Distribution of Vanadium Compounds in Petroleum Vacuum

Mar 11, 2015 - CnHmN5V1O2 species were found in the DMF extracts, which can be easily converted or removed under severe reaction conditions. The V com...
24 downloads 9 Views 1MB Size
Article pubs.acs.org/EF

Distribution of Vanadium Compounds in Petroleum Vacuum Residuum and Their Transformations in Hydrodemetallization Tingting Liu, Jincheng Lu, Xu Zhao, Yasong Zhou,* Qiang Wei, Chunming Xu, Yahe Zhang, Sijia Ding, Tao Zhang, Xiujuan Tao, Linqing Ju, and Quan Shi State Key Laboratory of Heavy Oil Processing, China University of Petroleum (Beijing), Beijing 102249, People’s Republic of China S Supporting Information *

ABSTRACT: The distribution of vanadium (V) compounds in the petroleum vacuum residuum (VR) and their transformations in hydrodemetallization (HDM) were investigated. V compounds in the VR and its hydrotreated products were extracted by different solvents in sequence to obtain methanol, dimethylformamide (DMF), and toluene extract fractions. The extracts were further separated into several subfractions using silica gel chromatography with various polar solvents. Positive-ion electrospray ionization (ESI) Fourier transform-ion cyclotron resonance mass spectrometry (FT-ICR MS) analyses was used to characterize V compounds before and after HDM. The contents of V compounds in the methanol, DMF, and toluene extracts were 9.04, 22.19, and 62.58%, respectively. The V compounds in the methanol extract were mainly porphyrin with a molecular formula of CnHmN4V1O1, which were found undergoing side-chain cracking and could be removed through hydrotreating. CnHmN5V1O2 species were found in the DMF extracts, which can be easily converted or removed under severe reaction conditions. The V compounds in the toluene extracts were most resistant for hydrotreating, which were speculated as vanadyl porphyrins with complex substituent groups attached to the core porphyrin structures. The results indicated that a highly active HDM catalyst should possess a highly active hydrogenesis property and macropore size distribution for the different V compounds removed. sequential HDM mechanism.12−17 However, the modal system cannot represent the hydroconversion behaviors of V and Ni compounds in the real oil system, such as VR, which is enriched with S, N, O, and other metal-containing compounds. These compounds exhibit high coking propensity. Some researchers have focused on the real oil system; however, they met many difficulties in this area as follows: First, using different measurement equipment and experimental methods, various porphyrin molecular structures in petroleum were hard to recognize.18−20 Second, the limitation of the analysis method for the metal compounds was still a challenge, although many new methods were used to study the metal compounds, such as X-ray absorption fine structure (XAFS)21 or gas chromatography−atomic emission detector (GC− AED)22 methods. Third, the analysis of metal compounds in the hydrotreating product oil was quite limited, because the metal compounds were easily removed in HDM. Kekalainen et al.23 found that there was little vanadyl porphyrin spectra in the hydrocracking product oils, which was hard to be identified in common because of a low metal concentration. In recent years, Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) has been used to characterize the vanadyl and nickel porphyrins in the petroleum because of its highly available broadband mass resolution together with mass resolving power as well as mass accuracy, which enables the assignment of a unique elemental composition to each peak in the mass spectrum.24−33 However, because of the complexity of metal compound molecule structures themselves, only a few

1. INTRODUCTION Recently, as the supply of conventional crude oil decreases worldwide, heavy oils [including heavy crude oil and vacuum residuum (VR)] become the significant resources to process for the oil refineries all over the world.1 However, it is a challenge for the faculties to use the heavy oils because of their inherently poor qualities, such as the highly enriched heteroatoms, like sulfur (S), nitrogen (N), oxygen (O), vanadium (V), and nickel (Ni). Among the heteroatoms found in the heavy oils, metal compounds (V and Ni) are detrimental to catalytic processes,2 especially for the corruption of fluidized catalytic cracking (FCC) catalysts3,4 and the deactivation of hydrotreatment catalysts.5−7 Therefore, the removal of metal compounds is essential for heavy oil processing. The hydrodemetallization (HDM) reaction is an efficient way to remove V and Ni in the vacuum residuum hydrotreating process.1 However, the accumulation of metal compounds depositing on the HDM catalysts causes a rapid deactivation of the HDM catalyst, which is the main reason why there is a high operating cost expected for residuum hydrotreating.8−11 Therefore, it is essential to design a long-lifetime HDM catalyst based on the better understanding of the structures and distributions of metal compounds in the VR and their catalytic transformation in HDM. Many researchers have studied the reaction behaviors of V and Ni compounds in HDM using different experimental methods and measurements. Their studies can be divided into two sections based on the feedstock that they used: (1) vanadyl and nickel porphyrin modal compounds and (2) the real oil system, such as VR. Kinetics studies of nickel and vanadyl porphyrin modal compounds and their catalytic HDM have been investigated in detail by Wei et al., who have proposed the © 2015 American Chemical Society

Received: October 17, 2014 Revised: February 8, 2015 Published: March 11, 2015 2089

DOI: 10.1021/ef502352q Energy Fuels 2015, 29, 2089−2096

Article

Energy & Fuels types of metal compounds were identified by FT-ICR MS unless the pretreatment of oil samples.29,33 Nickel porphyrins were more difficult than vandyl porphyrins to be ionized for FT-ICR MS analysis. In this work, VR and its hydrotreated products were extracted by different polar solvents in sequence to obtain methanol, dimethylformamide (DMF), and toluene extract fractions, respectively. To identify the structures of these V compounds, the extract fractions were further separated into several subfractions using silica gel chromatography with various solvents. Electrospray ionization (ESI) FT-ICR MS analyses were performed on the extracts and subfractions. The distribution and transformations of V compounds in the petroleum vacuum residuum as a result of HDM were investigated.

Figure 1. Pretreatment of oil samples before and after HDM in the thimble of the Soxhlet extractor using methanol, DMF, and toluene, sequentially.

2. EXPERIMENTAL SECTION 2.1. Sample Pretreatment and Characterization. The feed was Saudi Arabic VR from a refinary, and the properties are presented in Table 1. The hydrotreating reactions were carried out in the high-

Table 1. Properties of the Feed, P1, and P2 property

feed

P1

Physical Property ρ20 (g cm−3) 0.9921 0.9884 υ120 (mm2 s−1) 1632.9 351.3 CCR (wt %) 28.71 25.12 Element Composition (μg g−1) S 40930 32334 N 2100 1982 Ni 27.19 24.91 V 63.53 46.93 SARA Fraction (wt %) saturates 12.9 32.5 aromatics 52.2 48.0 resins 24.4 11.7 asphaltenes (n-C7) 10.5 7.8

P2 0.9801 57.7 20.41 14976 1806 17.54 19.30 40.1 44.7 12.6 2.6

Figure 2. Separation scheme of vanadyl porphyrins in feed, P1, and P2 (CH, DCM, CF, DMF, and n-C7 are corresponding to cyclohexane, dichloromethane, chloroform, and dimethylfomamide, and n-heptane, respectively.

pressure fixed-bed down-flow microreactor packed with 10 mL of commercial HDM catalysts. Prior to hydrotreating, the catalyst was presulfided with a solution of 2 wt % CS2/cyclohexane for 6 h at 320 °C and 4 MPa.34 After the sulfidation, the flow was switched to VR. The VR was fed continuously to the reactor for 30 h at the reaction conditions of 390 °C, 10 MPa, and 20 h−1, and the product oil was signed as P1. Similarly, another reaction was performed at 390 °C, 10 MPa, and 0.5 h−1 for 30 h after the sulfidation, and the product oil was signed as P2. The feed, P1, and P2 were subjected to solvent extraction first, as shown in Figure 1, which was modified from the scheme used by Zhao et al.29 A total of 3 g of oil sample was dissolved in 60 mL of chloroform, and then 40 g of silica gel was added and formed a slurry mixture. After the evaporation of chloroform, the mixture was moved into the Soxhlet extractor. Using different solvents sequentially (300 mL of methanol, 300 mL of DMF, and 300 mL of toluene) for 24 h each, methanol extracts, DMF extracts, and toluene extracts were obtained, respectively, after evaporating solvents by rotary evaporation. These extracts were further separated into various subfractions by silica gel chromatographic separation, as shown in Figure 2. The glass column (1 × 80 cm) was packed with 100/200 mesh silica gel (about 50 cm high). Solvents with various polarities were used for the elution. Different from the methanol and DMF extracts, the toluene extract fraction was first precipitated by n-hptane (n-C7), and then the n-C7 insolubles were separated into seven subfractions. The detailed steps were as follows: 50 mL of n-C7 was mixed with the toluene extract fraction for 30 min in the condition of reflux, and then they were

stored in the dark for 1 h. The n-C7-soluble (maltene) and n-C7insoluble (asphaltene) fractions were obtained by filtration using a 1− 3 μm filter paper (ShuangQuan no. 202 filter paper), and the insoluble fractions were refluxed with n-C7 for about 0.5 h. The n-C7-insoluble fractions on the filter paper were obtained by the 100 mL toluene solvent refuxing for 2 h. Solvents were removed by vacuum rotary evaporation. The contents of V in the feed, P1, and P2 were determined by an atomic emission spectrometer (ICP, Optima 7000 of PerkinElmer Company); the detailed steps were as follows: about 1 g of oil was ashed in a muffle furnace at 550 °C for 12 h after ignition. The ash was dissolved in 2 mL of nitric acid and 0.5 mL of hydrochloric acid and then transferred to a 25 mL volumetric flask, where it was evenly mixed for measurement. The solution from the volumetric flask was introduced into the spectrometer to determine the V and Ni contents. The total S content was analyzed by ultraviolet fluorescence, while nitrogen was measured by oxidative combustion and chemiluminescence at high-temperature combustion in an oxygen-rich atmosphere. The saturates/aromatics/resins/asphaltenes (SARA) composition of the oil sample was determined by SH/T0509-2010 (Chinese Standard Analytical Method for Petrochemical Industry). The determination of the carbon residue was the method of the Conradson carbon residue (CCR). 2.2. ESI FT-ICR MS Analysis and Data Processing. The different solvent extracts and their subfractions of feed, P1, and P2 were diluted with toluene to 10 mg/mL solution, and then they were diluted with toluene/methanol (1:1, v/v) solution to yield 0.02−0.15 2090

DOI: 10.1021/ef502352q Energy Fuels 2015, 29, 2089−2096

Article

Energy & Fuels mg/mL solution concentrations (resolved by the V/porphyrin content), which can minimize the molecular aggregation. To enhance the ESI ionization efficiency, 5 μL of formic acid was added for positive-ion ESI analysis in 1 mL diluted samples, and then the samples were analyzed by +ESI FT-ICR MS. The condition of FT-ICR MS was reported before.29 For Bruker apex-ultra FT-ICR MS equipped with a 9.4 T actively shielded superconducting magnet, the diluted samples were infused through an Apollo II electrospray source at 180 μL/h of a syringe pump. All of the samples are tested under the operation conditions with positive-ion formation of −4.0 kV emitter voltage, −4.5 kV capillary column frontend voltage, and 320 V capillary column end voltage. Ions accumulated for 0.1 s in a hexapole with 2.4 V direct-current (DC) voltage and 500 Vp−p radio-frequency (RF) amplitude. The quadrupole (Q1) was optimized to obtain a broad range for ion transfers. An argon-filled hexapole collision cell was operated at 5 MHz and 700 Vp−p RF amplitude, and ions accumulated for 0.6 s. The extraction period for ions from the hexapole to the ion cyclotron resonance cell was set to 1.5 ms. The RF excitation was attenuated at 11.75 dB and used to excite ions from 200 to 1000 Da. Data sets (4 M) were acquired, and 128 scans were co-added to enhance the signal-to-noise ratio and dynamic range. MS was internally calibrated using a N1 class homologous series [CnH2n + 17N1 + H]+ and [CnH2n + 19N1 + H]+. Internal quadratic calibration was also performed. Peaks with relative abundance greater than 6 times the standard deviation of the baseline noise level were exported to a spreadsheet. Data analysis was performed by selecting a two-mass scale-expanded segment in the middle of the mass spectrum, followed by the detailed identification of each peak. The peak of at least one of each heteroatom class species was arbitrarily selected as a reference. Species with the same heteroatom class and their homologues with different double bond equivalent (DBE) values and carbon numbers were searched within a set of 0.002 Kendrick mass defect tolerance. The details of the data analysis procedure used have been described elsewhere.35

feed, and the yields of DMF extract and toluene extract were 35.26 and 47.95 wt %, respectively. With the increasing reaction severity, the extract yields of methanol (13.10% in P1 and 13.88% in P2) and DMF (39.56% in P1 and 44.33% in P2) increase, while that of toluene (43.64% in P1 and 36.60% in P2) decreased. This was consistent with the difference of SARA data in Table 1 (the content of resins and asphaltenes decreased, while that of saturates increased), indicating that large molecules, such as asphaltenes, in the feed have been hydrogenated after HDM,36−38 and then more components can be extracted by methanol and DMF. On the basis of the similar compatibility theorem, extracts of different solvents have different polarity and metal compounds also have different polarity. Metal compounds in the methanol extracts have relatively high polarity, and that in DMF have relatively semipolarity, while that in the toluene extracts have relatively weak polarity. Table 2 shows the distribution of V compounds in different extracts of feed, P1, and P2. Only 9.04% V compounds can be Table 2. Distribution of V in Different Extracts of Feed, P1, and P2 sample

item

yield (g)

concentration of V (wppm)

feed

sample methanol DMF toluene total sample methanol DMF toluene total sample methanol DMF toluene total

2.967 0.362 1.046 1.423 2.831 3.106 0.407 1.229 1.355 2.991 2.990 0.415 1.325 1.094 2.835

63.53 47.07 39.99 82.90

P1

3. RESULTS AND DISCUSSION 3.1. Distribution of V Compounds in the VR before and after HDM. Table 1 shows properties of the feed, P1, and P2. It can be seen that after hydrotreating, the contents of V compounds decreased gradually (from 63.53 ppm in the feed to 46.93 ppm in P1 and 19.30 ppm in P2) with the increasing severity of reaction conditions. Simultaneously, the SARA composition shows that asphaltenes decreased (from 10.5 wt % in the feed to 7.8 wt % in P1 and 2.6 wt % in P2), which indicates that V removal is relevant to the conversion of asphaltenes. The extract yield of different solvents is shown in Figure 3. It can be seen that 12.21 wt % methanol extract was from the

P2

46.93 19.02 17.70 80.61 19.30 4.12 5.89 41.83

V (μg) 188.49 17.04 41.83 117.97 176.84 145.76 7.74 21.75 109.23 138.72 57.71 1.71 7.80 45.76 55.27

percentage of V yield (%) 9.04 22.19 62.58 93.81 5.31 14.92 74.94 95.17 2.96 13.52 79.29 95.77

extracted by methanol in the feed. The V compounds were slightly enriched in the toluene extract, which accounts for the major portion (62.58%). While for P1 and P2, the amount of V compounds in methanol and DMF extracts decreased. The V yields in the toluene extracts for P1 and P2 were 74.94 and 79.29%, respectively. It indicated that V compounds in the toluene extracts were hard to remove in HDM. Total yields in weight of all of the extracts in feed, P1, and P2 were 93.81, 95.17, and 95.77%, respectively. Other V could be lost because of the permanent absorption on the silica gel and/or the experimental error of determination. The distribution of the V content in different extracts is also shown in Table 2. It represented that the V content in the methanol and DMF extracts decreased to 19.02 and 17.70 wppm, respectively, under the mild reaction conditions, while the V concentration in the toluene extract was 80.61 wppm compared to 82.90 wppm in the feed. In P2, the V content of methanol and DMF extracts was 4.12 and 5.89 wppm, respectively, which indicated that V compounds with high polarity in the methanol and DMF extracts can be removed easily. However, V compounds in the toluene extract of P2 still remained 41.83 wppm, which indicates that these V

Figure 3. Extract yield of different solvents in feed, P1, and P2. 2091

DOI: 10.1021/ef502352q Energy Fuels 2015, 29, 2089−2096

Article

Energy & Fuels compounds were resistant to be removed, even under severe reaction conditions. The percentage of V removal in different extracts in P1 and P2 is shown in Figure 4. It can be seen that, in the mild reaction

Figure 4. Percentage of V removal in different extracts in P1 and P2. Figure 5. Relative abundances of heteroatom compounds in the feed, P1, and P2.

conditions, 56.62% and 50.36% V compounds can be removed in the methanol and DMF extracts, respectively, while only 11.57% V compounds can be removed in the toluene extracts. However, under severe reaction conditions, 90.01 and 81.49% V can be removed in the methanol and DMF extracts, respectively, but only 64.34% can be removed in the toluene extract, saying that V compounds in the toluene extracts were the most difficult to remove. It was reported that metals (both Ni and V) in the asphaltenes seem to be difficult to remove during hydrotreating because of the large molecules in this fraction.39 Miller et al.21 found that V compounds in the toluene-insoluble solid characterized by XAFS showed a similar coordination structure with vanadyl porphyrin. In our study, we speculate that V compounds in the toluene extracts may be large molecules of complex structures and hard to remove. 3.2. Characterization of V Compounds by +ESI FT-ICR MS and Their Transformation in HDM. To characterize the structure of V compounds before and after HDM, the oil samples of feed, P1, and P2 were analyzed by +ESI FT-ICR MS. Broadband and expanded +ESI FT-ICR MS spectra for the feed, P1, and P2 in Figure S1 of the Supporting Information show that, with the increasing severity of reaction conditions, the center of the spectrum peaks shifted to the left. The relative abundances of heteroatom compounds in the feed, P1, and P2 are shown in Figure 5. O1S1 class species were dominant (30.78%) in the feed. After hydrotreating, O1S1 class species decrease to 21.92% in P1 and 5.35% in P2, respectively. However, N4V1O1 class species in the feed were in lower content (0.27%), and they were not identified in the P1 and P2. Broadband and expanded +ESI FT-ICR MS spectra for the feed−M, P1−M, and P2−M are shown in Figure S2 of the Supporting Information, in which typical vanadyl porphyrins, such as C29H31N4V1O1 (502.19320) were found in the P1−M with low mass peak intensity, while they were absent in the P2−M. However, the vanadyl porphyrins can be found in the subfractions with high intensity, as shown in Figure 6, which indicates that pre-separation was significant for the vanadyl porphyrin characterization. Vanadyl porphyrins have abundant mass peaks in the subfractions of methanol extracts,29 and many types of vanadyl porphyrins have been found as reported previously.25,29,40

Figure 7 shows iso-abundant plots of DBE as a function of the carbon number for N4V1O1 vanadyl porphyrins derived from +ESI FT-ICR MS for feed−M4, feed−M5, P1−M3, P1−M4, P1−M5, P2−M3, P2−M4, and P2−M5. It can be seen that the ESI-detectable vanadyl porphyrins were enriched in the subfraction of feed−M4 and feed−M5. Eight types of identified petroleum CnHmN4V1O1 porphyrins (DBE = 17, 18, 19, 20, 21, 22, 23, and 24) were found in the feed methanol subfractions. The CnHmN4V1O1 porphyrins were mainly etioporphyrin (ETIO) structures (corresponding to DBE = 17), and the carbon number range was C32−C43 with a center mass at C31 in feed−M4 and C25−C41 with a center mass at C29 in feed−M5. Vanadyl porphyrins present in feed−M5 have a lower carbon number than those in feed−M4, indicating that these porphyrins have high polarity because they are eluted by the silica gel chromatography. After hydrotreating, CnHmN4V1O1 porphyrins can be found in the P1−M3, and one type of CnHmN4V1O1 porphyrin (DBE = 20) also has a comparatively strong intensity. The CnHmN4V1O1 porphyrins (DBE = 20) in P1−M3 may be formed from CnHmN5V1O2 species (DBE = 20) in the feed− DMF extract, which will be discussed below. In P1−M4 and P1−M5, the main structure of CnHmN4V1O1 porphyrins should be ETIO (DBE = 17), but the carbon number in the center mass (C29 in P1−M4 and C28 in P1−M5) was smaller than that in feed−M4 and feed−M5, respectively. This can be explained that these vanadyl compounds may be the vanadyl porphyrin intermediates with a shorter size chain attached to the main porphyrin structure formed during HDM. When the reaction condition was severe, only two types of vanadyl porphyrins (DBE = 17 and 20) with a smaller carbon number can be found in the P2−M3, while the types of CnHmN4V1O1 porphyrins in P2−M4 and P2−M5 were only ETIO. Thus, the CnHmN4V1O1 species in the methanol extracts may undergo a reaction of cracking the long side chains attached to the porphyrin core structure first, and then the porphyrin core could be hydrogenated, with V removed.14−16 Another factor that should be of concern is that only vanadyl porphyrins with a relatively high abundance could be identified after hydro2092

DOI: 10.1021/ef502352q Energy Fuels 2015, 29, 2089−2096

Article

Energy & Fuels

Figure 6. Broadband and expanded +ESI FT-ICR MS spectra of vanadyl porphyrins for the feed−M4, P1−M4, and P2−M4.

Figure 7. Iso-abundant plots of DBE as a function of the carbon number for N4V1O1 vanadyl porphyrins derived from +ESI FT-ICR MS for feed− M4, feed−M5, P1−M3, P1−M4, P1−M5, P2−M3, P2−M4, and P2−M5.

CnHmN5V1O2 species (DBE = 17, 18, and 20) can be found in the feed−D3, and there were four types of CnHmN5V1O2 species (DBE = 17, 18, 20, and 21) in the feed−D4 and feed−D5. In contrast, the CnHmN5V1O2 class species exhibited one type of compound in feed−M4. This implies that the composition and structures of CnHmN5V1O2 species were different from those of CnHmN4V1O1 species, leading to the weaker polarities of CnHmN5V1O2 species, which can be extracted by the DMF solvent after the extraction by methanol. After hydrotreating, one type of CnHmN5V1O2 species (DBE = 17) can be eluted to P1−M3 and P1−M4. However, only two types of CnHmN5V1O2 species (DBE = 17 and 20) were in the P1−D3, and the carbon number range for CnHmN5V1O2 species (DBE = 17) was C35−C39, shorter than that in feed−D3 (C31−C51). In comparison to the distribution of CnHmN4V1O1

treating because of the limited dynamic detection range of FTICR MS. It is reported that new V compounds in Venezuela heavy crude oil have been detected.33 In our study, the N5V1O2 species were found. Figure 8 shows an expanded view of the segmental spectra at m/z 631.10−631.50 of these proposed vanadyl porphyrins (C36H46N5V1O2 at the mass of 631.308 72) for the feed−M4, feed−D3, feed−D4, feed−D5, P1−M3, P1− M4, P1−D3. Figure 9 shows the iso-abundant plots of DBE as a function of the carbon number for CnHmN5V1O2 porphyrins derived from +ESI FT-ICR MS for feed−M4, feed−D3, feed− D4, feed−D5, P1−M3, P1−M4, and P1−D3. It can be seen that N5V1O2 species were mainly in the DMF extract subfractions, while there were CnHmN5V1O2 species only in one methanol extract subfraction (feed−M4). Three types of 2093

DOI: 10.1021/ef502352q Energy Fuels 2015, 29, 2089−2096

Article

Energy & Fuels

Figure 8. Expanded +ESI FT-ICR MS spectra of new vanadyl porphyrins (N5V1O2) for the feed−M4, feed−D3, feed−D4, feed−D5, P1−M3, P1− M4, and P1−D3.

Figure 9. Iso-abundant plots of DBE as a function of the carbon number for N5V1O2 vanadyl porphyrins derived from +ESI FT-ICR MS for feed− M4, feed−D3, feed−D4, feed−D5, P1−M3, P1−M4, and P1−D3.

that methanol can extract most of the CnHmN4V1O1 species. However, in the DMF extract subfractions, only the C n H m N 5 V 1 O 2 species can be characterized and no CnHmN4V1O1 species can be found in the feed−DMF and P1−DMF extracts and their subfractions. However, no V species in toluene extracts and subfractions can be detected by ESI FT-ICR MS in our study. It can be inferred that some other more complex V compounds existed in the toluene extracts, and they were hard to detect by +ESI FT-ICR MS, although the V concentration in toluene extracts was much higher than that in the methanol extract fraction (for feed, P1, and P2, they were 39.76, 35.16, and 14.18 ppm, respectively). These compounds likely have the main structure of porphyrin, with the addition of more aromatic rings, thiophene, and amino functional groups.32 We speculated that V compounds in toluene extracts may be more complex substituent groups attached to the vanadyl porphyrin structure, but they were hard to recognize in the

species in P1−M3 in Figure 7, it can be inferred that CnHmN5V1O2 species may be converted to CnHmN4V1O1 species, leading to a higher intensity of CnHmN4V1O1 species with DBE of 20 in P1−M3. Under severe reaction conditions, there were no CnHmN5V1O2 species in the P2 subfractions, which means that CnHmN5V1O2 species have been converted or hydrodemetallized, leaving the most complex V compounds in the DMF, and were hard to characterize. It also indicates that CnHmN5V1O2 species can be easier to remove or convert than CnHmN4V1O1 porphyrins in the methanol extract under severe reaction conditions. The relative abundances of CnHmN4V1O1 and CnHmN5V1O2 species in the feed−M4, P1−M3, and P1−M4 are shown in Figure 10. It can be seen that CnHmN4V1O1 species were the most abundant (92.41% in feed−M4, 96.09% in P1−M3, and 94.09% in P1−M4) compared to the CnHmN5V1O2 species detected in the methanol extract subfractions, which indicates 2094

DOI: 10.1021/ef502352q Energy Fuels 2015, 29, 2089−2096

Article

Energy & Fuels

Figure 10. Relative abundances of CnHmN4V1O1 and CnHmN5V1O2 species in the feed−M4, P1−M3, and P1−M4.

(Figure S2). This material is available free of charge via the Internet at http://pubs.acs.org.

+ESI FT-ICR MS spectra. Also, we speculate that, for the V compounds in toluene extracts, the complex substituent groups attached to the main porphyrin structures may hinder the interaction of V with the catalyst surface during hydrotreating, unless the complex substituent groups were converted and/or hydrocracked under the severe reaction conditions, just like the refractory of nitrogen compounds, which have long alkyl side chains in the hard-to-convert region.32 It also implied that a highly active HDM catalyst should possess different properties for the removal of different V species. For the vanadyl porphyrins in the methanol extracts, a catalyst with a highly active hydrogenesis property would promote the cracking of the long alkyl side chains attached to the porphyrin core structure. For the conversion of V compounds in the DMF and toluene extracts, a catalyst with a macropore size distribution would be more appropriate for the large molecule diffusion41 and conversion.39 It can be used to explain that a HDM catalyst with a large pore size distribution in the 10−30 nm region is the key point for the removal of V and asphaltenes.42−45 This study is significant for a highly active HDM catalyst design.



*Telephone: +86-10-8973-3501. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This work was supported by the National Natural Science Foundation of China (NSFC, U1362203 and 21376262). REFERENCES

(1) Rana, M. S.; Sámano, V.; Ancheyta, J.; Diaz, J. A. I. Fuel 2007, 86 (9), 1216−1231. (2) Furimsky, E.; Massoth, F. E. Catal. Today 1999, 52 (4), 381−495. (3) Baker, E. W.; Yen, T. F.; Dickie, J. P.; Rhodes, R. E.; Clark, L. F. J. Am. Chem. Soc. 1967, 89 (14), 3631−3639. (4) Wormsbecher, R. F.; Peters, A. W.; Maselli, J. M. J. Catal. 1986, 100 (1), 130−137. (5) Ledoux, M. J.; Hantzer, S. Catal. Today 1990, 7 (4), 479−496. (6) Dejonghe, S.; Hubaut, R.; Grimblot, J.; Bonnelle, J. P.; Des Courieres, T.; Faure, D. Catal. Today 1990, 7 (4), 569−585. (7) Mitchell, P. C. H.; Scott, C. E. Polyhedron 1986, 5 (1−2), 237− 241. (8) Vogelaar, B. M.; Berger, R. J.; Bezemer, B.; Janssens, J.; Langeveld, A. D.; Eijsbouts, S.; Moulijn, J. A. Chem. Eng. Sci. 2006, 61, 7463−7478. (9) Isaza, M. N.; Pachon, Z.; Kafarov, V.; Resasco, D. E. Appl. Catal., A 2000, 199 (2), 263−273. (10) Maity, S. K.; Blanco, E.; Ancheyta, J.; Alonso, F.; Fukuyama, H. Fuel 2012, 100, 17−23. (11) Vogelaar, B. M.; Eijsbouts, S.; Bergwerff, J. A.; Heiszwolf, J. J. Catal. Today 2010, 154 (3), 256−263. (12) Hung, C. W.; Wei, J. Ind. Eng. Chem. Process Des. Dev. 1980, 19, 250−257. (13) Hung, C. W.; Wei, J. Ind. Eng. Chem. Process Des. Dev. 1980, 19, 257−263. (14) Ware, R. A.; Wei, J. J. Catal. 1985, 93 (1), 100−121. (15) Ware, R. A.; Wei, J. J. Catal. 1985, 93 (1), 122−134. (16) Ware, R. A.; Wei, J. J. Catal. 1985, 93 (1), 135−151. (17) Chen, H. J.; Massoth, F. E. Ind. Eng. Chem. Res. 1988, 27 (9), 1629−1639. (18) Dechaine, G. P.; Gray, M. R. Energy Fuels 2010, 24 (5), 2795− 2808. (19) Xu, H.; Yu, D.; Que, G. Fuel 2005, 84 (6), 647−652.

4. CONCLUSION The distribution of V compounds for VR in the methanol, DMF, and toluene extracts was 9.04, 22.19, and 62.58%, respectively. The analysis of +ESI FT-ICR MS showed that V compounds in the methanol extract were mainly porphyrins with a molecular formula of CnHmN4V1O1, which were found undergoing side-chain cracking first and could be removed through hydrotreating. CnHmN5V1O2 species were found in the DMF extracts, which can be easily converted or removed under severe reaction conditions. The V compounds in the toluene extracts were most resistant for hydrotreating, which were speculated as vanadyl porphyrins with complex substituent groups attached to the core porphyrin structures. The results indicated that a highly active HDM catalyst should possess a highly active hydrogenesis property and macropore size distribution for the different V compounds removed.



AUTHOR INFORMATION

Corresponding Author

ASSOCIATED CONTENT

S Supporting Information *

Broadband and expanded +ESI FT-ICR MS spectra for the feed, P1, and P2 (Figure S1) and broadband and expanded +ESI FT-ICR MS spectra for the feed−M, P1−M, and P2−M 2095

DOI: 10.1021/ef502352q Energy Fuels 2015, 29, 2089−2096

Article

Energy & Fuels (20) Yin, C. X.; Stryker, J. M.; Gray, M. R. Energy Fuels 2009, 23 (5), 2600−2605. (21) Miller, J. T.; Fisher, R. B.; Van der Eerden, A. M. J.; Koningsberger, D. C. Energy Fuels 1999, 13 (3), 719−727. (22) Kim, T.; Ryu, J.; Kim, M. J.; Kim, H. J.; Shul, Y. G.; Jeon, Y.; Park, J. I. Fuel 2014, 117, 783−791. (23) Kekäläinen, T.; Pakarinen, J. M.; Wickström, K.; Lobodin, V. V.; McKenna, A. M.; Jänis, J. Energy Fuels 2013, 27 (4), 2002−2009. (24) Hsu, C.; Shi, Q. Sci. China Chem. 2013, 56 (7), 833−839. (25) McKenna, A. M.; Purcell, J. M.; Rodgers, R. P.; Marshall, A. G. Energy Fuels 2009, 23 (4), 2122−2128. (26) Qian, K.; Edwards, K. E.; Mennito, A. S.; Walters, C. C.; Kushnerick, J. D. Anal. Chem. 2009, 82 (1), 413−419. (27) Qian, K.; Mennito, A. S.; Edwards, K. E.; Ferrughelli, D. T. Rapid Commun. Mass Spectrom. 2008, 22 (14), 2153−2160. (28) McKenna, A. M.; Williams, J. T.; Putman, J. C.; Aeppli, C.; Reddy, C. M.; Valentine, D. L.; Rodgers, R. P. Energy Fuels 2014, 28 (4), 2454−2464. (29) Zhao, X.; Liu, Y.; Xu, C.; Yan, Y.; Zhang, Y.; Zhang, Q.; Shi, Q. Energy Fuels 2013, 27 (6), 2874−2882. (30) Zhang, Y.; Zhang, L.; Xu, Z.; Zhang, N.; Chung, K. H.; Zhao, S.; Xu, C.; Shi, Q. Energy Fuels 2014, 28, 7448−7456. (31) Zhang, L.; Xu, Z.; Shi, Q.; Sun, X.; Zhang, N.; Zhang, Y.; Chung, K. H.; Zhao, S. Energy Fuels 2012, 26 (9), 5795−5803. (32) Zhang, T.; Zhang, L.; Zhou, Y.; Wei, Q.; Chung, K. H.; Zhao, S.; Xu, C.; Shi, Q. Energy Fuels 2013, 27 (6), 2952−2959. (33) Zhao, X.; Shi, Q.; Gray, M. R.; Xu, C. Sci. Rep. 2014, 4, 5373. (34) Tao, X.; Zhou, Y.; Wei, Q.; Yu, G.; Cui, Q.; Liu, J.; Liu, T. Fuel Process. Technol. 2014, 118, 200−207. (35) Shi, Q.; Pan, N.; Long, H.; Cui, D.; Guo, X.; Long, Y.; Chung, K. H.; Zhao, S.; Xu, C.; Hsu, C. S. Energy Fuels 2013, 27 (1), 108−117. (36) Haulle, F. X. Dissertation, University of Paris VI, Paris, France, 2002. (37) Ancheyta, J.; Centeno, G.; Trejo, F.; Marroquin, G. Energy Fuels 2003, 17 (5), 1233−1238. (38) Verstraete, J. J.; Le Lannic, K.; Guibard, I. Chem. Eng. Sci. 2007, 62 (18), 5402−5408. (39) Ferreira, C.; Tayakout-Fayolle, M.; Guibard, I.; Lemos, F.; Toulhoat, H.; Ramôa Ribeiro, F. Fuel 2012, 98, 218−228. (40) Doukkali, A.; Saoiabi, A.; Zrineh, A.; Hamad, M.; Ferhat, M.; Barbe, J. M.; Guilard, R. Fuel 2002, 81, 467−472. (41) Marchal, C.; Abdessalem, E.; Tayakout-Fayolle, M.; Uzio, D. Energy Fuels 2010, 24 (8), 4290−4300. (42) Hardin, A. H.; Packwood, R. H.; Ternan, M. Prepr.Am. Chem. Soc., Div. Pet. Chem. 1978, 23 (4), 1470. (43) Hardin, A. H.; Ternan, M.; Packwood, R. H. The Effects of Pore Size in MoO3−CoO−A12O3 Hydrocracking Catalysts; CANMET, Energy, Mines and Resources Canada: Ottawa, Ontario, Canada, 1981; CANMET Report 81-4E. (44) Juarez, J. A.; Maity, S. K.; Rivera, G. B. Appl. Catal., A 2001, 216, 195−208. (45) Rana, M. S.; Ancheyta, J.; Rayo, P.; Maity, S. K. Catal. Today 2004, 98 (1), 151−160.

2096

DOI: 10.1021/ef502352q Energy Fuels 2015, 29, 2089−2096