Petroleum crude oil is a complex hydrocarbon mixture in which molecular aggregates ... formed by heavier components mainly constituting asphaltenes. During ...
Petroleum Science and Technology, 25:121-139, 2007 Copyright © Taylor & Francis Group, LLC ISSN: 1091-6466 print/1532-2459 online DO!: 10. 1080110916460601054263
Taylor & Francis Taylor & Francis Group
Structural Characterization of Asphaltenes and Ethyl Acetate Insoluble Fractions of Petroleum Vacuum Residues B. K. Sharma Department of Chemical Engineering, Pennsylvania State University, University Park, Pennsylvania, USA
C. D. Sharma, 0. S. Tyagi, and S. D. Bhagat Analytical Sciences Division, Indian Institute of Petroleum, Dehra Dun, India
S. Z. Erhan Food and Industrial Oil, NCAURIUSDA/ARS, Peoria, Illinois, USA
Abstract: Asphaltenes and insoluble fractions of vacuum residues (VRs) of two Indian crude oils (viz. Heera and Jodhpur) of different specific gravity were obtained by precipitation of VRs in n-hexane, n-heptane, and ethyl acetate, and also ' by subsequent reprecipitation of n-heptane and ethyl acetate soluble fractions by n-pentane. The effect of various solvents on average molecular structure of asphaltenes and insolubles was studied using nuclear magnetic resonance spectroscopy (NMR), Fourier transform infrared spectroscopy (FTIR), and size exclusion chromatography (SEC). The asphaltenes and insolubles of Jodhpur YR have higher amounts of high molecular weight species with a high concentration of condensed and substituted aromatic rings, branched and/or short alkyl side chains, oxygen and nitrogen functionalities, compared to that of l-Ieera y R. Ethyl acetate insolubles comprise a higher number of substituted aromatic structures, branched aliphatic structures, complex average unit structures, nitrogen and oxygen functional ities, and high molecular weight (MW) species as compared to hexane and heptane asphaltenes. Heptane insolubles consist of more naphthenic rings condensed with aromatic rings than C6A and EAI. Keywords: asphaltenes, FTIR, insolubles, NMR, petroleum, SEC, vacuum residue Names are necessary to report factually on available data; however, the USDA neither guarantees nor warrants the standard of the product, and the use of the name by USDA implies no approval of the product to the exclusion of others that may also be suitable. Address correspondence to Bràjendra K. Sharma, Food and Industrial Oil, NCAUR/USDAJARS 1815 N. University Street, Peoria, IL 61604. E-mail: sharmab@ ncaur.usda.gov 121
122
B. K. Sharma et at.
INTRODUCTION Petroleum crude oil is a complex hydrocarbon mixture in which molecular aggregates and/or micelles of different sizes and compositions are usually formed by heavier components mainly constituting asphaltenes. During the last decade, asphaltenes have drawn considerable attention due to problems caused by their detrimental effects in both production and refining of crude oils due to their viscous and flocculating nature (Creek, 2005; Sirota, 2005). The asphaltene precipitates settle to the bottom of crude oil leading to excessive sedimentation during refinery operations. This sedimentation reduces the permeability of oil reservoirs, resulting in the reduction of oil production. To understand this precipitation a number of studies are reported on asphaltene solubility and precipitation using a variety of diluent composition (Buenrostro-Gonzalez et al., 2002; Hong and Watkinson, 2004; Angle et al., 2006). In view of such severe implications, it is expedient to develop an understanding of their chemical composition (Michael et al., 2005). For this purpose, a consortium of analytical techniques is required, such as elemental analysis (Ali and Saleem, 1994), spectroscopic (Ali et al., 1990, 2006; Calenima et al., 1995; Yen, 1984a, 1984b; Zajac et al., 1994), and chromatographic techniques (Sharma, 1999). Bunger and Li (1981) compiled structural information on asphaltenes derived from their elemental and spectroscopic (nuclear magnetic resonance spectroscopy (NMR), infrared spectroscopy.(IR), mass spectroscopy (MS), and x-ray) analysis and gentle degradation followed by analysis of their fragments. In a review, Speight (1994) provided details of the types of molecular species identified in the asphaltene matrix of petroleum crude oil. Andersen (1994) fractionated asphaltenes into soluble and insoluble parts using solvent mixtures having varying solubility power and characterized the separated fractions by ultraviolet-visible spectroscopy (UV-VIS), IR, size exclusion chromatography (SEC), vapor pressure osmometer (VPO), and elemental analysis. In another similar study, the asphaltenes were separated by solubility in polar and nonpolar precipitating solvent and were characterized using SEC, elemental analysis, Fourier transform infrared spectroscopy (F1'IR), 'H NMR, and synchronous fluorescence spectroscopy (BuenrostroGonzalez et al., 2002). Chemical derivatization along with suitable analytical techniques has been used in a study by Juyal et al. (2005) to study the influence of asphaltenes heteroatom groups on molecular interactions within asphaltenes. Vapor pressure osmometery gives inflated values of molecular weight (MW) when toluene and other nonpolar solvents are used, but considerably lower values when polar solvents like pyridine and nitrobenzene are used. The optimum conditions for obtaining true molecular weight of insolubles free from aggregation and other interference are not established yet. Other methods for MW determination, like field-ionization mass spectroscopy (FIMS), field desorption mass spectroscopy (FDMS), laser-desorpton ionization-timeof-flight MS, and small angle neutron scattering (SANS) (Sirota, 2005) have
Structural Characterization of Asphaltenes
123
limited use due to several reasons (Acevedo et al., 2005). Physical methods, viz, ultra-centrifuge, osmometry, ebullioscopy, viscosity, cryoscopy, and light scattering techniques yield unusually high molecular weights of insolubles. Although SEC suffers due to aggregation and calibration problems, it has an advantage. In principle, it gives the molecular weight distribution in a sample (Dettman et al., 2005). Nali and Manclossi (1995) studied the problems of molecular weight determination of asphaltenes by SEC and VPO. Shirokoff et al. (1997) reported their SEC profiles which were consistent with NMR and x-ray diffraction (XRD) results. Masuda et al. (1996) used polystyrene and fractionated novolak phenol resin as the calibration standards for the determination of the average MW of asphaltenes and preasphaltenes by SEC employing trimethylsilylated silica gel column and an N-methyl-2-pyrrolidone (NMP) mobile phase. Sato et al. (2005) reported a molecular weight calibration derived from the molecular weights of asphaltenes obtained using gel permeation chromatography (GPC) and gel permeation chromatography-mass selective detector (GPC-MSD). Since an insoluble is a mixture of different chemical species with a wide range of molecular weights, and there are no commercially available standards that would reflect such heterogeneity and be used as reliable calibration standards. This difficulty of SEC analysis has been resolved in the present study. For a country like India, oil is critical to its economy as well as the import bill. Therefore, India cannot afford to waste these vacuum residues to be used as a less valuable product. This vacuum residue, which is around 15 wt% of the total- processable crude oil, i.e., around 0.16 million bid, can be efficiently converted into high value fuels. Such conversion requires the knowledge of chemical composition and structural features of vacuum residues. Moreover, a detailed study on the compositional aspects of heavy ends of Indian crude oils, especially vacuum residues, has not been done extensively. Keeping this in mind, a detailed compositional study on vactium residues (>530°C) of Jodhpur and Heera crude oil (satellite crude oil of Bombay High) was undertaken. Jodhpur crude oil is the heaviest Indian crude oil having an API of 13.92, which represents an inland oil field. The other represents an offshore oil field in the Western region of India which has an API of 37.88. These two vacuum residues (VR), i.e., Jodhpur (JVR) and Heera vacuum (HVR) residue were then subjected to further fractionation using nonpolar and polar solvents. Fractionation of VR using a nonpolar solvent may not yield a true representative sample of its insoluble species which may contribute to the precipitation phenomenon in the crude oil. In the present study, therefore, a polar solvent is also employed for the precipitation of vacuum residues along with various nonpolar solvents. Hence, the general term "insolubles" instead of "asphaltenes" is given to the precipitated fractions of VRs. Scanty characterization studies are reported on insoluble fractions of VRs of Indian crude oils. Therefore, this study reports fractionation of VRs of two Indian crude oils in different solvents followed by the structural characterization of insolubles using NMR, FFIR, and size exclusion chromatography (SEC).
124
B. K. Sharma et al.
EXPERIMENTAL Materials The 530°C+ cuts (VRs) of two Indian crude oils (viz., Jodhpur and Heera crude oils) were dissolved separately in n-hexane, n-heptane, and ethyl acetate. These were recovered by filtration through a G4 sintered glass funnel after washing with respective solvents and dried to a constant weight in an oven at 110°C to give n-hexane asphaltenes (C6A), n-heptane asphaltenes (C7A), and ethyl acetate insolubles (EAT), respectively. Subsequently n-pentane was added to the n-heptane maltene and ethyl acetate soluble fraction and the resultant precipitates (termed as C7C5A and EAC51 insolubles) were recovered by a similar procedure.
NMR Spectroscopy 'H NMR spectra of four Jodhpur insolubles were recorded on a Jeol FX100 FT-NMR (JEOL Active Co., Ltd., Japan) spectrometer at an observing frequency of 99.5 MHz using a 10% w/v solution of the sample in CDCI3 (in a 5-mm o.d. NMR tube) with tetramethylsilane (TMS) as an internal standard. Experimental and instrumental conditions were selected to obtain quantitative spectra. The integral intensities of resulting signals were used to determine the percentage of aromatic protons (H) and aliphatic protons (Hsat ) attached to different carbons (viz. 11sat-a, Hsat-fl+y and Hsat-Me). The obtained data are given in Table 1. The solid state 13 C-NMR spectra of the same insolubles were recorded on a Bruker DRX-300 AVANCE CP-MAS-NMR (Bruker BioSpin AG, Fällanden, Switzerland) spectrometer at a 13 C resonance frequency of 75.47 MHz using 150-mg sample packed into the tube of a 7-mm CP-MAS probe spinning at 5 KHz to provide a contact time of 500 Its. The spectra were obtained by applying three different pulse sequences, i.e., standard CP-MAS, total suppression of spinning sidebands (TOSS), and CPD-MAS. The standard CP-MAS experiment was used for quantitative analysis. The CP-MAS spectra always showed significant spinning side bands. Therefore, TOSS experiments were performed in order to separate compound signals from the spinning side bands (Figure 1). The CPD-MAS experiment was carried out for spectral editing and was performed with different evolution delays (contact times of 25, 50, 75, 100, 150, and 250 its) in order to separate signals in the aliphatic region (Figure 2). The quaternary aromatic signals could thus easily be cancelled with evolution periods larger than 75 ps (Figure 2). It is pertinent to mention here that, although the sensitivity of measurement is better in CP mode than in single-pulse excitation (SPE) mode, CP underestimates the carbon aromaticity using relatively short contact times. However,
ri
r
Structural Characterization of Asphaltenes
125
Table 1. Percentage distribution of hydrogen and carbon atoms in insolubles of JVR estimated by NMR spectroscopy JEAI JC7A JC6A JC7C5A 1.8 Har 3.4 3.3 1.1 98.2 96.7 Hsat 96.6 98.9 10.3 4.9 11.2 Hsat-a 8.0 56.7 59.2 58.7 59.5 Hsatfl+y 31.2 32.6 26.7 31.4 Hsat_Me B! 0.55 0.55 0.45 0.53 (J.14 U.4L 0.53 0.78 °H ( H sat .. +y + Hsat.. Me)/ Har 49.4 27.6 25.3 79.7 86.2 .89.3 77.4 91.2 Csat 21.2 5.1 3.8 7.8 4.5 30.2 30.8 - 31.5 27.4 34.9 23.5 19.3 38.1 28.1 39.09 26.8 36.6 31.2 C (131 + 140) 13.8 10.7 22.6 8.8 3.3 6.3 6.0 2.2 Car..H 5.0 3.4 7.4 2.6 C&.alk (140.6) 3.0 0.8 1.7 3.4 C ar-Me 11.3 10.5 15.1 8.2 Car-per 2.5 0.2 7.5 0.6 Car..b C/H 0.54 0.53 0.56 0.52 Parameters
at longer contact time, the discrimination against aromatic carbons is not that acute. The contact time of 500 Its was thus found optimum and hence used in these measurements. Various structural parameters of the insolubles derived from their solid state NMR spectra are given in Tables 1 and 2. FTJ.R Spectroscopy Infrared spectroscopy (IR) spectra were recorded on pressed 13-mm diameter pellets of a 2.0 wt% mixture of insolubles in spectroscopy-grade KBr, which were dried at 150°C for 5 hr in an oven and stored over silica gel in a desiccator until use. A Perkin Elmer 1760X FTIR spectrometer (Perkin Elmer Life and Analytical Sciences, Inc., Wellesley, MA, USA) with IRDM 3.3 software was employed to record the spectra, by coadding eight scans at 4 cm resolution in the 4,000-600 cm -1 region. Eight scans were optimum as no useful improvement in the data was further achieved in a test coadding up to 300 scans. All spectra were recorded in a linear absorbance scale and the resulting spectra were normalized to an equivalent of I mg of the sample spread uniformly over the surface of a 13-mm diameter pellet. The absorbance
126
B. K. Sharma et al.
ppm 250 200 150 100 50 0
Figure 1. CP-MAS and CP-TOSS 13 C NMR spectra of JVR MA insoluble. SSBspinning side band.
values of selected IR bands of different functional groups identified in the spectra of each insoluble are given in Table 3, wherein each reported value is an average of at least three measurements.
Size Exclusion Chromatography The SEC system consisted of a Waters 510 pump (Waters Corporation, Milford, MA, USA), DuPont Rheodyne sample injector (DuPont Analytical Instruments Division, Wilmington, DE, USA), four Zorbax columns (Agilent Technologies, Inc., Santa Clara, CA, USA) in series (PSM 60 x 2 and PSM 300 x 2), a DuPont RI (refractive index) detector (sensitivity 0.10 x 10 RI units), and a PC-based Waters 820 maxima data station. The calibration and linearity of the column was evaluated using a quantitative mixture of polystyrene samples having molecular weights as 7,000, 5,970, 5,050, 3,250, 2,450, 1,700, and 800 and two hydrocarbon standards (viz., tetraphenylpyrene and benzopyrene having molecular weights of 506 and 252, respectively)
Structural Characterization of Asphaltenes
127
75 uS
50 pS
SSB
V
SSB
. p 1 p ppm 250 200 150 100 50 0
Figure 2. CPD-MAS 13 C NMR spectra of JVR C7A insoluble at different contact
times. SSB—spinning side band.
dissolved in tetrahydrofuran (THF). The sample was prepared by dissolving
0.5 wt% of insolubles into THF and filtering the solution through a Millipore filter (pore size 0.45 /Lm) made of mixed cellulose acetate and nitrate.
The filtered and degassed THF was used as a mobile phase at 1.0 ml min flow rate. A 20 jzL of sample solution was injected into the SEC column to obtain its SEC chromatogram. The retention volumes (R) were measured in triplicate for each sample and their average R-value was used for calculation of MW. The following calibration equation between the retention volume (R)
128
B. K. Sharma et al.
Table 2. Average number of carbons and hydrogens per molecule and per unit sheet from NMR data and elemental composition of insolubles Average number of C and H per molecule JEAI JC7A JC6A JC7C5A C Cn = C. fa ar
C
H
Czar-per C' ar-sub C" ar-b C" -'C"-) Car sat" H" H 'at = H".(%Hsat) H" ar (H/C)sat = H 'at /C" sat
165.6 167.4 158.9 220.9 22.8 17.9 35.9 19.4 5.5 10.5 9.5 4.9 18.7 17.6 24.0 18.1 13.2 7.1 14.5 13.2 4.1 0.3 11.9 1.3 142.8 149.5 123.0 201.5 231.8 224.4 217.9 290.9 227.7 216.9 210.6 287.6 4.1 7.5 7.3 3.3 1.60 1.45 1.71 1.43
Cxi. = 7(Cr/Cr per)2 - 1 9.4 6.2 14.7 7.0 - 'C 2.4 2.9 2.5 2.8 - ar 'I ar 69.0 57.7 63.6 78.9 C* = C"/G 1.7 0.1 4.8 0.5 = C b / G 5.5 2.4 5.8 4.7 ar-sub- -C" /G arar-sub C H = C* - ar-b + Car sub) 1.6 3.7 4.1 1.8 Cper C H + C sUb 7.1 6.1 9.9 6.5 a = C* b /C per 0.8 0.4 0.6 0.7 0.18 0.02 0.33 0.07 Y = C b /C Cs' C" " 59.6 51.5 48.9 71.9 sat - '- ar 96.6 77.4 87.2 103.9 H* = H/G ar H* '-" "- 1.7 2.6 2.9 1.2 ar -- arI_J 94.9 74.8 84.3 102.7 H* - H s' - H ar s' sat 1.8 1.05 3.4 1.2 R = (C:r b/2) + 1 RT (2Cs' + 2 - H' - C) 17 16.9 13.7 24.5 ar RN = RT - RA
15.2 15.8 10.3 23.3
and the molecular weight (M) was established based on the data generated using reference standards. Log = 9.237878 - (0.30037)R (with r2 = 0.99) The number average (Ma) and weight average (Mm) molecular weights of a sample were computed from their retention volumes and respective percent area of sliced peaks of its SEC chromatogram using a computer program.
r
Table 3. JR absorbance values (per gram sample) and structural ratios of various insolubles of Jodhpur (JYR) and Heera (HVR) vacuum residue JVR Frequency (cm- 1 ) C7A EM C7C5A EACH C7A EM
HYR C7C5A EAC5I C6A
3,627 [v(O-H)Fl 77 - - 32 34 24 66.5 63 3,545 [v(O-H),] - - 89.5 - 97.7 82.5 - 116 104.3 3,469 [v(N-H)] - - 132.7 - 126 128.5 - 172 168.5 3,412 [v(O-H') bJ 73 105.5 156 158 201.3 132.3 91.5 162 175.7 - 1,734 [v(C=O)EJ - - 33 - - - 29.5 - 42 1,699 [v(C=O)A] - - 58.7 - 67.7 68.3 64 223.3 90.3 1,647 [v(C=0) 0 ] 196 - 95 167 - - - 135.7 69.5 1,033 [v(C-O)] - 71 55.5 138.5 - - 61.3 - 152 882 [y (CH)ar1 ] 70.5 33 25.5 - 29.3 24.5 29.3 59.3 54.5 825 [y (CH)ar23 1 146 - - - - - - - 745 [y (CH)ar] 92.5 - - - 42.5 51.7 94.7 77.3 30.5 720 [r (CH2)], n > 4 102 50 33.25 - 58.7 65 110.3 - 49.7 Hme/Hsa t 1.57 0.64 0.51 2.13 0.39 0.28 0.16 0.73 0.36 (C=O) index 1.0 0.0 1.0 1.0 1.0 1.0 1.0 0.80 0.81 8 (CH3 )/8 (CH 3 + CH2 ) 0.54 0.71 0.68 0.78 0.57 0.51 0.34 0.76 0.61 P - y (CH)ar1 /y (CH)ar4 0.76 - - - 0.69 0.47 0.31 0.77 1.79
130
B. K. Sharma et al.
RESULTS AND DISCUSSION NMR Spectroscopy The 'H NMR spectra of insolubles showed strong peaks of methyl, methylene, and methine hydrogens of alkyl chains fi or further to aromatic ring (Hsat -p+y; 2.0-1.0 ppm). The methyl hydrogens present in aliphatic compounds or attached to aromatic rings in a y or further position (Hsat-me; 1.0-0.5 ppm) are quite intense in all the spectra. However, the hydrogens of methyl, methylene, and methine attached to aromatic rings in an a position (Hsat. a; 4.0-2.0 ppm) showed very weak signals. The aromatic hydrogens (H) are less in ethyl acetate insolubles (JEAI) as compared to heptane insolubles (JC7A) and hexane insolubles (JC6A), showing that ethyl acetate could effectively dissolve a larger number of substituted and condensed aromatic structures than othersolvents. This fact is also corroborated by a higher value of Hsa t -a in JEAI than in JC7A, and by a higher value for the degree of substitution (aH) in JEAI than in JC7A and JC6A. The branchiness index (BI) data indicate that, as compared to JC6A, JEAI and JC7A contain larger amounts of compounds having a higher degree of branching and/or shorter chain length. The spectral resolution obtained in the aliphatic region of 13 C NMR spectra produced partially resolved signals of four types of carbons; namely, aliphatic carbons of a methyl group attached at an a-position to an aromatic ring (a-CH3 ; '-21 ppm); aliphatic carbons of methylene groups in long alkyl chains (CH2 ; '-30 ppm); aliphatic carbons near a branched point (CH; '—p34 ppm); and aliphatic carbons of bridged methylene (CH; '39 ppm). The aromatic region (C; 160-100 ppm) was well separated from the aliphatic region and showed a partially resolved band of the aromatic carbon attached to carbons of alkyl groups (C-alk; '--140 ppm). The percent distribution of various carbon types derived from solid state CP-MAS 13 C-NMR spectra is given in Table 1. There are more aromatic carbons in JC6A (22.6) followed by JEAI (13.8) and JC7A (10.7). This shows that hexane removes the highest amount of aromatic carbons from VR, while heptane removes the least. in addition, the data on C7C5A insolubles reveal that n-pentane precipitated approximately 9% of the aromatic carbons present in the heptane-soluble fraction. The concentrations of a-methyl groups on aromatic ring (a-CH 3 ; '21 ppm) were found to be maximum in JC6A followed by JEAI and JC7A. The observation was also corroborated by a similar decrease in Hsat-a value from JC6A to JEAI followed by JC7A. The concentration of protonated aromatic carbons (C-H) decreases in the following order: JC7A, JC6A, and JEAI; but a reverse trend is seen in the values of C ar-Me, C ar-alk (methyl and alkyl substituted aromatic carbons, respectively) and ac (degree of substitution on aromatics calculated from the peaks at 130 and 140 ppm in 13 C NMR). The above results suggest that heptane dissolves a greater amount of substituted aromatics than hexane followed by ethyl acetate.
Structural Characterization of Asphaltenes
131
This inference is similar to that drawn from 0H (degree of substitution) data based on 'H NMR spectra. The values of Cb (bridgehead aromatic carbons) and a CIH ratio decrease from JC6A to JEAI to JC7A, confirming that JC6A contains a higher content of condensed aromatic structures than either JEAI or JC7A. The average number of carbon and hydrogen atoms per average molecule (shown with ") and some average molecular parameters derived from 13C NMR, elemental composition, and molecular weight data are reported in Table 2. These parameters suggest that there are more bridgehead carbon atoms in JC6A followed by JEAI and JC7A. The relative distribution of various aromatic carbons suggests that ethyl acetate removes more of the condensed aromatic compounds having a higher number of substituents, while n-heptane precipitates the aromatic structures having the least number of substituents and condensed rings. A high (H/C)sa t ratio and RA value (number of aromatic rings) of JC6A as compared to JEAI and JC7A further indicate the presence of a larger number of open alkyl chains and aromatic ring structures in the average molecule, but the RN value (number of naphthenic rings) decreases in the reverse order. A high molecular weight average molecule may be considered as constituted by the G number of elementary unit structures, which can be calculated by the following equation:
G = C"r/Cr C = 7(C/C'ar-per )2 - 1 C is the number of aromatic carbons per unit structure; hence, the other carbons of unit structure can be obtained by dividing the average molecular parameters by G. The number of carbons and hydrogens per unit sheet (shown with an *) and the average unit structural parameters (USP) thus obtained are listed in Table 2. These data reveal many differences in the structure of insolubles. JC7A insolubles have a maximum number of structural units per average molecule (2.9), followed by JC6A (2.5) and JEAI (2.4).' The maximum number of C* (total number of carbon atoms per unit sheet) is in JC7C5A (79), followed by JEAI (69), JC6A (64), and JC7A (58). JC6A has the maximum number of aromatic rings (3.4), followed by JEAI (1.8), JC7C5A (1.2), and JC7A (1.05). The small value of CH (protonated aromatic carbons) and a large value of CSUb (substituted aromatic ar carbons) and a (degree of substitution) together illustrate a higher concentration of substituted aromatic structures in the average molecule of JEAI than that of JC6A and JC7A. This observation further supports the fact that ethyl acetate removed higher amounts of substituted aromatic structures than heptane and hexane. The degree of condensation (y) is maximized in C6A (0.33), followed by EAI (0.18), C7C5A (0.07), and C7A (0.02). The inference is also supported by overall C/H ratio. The (H/C)sat ratio increases in the order:
132
B. K. Sharma et al.
JEAC5I < JC7A < JEAI
