A molecular insight into deoiled asphalt’s slurry-phase hydrocracking process

3.1.1 Product distribution and upgrading results for solvent deasphalting process

All the property data were analyzed at least two times till the parallel of the assay was quite well. For SDA1 test, it can be seen from the Table 3 that the yield of DAO was 32.1 % and the corresponding yield of DOA was 67.9 %. If decreasing the extraction temperature (Cao et al., 2010), the yield of DAO for SDA2 test was increased to 48.3 % and the corresponding yield of DOA was reduced to 51.7 %. It can be seen from Table 3 that the hydrogen content of SDA1-DAO was slightly higher than that of SDA2-DAO (12.18 % vS 11.96 %), with the opposite trend for S and N content. The hydrogen content of SDA1-DOA was higher than that of SDA2-DOA (10.08 % vS 9.68 %), with the opposite trend for S and N content. DAO has much higher hydrogen and much lower metal and asphaltene content than that of DOA, which could prove the high separation efficiency for SDA process.

It can be seen from Table 3, in DAO the metal content for Ni + V + Ca was<15 ppm and the asphaltene content was<0.1 %. DAO could be good feedstock for FCC or hydrocracking process, and the DOA could be taken as the feed of slurry-phase hydrocracking process in this work.

3.1.2 FT-ICR analysis of DAO and DOA derived from solvent deasphalting process

FT-ICR MS is an effective method (Li et al., 2021; Chacón-Patiño et al., 2020) to characterize the heavy oil even for the heaviest fractions rich in asphaltenes. As illustrated in Fig. 2, typical aromatic compounds including HC, S1, S2, S3, N1, O1, et al. classes were detected by FT-ICR MS (Chacón-Patiño et al., 2020; Hsu et al., 2011; Wang et al., 2016). Relative abundance of HC and S1 classes of DAO was much higher than that of DOA, which could prove HC and S1 classes with lower polarity have higher selectivity in solvent extraction for SDA process. Relative abundance of HC and S1 classes of SDA1-DAO was a little higher than that in SDA2-DAO, which could prove SDA1 test have higher selectivity for HC and S1 classes with lower polarity than that of SDA2 test in solvent extraction for process.

Close

Fig. 2

The aromatic species of SDA1-DAO and SDA1-DOA by FT-ICR MS.

Export to PPT

The comparisons of DBE distribution of HC class species of DAO and DOA were illustrated in Fig. 3. As seen in Fig. 3, in all DAO and DOA, the DBE of HC class species distributed in a broad range of 4 ∼ 29 for DAO and 4 ∼ 36 for DOA, with corresponding distribution centers at 8–12 for DAO and 10–19 for DOA respectively. The relative abundance of DAO was much higher than that of DOA when DBE < 14, with the opposite trend for DBE greater than 14. The relative abundance of SDA1-DAO was also higher than that of SDA2-DAO when DBE < 14, with the opposite trend for DBE ≥ 14. The relative abundance of SDA1-DOA was higher than that of SDA2-DOA when DBE < 20, with the opposite trend for DBE ≥ 20.

Close

Fig. 3

DBE distribution of HC class species of DAO and DOA by FT-ICR MS.

Export to PPT

As seen in Fig. 3, the abundance of SDA1-DOA was higher than that of SDA2-DOA when DBE < 20, with the opposite trend for DBE ≥ 20. The most abundance HC classes were located at DBE = 10–19, and the adjusted relative abundance as a function of CN for HC class species with DBE = 10, 13, 16 and 19 was illustrated in Fig. 4. For both SDA1-DOA and SDA2-DOA, the CN distribution was centered at 45 ∼ 46, and the relative abundance of HC class species of SDA1-DOA shows a little higher around the CN center at 45 ∼ 46 than that of SDA2-DOA.

Close

Fig. 4

HC compounds of SDA1-DOA and SDA2-DOA with DBE = 10, 13, 16, and 19 distribution. The relative abundance was adjusted by its absolute HC content, and the skeletons are shown only for illustrative purposes.

Export to PPT

The comparisons of CN distribution of HC class species of DAO and DOA were illustrated in Fig. 5. As seen in Fig. 5, in all DAO and DOA, the HC class species distributed in a broad range of CN (28 ∼ 73), with corresponding distribution centers at 42–47. The relative abundance of DAO was higher than that of DOA when CN < 46, with the opposite trend for CN ≥ 46. The relative abundance of SDA1-DAO was also higher than that of SDA2-DAO when CN < 45, with the opposite trend for CN ≥ 45. The abundance of SDA1-DOA was also higher than that of SDA2-DOA when CN < 46, with the opposite trend for CN ≥ 46.

Close

Fig. 5

CN distribution of HC class species of DAO and DOA by FT-ICR MS.

Export to PPT

It can be concluded that the SDA process in this work not only can obtain different DAO and DOA yields but also obtain measurably different DAO and DOA properties. The n-butane solvent in SDA process has higher selectivity for HC class, and correspondingly the relative abundance of HC class of DAO was much higher than that of DOA. The abundance of DAO was higher than that of DOA when DBE < 14 or CN < 46. The relative abundance of SDA1-DAO was higher than that of SDA2-DAO when DBE < 14 or CN < 45. The relative abundance of SDA1-DOA was higher than that of SDA2-DOA when DBE < 20 or CN < 46.

3.2.1 Product distribution and upgrading results for DOA and VR in slurry-phase hydrocracking process

Taking the conversion ratio and the stability control into account, the operating parameters listed in Table 4 were taken in SHC process. The distillation range of the products were obtained by ASTM D7169 method. The conversion depth for each experiment could be evaluated by CR_SHC in Equation (1).

CR_SHC represents the conversion ratio of DOA in slurry-phase hydrocracking reaction, %. ${\mathrm{Y}}_{\mathrm{S}\mathrm{H}\mathrm{C}}^x$ in this work represents the yield of × in slurry-phase hydrocracking process, %. The CR_SHC of SDA1-DOA for SHC1-120 min and SHC1-240 min reactions were 66.92 % and 81.64 % respectively, and the CR_SHC of SDA2-DOA for SHC2-120 min and SHC2-240 min reactions were 70.19 % and 85.71 % respectively. Under the same experimental conditions, the reaction time has a great influence on the conversion, and the longer the reaction time was, the higher conversion ratio could be. The conversion ratio of DOA could be more than 80 % for 240 min reaction.

3.2.2 Comparisons of product distribution and upgrading results for different DOA in slurry-phase hydrocracking process

There is an interesting phenomenon that under the same experimental conditions, heavier DOA has higher CR_SHC in SHC process. The CR_SHC of SDA2-DOA was higher than that of SDA1-DOA for 120 min and 240 min reactions respectively, while SDA2-DOA has lower H content than that of SDA1-DOA. Because dispersed catalysts in slurry-phase hydrocracking process is lack of acidic sites for cracking (Kang et al., 2021; Cui et al., 2021), slurry-phase hydrocracking process still follows the mechanism of free radicals (Du et al., 2015). Hydrocarbons are decomposed through thermal cracking by free radicals because dispersed catalyst shows no catalytic function for C—C cleavage (Kang et al., 2021), so the conversion from heavy oil to light products is driven by thermal cracking. During this reaction, H₂ molecule splits to hydrogen free radicals through homolytic and heterolytic ways on dispersed catalysts, and those hydrogen free radicals could quench with macromolecular free radicals (Zhang et al., 2008; Du et al., 2015). This process can greatly inhibit secondary radical reactions that lead to the formation of coke precursors and gases (Prajapati et al., 2021; Kang et al., 2021).

In slurry-phase hydrocracking process, the macromolecules convert to smaller molecules by hydrogenlysis and side-chain breaking (Ortiz-Moreno et al., 2012). The bond energy of C—S is much lower, and sulfur can abstract H from hydrocarbon molecule to form free radicals even at a lower temperature (Nguyen et al., 2021; Chang et al., 2001). Sulfur was an effective free-radical initiator to enhance the reaction (Chang et al., 2001).

As seen in Table 3, S content of SDA2-DOA was higher than that of SDA1-DOA. For further analyzing sulfur species and distribution, XPS spectra were recorded as seen in Fig. 6 and Table 5, and the sulfur functional groups contain alkyl sulfur and thiophene sulfur and sulfur oxide (Wang et al., 2021). SDA2-DOA had higher sulfur content but a little lower thiophene sulfur content than that of SDA1-DOA, which could be the reason that SDA2-DOA had higher conversion ratio under the same reaction conditions. As illustrated in Fig. 2, the sum of S-containing structures (S1, S2, S3, S1O1, S1O2, S1O3, S1O4, N1S1, and N1S1O1) of SDA2-DOA was a little lower than that of SDA1-DOA, which could give an indirect identification of lower thiophene sulfur content for SDA2-DOA.

Close

Fig. 6

XPS spectra curve-resolution of S in SDA1-DOA (left) and SDA2-DOA (right).

Export to PPT

3.2.3 Stability analysis for DOA in slurry-phase hydrocracking process

m_Tot means total mass of the feed.

Coking resulting from asphaltene precipitation is still one of the biggest concerns in heavy oil processing industry and causes equipment blocking and catalyst deactivation (Wiehe, et al., 2000; Wiehe, et al., 2001; Xu, 2018; Wang et al., 2009; Liu et al., 2019); and toluene insoluble (TI) content obtained by GB/T2292 method is an important coking index for SHC process. As seen in Table 6 for SDA1-DOA, ${\mathrm{Y}}_{\mathrm{S}\mathrm{H}\mathrm{C}}^{\mathrm{T}\mathrm{I}}$ increased to 2.39 % from 1.62 % and ${\mathrm{Y}}_{\mathrm{S}\mathrm{H}\mathrm{C}}^{\mathrm{A}\mathrm{s}\mathrm{p}}$ decreased to 1.18 % from 4.13 % when reaction time increased to 240 min from 120 min respectively. For SDA2-DOA, ${\mathrm{Y}}_{\mathrm{S}\mathrm{H}\mathrm{C}}^{\mathrm{T}\mathrm{I}}$ increased to 3.10 % from 2.14 % when reaction time increased to 240 min from 120 min respectively, with opposite trend for ${\mathrm{Y}}_{\mathrm{S}\mathrm{H}\mathrm{C}}^{\mathrm{A}\mathrm{s}\mathrm{p}}$ . The overall conversion of asphaltene decomposition including coke formation has been found to follow first-order reaction (Phama et al., 2020), and the quantity and quality of asphaltenes play main factor for coke formation in residue hydroprocessing (Almutairi et al., 2007). The quality of asphaltenes of SDA1-DOA and SDA2-DOA was the same, but the quantity of asphaltenes followed the order SDA2-DOA > SDA1-DOA. SDA2-DOA (C7-Asphaltene% = 11.1 %) had higher asphaltene content than that of SDA1-DOA (C7-Asphaltene% = 8.5 %), which could be the main reason that SHC2-120 min and SHC2-240 min reactions have higher ${\mathrm{Y}}_{\mathrm{S}\mathrm{H}\mathrm{C}}^{\mathrm{T}\mathrm{I}}$ than that of SHC1-120 min and SHC1-240 min reactions respectively.

Except for the quantity of ${\mathrm{Y}}_{\mathrm{S}\mathrm{H}\mathrm{C}}^{\mathrm{A}\mathrm{s}\mathrm{p}}$ and ${\mathrm{Y}}_{\mathrm{S}\mathrm{H}\mathrm{C}}^{\mathrm{T}\mathrm{I}}$ , it’s important to investigate asphaltene structure changing rules in high conversion process to evaluate the reaction. The comparisons of asphaltenes analyzed by FT-ICR MS were illustrated in Fig. 7. As illustrated in Fig. 7, typical aromatic compounds including HC, S-containing, O-containing, and N-containing structures of secondary asphaltenes were detected, and N1 compounds were the most abundant aromatic compounds and had the most significant increase with increasing the reaction time. Furthermore, the organic nitrogen compounds present in the feed in hydrocracking process, act as gum and coke precursors that cause asphaltene precipitation and catalyst inhibition and deactivation (Sau, et al., 2005; Liu, et al., 2012; Bello et al., 2021) DBE versus CN distributions of aromatic N1 compounds of asphaltenes were shown in Fig. 8, and the size of the bubble indicates the relative abundance of the compound. For those asphaltenes, the carbon number distribution varied from about 30 to 70, and the DBE distribution varied from about 15 to 45. The size of the center got smaller, which could be derived from lighter components or coke formation by asphaltene conversion. For further analyzing the asphaltene structure, DBE distribution of N1 compounds of different asphaltenes was shown in Fig. 9. As a result, in all asphaltenes, the N1 class species distributed in a broad range of DBE (15 ∼ 40), with corresponding distribution centers at 26–31. When DBE greater than 31, the relative abundance of N1 compounds followed the order SHC2-240 min-Asp > SHC1-240 min-Asp, which agreed with the trend for TI yields.

Close

Fig. 7

The aromatic species of the different asphaltenes by FT-ICR MS.

Export to PPT

Close

Fig. 8

DBE vS carbon number distribution of aromatic N1 compounds of different asphaltenes.

Export to PPT

Close

Fig. 9

DBE distribution of N1 compounds of different asphaltenes.

Export to PPT