESI tuning mix (Agilent, Santa Clara, CA). The m/z values of the most abundant peaks were listed and used to calibrate the spectra obtained without added standards. Typically, homologous series of oxygenated compounds were used as internal calibrants. Elemental formulas were calculated from the calibrated peak list and assigned based on m/z values within a 1 ppm error range. Normal conditions for petroleum data (CcHhNnOoSs, c unlimited,

Figure 1. A simulation of planar limits on a plot of DBE versus carbon number was generated by adding a series of (a) saturated cyclic rings, benzene rings either (b) linearly or (c) nonlinearly.

Figure 2. A simulation of planar limits was generated by adding CH3 functional groups to a series of saturated cyclic ring additions (a). The resulting DBE versus carbon number plot is shown in part b.

h unlimited, 0 e n e 5, 0 e o e 5, 0 e s e 4) were used for these calculations. Double-bond equivalences (DBE) represent the number of rings plus the number of double bonds in a given molecular formula. DBE values can be calculated by the following equation: DBE ¼ c  h=2 + n=2 + 1 ðfor elemental formulae of Cc Hh Nn Oo Ss Þ

ð1Þ

’ RESULTS AND DISCUSSION Simulation of Planar Limits for Proof-of-Concept. Figure 1 shows plausible progressive changes in the molecular structures of S1 class compounds typically found in crude oils. A series of saturated cyclic rings were added to thiophene in Figure 1a. Each molecule shown in Figure 1 can be represented by a dot in a Cartesian coordinate system having carbon number and DBE values as the abscissa and ordinate, respectively, and lines can be drawn to connect the dots for each series. The slope and intercepts of these lines were calculated by linear regression, and the values 6069

dx.doi.org/10.1021/ac2011685 |Anal. Chem. 2011, 83, 6068–6073

Analytical Chemistry

ARTICLE

Figure 3. (a) DBE versus carbon number plots were generated from S1 class compounds observed in (1) saturates, (2) aromatics, (3) resins, and (4) asphaltenes fractions. The coordination of molecules existing at the planar limits is shown in part b. The slopes and y-intercepts of the lines are listed.

are shown in Figure 1d-1. The slope of these lines was 0.25 because the carbon number was increased by 4 while the DBE was increased by 1 following the addition of a saturated cyclic ring. Therefore, the planar limit determined by the serial addition of saturated cyclic rings will have a slope of 0.25 in a plot of DBE versus carbon number. Figure 1b,c shows cases in which planar limits are determined by the serial addition of aromatic rings. The difference between these two figures depends on whether an aromatic ring is linearly or nonlinearly added (nonlinear addition is a fusion between existing aromatic rings). When aromatic rings are linearly added to an existing structure, the carbon number is increased by 4 and the DBE is increased by 3 for each addition. Therefore, the ratio of change represented by the slope of the line in Figure 1d-2 is

0.75 (3/4). Accordingly, the carbon number is increased by 2 and the DBE by 2 for each nonlinear addition of an aromatic ring to an existing structure, and the ratio of change is 1 (2/2) (refer to Figure 1d-3). Therefore, the slope of the planar limits shown in Figure 1 is determined by the DBE versus carbon number ratio of functional groups that are serially added to a given initial structure. The addition of functional groups to all of the molecular structures in a given series changes the y-intercept, not the slope. For example, in Figure 2a, the addition of a methyl group to the starting structure, marked by a red-dotted rectangle, shifted the carbon number and hence reduced the y-intercept of the line from 2 to 1.75 (compare line 1 in Figure 1d with line 1 in Figure 2b). Therefore, intercept values were systematically altered by adding functional groups to the starting structure within a given 6070

dx.doi.org/10.1021/ac2011685 |Anal. Chem. 2011, 83, 6068–6073

Analytical Chemistry

ARTICLE

Figure 4. Proposed molecular structures of S1 class compounds found in (a) saturates, (b) aromatics, and (c) asphaltenes fractions.

series. An empirical equation to calculate y-intercepts for the lines in Figures 1 and 2 is displayed below   ΔDBE y intercept ¼ k  ΔC + +m ð2Þ k where ΔC and ΔDBE represent the difference in carbon number and DBE between starting forms of a modified structure prior to the addition of functional groups; starting forms are indicated by red rectangular boxes in Figures 1 and 2. Note that for saturated cyclic ring additions, k = 0.25 and m = 2. For linear benzene ring additions, k = 0.75 and m = 0, and for nonlinear benzene ring additions, k = 1 and m = 5. For example, the carbon number and DBE difference between thiophene (unmodified starting molecule) and methylthiophene

(modified by the addition of a methyl group) are +1 and 0, respectively. For a saturated cyclic ring addition, k = 0.25 and m = 2. Therefore, the y-intercept of the series starting from methylthiophene can be calculated using eq 1 to yield a value of 1.75, which agrees with the value shown in Figure 2b. Observation and Interpretation of Planar Limits in Real Samples. DBE versus carbon number distributions of S1 class compounds observed in SARA fractions of Arabian heavy crude oil are shown in Figure 3a. Each fraction exhibits a distinctive DBE versus carbon number distribution. For example, the S1 class compounds in the asphaltenes fraction had DBE values that fell mostly between 24 and 40. However, S1 class compounds in the saturates fraction had DBE values distributed between 0 and 18, which agrees with well-known properties of SARA fractions. 6071

dx.doi.org/10.1021/ac2011685 |Anal. Chem. 2011, 83, 6068–6073

Analytical Chemistry The saturates fraction is known to contain mostly saturated compounds (as the name indicates), which agrees with the relatively low DBE distribution in Figure 3a. In contrast, the asphaltenes fraction is known to contain polycondensed aromatic compounds as evidenced by the higher DBE distribution observed in Figure 3a. Another noticeable difference observed in each SARA fraction is in the slopes and y-intercepts of lines formed by compounds with the highest DBE values within a given carbon number. Red lines are drawn in Figure 3a to visualize the planar limits. The term “planar aromatic limit” was previously used to describe the line observed in the asphaltene sample.14 However, the data in this study clearly show that planar limits are observed in other fractions. Therefore, the terms planar saturated limit (PSL) for the saturates fraction, planar aromatic limit (PAL) for the aromatics fraction, planar resin limit (PRL) for the resins fraction, and planar asphaltene limit (PASL) for the asphaltenes fraction would be preferable. The dots forming each line were selected and moved into Cartesian coordinates with carbon number and DBE values as the abscissa and ordinate, respectively (Figure 3b). Linear regression was used to calculate the slope and y-intercepts of the planar limits, and the values are presented in Figure 3a. Each planar limit exhibits its own unique slope. The PSL had a slope of ∼0.28, the PAL and PRL had one of ∼0.73, and the PASL had a slope of ∼0.95. Fractions containing molecules with more condensed molecular structures exhibited higher slopes, implying that the slope is closely related to the molecular structure. The slope of PSL (0.28) agrees well with the slope (0.25) calculated from the addition of a saturated cyclic ring in Figure 1a. Therefore, the planar limit of PSL can be attributed to a serial addition of cyclic ring structures. The slopes of the PAL and PRL (0.73 and 0.69, respectively) match well with that calculated from the serial linear addition of aromatic rings (0.75; refer to Figure 1b). Accordingly, PAL and PRL were generated by the linear addition of benzene rings. The slope of the PASL (0.95) is very close to unity in Figure 1c. Therefore, the PASL represents a serial, nonlinear addition of benzene rings. Prediction of Molecular Structures Based on Planar Limits. The simulated and experimental data in Figures 13 show that the slope and y-intercept of a planar limit are closely related to the molecular structures existing within each fraction. The proposed chemical structures of S1 class compounds in (a) the saturates fraction, (b) resins and aromatics fractions, and (c) asphaltenes fractions are shown in Figure 4. The structures follow the trends in slope and y-intercept shown in Figure 3. All of the elemental formulas in Figure 4 were identified in spectra obtained from SARA fractions. The saturates fraction contains a series of chemical compounds separated by saturated cyclic rings (Figure 4a). The aromatics and asphaltenes fractions contain series of molecules separated by benzene rings linearly or nonlinearly (Figure 4b,c, respectively) added. The suggested structures of asphaltene compounds are very similar to the one proposed in a previous study.3 Note that the structures shown in Figure 4 are intended to illustrate the observed trend. Isomers of the suggested structures are possible, and the exact molecular structures can be more complex. The maximum number of carbon atoms observed for saturated S1 class compounds with DBE values of 12 in Figure 3a-1 is approximately 90. The compound with a DBE at the planar limit comprises about 30 carbon atoms contained in cyclic rings (refer to the structures in Figure 4a). Therefore, the saturated compounds

ARTICLE

most likely consisted of alkyl chains containing up to ∼60 carbon atoms. In the aromatics fraction, the compound with a DBE value of 12 has a maximum number of carbon atoms around 70 (refer to Figure 3a-2). The compound with a DBE at the planar limit has about 20 carbon atoms contained in benzene rings (refer to the structures in Figure 4b). Calculations indicate that alkyl chains with up to 50 carbon atoms are attached to the aromatic compounds. The resin S1 class compounds contained up to ∼50 carbon atoms (refer to Figure 3a-3). The planar limit of the resin compounds was similar to that of the aromatics fraction. Therefore, the maximum number of carbon atoms in the alkyl chain of the resin compounds is approximately 30. For the asphaltenes fraction, the S1 class compounds with DBE values of 25 contain a maximum number of carbon atoms up to ∼60 (refer to Figure 3a-4). The S1 compound at the planar limit has around 40 carbon atoms (refer to Figure 4c). The number of carbon atoms in the alkyl chains of asphaltene compounds was calculated to be around 20. Therefore, each fraction contains compounds with different alkyl chain lengths. The length of the average alkyl chain decreases as saturates > aromatics > resins > asphaltenes.

’ CONCLUSIONS Planar limits defined by maximum DBE values at a given carbon number are proposed as a means of predicting and understanding the molecular structure of compounds in crude oil. Classes of compounds differing by certain functional groups can be drawn as a line on a plot of DBE versus carbon number. The slope and y-intercept of this line depends on the DBE/ carbon number ratio of the functional group. The concept of planar limits was demonstrated by simulation and exemplified using SARA fractions of crude oil. The molecular structures of compounds existing in each fraction were predicted based on the slopes and y-intercepts of the planar limits. In this study, S1 class compounds were used as an example to demonstrate the usefulness of the planar limit. However, the planar limit works well with other classes and it is not limited to S1 class compounds. We believe that the application of planar limits for structural interpretation is not limited to crude oil compounds but can also be applied to study other organic mixtures such as humic substances or metabolites. Research is currently under way to explore wider applications of this concept for structural elucidation. ’ AUTHOR INFORMATION Corresponding Author

*Phone: 82-53-950-5333. Fax: 82-53-950-6330. E-mail: sunghwank@ knu.ac.kr.

’ ACKNOWLEDGMENT The authors thank the Korea Basic Science Institute for instrument time of 15T FT-ICR MS instrument. This work was supported by the Industrial Strategic technology development program (10038662, MALDI-TOF for the diagnosis of BRCA mutation and genitourinary infection pathogen) funded by the Ministry of Knowledge Economy (MKE, Korea) and by the Korea Science and Engineering Foundation (KOSEF) grant funded by the Korean government (MEST) (20110003796). 6072

dx.doi.org/10.1021/ac2011685 |Anal. Chem. 2011, 83, 6068–6073

Analytical Chemistry

ARTICLE

’ REFERENCES (1) Marshall, A. G.; Rodgers, R. P. Acc. Chem. Res. 2004, 37, 53–59. (2) Hur, M.; Yeo, I.; Kim, E.; No, M. H.; Koh, J.; Cho, Y. J.; Lee, J. W.; Kim, S. Energy Fuels 2010, 24, 5524–5532. (3) Groenzin, H.; Mullins, O. C. Energy Fuels 2000, 14, 677–684. (4) Michael, G.; Al-Siri, M.; Khan, Z. H.; Ali, F. A. Energy Fuels 2005, 19, 1598–1605. (5) Ali, F. A.; Ghaloum, N.; Hauser, A. Energy Fuels 2006, 20, 231–238. (6) Comisarow, M. B.; Marshall, A. G. Chem. Phys. Lett. 1974, 26, 489–490. (7) Kim, S.; Rodgers, R. P.; Marshall, A. G. Int. J. Mass Spectrom. 2006, 251, 260–265. (8) Pinkston, D. S.; Duan, P.; Gallardo, V. A.; Habicht, S. C.; Tan, X.; Qian, K.; Gray, M.; Mllen, K.; Kenttmaa, H. I. Energy Fuels 2009, 23, 5564–5570. (9) Bae, E.; Na, J.-G.; Chung, S. H.; Kim, H.; Kim, S. Energy Fuels 2010, 24, 2563–2569. (10) Borton, D.; Pinkston, D. S.; Hurt, M. R.; Tan, X.; Azyat, K.; Scherer, A.; Tykwinski, R.; Gray, M.; Qian, K.; Kenttmaa, H. I. Energy Fuels 2010, 24, 5548–5559. (11) Habicht, S. C.; Amundson, L. M.; Duan, P. G.; Vinueza, N. R.; Kenttamaa, H. I. Anal. Chem. 2010, 82, 608–614. (12) Hur, M.; Yeo, I.; Park, E.; Kim, Y. H.; Yoo, J.; Kim, E.; No, M. H.; Koh, J.; Kim, S. Anal. Chem. 2010, 82, 211–218. (13) Pinkston, D. S.; Duan, P. G.; Gallardo, V. A.; Fu, M. K.; Habicht, S. C.; Kenttamaa, H. I. Energy Fuels 2010, 24, 3119–3124. (14) Purcell, J. M.; Merdrignac, I.; Rodgers, R. P.; Marshall, A. G.; Gauthier, T.; Guibard, I. Energy Fuels 2010, 24, 2257–2265. (15) Yeo, I.; Lee, J. W.; Kim, S. Bull. Korean Chem. Soc. 2010, 31, 3151–3155. (16) Zhu, X.; Shi, Q.; Zhang, Y.; Pan, N.; Xu, C.; Chung, K. H.; Zhao, S. Energy Fuels 2011, 25, 281–287. (17) Kim, Y. H.; Kim, S. J. Am. Soc. Mass Spectrom. 2010, 21, 386–392. (18) Bae, E.; Yeo, I. J.; Jeong, B.; Shin, Y.; Shin, K.-H.; Kim, S. Anal. Chem. 2011, 83, 4193–4199. (19) Hsu, C. S.; Lobodin, V. V.; Rodgers, R. P.; McKenna, A. M.; Marshall, A. G. Energy Fuels 2011, 25, 2174–2178. (20) Klein, G. C.; Angstrom, A.; Rodgers, R. P.; Marshall, A. G. Energy Fuels 2006, 20, 668–672. (21) Hur, M.; Oh, H. B.; Kim, S. Bull. Korean Chem. Soc. 2009, 30, 2665–2668.

6073

dx.doi.org/10.1021/ac2011685 |Anal. Chem. 2011, 83, 6068–6073