Feasibility Assessment on Remanufacturing of Ni–Mo/ γ –Al 2 O 3 Catalyst for Residue Hydrodesulfurization

: Residue hydrodesulfurization (RHDS) is a critical process in the petroleum reﬁning industry for removing sulfur compounds from heavy residual oils. However, catalysts used in RHDS can easily be deactivated by numerous factors, leading to reduced process efﬁciency and economic beneﬁts. The remanufacturing of spent catalysts can be a useful strategy for extending the lifespan of catalysts, reducing waste, and improving process sustainability. This paper proposes an effective catalyst remanufacturing process for commercial RHDS catalysts. In detail, sequential unit processes including oil washing (OW), complete incineration (CI), and acid leaching (AL) were conducted to remanufacture the spent RHDS catalysts. We also highlight some of the key challenges in remanufacturing catalysts, such as the key factors involved in catalyst deactivation. Finally, we provide future perspectives on the development of an effective catalyst remanufacturing process for RHDS, with the goal of improving the efﬁciency, sustainability, and competitiveness of the petroleum reﬁning industry.


Introduction
Currently, the global consumption of catalysts surpasses 800,000 t/y, with approximately 30% being hydrodesulphurization (HDS) catalysts utilized in the petroleum refining sector [1]. HDS catalysts are mainly employed in the oil refining sector to eliminate sulfur heteroatoms, which in turn, improves the quality of intermediate and final products [2]. In the process of heterogeneous catalysis, compounds with heteroatoms such as Ni and V, as well as polycyclic aromatic hydrocarbons, colloids, and asphaltenes present in crude oil, can be accumulated on the catalyst surface in the form of NiS x , VS 2 , VS 4 , coke, and other substances [2]. The accumulation by these metal sulfates and coke deposition induced significant catalyst deactivation [3,4]. Accordingly, the HDS catalysts have a short lifespan of approximately 6-12 months [5]. This is the reason that the amount of spent catalyst discarded from the hydrotreating unit is always greater than the amount of fresh catalyst loaded in the reactor.
The hydroprocessing units in refineries are responsible for a significant portion of spent catalysts due to the substantial use of catalysts in the hydrotreating process, which is utilized to refine and enhance various petroleum streams and residues. In addition, the amount of spent catalysts discharged as solid waste has increased significantly in recent years due to the rapid growth of distillate hydrotreating capacity to meet the growing demand for ultra-low sulfur transportation fuels, the increase in the processing of heavier feedstocks containing high sulfur and metal content, and the rapid deactivation and unavailability of reactivation processes for residue hydroprocessing catalysts. In addition, as the growth engine for the rebound from the economic recession caused by the COVID-19 pandemic in the last three years is based on fossil fuels, this prospect is expected to continue [6,7]. From an environmental point of view, spent catalysts have been designated as hazardous waste by U.S. Environmental Protection Agency because of their self-heating and toxic nature. Metal components (Co, Ni, V) from the spent catalysts can be leached into water, thereby contaminating the aquatic environment. Additionally, there is a risk of releasing harmful gases such as HCN when spent catalysts with deposited carbon come into contact with water [8]. Therefore, including the above reasons, the strict environmental regulations concerning the handling and disposal of spent catalysts have led to significant interest in the research and development of processes for recycling and reutilizing spent hydrotreating catalysts.
Three methods are generally employed to minimize the amount of spent catalyst: (1) utilizing in less challenging hydrotreating or alternative processes, (2) rejuvenation and regeneration, and (3) decreasing the consumption of catalysts by utilizing improved catalysts with extended lifespans. However, research into the development of new catalysts that outperform commercial catalysts is still in progress. Therefore, research on alternative ways to utilize spent catalysts or reactivate them is currently being actively pursued [9].
Several recent studies have investigated various ways to utilize or regenerate spent hydrotreating catalysts. Srour et al. reported the applicability of the non-thermal plasma regeneration process of the spent CoMoP/Al 2 O 3 catalyst for HDS. It has been found that, despite the presence of CoMoO 4 and other oxide species that impede the promotional role of cobalt, HDS catalysts regenerated through non-thermal plasma exhibit higher hydrogenation activity than those regenerated through conventional thermal means in industrial units [10]. Gao et al. applied a roasted spent HDS ash as a catalytic promoter. Although there is a problem of gaseous pollutants such as S and N being released during the thermal decomposition process, the potential for self-circulation of the spent oily HDS catalyst and value-added energy recovery through thermal cracking has been verified [11]. Yang et al. proposed a microwave pyrolysis method as a strategy for removing oil content from the spent HDS catalyst, and confirmed that it was effective in removing elemental carbon including oil [12]. In addition, various studies have been conducted on the recovery of valuable metals from spent catalysts through leaching [2,[13][14][15][16]. However, most of the studies have focused only on one process of treating the spent catalyst [17][18][19][20], and there have been few studies examining the reusability of the catalyst through remanufacturing.
Overall, the remanufacturing of catalysts for RHDS is an important area of research and development, with potential benefits for improving the efficiency and sustainability of the petroleum refining industry. Further studies are needed to optimize the remanufacturing processes and evaluate their economic and environmental impacts.
In this study, we proposed an effective catalyst remanufacturing process consisting of oil washing (OW), complete incineration (CI), and acid leaching (AL). The physicochemical properties of each catalyst in each unit step were analyzed by various characterization techniques and correlated with their activity results in the HDS of the dibenzothiophene (DBT) reaction. accumulation of vanadium prevents activity recovery by blocking the pore structure and active sites, and sulfur is known as a representative catalyst poison due to its strong adsorption to reactants [21][22][23][24]. After complete incineration (NMA -CI), it was confirmed that approximately 87% of the sulfur component was successfully removed as planned in the catalyst remanufacturing process proposed in this study. For the leaching of vanadium remaining after complete incineration, acid leaching (NMA -AL) was performed using oxalic acid solution. It was confirmed that most of the vanadium was removed by showing 0.27 wt.%. However, it was confirmed that Ni (↓ 47%) and Mo (↓ 89%), which can act as active species, also significantly decreased during this process. As a result, vanadium and sulfur corresponding to contaminants were successfully removed through the catalyst remanufacturing process, but Ni and Mo corresponding to active species were also leached.

Textural Properties of the Ni-Mo/γ-Al 2 O 3 Catalyst during the Remanufacturing Process
As mentioned above, it is well known that vanadium and sulfur contained in residue oil affect the pore structure to deactivate the catalyst. Therefore, in this section, the N 2 physisorption results were presented to examine the change in the pore structure by the catalyst remanufacturing process. The textural properties of the catalysts during the remanufacturing process are given in Table 2. The specific surface area (S.A.) decreases after the RHDS reaction. This is due to the pore blocking phenomenon by the accumulation of residual contaminants (V, S, C) [21][22][23]. The specific surface area of the catalyst was restored to almost the same level as that of the fresh catalyst (NMA -F) after complete incineration (NMA -CI) through the catalyst remanufacturing process consisting of sequential processes of complete incineration and acid leaching. In addition, the specific surface area of NMA -AL, which is treated by acid leaching, showed a higher value than that of NMA -F. As shown in the ICP-OES results, this can be explained by the simultaneous leaching of V and S contaminants as well as active metal components. The total pore volume also showed the same trend as the specific surface area. Figure 1 depicts the N 2 adsorption/desorption isotherms of the catalysts. Catalysts in the entire catalyst remanufacturing process, including the NMA -F catalyst, showed adsorption branches of type IV classified by the International Union of Pure and Applied Chemistry (IUPAC) [25]. These results indicate that all catalysts have mesoporous structures [26][27][28]. In this case, if the diameter of the pore is wider than 4 nm, then hysteresis will occur alongside capillary condensation. All catalysts except the NMA -OW catalyst showed an H1 hysteresis loop. The presence of the H1 hysteresis loop is indicative of materials with a narrow range of uniform mesopores [25]. On the other hand, the NMA -OW catalyst exhibited an H2 hysteresis loop. Pore structures that are more complex are associated with the H2 hysteresis loop. This phenomenon can be caused by pore blocking [21][22][23]29]. Since the NMA -OW catalyst is in almost the same state as the catalyst after the RHDS reaction, it can be expected to exhibit such a pore structure for pore blocking by contaminants such as V and S in the residue oil. teresis loop. Pore structures that are more complex are associated with the H2 hysteresis loop. This phenomenon can be caused by pore blocking [21][22][23]29]. Since the NMA -OW catalyst is in almost the same state as the catalyst after the RHDS reaction, it can be expected to exhibit such a pore structure for pore blocking by contaminants such as V and S in the residue oil.  To further investigate the changes in the pore structures during the catalyst remanufacturing process, the pore size distribution was analyzed by the desorption branch using the Barrett-Joyner-Halenda (BJH) theory, as shown in Figure 2 and Table 3. During the catalyst remanufacturing process, the catalysts exhibited uniform pore diameters of 7.6 to 10.9. Notably, the NMA -AL catalyst showed a main pore diameter distribution peak located at 11.2 nm, which is larger than that of the NMA -F catalyst. As revealed in the ICP-OES results, this is because other metals including V were also leached by the acid leaching treatment. This is in agreement with the results reported by Liu et al. [30]. They revealed that the pore volume and average pore diameter decreased as the metal content increased in the NiMo-Al2O3 catalyst. In addition, the NMA -OW catalyst showed a main pore diameter distribution peak of 7.1 nm, which is the smallest size. Bore et al. claimed that when metal particles outgrow the pore size, the nearby pore structure can collapse To further investigate the changes in the pore structures during the catalyst remanufacturing process, the pore size distribution was analyzed by the desorption branch using the Barrett-Joyner-Halenda (BJH) theory, as shown in Figure 2 and Table 3. During the catalyst remanufacturing process, the catalysts exhibited uniform pore diameters of 7.6 to 10.9. Notably, the NMA -AL catalyst showed a main pore diameter distribution peak located at 11.2 nm, which is larger than that of the NMA -F catalyst. As revealed in the ICP-OES results, this is because other metals including V were also leached by the acid leaching treatment. This is in agreement with the results reported by Liu et al. [30]. They revealed that the pore volume and average pore diameter decreased as the metal content increased in the NiMo-Al 2 O 3 catalyst. In addition, the NMA -OW catalyst showed a main pore diameter distribution peak of 7.1 nm, which is the smallest size. Bore et al. claimed that when metal particles outgrow the pore size, the nearby pore structure can collapse [31,32]. Therefore, it can be seen that the clogging of the pore structure occurred due to complex phenomena, including sintering of the active metal components and deposition of contaminants (V, S, C) present in the reactants during the RHDS reaction. Moreover, disordered accumulation of the active species induced by the reaction temperature of RHDS reaction can be another reason for pore blocking. The resulting collapse of the pore structure due to the low surface area lowers the mass transfer rate and acts as a major cause of catalyst deactivation [33,34]. The results of the metal surface area calculated by CO-chemisorption support this hypothesis. The NMA -OW catalyst exhibited the lowest metal surface area (1.19 m 2 /g) among the treated catalysts and increased in the following order: NMA -OW (1.19 m 2 /g) < NMA -AL (3.87 m 2 /g) < NMA -F (8.73 m 2 /g) < NMA -CI (12.46 m 2 /g). The number of active sites estimated from the CO uptake shows the same trend with metal surface area. The lowest value of NMA -OW can be ascribed as the sintering of active species, and is consistent with N 2 physisorption results. Interestingly, NMA -CI showed a higher value than that of NMA -F. This is expected to result from the enhancement of the metal-support interaction due to the additional heat treatment. Additionally, NMA -AL exhibited a value comparable to NMA -OW, despite having a surface area close to five times larger. Generally, a high number of active sites is achieved by a high surface area [35]. In this case, it is determined that sintering occurred due to the metal-support interaction weakened by acid leaching process. Another factor may be sintering caused by a relatively high Ni/Mo ratio, which is due to Mo leaching [36]. structure due to the low surface area lowers the mass transfer rate and acts as a major cause of catalyst deactivation [33,34]. The results of the metal surface area calculated by CO-chemisorption support this hypothesis. The NMA -OW catalyst exhibited the lowest metal surface area (1.19 m 2 /g) among the treated catalysts and increased in the following order: NMA -OW (1.19 m 2 /g) < NMA -AL (3.87 m 2 /g) < NMA -F (8.73 m 2 /g) < NMA -CI (12.46 m 2 /g). The number of active sites estimated from the CO uptake shows the same trend with metal surface area. The lowest value of NMA -OW can be ascribed as the sintering of active species, and is consistent with N2 physisorption results. Interestingly, NMA -CI showed a higher value than that of NMA -F. This is expected to result from the enhancement of the metal-support interaction due to the additional heat treatment. Additionally, NMA -AL exhibited a value comparable to NMA -OW, despite having a surface area close to five times larger. Generally, a high number of active sites is achieved by a high surface area [35]. In this case, it is determined that sintering occurred due to the metal-support interaction weakened by acid leaching process. Another factor may be sintering caused by a relatively high Ni/Mo ratio, which is due to Mo leaching [36].  Table 3. Pore volume distribution of the Ni-Mo/γ-Al2O3 catalysts during the remanufacturing process.

Catalyst
Pore Volume (cm 3 /g) Micropore (dp < 2 nm) Mesopore (2 nm < dp < 50 nm) Macropore (dp > 50 nm)  As a result, most of the textural properties were recovered through the catalyst remanufacturing process, showing about 85% of mechanical strength compared to the fresh catalyst.

Crystal Structures of the Ni-Mo/γ-Al 2 O 3 Catalyst during the Remanufacturing Process
Changes in the crystal structure of the catalyst by the catalyst remanufacturing process were investigated through powder X-ray diffraction (PXRD) analysis, and the results are shown in Figure 3. During the remanufacturing process, no particular crystal structure changes were observed, but some samples showed slight deformation. As shown in Figure 3, all treated catalysts mainly show diffraction peaks characteristic of γ-Al 2 O 3 (PDF-ICDD # 29-0063) [37,38]. For all treated catalysts, diffraction peaks corresponding to the Ni and Mo compounds were not detected due to the low content, as confirmed by the ICP-OES results. Compared to other treated catalysts, the NMA -OW catalyst show a characteristic peak corresponding to NiV 2 S 4 (PDF-ICDD # 36-1132) at 34.966 • [11]. This diffraction peak disappeared after the subsequent remanufacturing process. This result indicates that the metal contaminants mainly exist in the form of NiV 2 S 4 after the RHDS reaction. In addition, in the case of NMA -CI, it was confirmed that the diffraction peaks of MoO 2 (PDF-ICDD # 50-0739) and MoO 3 (PDF-ICDD # 05-0508) were more prominent than those of other treated catalysts [39]. This suggests that MoO 2 and MoO 3 species coexist due to the increase in Mo-Al 2 O 3 interaction by additional heat treatment. These peaks disappeared after acid leaching, confirming that the metal-support interaction is weakened by acid leaching. This result is consistent with the CO-chemisorption results.
Ni and Mo compounds were not detected due to the low content, as confirmed by the IC OES results. Compared to other treated catalysts, the NMA -OW catalyst show a chara teristic peak corresponding to NiV2S4 (PDF-ICDD # 36-1132) at 34.966° [11]. This diffra tion peak disappeared after the subsequent remanufacturing process. This result indicat that the metal contaminants mainly exist in the form of NiV2S4 after the RHDS reaction. addition, in the case of NMA -CI, it was confirmed that the diffraction peaks of MoO (PDF-ICDD # 50-0739) and MoO3 (PDF-ICDD # 05-0508) were more prominent than tho of other treated catalysts [39]. This suggests that MoO2 and MoO3 species coexist due the increase in Mo-Al2O3 interaction by additional heat treatment. These peaks disa peared after acid leaching, confirming that the metal-support interaction is weakened b acid leaching. This result is consistent with the CO-chemisorption results.

Reduction Properties of the Ni-Mo/γ-Al2O3 Catalyst during the Remanufacturing Process
As the HDS reaction applies H2 gas as a co-fed reactant, the reduction property of th applied catalyst can be an important indicator for determining the reaction efficiency. Fi ure 4 presents H2-temperature programmed reduction (H2-TPR) results of catalysts du ing remanufacturing process. All treated catalysts exhibited two reduction peaks at 40 °C and 800 °C, respectively. Generally, reduction of MoO3 has two sequential steps: MoO → MoO2 → Mo. The first low reduction peak ascribed the transformation of the polymer octahedral Mo species from Mo 6+ to Mo 4+ [40,41]. This kind of Mo has amorphous stru ture, and is a high defect multilayer Mo oxide or heteropoly acid. This octahedral Mo

Reduction Properties of the Ni-Mo/γ-Al 2 O 3 Catalyst during the Remanufacturing Process
As the HDS reaction applies H 2 gas as a co-fed reactant, the reduction property of the applied catalyst can be an important indicator for determining the reaction efficiency. Figure 4 presents H 2 -temperature programmed reduction (H 2 -TPR) results of catalysts during remanufacturing process. All treated catalysts exhibited two reduction peaks at 400 • C and 800 • C, respectively. Generally, reduction of MoO 3 has two sequential steps: MoO 3 → MoO 2 → Mo. The first low reduction peak ascribed the transformation of the polymeric octahedral Mo species from Mo 6+ to Mo 4+ [40,41]. This kind of Mo has amorphous structure, and is a high defect multilayer Mo oxide or heteropoly acid. This octahedral Mo is considered as active precursor in the HDS reaction [42]. The second high temperature reduction peak indicates the reduction of the polymeric octahedral Mo 4+ and intensive reduction of the tetrahedral Mo species [43,44]. This type of Mo is highly dispersed Mo and MoO 3 . Mo 4+ species has a negative effect on sustainability for the Ni-Mo-S phase [45]. In particular, the first reduction peak of NMA -OW shifted to a lower temperature range. This is due to the enhanced reducibility by the additional reduction of the sulfur compounds present in the spent catalyst, which is consistent with the result of Liu et al. [46]. In addition, it can be seen that the Mo 4+ reduction peak near 800 • C is also rarely observed, and the metal-support interaction has been weakened due to the improved reducibility. On the other hand, after the complete incineration, the first reduction peak shifted to a high temperature and a new reduction peak was observed at 500 • C. The H 2 consumption peak at 500 • C is known to be due to the reduction of NiMoO 4 [47,48]. In addition, the intensity of the reduction peak corresponding to Mo 4+ was also emphasized.
It suggests that the strong metal to support interaction (SMSI) of the catalyst is promoted by the complete incineration process. This result is well matched with the CO-chemisorption and PXRD analysis results. After the acid leaching process, the H 2 consumption of all reduction peaks was reduced. This is the result of an additional metal (Ni, Mo) leaching and is consistent with the results of ICP-OES and CO-chemisorption.
On the other hand, after the complete incineration, the first reduction peak shifted high temperature and a new reduction peak was observed at 500 °C. The H2 consum peak at 500 °C is known to be due to the reduction of NiMoO4 [47,48]. In addition intensity of the reduction peak corresponding to Mo 4+ was also emphasized. It sug that the strong metal to support interaction (SMSI) of the catalyst is promoted by the plete incineration process. This result is well matched with the CO-chemisorption PXRD analysis results. After the acid leaching process, the H2 consumption of all re tion peaks was reduced. This is the result of an additional metal (Ni, Mo) leaching a consistent with the results of ICP-OES and CO-chemisorption.

FT-IR Analysis of the Ni-Mo/γ-Al2O3 Catalyst during the Remanufacturing Process
In order to clarify the structural changes and residual contents during the cat remanufacturing process, Fourier-transform infrared spectroscopy (FT-IR) spectra o Ni-Mo/γ-Al2O3 catalysts during the remanufacturing process were collected in the r of 1000-4000 cm −1 . According to the results presented in Figure 5, all treated catalys hibit the broad absorption band located at 3433 cm −1 and an absorption band with intensity located at 1630 cm −1 , which are due to the stretching vibration of the O-H and the bending vibration of the H-O-H bond of the adsorbed H2O molecule, respect [49]. These absorption bands indicate that H2O molecules are linked to nanopartic the treated catalysts. The NMA -OW catalyst exhibited a slightly different spectrum pared to other catalysts. In particular, the absorption bands located at 1400, 1450, and cm −1 were further detected. These peaks correspond to branched-chain and cyclic par (1400 cm −1 ), straight chain olefins (1450 cm −1 ), and the C-H bond of organic residues

FT-IR Analysis of the Ni-Mo/γ-Al 2 O 3 Catalyst during the Remanufacturing Process
In order to clarify the structural changes and residual contents during the catalyst remanufacturing process, Fourier-transform infrared spectroscopy (FT-IR) spectra of the Ni-Mo/γ-Al 2 O 3 catalysts during the remanufacturing process were collected in the range of 1000-4000 cm −1 . According to the results presented in Figure 5, all treated catalysts exhibit the broad absorption band located at 3433 cm −1 and an absorption band with low intensity located at 1630 cm −1 , which are due to the stretching vibration of the O-H bond and the bending vibration of the H-O-H bond of the adsorbed H 2 O molecule, respectively [49]. These absorption bands indicate that H 2 O molecules are linked to nanoparticles in the treated catalysts. The NMA -OW catalyst exhibited a slightly different spectrum compared to other catalysts. In particular, the absorption bands located at 1400, 1450, and 2924 cm −1 were further detected. These peaks correspond to branchedchain and cyclic paraffin (1400 cm −1 ), straight chain olefins (1450 cm −1 ), and the C-H bond of organic residues (2924 cm −1 ), respectively. It also showed a distinct absorption band located at 1088 cm −1 corresponding to the asymmetrical and symmetrical C-O-C ring stretching vibration. These findings confirm that residual heavy oil components resulting from the RHDS reaction remained in the spent catalyst [50][51][52]. Catalysts from the subsequent sequential remanufacturing process did not show an absorption band for heavy oil-containing materials. Therefore, it was confirmed that the impurities were successfully removed through the proposed remanufacturing process in this study. remained in the spent catalyst [50][51][52]. Catalysts from the subsequent sequential rem facturing process did not show an absorption band for heavy oil-containing mater Therefore, it was confirmed that the impurities were successfully removed through proposed remanufacturing process in this study.

Quantitative Investigation of Impurities in the Ni-Mo/γ-Al2O3 Catalyst during the Rem ufacturing Process
A quantitative investigation of impurities in the Ni-Mo/γ-Al2O3 catalyst during remanufacturing process was achieved through thermogravimetric analysis (TGA). ure 6 depicts the TG weight loss curves as a function of temperature for catalysts for remanufacturing process. The weight loss in the first region corresponding to the tem ature range of 20-200 °C is due to residual moisture. The weight loss in the second reg corresponding to the temperature range of 200-400 °C, is attributed to the devolatiliza of residual oil components. Accordingly, the spent catalyst (NMA -OW) showed the est weight loss. It is well known that the weight loss observed after 400 °C is cause the deposited coke and decomposable inorganic compounds, such as sulfates [37]. weight loss that occurs after 600 °C is due to heavier carbon deposition and sulfate composition. The deposition of coke is usually associated with the diffusion length reactants and products must cover within the porous catalysts. If the diffusion leng longer, it can result in the generation of more coke, which may be of a heavier type For the above reasons, the weight loss in the temperature range of 400-1100 °C was s a value for quantitative investigation of impurities. The weight loss by carbon-and su containing compounds decreased in the following order: NMA -OW (23%) > NMA (8%) > NMA -AL (4%). This result can be explained by the recovery of the pore struc through the remanufacturing process revealed by the N2 physisorption analysis. In a tion, it is consistent with the FT-IR and ICP-OES results, confirming that impuritie removed by the remanufacturing process.

Quantitative Investigation of Impurities in the Ni-Mo/γ-Al 2 O 3 Catalyst during the Remanufacturing Process
A quantitative investigation of impurities in the Ni-Mo/γ-Al 2 O 3 catalyst during the remanufacturing process was achieved through thermogravimetric analysis (TGA). Figure 6 depicts the TG weight loss curves as a function of temperature for catalysts for each remanufacturing process. The weight loss in the first region corresponding to the temperature range of 20-200 • C is due to residual moisture. The weight loss in the second region, corresponding to the temperature range of 200-400 • C, is attributed to the devolatilization of residual oil components. Accordingly, the spent catalyst (NMA -OW) showed the largest weight loss. It is well known that the weight loss observed after 400 • C is caused by the deposited coke and decomposable inorganic compounds, such as sulfates [37]. The weight loss that occurs after 600 • C is due to heavier carbon deposition and sulfate decomposition. The deposition of coke is usually associated with the diffusion length that reactants and products must cover within the porous catalysts. If the diffusion length is longer, it can result in the generation of more coke, which may be of a heavier type [53]. For the above reasons, the weight loss in the temperature range of 400-1100 • C was set as a value for quantitative investigation of impurities. The weight loss by carbon-and sulfur-containing compounds decreased in the following order: NMA -OW (23%) > NMA -CI (8%) > NMA -AL (4%). This result can be explained by the recovery of the pore structure through the remanufacturing process revealed by the N 2 physisorption analysis. In addition, it is consistent with the FT-IR and ICP-OES results, confirming that impurities are removed by the remanufacturing process.

HDS Activity Test for the Remanufactured Ni-Mo/γ-Al2O3 Catalyst
In order to examine the reusability of the remanufactured catalyst, fre spent (NMA -OW), and remanufactured catalysts (NMA -AL) were appl of DBT, and their performance was compared and evaluated at the liquid velocity (LHSV) of 50 h −1 . The results are presented in Figure 7 and Tabl illustrates the DBT conversion as a function of the reaction temperature from The catalytic activity decreased by approximately 35-40% when the spen reused in the HDS of the DBT reaction. Compared to the spent catalyst, t tured catalyst demonstrated an approximately 10% improvement in catalyt reaction rate and HDS kinetic constant exhibited a similar trend to that of D and decreased in the following order: NMA -F (6.37 μmolDBT/gcat•s, 14.3 × NMA -AL (4.59 μmolDBT/gcat•s, 4.3 × 10 −7 mol/g•s) > NMA -OW (4.06 μm × 10 −7 mol/g•s). For an in-depth investigation of the recovery of catalytic remanufacturing process, turnover frequency (TOF) results were obtai chemisorption and the apparent activation energies (Ea) were calculated fr nius equation plot, as shown in Figure 7C. The remanufactured catalyst exh TOF value compared to the spent catalyst, even though it had a greater nu sites. To the contrary, the remanufactured catalyst exhibited a higher appa energy compared to the spent catalyst. This result indicates that the capac recovery is dependent on the performance of a single active site and the nu sites. It is noteworthy that the NMA -AL catalyst has a higher number

HDS Activity Test for the Remanufactured Ni-Mo/γ-Al 2 O 3 Catalyst
In order to examine the reusability of the remanufactured catalyst, fresh (NMA -F), spent (NMA -OW), and remanufactured catalysts (NMA -AL) were applied to the HDS of DBT, and their performance was compared and evaluated at the liquid hourly space velocity (LHSV) of 50 h −1 . The results are presented in Figure 7 and Table 4. Figure 7A illustrates the DBT conversion as a function of the reaction temperature from 350 to 395 • C. The catalytic activity decreased by approximately 35-40% when the spent catalyst was reused in the HDS of the DBT reaction. Compared to the spent catalyst, the remanufactured catalyst demonstrated an approximately 10% improvement in catalytic activity. The reaction rate and HDS kinetic constant exhibited a similar trend to that of DBT conversion and decreased in the following order: NMA -F (6.37 µmol DBT /g cat ·s, 14.3 × 10 −7 mol/g·s) > NMA -AL (4.59 µmol DBT /g cat ·s, 4.3 × 10 −7 mol/g·s) > NMA -OW (4.06 µmol DBT /g cat ·s, 3.4 × 10 −7 mol/g·s). For an in-depth investigation of the recovery of catalytic activity by the remanufacturing process, turnover frequency (TOF) results were obtained from CO-chemisorption and the apparent activation energies (E a ) were calculated from the Arrhenius equation plot, as shown in Figure 7C. The remanufactured catalyst exhibited a higher TOF value compared to the spent catalyst, even though it had a greater number of active sites. To the contrary, the remanufactured catalyst exhibited a higher apparent activation energy compared to the spent catalyst. This result indicates that the capacity for activity recovery is dependent on the performance of a single active site and the number of active sites. It is noteworthy that the NMA -AL catalyst has a higher number of active sites compared to the NMA -OW catalyst, even though the content of active species identified from ICP-OES is absolutely low.  This simultaneous improvement in the number of active sites and the TOF value can be explained by enhanced mechanical properties such as pore diameter and mechanical strength. As shown in Figure 8A,B, the DBT conversion increased linearly with the mean pore diameter and mechanical strength. In addition, this tendency appeared more prominent as the reaction temperature increased. As discussed above, it is possible to recover the pore structure that can come into contact with the reactants through the removal of impurities blocking the pore structure of the catalyst. The enhanced specific surface area has the advantage of lowering the diffusion length and thus limiting the potential for carbon deposition. The greater the specific surface area of a catalyst, the more active sites are available to react with the reactants, which generally results in improved catalytic performance [54]. This may be the reason for having a high number of active sites despite the absolutely low content of active species. In addition, the acquisition of SMSI by additional heat treatment (complete incineration) played an additional role in obtaining a high number of active sites. Figure 8C,D shows the DBT conversion according to the residual vanadium and sulfur content. As expected, these results showed a similar trend to the relationship between mechanical properties and DBT conversion. In summary, impurities such as V, S, and C were successfully removed through the catalyst remanufacturing process to secure improved mechanical properties, and the number of active sites and performance per active site were improved by securing SMSI through additional heat treatment.  This simultaneous improvement in the number of active sites and the TOF value can be explained by enhanced mechanical properties such as pore diameter and mechanical strength. As shown in Figure 8A,B, the DBT conversion increased linearly with the mean pore diameter and mechanical strength. In addition, this tendency appeared more prominent as the reaction temperature increased. As discussed above, it is possible to recover the pore structure that can come into contact with the reactants through the removal of impurities blocking the pore structure of the catalyst. The enhanced specific surface area has the advantage of lowering the diffusion length and thus limiting the potential for carbon deposition. The greater the specific surface area of a catalyst, the more active sites are available to react with the reactants, which generally results in improved catalytic performance [54]. This may be the reason for having a high number of active sites despite the absolutely low content of active species. In addition, the acquisition of SMSI by additional heat treatment (complete incineration) played an additional role in obtaining a high number of active sites. Figure 8C,D shows the DBT conversion according to the residual vanadium and sulfur content. As expected, these results showed a similar trend to the relationship between mechanical properties and DBT conversion. In summary, impurities such as V, S, and C were successfully removed through the catalyst remanufacturing process to secure improved mechanical properties, and the number of active sites and performance per active site were improved by securing SMSI through additional heat treatment.

Catalyst
To confirm the degree of catalyst activity recovery by the catalyst remanufacturing process, a Ni-Mo/γ-Al2O3 catalyst purchased from Advanced Refining Technologies LLC (Richmond, California, United States) was used. The catalysts were labeled according to the remanufacturing process as follows: NMA -F (fresh catalyst), NMA -OW (spent catalyst washed by n-heptane), NMA -CI (catalyst after complete incineration), and NMA -AL (remanufactured catalyst with acid leaching as the final process). The elemental composition of the catalyst during the remanufacturing process was analyzed using ICP-OES, and the results are shown in Table 1.

Catalyst Remanufacturing Process
A series of processes for remanufacturing the Ni-Mo/γ-Al2O3 catalyst is shown in Scheme 1. Detailed experimental conditions and information for each process are provided in the following sections.

Catalyst
To confirm the degree of catalyst activity recovery by the catalyst remanufacturing process, a Ni-Mo/γ-Al 2 O 3 catalyst purchased from Advanced Refining Technologies LLC (Richmond, California, United States) was used. The catalysts were labeled according to the remanufacturing process as follows: NMA -F (fresh catalyst), NMA -OW (spent catalyst washed by n-heptane), NMA -CI (catalyst after complete incineration), and NMA -AL (remanufactured catalyst with acid leaching as the final process). The elemental composition of the catalyst during the remanufacturing process was analyzed using ICP-OES, and the results are shown in Table 1.

Catalyst Remanufacturing Process
A series of processes for remanufacturing the Ni-Mo/γ-Al 2 O 3 catalyst is shown in Scheme 1. Detailed experimental conditions and information for each process are provided in the following sections.

Catalyst Remanufacturing Process
A series of processes for remanufacturing the Ni-Mo/γ-Al2O3 catalyst is shown in Scheme 1. Detailed experimental conditions and information for each process are provided in the following sections. Scheme 1. Catalyst remanufacturing process of the spent Ni-Mo/γ-Al2O3 catalysts.

Oil Washing
To remove residual oil components contained in the spent catalyst, 100 g of spent catalyst after HDS of heavy oil was washed with 800 mL of n-heptane (98%, DAEJUNG CHEMICALS & METALS, Siheung-si, Republic of Korea). Oil washing treatment was performed for 1 h. Then, the washed spent catalyst was naturally dried in an atmosphere of room temperature and 1 atm for 4 h. Finally, a drying process was performed in a drying oven at 120 °C for 6 h to remove all residual solvents.

Complete Incineration
After residual oil component removal treatment, the catalyst was roasted at 550 °C for 2 h in contact with air using an aspirator. Inside the roasting furnace, 4 trays for mounting catalysts were installed, and metal foam was cut and applied as a mesh network for stable loading of catalysts.

Acid Leaching
Acid leaching was performed using an aqueous solution of 5 wt.% oxalic acid (98%, Sigma-Aldrich, St. Louis, MI, USA) to remove vanadium-containing materials present in the spent catalyst, which had undergone sequential treatments of oil washing and complete incineration. A suspension of a mixture of 5 wt.% oxalic acid aqueous solution and spent catalyst was sonicated for 7 min at 50 °C at an amplitude of 40 kHz. Then, the resulting catalyst in suspension was washed several times with deionized water and dried at 150 °C for 2 h.

Catalyst Characterization
The elemental composition of each catalyst in each unit step of remanufacturing was determined by ICP-OES using PerkinElmer Optima 5300 DV equipment (PerkinElmer, Waltham, MA, USA).
The textural properties of the catalysts during the remanufacturing process were analyzed by N2 physisorption isotherms at −196 °C using an ASAP 2010 apparatus (Micromeritics, Norcross, GA, USA). To eliminate the contaminants and remaining gaseous components from the sample, the samples were subjected to a degassing process for 12 h at a temperature of 110 °C under a vacuum of less than 0.5 mmHg.
CO-chemisorption was conducted using an Autochem II 2920 (Micromeritics, Norcross, GA, USA) apparatus. Prior to the analysis, 0.1 g of the sample was in situ reduced in a 5% H2/Ar atmosphere at 395 °C for 1 h and then cooled to 60 °C in an Ar atmosphere. Next, CO pulses with 10% CO/He were introduced onto the catalyst surface until the catalyst reached its saturation point with CO. CO uptake was estimated using a thermal conductivity detector and utilized for the determination of active metal dispersion.

Oil Washing
To remove residual oil components contained in the spent catalyst, 100 g of spent catalyst after HDS of heavy oil was washed with 800 mL of n-heptane (98%, DAEJUNG CHEMICALS & METALS, Siheung-si, Republic of Korea). Oil washing treatment was performed for 1 h. Then, the washed spent catalyst was naturally dried in an atmosphere of room temperature and 1 atm for 4 h. Finally, a drying process was performed in a drying oven at 120 • C for 6 h to remove all residual solvents.

Complete Incineration
After residual oil component removal treatment, the catalyst was roasted at 550 • C for 2 h in contact with air using an aspirator. Inside the roasting furnace, 4 trays for mounting catalysts were installed, and metal foam was cut and applied as a mesh network for stable loading of catalysts.

Acid Leaching
Acid leaching was performed using an aqueous solution of 5 wt.% oxalic acid (98%, Sigma-Aldrich, St. Louis, MI, USA) to remove vanadium-containing materials present in the spent catalyst, which had undergone sequential treatments of oil washing and complete incineration. A suspension of a mixture of 5 wt.% oxalic acid aqueous solution and spent catalyst was sonicated for 7 min at 50 • C at an amplitude of 40 kHz. Then, the resulting catalyst in suspension was washed several times with deionized water and dried at 150 • C for 2 h.

Catalyst Characterization
The elemental composition of each catalyst in each unit step of remanufacturing was determined by ICP-OES using PerkinElmer Optima 5300 DV equipment (PerkinElmer, Waltham, MA, USA).
The textural properties of the catalysts during the remanufacturing process were analyzed by N 2 physisorption isotherms at −196 • C using an ASAP 2010 apparatus (Micromeritics, Norcross, GA, USA). To eliminate the contaminants and remaining gaseous components from the sample, the samples were subjected to a degassing process for 12 h at a temperature of 110 • C under a vacuum of less than 0.5 mmHg.
CO-chemisorption was conducted using an Autochem II 2920 (Micromeritics, Norcross, GA, USA) apparatus. Prior to the analysis, 0.1 g of the sample was in situ reduced in a 5% H 2 /Ar atmosphere at 395 • C for 1 h and then cooled to 60 • C in an Ar atmosphere. Next, CO pulses with 10% CO/He were introduced onto the catalyst surface until the catalyst reached its saturation point with CO. CO uptake was estimated using a thermal conductivity detector and utilized for the determination of active metal dispersion.
To examine the reduction properties of the catalysts during the remanufacturing process, H 2 -TPR was conducted at 25 • C up to 1000 • C at 10 • C/min in a 10% H 2 /Ar atmosphere using an Autochem II 2920 (Micromeritics) instrument.
FT-IR spectra were obtained using a PerkinElmer L1600300 Spectrum Two LiTa (PerkinElmer, Waltham, MA, USA) in the range of 650-4000 cm −1 with 1 cm −1 resolution.
The quantitative analysis of carbon deposition and residual sulfur containing components on the catalyst was conducted with a thermogravimetry analyzer (TGA, Hitachi STA 7300, Minato City, Tokyo, Japan). The catalyst sample of 5-10 mg was heated from 25 to 1100 • C with a heating rate of 10 • C/min in an air atmosphere.
The mechanical strength of each catalyst in each unit step of the remanufacturing was confirmed through compression stress at maximum compressive load using an Instron 5848 (INSTRON, Norwood, MA, USA) instrument.

HDS Performance Test
The lab-scale HDS of DBT reaction was performed using a micro-pilot CATATEST ; m represents the loading amount of catalyst (g); x is the DBT conversion (%); AS M is the molar mass of active species; F is the total flow rate (mol·s −1 ); D M is the dispersion of active species; W is the mass of catalyst (g cat ); and X M is the content of active species (g M ·g cat −1 ).

Conclusions
We examined the feasibility of the remanufacturing process for spent heavy oil desulfurization catalysts, which are currently being pulverized for the recovery of valuable metals or landfilled as designated wastes in case of low economic feasibility. It was confirmed that catalyst deactivation factors such as vanadium, sulfur, and carbon-containing materials can be effectively removed through the remanufacturing process. The largest deactivation factor of the RHDS catalyst was the deformation of the pore structure by contaminants and the reduction of the specific surface area. When the contaminants in the pores were effectively removed through the remanufacturing process, it was confirmed that approximately 75% of the activity was recovered compared to the fresh catalyst at a reaction temperature of 395 • C. Above all, the HDS activity was recovered even though most of the catalytic active species were removed by acid leaching, and it was revealed that the most important factor for the RHDS reaction was the mechanical properties (mean pore diameter and mechanical strength) of the catalyst. It is believed that a remanufactured catalyst with performance equivalent to that of the fresh catalyst can be made through research optimization of the remanufacturing process.