Influence of roasting on the structure and leaching behavior of spent residue hydrotreating catalysts

Abstract

Residue hydrotreating is a critical process in petroleum refining. The spent catalysts typically contain Ni (Co) and Mo (W) as active metals and accumulated Ni and V from processed residues during the operation. Substantial amounts of spent residue hydrotreating catalysts (sRHCs) are generated, which pose severe environmental threats and require proper treatment. After roasting, sRHCs can be further utilized efficiently, such as metal leaching and recovery. Herein, two typical sRHCs, spent hydrodesulfurization and hydrodemetallization catalysts, were selected to investigate the phase transformations and structural changes during roasting. It was found that roasting at 600 °C promoted crystal growth and the formation of composite oxides through solid-state reactions compared to roasting at 430 °C, which hindered the leaching of Ni and Fe in H₂SO₄ and NaOH. However, the elevated oxidation states of V induced by high-temperature roasting facilitated V leaching, while aggregated deposited metals improved the leaching of Mo and Al. The spent hydrodemetallization catalysts predominantly formed Ni/Fe–V–O composite oxides, VO₂ and metal sulfates during roasting, whereas the spent hydrodesulfurization catalysts tended to generate Ni–Mo–O composite oxides and V₂O₅. In addition, the influence of the phase structures of the roasted sRHC on their leaching efficiency in H₂SO₄, water, and NaOH solutions was studied. Fe₂O₃ and VO₂ exhibited a superior leaching performance in H₂SO₄, MoO₃, metal sulfates, NiMoO₄, and AlVO₄ demonstrated a higher leaching efficiency in NaOH, and V₂O₅ possessed good leaching efficiency in both H₂SO₄ and NaOH.

---

Introduction

Residue hydrotreating is a critical measure for the efficient utilization of petroleum resources.^1,2 To enhance the hydrogenation and impurity removal performance, a large amount of catalysts are used in the residue hydrotreating unit. However, these catalysts inevitably lose their activity during industrial operation.^3,4 Coke and deposited metal lead to pore blockage and coverage of their active sites. Currently, in countries such as the USA and China, most spent catalysts are classified as hazardous waste. Spent residue hydrotreating catalysts (sRHCs) are produced in quantities over 60 000 tons annually in China. Consequently, the disposal of these sRHCs has become a significant challenge.⁵

sRHCs mainly contain coke, Al₂O₃ and metal sulfides. Metal sulfides are sulfides of Ni (or Co) and Mo (or W) active metals, as well as sulfides of deposited metals such as V, Ni, Fe, and Na.^6,7 The structure and composition of sRHCs highly depend on the type of corresponding fresh catalyst and the operating conditions.⁴ In the case of the most widely used fixed-bed residue hydrotreating catalysts in industry at present, sRHCs can be classified into two types based on the type of fresh catalysts, i.e., spent hydrodemetallization (HDM) catalysts and spent hydrodesulfurization (HDS) catalysts.^4,8 The main components of fresh HDS and HDM catalysts are similar, consisting of Al₂O₃-supported Ni(Co)–Mo(W). Due to the functional differences, fresh HDS and HDM catalysts exhibit distinct metal loadings and pore structures, with fresh HDS catalysts generally possessing higher Mo and Ni loadings. In industrial applications, fresh HDM and HDS catalysts are used together. The HDM catalysts are filled in the front section of the fixed-bed reactor, while the HDS catalysts are filled in the rear section. HDM catalysts adsorb and accumulate most of the metals existing in the residue oil, mainly including V, Ni and Fe.^4,6,9,10 In contrast, the amount of deposited metal on the spent HDS catalyst is relatively lower. Therefore, the metal content and distribution on spent HDM catalysts and HDS catalysts are different, which have a potential impact on the structure of sRHCs, as well as the further treatment of sRHCs.

In terms of treatment methods, a large amount of spent catalysts from decades ago have been landfilled, posing serious environmental risks. Among the current methods, the recovery of metals from sRHCs by chemical methods is considered highly promising.^7,11 On the one hand, the valuable metals on spent catalysts can be recycled and further sold. Notably, the content of metals such as Mo and V on sRHCs is far greater than that of natural minerals.¹² On the other hand, metals such as Al, Ni, and Mo can also be used in the preparation of hydrogenation catalysts to reduce production costs. Therefore, this approach not only reduces hazardous waste but also generates economic benefits. The existing technologies for metal recovery of spent hydrogenation catalysts mainly include hydrometallurgy, pyrometallurgy, chlorination, and bioleaching.^12–14 At present, the most widely studied and applied route is the hydrometallurgy route due to is advantages of low energy consumption, low equipment cost, and mild operation of the solvent process.^3,15–17

The hydrometallurgy route mainly involves sRHC roasting, followed by solvent leaching, and finally metal separation and recovery.¹⁸ Roasting involves removing the covered organic compound including oil and carbon deposition, and converting the metal on sRHCs into species and structures with low difficulty to undergo leaching.^19,20 The organic substances on the surface of sRHCs, such as oil and carbon deposition, can be removed through roasting, while the metal sulfides in sRHCs are converted into metal oxides.³ Then, the target metal will be leached into the solvent to achieve separation from the sRHCs. Extensive works have conducted to improve the metal leaching efficiency.^15,21,22 The leaching performance of various types of solvents containing organic acids, inorganic acids, or ammonium salts has been screened.²³ Currently, the most studied leaching solvents are inorganic acids represented by sulfuric acid (H₂SO₄) and alkalis represented by NaOH and ammonia.^24,25 According to literature findings, sulfuric acid leaching has been proven to be highly effective for complete metal extraction at lower costs,²⁶ while selective extraction of Mo and V can be achieved in alkaline media.²⁷ After the leaching of roasted sRHCs, metals including V, Ni, and Mo can be separated and utilized through methods such as chemical precipitation and solvent extraction.^12,28–30

The performance of metal recovery by the hydrometallurgy route does not entirely depend on the leaching and extraction processes. Previous studies have found that the roasting process can affect the metal recovery performance. It was reported by Zhang et al. that high-temperature roasting facilitates the ammonia leaching of V and Mo from spent hydrotreating catalysts.²⁷ Similarly, Fang observed the positive effect of high-temperature roasting on the leaching rates of Mo and V in NaOH solvent.²⁵ Theoretically, the influence of roasting conditions on the leaching efficiency is mediated by structural modification of the roasted spent catalysts. In the literature, the combustion of metal sulfides under an oxygen-containing atmosphere has been shown to involve several possible sub-processes including conversion of metal sulfides to metal sulfates,³¹ transformation of metal sulfides into other metal sulfides,³² oxidation of metal sulfides to metal oxides, formation of composite oxides through solid-state reaction,²⁵ and crystal aggregation.²⁵ The regulation of these reactions and processes may alter both the chemical composition and physical structure of the roasted catalysts. Besides, metals on sRHCs have different leaching behaviors. Ni and Fe are primarily dissolved as cationic species during leaching, and Mo usually dissolves as anionic species, whereas V and Al can be leached through both cationic and anionic dissolution pathways. The complex interactions among these metals with differing properties further influence the leaching process. The identification of favorable physicochemical structures for specific leaching methods is of critical importance. Therefore, it is necessary to have a clear understanding of the reactions and changes in various sRHCs during roasting, as well as the influence of the catalyst structure after roasting on its leaching behavior in common solvents.

Herein, typical spent HDM and HDS catalysts were roasted at 430 °C or 600 °C. The effects of temperature and the properties of sRHCs on the reactions and processes during roasting were investigated by characterizing the roasted sRHCs using various bulk and surface analysis techniques. The roasted sRHC powders were leached in H₂SO₄ solution, deionized water, and NaOH solution and the leaching residues were further analyzed for their phase compositions. The relationship between the physicochemical structures of the roasted sRHCs and leaching behaviors of the metal elements in different solvents were correlated. The findings enhance the understanding of the processes during sRHC roasting and their impact on metal leaching, contributing to the design of methods for the recovery of metals from sRHCs and other spent petrochemical catalysts.

Experimental

Materials and regents

The sRHCs were provided by a refinery in Sinopec Group, which were run for more than 1.5 years in a residual oil hydrogenation unit. Both sRHCs were Ni–Mo@Al₂O₃ catalysts when they were produced. H₂SO₄ (98.0%) and NaOH (99.5%) were of analytical grade and received from Modern Oriental Fine Chemistry, China. All reagents were used as received.

Experimental procedures

The sRHCs were roasted prior to leaching. Approximately 50 g of sRHC was placed in a muffle furnace, where the temperature was increased at a rate of 2 °C min⁻¹ to the target temperature of 430 °C or 600 °C and maintained for 3 h. A continuous airflow of 70 mL min⁻¹ was supplied to ensure sufficient oxygen. The roasted sRHC was then ground and sieved, with particles below 60 mesh used for subsequent leaching experiments. The roasted samples were labeled as HDM-430-O, HDM-600-O, HDS-430-O and HDS-600-O, where 430 and 600 represent the roasting temperature and O represents oxidized.

The leaching experiments were conducted in beakers. The leaching behaviors of metals under 1 mol L⁻¹ H⁺ or OH⁻ solvents were investigated, and deionized water was also tested as a reference. A mixture of 2 g roasted sRHC and 60 mL leaching solvent was reacted at 70 °C for 20 min with stirring at a speed of 120 rpm. Three leaching solvents were employed including 0.5 mol L⁻¹ H₂SO₄ solution, deionized water, and 1 mol L⁻¹ NaOH solution. Upon completion of the leaching process, all leachates were collected for analysis. The mass fraction of metal irons in the leachates was determined by inductively coupled plasma optical emission spectroscopy (ICP-OES). The metal leaching efficiency (E_f) was calculated according to the following equation:

E_f = (W₁ × F₁)/(W₂ × F₂) × 100%

where W₁ is the weight of the leachate, F₁ is the mass fraction of metal ion in the leachate (mg kg⁻¹), W₂ is the weight of roasted catalyst, and F₂ is the mass fraction of metal in the roasted catalyst (mg kg⁻¹).

After leaching, the solid residues were collected by centrifugation, and then washed three times with deionized water. The leaching residues were dried at 120 °C and weighed.

Measurements and characterization

The phase structures of the spent catalyst, roasted products and leaching residues were characterized by X-ray diffraction (XRD). The XRD patterns were collected on an X’Pert diffractometer (PANalytical, Netherlands) equipped with a secondary graphite monochromator operating at 40 kV and 30 mA and employing nickel-filtered Cu Kα radiation (λ = 0.154 nm). The contents of C and S in the sRHCs were measured using an EMIA carbon/sulfur analyzer (Horiba, Japan). Transmission electron microscopy (TEM) images of the catalysts were recorded on a Tecnai G2 F20 S-TWIN microscope (Hillsboro, USA). The elemental distribution was examined using a JXA-8230 electron probe microanalyzer (EPMA) (JEOL, Japan). The analyses were executed at an accelerating voltage of 15 kV. A TA-SDT Q600 (TA, USA) thermogravimetric analysis system assembled with a mass spectrometry detector (TG–MS) was adjusted to operate under the following conditions: dynamic atmosphere of 10 mL min⁻¹, heating rate of 10 °C min⁻¹, and endpoint temperature of 800 °C. The mass spectrometry signal m/z = 44 represents CO₂ and m/z = 64 represents SO₂. The Fourier transform infrared spectrometry (FT-IR) was performed using a VERTEX70 (BRUKER, Germany) spectrometer to detect the functional groups in the roasted sRHC. The X-ray photoelectron spectroscopy (XPS) spectra of the roasted sRHC were recorded on a VG ESCALAB 250 XPS spectrometer (Thermo Fisher, USA). The C 1s contamination peak at 284.8 eV was used for binding energy calibration. The metal contents of the roasted sRHC were determined on a 3271E X-ray fluorescence (XRF) analyzer (Rigaku, Japan). The pore characteristics of the roasted sRHC were detected by N₂ adsorption and desorption tests using an ASAP 2400 instrument (Micromeritics, USA). N₂ adsorption isotherms were measured at 77 K after degassing the samples under a vacuum of 6 mTorr at 200 °C for 4 h. The pore size distribution was calculated using the BJH method.

Results and discussion

Characterization of spent catalysts

The structure of spent HDM and HDS catalysts may have significant differences, and thus selected as the typical sRHC for this study. The HDS and HDM catalysts selected for this study were both Ni–Mo@Al₂O₃-type catalysts. The main phase composition of the two sRHCs were characterized using XRD technology (Fig. 1(a)). The XRD pattern of the spent HDM catalyst mainly corresponded to Al₂O₃ and well-crystallized V₃S₄ or NiV₂S₄. The diffraction peaks of V₃S₄ and NiV₂S₄ positioned close to each other, making them difficult to distinguish. The crystallite size of V₃S₄ or NiV₂S₄ was determined to be 23.9 nm, indicating the significant aggregation of V on the spent HDM catalyst. In contrast, the XRD pattern of the spent HDS catalyst primarily showed diffraction peaks from Al₂O₃, with only minor peaks attributed to V₃S₄ or NiV₂S₄. No other metal sulfides were detected, suggesting that they existed as extremely small crystallites or in an amorphous phase.

Fig. 1 Characterization results of the spent catalysts. (a) XRD diffraction patterns and (b) contents of C and S and (c) and (d) TEM images.

Coke and the sulfide state of the metals on the spent catalysts were the key factors hindering metal leaching. The spent HDM catalyst exhibits approximately 7.1 wt% carbon deposition (Fig. 1(b)), which is slightly higher than that of the spent HDM catalyst (6.6 wt%). Notably, the sulfur content on the spent HDM catalyst was about three-times that of the spent HDS catalyst. This was primarily due to the fact that the HDM catalyst adsorbed more metals in the residue oil during hydrotreating than the HDS catalyst, and the metals on the spent catalysts were in the form of metal sulfides. As shown in the TEM images (Fig. 1(c) and (d)), the morphology of both spent catalysts were similar, revealing layered MoS₂ structures in addition to the Al₂O₃ support. MoS₂ in the HDM catalyst was less abundant compared to that in the HDS catalyst.

EPMA was employed to analyze the elemental distribution in the spent catalysts (Fig. 2). In the images, red dots represent the positions where the corresponding elements were most distributed, green dots are moderate, blue dots are less, and black dots are almost non-existent. V, Fe, and S were mostly located near the particle outside, indicating that V and Fe preferentially deposited in the form of metal sulfides on the outer surface of the HDM and HDS catalysts during the hydrotreating of residue oil, and gradually deposited within the catalyst particles. Compared with V, Fe was distributed almost all over the outer surface of the particles. This indicated that Fe in the residue oil was more easily removed compared to V. Ni was distributed both inside and outside the particles, and it could be observed that there was a slight enrichment of Ni on the outer layer. The Ni in outer layer of particles mainly originated from the deposition of Ni in the residue oil, while the Ni located inside of the particles originated from the fresh catalyst itself. The distribution differences of Ni inside and outside of the particles were less than that of V and Fe, indicating that the removal of Ni from the residue oil was more difficult, and thus its deposition on the catalyst particles was deeper. Mo and Al were predominantly distributed within the catalyst particles, which was because Mo was uniformly dispersed on the Al₂O₃ support during the preparation of fresh catalyst. The results showed that the distribution patterns of elements on the spent HDM and HDS catalysts were different. These distribution patterns resulted in V being in closer proximity to co-deposited elements such as Fe and Ni, increasing the likelihood of their solid-state reactions during the roasting processes. Although the elemental distributions of the spent HDM and HDS catalysts were similar, it could be seen from the intensity of the EPMA results that the spent HDM catalyst contained more S and V, while the spent HDS catalyst contained more Al and Mo.

Fig. 2 EPMA images of the (a) spent HDM catalyst and (b) spent HDS catalyst.

The combustion of both spent catalysts under an oxygen-containing atmosphere was further analyzed by TG–MS. The thermogravimetry curve of the spent HDM catalyst displayed three distinct mass loss stages (Fig. 3(a)). The initial mass loss ranging from room temperature to 240 °C was attributed to the desorption of water and partial combustion of reactive metal sulfides. A slight mass increase emerged at 240 °C. One reasonable explanation for this was that the slow oxidation of metal sulfides covered by coke formed metal sulfates, thus resulting in net mass gain. The second mass loss stage initiated at 350 °C, where coke and exposed metal sulfides combusted and predominantly released CO₂ and SO₂. Around 470 °C, the mass loss rate decelerated, forming a second plateau. The subsequent mass loss after 570 °C may be ascribed to the combustion of polymeric sulfur S_x. The spent HDS catalyst exhibited a simpler thermogravimetry curve, with only two continuous mass loss peaks from room temperature to 248 °C and from 248 °C to 547 °C (Fig. 3(b)), which can be attributed to impurity desorption and the combustion of coke and metal sulfides, respectively. No plateau or significant mass increase was observed, indicating that oxidation of the metal sulfides in the spent HDS catalyst mainly generated SO₂.

Fig. 3 (a) and (b) Thermogravimetric analysis results and (c) and (d) the tail gas analysis in the thermogravimetric tests by MS detector of the spent HDM and HDS catalysts.

The composition changes in the exhaust gases during thermogravimetric analysis provided a clearer understanding of the thermal reaction behaviors during roasting (Fig. 3(c) and (d)). The CO₂ generation curves of both spent catalysts exhibited a single peak. In the case of the spent HDM catalyst, CO₂ was generated at approximately 187 °C with its maximum at 414 °C, while for the spent HDS catalyst, it started at about 203 °C with a peak center at 450 °C. This indicated that roasting temperatures below 400 °C were difficult to burn off most of the coke on the sRHC.

The SO₂ generation curves of the spent catalysts could be divided into three regions. The first region in the range of 110 °C–350 °C was attributed to highly dispersed metal sulfides not covered by coke. The second region in the range of 350 °C–580 °C corresponded to most metal sulfides that became exposed as coke burned off. The third region within 600 °C–667 °C was possibly associated with the combustion of aggregated S_x species. Although all three SO₂ regions were observed for the spent HDM catalyst, only the first two regions appeared for the spent HDS catalyst, suggesting fewer S_x species existed. In the second region, the spent HDM catalyst exhibited three SO₂ peaks. The peak at 387 °C and the shoulder peak at 423 °C corresponded to metal sulfides with weak and strong binding to the Al₂O₃ support, respectively. Alternatively, the third shoulder peak at 523 °C corresponded to metal sulfides with high oxidation difficulty, mainly including NiS₂ according to the subsequent XRD results. There is only one SO₂ peak in the second SO₂ region of the spent HDS catalyst due to the scarcity of metal sulfide types, mainly MoS₂, which has strong interaction with the Al₂O₃ support. The results suggest that the spent HDM catalyst required a slightly higher temperature to complete the reaction of metal sulfides during roasting.

Physicochemical properties of the roasted spent catalysts

The processes and reactions during the roasting of sRHC were inferred by analyzing the differences in the physicochemical properties of the roasted spent catalysts. The roasting temperature was selected as 430 °C and 600 °C to represent low- and high-temperature conditions, respectively.

The morphology of the roasted spent catalysts was examined by TEM (Fig. 4). Lattice fringes corresponding to the (2 2 2) facet of Al₂O₃ (d = 0.226 nm) were observed on the rod-like structures that formed the basic framework of the spent catalysts. On the HDM-430-O sample, several 1–2 nm spherical particles were found attached to Al₂O₃ rods, which were attributed to oxides of other metals. Within the yellow dashed regions in Fig. 4, lattice fringes corresponding to the (1 0 1) facet of NiV₂O₆ (d = 0.380 nm) were identified. In the TEM image of HDM-600-O, the metal oxide particles appeared visibly larger. Additionally, NiV₂O₆ and V₂O₅ aggregated into distinct spherical grains with diameters of over ten nanometers (yellow dotted line). The results indicated that high-temperature roasting promoted the migration and agglomeration of metal oxides into larger crystalline grains. In the HDS-430-O sample, no significant spherical particles assigned to metal oxides were detected, and lattice fringes were scarcely observable due to the low metal content and poor crystallinity of the metal oxides. However, in the HDS-600-O sample, spherical metal oxide grains with sizes of 1–3 nm were observed. Lattice fringes corresponding to the (3 1 0) facet of NiMoO₄ were identified, demonstrating that high-temperature roasting promoted the formation of metal composite oxides.

Fig. 4 TEM images of the spent catalyst after roasting (a) HDM-430-O, (b) HDM-600-O, (c) HDS-430-O, and (d) HDS-600-O.

As shown in Fig. 5, the pore structure characteristics of the roasted spent catalysts were further analyzed. Fig. 5(a) showed the pore distribution of the roasted spent catalyst, indicated that the pore structure decreased a lot for HDM catalyst. The N₂ adsorption–desorption isotherms of the roasted catalysts were identified as Type IV based on the IUPAC classification, indicating the presence of mesoporous structures (Fig. 5(b)).^33,34 The roasted HDM catalyst exhibited H3-type hysteresis loops, corresponding to slit-shaped pores formed by platelet particle aggregates.³⁵ Alternatively, the roasted HDS catalyst showed H1-type hysteresis loops, associated with cylindrical pore channels formed by particle aggregation.³⁴ An increase in the roasting temperature resulted in an enhanced N₂ adsorption capacity without significantly altering the shape of the adsorption–desorption isotherms for both roasted HDM and HDS catalysts, suggesting an increase in pore volume with minimal changes in pore morphology. The calculated textural properties are summarized in Table 1. The roasted HDM catalyst possessed a relatively low specific surface area of approximately 50 m² g⁻¹, whereas the roasted HDS catalyst displayed a higher specific surface area of around 130 m² g⁻¹. In the case of both spent HDM and HDS catalysts, the samples roasted at 430 °C exhibited slightly reduced surface areas compared to that roasted at 600 °C, while their pore diameters and volumes increased. Given that the spent catalysts primarily consisted of particle-aggregated pores, this observation implied that high-temperature roasting promoted the grain growth and aggregation of Al₂O₃. The limited expansion of the pore diameters and reduction in specific surface area suggested that the textural changes induced by the roasting temperature had negligible effects on the metal leaching.

Fig. 5 N₂ adsorption–desorption isotherms of the roasted spent catalysts and their BJH pore size distribution.

Table 1 Texture characteristics of the roasted spent catalysts

Sample	HDM-430-O	HDM-600-O	HDS-430-O	HDS-600-O
BET surface area (m^2 g^−1)	55	49	138	129
BJH volume of pores (cm^3 g^−1)	0.12	0.14	0.30	0.39
BJH average pore diameter (nm)	10.8	12.9	10.1	11.8

---

The elemental composition of the roasted spent catalysts is presented in the Table 2. It could be observed that the spent catalysts primarily contained Al, V, Fe, Ni, Mo, and S, along with trace amounts of Na, Si, and P. Regarding metal content, the roasted HDM catalysts exhibited a higher amount of deposited V, Fe and Ni metals compared to the roasted HDS catalysts, whereas the roasted HDS catalysts retained more Al and Mo metals. S was not completely removed after oxidative roasting, with residual levels following the order of HDM-430-O > HDM-600-O > HDS-430-O > HDS-600-O. It should be noted that the remaining S may not exist solely as metal sulfides but could also be present in the form of metal sulfates or S_x. The findings indicated that higher roasting temperatures promoted S removal, and the spent catalysts with a high deposited metal content tended to retain more S.

Table 2 The elemental content of the spent catalyst after roasting

Sample	Content (wt%)
	Na	Al	Si	P	S	V	Fe	Ni	Mo
Note: the results were measured by XRF technology. Results of the elements with a content of less than 0.05 wt% were ignored.
HDM-430-O	0.1	11.4	0.7	0.1	9.2	20.0	2.7	7.9	2.5
HDM-600-O	0.2	17.3	0.6	0.1	4.4	20.5	2.4	8.1	2.5
HDS-430-O	0.1	30.8	0.2	1.0	2.6	3.0	0.5	5.4	13.1
HDS-600-O	0.1	34.0	0.1	0.9	2.2	2.3	0.4	5.0	11.2

---

The chemical bond information of the roasted sRHC was analyzed using the FT-IR technique (Fig. 6). All the samples exhibited characteristic vibration bands corresponding to tetrahedrally coordinated Al–O in alumina (816 cm^−1 36), SO₄²⁻ in sulfate (1110 cm^−1 37,38), and water molecules (1640 cm⁻¹ and 3440 cm⁻¹). The presence of SO₄²⁻ vibration peaks indicated that partial sulfur remained on the catalyst in sulfate form after oxidative roasting. Additionally, peaks of V–O in V₂O₅ (980 cm^−1 39) were observed on the HDM-430-O and HDM-600-O samples. The V₂O₅ peaks on HDM-430-O appeared sharper and well-defined, while HDM-600-O showed smaller shoulder peaks. This suggested that higher roasting temperatures reduced the amount of V₂O₅, possibly due to the enhanced transformation of V₂O₅ into vanadium-containing composite metal oxides.

Fig. 6 FT-IR spectra of the catalyst after roasting.

The chemical states of the Al, Mo, Ni, V, O, and S elements were analyzed using XPS technology. The Al 2p peaks for all the roasted samples were observed at the binding energy of 74.7 eV, corresponding to Al(III) in Al₂O₃ (Fig. 7(a)). The intensity of the Al 2p peak for HDM-430-O was lower than that of HDM-600-O, indicating that after high-temperature roasting, the exposure of Al increased, which might be related to the aggregation of other metal oxides.

Fig. 7 XPS spectra of the catalyst after roasting: (a) Al 2p, (b) Mo 3d, (c) Ni 2p, (d) V 2p 3/2, (e) O 1s, and (f) S 2p.

The V 2p3/2 spectra of all the roasted samples (Fig. 7(d)) exhibited XPS peaks corresponding to both V(IV) and V(V) with binding energies at 516.4 eV and 517.5 eV, respectively.⁶ The relative content of V(IV) followed the order of HDM-430-O > HDM-600-O > HDS-430-O > HDS-600-O. The results demonstrated that an increase in the roasting temperature facilitated the conversion of V(IV) to V(V). Under identical roasting conditions, the roasted HDM samples contained more V(IV) and less V(V), primarily due to the higher content of deposited V hindering the oxidation of V located within the grains. The intensity of the V 2p3/2 peaks followed the order of HDM-600-O > HDM-430-O > HDS-600-O > HDS-430-O. This confirmed that more V was deposited on the HDM catalyst. The enhanced XPS signal of V on the samples roasted at higher temperature implied that the migration and aggregation of V were promoted, leading to greater exposure outside the support pores after grinding.

The Ni 2p spectra of HDM-430-O exhibited five peaks (Fig. 7(c)), which are assigned to Ni(II)–S 2p3/2 at 854.2 eV, Ni(II)–O 2p3/2 at 856.3 eV, a 2p3/2 satellite peak at 862.4 eV, Ni(II)–O 2p1/2 at 874.1 eV, and a 2p1/2 satellite peak at 880.4 eV.³³ The HDM-600-O, HDS-430-O, and HDS-600-O samples showed no evidence of sulfur-bound Ni, displaying only Ni(II)–O related XPS peaks and their satellites. No peaks corresponding to Ni(II) in NiO at 853.7eV were detected in any sample, indicating that Ni primarily existed as metal composite oxides after roasting. The intensity of the Ni(II)–O 2p3/2 peaks followed the order of HDM-600-O > HDM-430-O > HDS-600-O > HDS-430-O, which was similar to that observed for V. This indicated that the spent HDM catalysts had more deposited Ni species than the spent HDS catalysts, and Ni species underwent aggregation and migration during high-temperature roasting.

The Mo 3d spectra of the roasted samples contained two peaks at binding energies of 233.1 eV and 236.3 eV (Fig. 7(b)), which were assigned to the Mo(VI) 3d5/2 and Mo(VI) 3d3/2 levels, respectively.^6,33 Peaks at 229.0 eV and 228.3 eV were observed in the spectra of the HDM-430-O and HDM-600-O samples, corresponding to the S 2s orbital. The S 2s orbital was not detected in the roasted HDS catalysts, which was attributed to their lower sulfur content. No significant variation in the Mo peak intensity with roasting temperature was observed, suggesting the limited aggregation and migration of Mo. This could be attributed to the fact that Mo has strong interaction with Al₂O₃ during the preparation of fresh catalysts, making it less prone to aggregation during spent catalyst roasting compared to the deposited metals such as V.

The O 1s spectra showed broad peaks at 531.6 eV and 532.1 eV (Fig. 7(e)), corresponding to oxygen in metal oxides and sulfate groups, respectively.⁴⁰ Among the roasted samples, HDM-430-O exhibited the highest binding energy for the O 1s orbital, indicating its highest sulfate content.

Three S2p peaks were identified at binding energies of 162.7 eV, 164.0 eV, and 169.5 eV (Fig. 7(f)), corresponding to S₂²⁻, polysulfides S_x, and sulfur in SO₄²⁻, respectively.⁴¹ The distinct spin–orbit splitting was not observed, as the small energy difference Δ = 1.16 eV between the S 2p1/2 and S 2p3/2 peaks resulted in overlapping broad peaks. The results demonstrated that after oxidative roasting, the S atoms from the metal sulfides were not completely removed as SO₂. Instead, some S²⁻ was oxidized to S₂²⁻, S_x⁰, or SO₄²⁻ and remained on the catalyst. HDM-430-O contained all three sulfur species, while HDM-600-O contained only S_x and SO₄²⁻. The HDS-430-O and HDS-600-O samples retained only SO₄²⁻. The signal intensity of the SO₄²⁻ peak in the spectra followed the order of HDM-430-O > HDM-600-O > HDS-430-O > HDS-600-O, indicating a higher sulfate content on the roasted HDM catalysts compared to the roasted HDS catalysts. More sulfate formation was observed on the catalysts roasted at 430 °C than at 600 °C, suggesting that oxidation at lower temperatures was conducive to the formation of sulfate.

The phase structures of the roasted catalysts were characterized. In the XRD pattern of the spent HDM catalyst roasted at 430 °C, the peaks were attributed to Al₂O₃, Al₂(SO₄)₃, V₂O₅, NiSO₄, and NiS₂ (Fig. 8(a)). Notably, except for Al₂O₃, none of these species were observed in the XRD pattern of the spent catalysts (Fig. 1(a)), indicating that they were formed during the roasting process. No diffraction peaks corresponding to oxides of Ni, Mo, or Fe were detected, suggesting their existence as small crystallites or amorphous phases. The XRD pattern of HDM-600-O exhibited more complex and sharper diffraction peaks compared to the pattern of HDM-430-O (Fig. 8(b)). No peaks related to NiS₂ were detected, while peaks corresponding to NiV₂O₆, MoO₃, VO₂, Fe₁₃V₂₃O₇₀ and Fe₂O₃ were found. This indicated the following: (1) metal sulfides or disulfides were mostly converted during roasting at 600 °C compared with 430 °C; (2) more metal composite oxides were formed; and (3) the crystallinity of VO₂, Fe₂O₃ and MoO₃ improved.

Fig. 8 XRD patterns of the catalysts after roasting and the corresponding leached residue.

In the case of the spent HDS catalyst roasted at 430 °C, only a few broad XRD peaks were observed (Fig. 8(c)), which are primarily assigned to Al₂O₃, Al₂(SO₄)₃ and NiMoO₄. The pattern of HDS-600-O showed additional diffraction peaks corresponding to MoO₃, NiV₂O₆ and AlVO₄, indicating the formation of composite oxides and improved crystallinity (Fig. 8(d)). Compared to the roasted HDM catalysts, the intensity of the Al₂O₃ peaks for the roasted HDS catalysts was higher, indicating the presence of more Al₂O₃. The XRD peaks of NiMoO₄ and AlVO₄ could be observed in the patterns of the roasted HDS catalysts, but not in that of the roasted HDM samples. This indicated that in the spent HDS catalysts, Mo was prone to contact with Ni, forming composite oxides, while V had a higher probability of contact with Al. This could be explained by the lower content of V and higher content of Mo and Al in the HDS catalysts. In addition, both HDS-600-O and HDM-600-O exhibited crystallographic NiV₂O₆, indicating that V on the spent sRHC tended to form composite oxides with Ni. No XRD peaks assigned to Fe species were detected on the two roasted HDS catalysts due to their low Fe content.

Overall, an increase in roasting temperature accelerated both the solid–gas reaction between the metal sulfides and O₂ and the solid-phase reactions among metal oxides. The elevated roasting temperature induced the following changes in the chemical properties and physical structure of the roasted catalysts: (1) metal sulfides were more extensively converted into oxides, thereby reducing the production of metal persulfides and sulfates; (2) the valence states of V were elevated; (3) the formation of composite metal oxides was promoted; (4) the crystallites of metal oxides grew larger; (5) migration and aggregation of V and Ni oxides.

Metal leaching performance of roasted spent catalysts

To investigate the influence of the structures of the roasted catalysts on the leaching performance of their metals, the roasted catalysts were leached at 70 °C in 1 mol L⁻¹ H⁺ (0.5 M H₂SO₄) or OH⁻ (1 M NaOH) solutions using deionized water as the reference.

As can be seen in Fig. 9, the leaching efficiency of Ni from HDM-430-O in H₂SO₄ was higher than in H₂O, while almost no Ni leaching occurred in NaOH, indicating that H⁺ facilitated Ni leaching. The Ni leaching efficiency of HDM-600-O in H₂SO₄ was slightly higher than that of HDM-430-O, which was attributed to its lower content of NiS₂. However, the leaching efficiency of HDM-600-O in H₂O decreased slightly, which could be attributed to the combined effect of sulfate content reduction and crystal aggregation during roasting at 600 °C, as shown by the XRD and XPS results. The Ni leaching behavior of the roasted HDS catalysts followed a trend similar to that of the HDM catalysts, with stronger leaching in H₂SO₄ than in H₂O and negligible leaching in NaOH. Additionally, the Ni leaching efficiencies of HDS-430-O in both H₂SO₄ and H₂O were higher than that of HDS-600-O. This indicated that the favorable formation of sulfate and lower crystallinity resulting from low-temperature roasting were more suitable to Ni leaching. Besides, upon roasting at 430 °C, the HDS catalyst exhibited higher Ni leaching efficiency that the HDM catalyst due to its lower NiS₂ content, whereas upon roasting at 600 °C, the HDM catalyst showed a higher leaching rate, primarily because of its higher NiSO₄ content.

Fig. 9 Leaching efficiencies of Ni, Fe, Mo, V and Al on the (a)–(e) roasted spent HDM catalysts and (a′)–(e′) roasted spent HDS catalysts under the following leaching conditions: 70 °C, 20 min, 1 mol L⁻¹ H⁺ or OH⁻ solvent, and solid–liquid ratio of 0.05 g mL⁻¹.

The Fe leaching efficiency of the roasted samples in different solvents followed the order of NaOH > H₂SO₄ > H₂O. Similar to Ni, Fe was more inclined to be leached by H⁺, with almost no leaching in NaOH. The ease of Fe leaching from the different roasted spent catalysts decreased in the order of HDM-430-O > HDS-430-O > HDM-600-O > HDS-600-O. High-temperature roasting hindered Fe leaching due to crystal growth and the formation of composite oxides that are difficult to dissolve, e.g., Fe₁₃V₂₃O₇₀. Fe leaching from the HDS catalysts was more difficult than from the HDM catalysts, likely because of their lower sulfate content.

The leaching behavior of Mo was relatively complex. The leaching rate for the roasted HDM catalysts under different leaching solvents followed the order of NaOH > H₂SO₄ > H₂O. Specifically, the Mo leaching efficiency of HDM-600-O was similar to that of HDM-430-O in H₂SO₄ and H₂O, while it was higher in NaOH. In the case of the roasted HDS catalyst, the leaching efficiencies of HDS-600-O were higher than that of HDS-430-O under all three leaching solvents. This phenomenon is consistent with the results reported by Fang et al.,²⁵ where high-temperature roasting favored the NaOH leaching of the spent HDS catalysts. In their report, this was mainly due to the conversion of Mo(IV) to Mo(VI) by high-temperature roasting. However, in this study, XPS and XRD characterization confirmed that all the roasted samples contained Mo(VI). Therefore, this phenomenon is due to other reasons besides the valence state of Mo. This might be attributed to the aggregation of Ni and V oxides at high roasting temperatures, as shown by the TEM and XRD results, which increased the exposure of Mo and Al species, thereby facilitating Mo leaching. This promoting effect of high-temperature roasting on the HDM catalysts was not obvious enough, which may be due to the excessive deposition of Ni and V metal and insufficient increased Mo exposure. The leaching efficiencies of the roasted HDS catalysts under all three solvents were significantly higher than that of the roasted HDM catalysts. This was associated with the higher amount of NiMoO₄ composite oxides in the roasted HDM catalysts and the lower amounts of deposited metals (Ni, V and Fe) competing with Mo for leaching. In the case of the NiMoO₄ composite oxides, the incorporation of Ni made them more reactive with acids compared to MoO₃.

In the case of V, the leaching efficiency of the roasted HDM catalysts followed the order of H₂SO₄ > NaOH > H₂O. The higher leaching efficiency of H₂SO₄ compared to NaOH was mainly attributed to the fact that H₂SO₄ could leach both V(IV) and V(V), while NaOH primarily leached V(V) in the form of NaVO₃. The leaching efficiencies of HDM-600-O in both H₂SO₄ and NaOH were observed to be higher than that of HDM-430-O. Although the high-temperature roasting at 600 °C caused the crystal growth of V oxides, which is a negative factor in leaching, it simultaneously increased the proportion of V(V) which was a positive factor for V leaching. The results demonstrated that the latter effect was more dominant for V leaching. Among the HDS catalysts, HDS-600-O exhibited higher V leaching efficiencies in all three solvents compared to HDS-430-O, which was similarly related to its higher V(V) proportion in V. In contrast to the roasted HDM catalysts, HDS-600-O showed a higher V leaching efficiency in NaOH than in H₂SO₄. This difference was associated with the formation of composite oxides, e.g. AlVO₄, during high-temperature roasting, which were more soluble in alkaline solutions, whereas the HDM catalysts predominantly produced composite oxides that are difficult to leach (e.g. NiV₂O₆).

The Al leaching efficiency for the roasted HDM and HDS catalysts followed similar patterns as V, showing the order of H₂SO₄ > NaOH > H₂O. The samples roasted at 600 °C consistently displayed higher leaching efficiencies in all solvents compared to those roasted at 430 °C. This phenomenon can possibly be explained by the formation of Al-containing composite oxides on the HDS catalysts, such as NaAlO₂ and AlVO₄, which are more soluble in the solvents than the Al₂O₃ support. Besides, the aggregation of Ni and V oxides at higher roasting temperatures could also increase the exposure of Al, which benefited Al leaching. Notably, the HDS catalysts roasted at different temperatures showed higher leaching efficiencies in both H₂SO₄ and NaOH than the HDM catalysts, which was correlated with their larger surface area and lower content of non-aluminum metals.

In summary, the metal leaching efficiencies in different solvents followed general order of V and Al leaching of H₂SO₄ > NaOH > H₂O; Ni and Fe leaching of H₂SO₄ > H₂O > NaOH; and Mo leaching of NaOH > H₂SO₄ > H₂O.

Influence of phase structure in the roasting products on the leaching performance

The phase structure of the leaching residue was analyzed using XRD technology to further investigate the influence of phase structure in the roasting products on their leaching performance (Fig. 8). In the case of the HDM-430-O sample, after H₂SO₄ leaching, the XRD diffraction peaks corresponding to V₂O₅ disappeared, while the intensity of the NiSO₄ peaks became stronger (Fig. 8(a)). This indicated that H₂SO₄ effectively leached V₂O₅, but the solubility of sulfates was weak. In the H₂O-leached residue, the peaks of Al₂(SO₄)₃ disappeared, and the peaks of NiS₂ weakened. After NaOH leaching, the peaks of V₂O₅, NiSO₄, and Al₂(SO₄)₃ on HDM-430-O disappeared, but the NiS₂ peaks intensified. This suggested that NaOH was effective in reducing both sulfates and V₂O₅. It should be noted that the Ni leaching efficiency in NaOH was nearly zero. This suggests that nickel may form sparingly soluble phases under alkaline conditions (e.g., Ni(OH)₂). These Ni substances were not detected in the XRD results of the NaOH leaching residue (HDM-430-NaOH), indicating that they were possibly amorphous or very fine crystalline precipitates.

In the case of HDM-600-O, after H₂SO₄ leaching, the XRD diffraction peaks of V₂O₅ and Fe₂O₃ disappeared, while that of VO₂, NiV₂O₆, NiSO₄, MoO₃, and Al₂(SO₄)₃ weakened (Fig. 8(b)). In the H₂O-leached residue, the NiSO₄ peaks disappeared, and the peaks of Al₂(SO₄)₃, NiV₂O₆, V₂O₅, and Fe₂O₃ weakened. After NaOH leaching, the peaks of V₂O₅, MoO₃, NiSO₄, and Al₂(SO₄)₃ on HDM-600-O disappeared, but the peaks of Fe₁₃V₂₃O₇₀ significantly intensified. The results demonstrated that compared with HDM-430-O, sulfates in HDM-600-O were more easily removed, possibly due to their lower content. VO₂ could be leached in H₂SO₄ but difficult to leach in NaOH, whereas MoO₃ was more readily leached in NaOH. Ni–V–O and Fe–V–O composite oxides such as NiV₂O₆ and Fe₁₃V₂₃O₇₀ were difficult to dissolve completely in either H₂SO₄ or NaOH, respectively. The solid-phase reactions generating composite oxides during the roasting of the spent HDM catalysts were unfavorable for subsequent leaching.

The XRD diffraction peaks of HDS-430-O were relatively broad and diffuse, and therefore almost no diffraction peaks other than Al₂O₃ were observed in the leached residue (Fig. 8(c)). The diffraction peaks of Al₂(SO₄)₃ and β-NiMoO₄ were weakened or disappeared after leaching in all three solvents (H₂SO₄, H₂O and NaOH), indicating that these solvents exhibited certain dissolution effects on both Al₂(SO₄)₃ and β-NiMoO₄.

In the case of HDS-600-O, the XRD diffraction peaks of β-NiMoO₄ and Al₂(SO₄)₃ disappeared after H₂SO₄ leaching, while the NiV₂O₆ peaks were weakened. In the H₂O-leached residue, the Al₂(SO₄)₃ peaks disappeared and the NiV₂O₆ peaks were weakened (Fig. 8(d)). In the NaOH-leached residue, the peaks of β-NiMoO₄, Al₂(SO₄)₃, MoO₃, and AlVO₄ disappeared. Additionally, new peaks assigned to α-NiMoO₄ were detected in all three types of HDS-600-O leached residues, suggesting a phase transformation from β-NiMoO₄ to α-NiMoO₄ during the leaching process. Among them, the α-NiMoO₄ peaks in the NaOH-leached residue showed the weakest intensity, while β-NiMoO₄ was almost absent, indicating that NaOH had the strongest leaching efficiency on NiMoO₄. Well-crystallized NiMoO₄ dissolves poorly in H₂SO₄ and H₂O. The formation of composite oxides during the roasting of the spent HDS catalysts was found to have relatively lower negative impacts on leaching compared with the spent HDM catalysts. This was attributed to the fact that the main metals participating in solid-state reactions in the HDS catalysts were Ni, Mo, V, and Al, which differed from the Ni, Fe, and V components in the HDM catalysts. Mo and Al were highly dispersed on the HDS catalyst, and the difficulty of dissolution of the formed composite oxides was low.

The leaching performance of the roasted spent catalysts was found to vary across different solvent systems. The common patterns regarding the influence of the spent catalyst structure on leaching processes found in this study are summarized as follows.

1. In the H₂SO₄ leaching process, most metals could be leached, with V₂O₅ and Fe₂O₃ being extensively extracted. Moderate leaching effects were observed for VO₂, NiV₂O₆, NiMoO₄, and sulfates, while no significant leaching occurred for well-crystallized MoO₃ or V-containing composite oxides such as AlVO₄ and Fe₁₃V₂₃O₇₀.

2. The H₂O leaching process dissolved most metal sulfates and oxides, albeit with lower efficiency. Ni and Mo exhibited higher leaching rates compared to other metals, suggesting that the H₂O leachate had potential for use in the preparation of hydrotreating catalysts.

3. The NaOH leaching process selectively extracted metals of V and Mo without dissolving Fe or Ni. Significant dissolution was achieved for V₂O₅, MoO₃, metal sulfates, NiMoO₄, and AlVO₄, whereas V(IV)-containing oxides e.g., VO₂, NiV₂O₆, Fe₁₃V₂₃O₇₀ and Ni/Fe–V–O composite oxides, showed a negligible leaching performance.

The high selectivity of alkaline solvents towards valuable metals (Mo and V) could reduce the difficulty of subsequent extraction and separation of the leachate. Meanwhile, the wide adaptability of the leaching ability of most metals made it possible to use acidic solvents to further separate Ni from the solid residue after alkaline solvent leaching. Therefore, a multi-stage leaching strategy could be considered, combining alkaline solvent leaching and acidic solvent leaching, which could reduce the separation cost of metals in the leachate and recover as much valuable metal as possible.

The influence of temperature on the roasting reactions and processes was complex. Elevated temperatures promoted crystallization and crystal size growth, solid-phase reactions forming composite oxides, V(V) oxide formation, and Ni/V oxide migration-aggregation, while lower temperatures favored the conversion of metal sulfides to sulfates or other sulfides (such as NiS₂). Although high-temperature treatment enhanced grain growth and composite oxide formation, which are generally considered as factors detrimental to metal leaching, the experiments in this work demonstrated that increased oxidation states and aggregated deposited metals paradoxically improved V, Al, and Mo leaching. In the results of NaOH leaching of HDS-600-O, some composite oxides (NiMoO₄ and AlVO₄) were also found to have no negative effects on leaching. The results exhibited that combined physicochemical structural changes led to divergent leaching behaviors of metals at different roasting temperatures. Typically, high-temperature roasting hindered acid/alkali leaching of Ni and Fe but benefited V, Mo, and Al leaching. Thus, the roasting parameters could be adjusted to optimize the selectivity for target metals to a certain extent. A slightly higher temperature could be beneficial for the recovery of high-valued metals from sRHCs and other spent petrochemical catalysts. It should be noted that excessively high roasting temperatures may not be conducive to metal recovery, as molybdenum oxides undergoes rapid sublimation above 650 °C, and the metal oxides may be sintered severely.

Additionally, the spent catalyst metal content modulated reactions and processes during roasting. High contents of deposited V, Ni and Fe metals facilitated sulfate/NiS₂ formation, V(IV) oxide generation, Ni–V–O and Fe–V–O composite oxide production, and crystal growth of metal oxides. The structural evolution patterns during roasting of the HDS and HDM spent catalysts differed due to their varying metal contents and distributions, posing challenges for industrial metal recovery from residual hydrotreating catalysts. Comparatively, the HDS catalysts exhibited superior Mo and Al leaching efficiencies, while the HDM catalysts exhibited better Fe leaching efficiencies.

Conclusion

The roasting processes of two typical types of spent residue hydrotreating catalysts were systematically investigated, along with the physicochemical structures of the roasting products and their corresponding leaching behaviors. High-temperature roasting was found to cause crystal growth and the formation of composite oxides, which negatively affected the leaching efficiency of Ni and Fe in acid solvent. However, the increased oxidation state of V induced by high-temperature roasting facilitated V leaching, while the aggregation of deposited metals benefited Mo and Al leaching. Differences in metal content also led to varying degrees of reactions during the roasting process. At the same combustion temperature, significant differences were observed between the roasting products of the HDM and HDS spent catalysts. The roasted HDM catalysts tended to form composite oxides of Ni, Fe, and V, along with V(IV) oxides and metal sulfates, whereas the HDS catalyst predominantly formed Ni–Mo–O and Al–V–O composite oxides and V(V) oxides. Consequently, better leaching efficiencies were generally obtained for Mo and Al from the HDS catalysts, while higher Fe leaching efficiencies were typically achieved from the HDM catalysts. Moreover, the HDM catalyst required higher temperatures to completely convert Ni sulfides into oxides, resulting in poor Ni leaching efficiency under 430 °C roasting. The leaching behaviors of metal species varied with different solvents. The H₂SO₄ leaching procedure proved effective for dissolving V₂O₅, Fe₂O₃, and VO₂, and had a moderate leaching capacity for MoO₃, but showed poor or negligible leaching efficiency for NiMoO₄, AlVO₄, and Fe₁₃V₂₃O₇₀. The H₂O leaching procedure exhibited generally low efficiency overall, with only Ni and Mo showing higher leaching rates than other metals. The NaOH leaching procedure could dissolve V₂O₅, MoO₃, NiMoO₄, AlVO₄ and metal sulfates, but performed poorly for VO₂, NiV₂O₆, Fe₁₃V₂₃O₇₀, and other V(VI)-containing oxides as well as Ni- and Fe-containing V composite oxides.

Conflicts of interest

There are no conflicts to declare.

Data availability

Data available from the authors on request.

Acknowledgements

The authors gratefully acknowledge the funding of the National Key R&D Program of China (2024YFC3907701), the Young Elite Scientists Sponsorship Program by BAST (BYESS2024255), and the Project of Sinopec (PR20232509 and 124067).

References

1. E. V. Parkhomchuk, K. V. Fedotov, A. I. Lysikov, A. V. Polukhin, E. E. Vorobyeva, I. A. Shamanaeva, N. N. Sankova, D. O. Shestakova, D. M. Reshetnikov, A. V. Volf, A. V. Kleymenov and V. N. Parmon, Fuel, 2023, 341, 127714.
2. E. E. Vorobyeva, I. A. Shamanaeva, A. V. Polukhin, A. I. Lysikov and E. V. Parkhomchuk, Fuel, 2023, 334, 126676.
3. L. Zhao, X. Zhang, G. Jiang, Z. Tan, R. Tao, Y. Chen and D. A. Pan, Resour. Environ. Sustain., 2025, 21, 100251.
4. L. Duarte, L. Garzón and V. G. Baldovino-Medrano, Catal. Today, 2019, 338, 100–107.
5. A. Pathak, H. Al-Sheeha, A. A. Ali and M. S. Rana, J. Environ. Manage., 2022, 322, 116055.
6. Y. Jia, Q. Yang, S. Sun, H. Nie and D. Li, Energy Fuels, 2016, 30, 2544–2554.
7. M. Marafi and A. Stanislaus, Resour., Conserv. Recycl., 2008, 52, 859–873.
8. T. Nguyen, Q. Nguyen, A. N. T. Cao, T. Ernest, T. B. Nguyen, P. T. H. Pham and T. M. Nguyen, J. Pet. Sci. Eng., 2022, 216, 110762.
9. T. Liu, L. Ju, Y. Zhou, Q. Wei, S. Ding, W. Zhou, X. Luo, S. Jiang and X. Tao, Catal. Today, 2016, 271, 179–187.
10. R. Agrawal and J. Wei, Ind. Eng. Chem. Process Des. Dev., 1984, 23, 515–522.
11. J. Wen, X. Wang, F. Yu, M. Tian, C. Wang, G. Huang and S. Xu, J. Cleaner Prod., 2023, 419, 138295.
12. R. Chen, C. Feng, J. Tan, C. Zhang, S. Yuan, M. Liu, H. Hu, Q. Li and J. Hu, Hydrometallurgy, 2022, 213, 105910.
13. H. Yu, S. Liu, A. Yaraş, B. Enkhchimeg, L. Hu, W. Zhang, M. Peng, H. Arslanoğlu and L. Mao, J. Environ. Manage., 2025, 379, 124920.
14. I. Asghari, S. M. Mousavi, F. Amiri and S. Tavassoli, J. Ind. Eng. Chem., 2013, 19, 1069–1081.
15. Y. Huang, K. Shi, S. Su, B. Liu, H. Sun and G. Han, J. Environ. Chem. Eng., 2022, 10, 108462.
16. B. Padh, P. C. Rout, G. K. Mishra, K. R. Suresh and B. Ramachandra Reddy, Hydrometallurgy, 2019, 184, 88–94.
17. Q. Tang, S. Liu, H. Hao, X. Dong, Y. Guo, Y. Lu, Z. Wang and C. Xu, Chem. Bio Eng., 2025, 2, 431–436.
18. L. Zhao, X. Zhang, Z. Tan, G. Jiang, Y. Chen and D. Pan, Sci. Total Environ., 2024, 953, 176127.
19. X. Hong, J. Chen, J. Wei, W. Wu, Z. Yang, J. Chen, S. Sun, F. Xiao and G. Tu, Recycling, 2025, 10, 10.
20. C. Feng, C. Zhang, S. Yuan, M. Liu, R. Chen, H. Hu and J. Hu, J. Cleaner Prod., 2022, 338, 130564.
21. D. Zhu, T. Dong, J. Pan, W. Ma, Z. Guo and S. Li, J. Environ. Chem. Eng., 2025, 118208.
22. R. Adavodi, P. Romano, S. Rahmati, I. Birloaga, A. Mancini and F. Vegliò, Sep. Purif. Technol., 2025, 363, 132020.
23. H. Al-Sheeha, M. Marafi, V. Raghavan and M. S. Rana, Ind. Eng. Chem. Res., 2013, 52, 12794–12801.
24. W. Cheng, J. Li, J. Deng, Y. Li and F. Cheng, ACS Omega, 2024, 9, 9286–9294.
25. K. Fang, G. Zhang, Q. Li, W. Guan, M. Wang, S. Wu, X. Xia and C. Shi, Hydrometallurgy, 2024, 229, 106381.
26. J. Wang, S. Wang, A. Olayiwola, N. Yang, B. Liu, J. J. Weigand, M. Wenzel and H. Du, J. Hazard. Mater., 2021, 416, 125849.
27. D. Zhang, Y. Liu, Q. Hu, X. Ke, S. Yuan, S. Liu, X. Ji and J. Hu, J. Cleaner Prod., 2020, 252, 119763.
28. H. Liu, Y. Zhang, T. Liu, J. Huang, L. Chen and Y. Hu, Trans. Nonferrous Met. Soc. China, 2023, 33, 1594–1608.
29. J. P. Edaugal, D. Zhang, D. Liu, V. Glezakou and N. Sun, Chem. Bio Eng., 2025, 2, 210–228.
30. Y. Chen, X. Wu, W. Guan, S. Xiao, K. Fang, J. Qing, R. Xie, S. Wu, Q. Li, Z. Cao, M. Wang and G. Zhang, J. Environ. Chem. Eng., 2024, 12, 113485.
31. R. Iwamoto, J. Jpn. Pet. Inst., 2013, 56, 109–121.
32. P. G. Coombs and Z. A. Munir, Metall. Trans. B, 1989, 20, 661–670.
33. Z. Liu, W. Han, D. Hu, S. Sun, A. Hu, Z. Wang, Y. Jia, X. Zhao and Q. Yang, J. Catal., 2020, 387, 62–72.
34. P. A. Monson, Microporous Mesoporous Mater., 2012, 160, 47–66.
35. S. Lee, Y. Son, J. Choi, S. Kim and S. Park, Catalysts, 2021, 11, 1229.
36. R. Hou, Q. Yang, S. Zeng, J. Bao, H. Nie, C. Yang, Y. Jia, A. Hu and Q. Dai, Catalysts, 2024, 14, 287.
37. D. A. Beattie, J. K. Chapelet, M. Gräfe, W. M. Skinner and E. Smith, Environ. Sci. Technol., 2008, 42, 9191–9196.
38. D. Stoilova, M. Georgiev and D. Marinova, Vib. Spectrosc., 2005, 39, 46–49.
39. A. Šurca and B. Orel, Electrochim. Acta, 1999, 44, 3051–3057.
40. N. Singh, R. Kumar Jyothi, K. S. Pasupuleti and J. Lee, J. Ind. Eng. Chem., 2025, 143, 632–644.
41. C. Yang, A. Hu, Q. Dai, Q. Yang, R. Hou and Z. Liu, ACS Omega, 2023, 8, 41182–41193.

---