Ni–Mo sulfide nanosized catalysts from water-soluble precursors for hydrogenation of aromatics under water gas shift conditions

Chemicals

The following reagents were used without additional purification: ammonium heptamolybdate tetrahydrate (NH₄)₆Mo₇O₂₄·4H₂O (99%, “Alfa Aesar”),nickel nitrate hexahydrate Ni(NO₃)₂·6H₂O (≥97%, “Sigma-Aldrich”), 1-methylnaphthalene С₁₁Н₁₀ (99%, “Sigma-Aldrich”), 2-methylnaphthalene С₁₁Н₁₀ (99%, “Sigma-Aldrich”), anthracene С₁₄Н₁₀ (98%, “Sigma-Aldrich”), 1,8-dimethylnaphtalene С₁₂Н₁₂ (95%, “Sigma-Aldrich”), 1,3-dimethylnaphtalene С₁₂Н₁₂ (96%, “Sigma-Aldrich”), 2,6-dimethylnaphtalene С₁₂Н₁₂ (99%, “Sigma-Aldrich”), 2,3-dimethylnaphtalene С₁₂Н₁₂ (97%, “Sigma-Aldrich”), 2-ethylnaphthalene С₁₂Н₁₂ (99%, “Sigma-Aldrich”), sorbitan monooleate (SPAN-80) C₂₄H₄₄O₆ (98%, “Merck”), elemental sulfur S (98%, “Chemical reactant”), toluene C₇H₈ (99%, “Khimmed”).

Characterization techniques

X-ray powder diffraction (XRD) patterns were recorded between 2θ = 1.5°–70.0° by 0.05, using nickel filter and Cu Kα = 1.54 Å (λ = 0.154 nm) radiation on the Bruker D2 PHASER diffractometer at room temperature. XRD data obtained were interpreted using Rigaku PDXL software with a powder database.

The dispersions of the Ni–Mo sulfide catalysts were investigated with Transmission electron microscopy (TEM) using a JEOL JEM-2100 microscope with an accelerating voltage 200 kV. The samples were prepared by dispersing in n-hexane and then put onto formvar grids sample holder. The samples covered with liquid hexane for protection from oxidation by air were introduced into the vacuum chamber. More detailed information about the active sulfide phase (distributions of slab length and layers stacking) was revealed by the statistical evaluation of around 300 particles from different zones of a number of various TEM images using Image-Pro Plus 7.0 software. The characteristics of the active phase were calculated assuming that the MoS₂ slabs were perfect hexagons and using the equations in Table 1.

The elemental composition of catalysts was determined by X-ray fluorescence (XRF) spectrometry on energy dispersive Thermo Scientific ARL Quant’X. The measurements were carried out in vacuum and results were processed using the standard-less UniQuant method.

The XPS spectra of the sulfide catalysts were recorded on a PHI5500VersaProbeII spectrometer equipped with a hemispherical electron analyzer and Al Kα (hν = 1486,6 eV) X-ray source. Binding energy (BE) scale was preliminary calibrated using the peaks position for the Au 4f7/2 (84.0 eV) and Cu 2p3/2 (932.6 eV) core levels. The pass energy was set at 117.4 eV for the survey scan and 23.5 eV for the narrow scan of Mo3d, S2p, O1s, C1s, while for the narrow scan of Ni2p the pass energy was set at 29.35 eV. The BE were referenced to the C1 s peak (284.9 eV) to account for the charging effects. The accuracy of the BE values is ± 0.1 eV. After fitting the instrument, peak areas were approximated by Gaussian/Lorentzian curves and removal of the background (Shirley function). The analysis of the individual spectral regions allowed identifying chemical state of the elements according to binding energies. Surface atomic ratios were calculated from the peak area ratios normalized by the corresponding atomic sensitivity factors. Metal distribution for Ni and Mo species at the surface of sulfided catalysts were calculated using the equations in Table 2.

The Ni/Mo atomic ratio on the edges of NiMoS active phase was calculated from XPS and TEM as follows:

Catalytic experiments

The activity of the catalysts was evaluated in a batch reactor (50 cm³) equipped with a magnetic stirrer. In the typical experiment the catalyst precursors (ammonium molybdate and nickel nitrate) were dissolved in distilled water (20 wt.%) and placed in autoclave. Then, solution of aromatic hydrocarbon (10 wt.%) in toluene containing surfactant (SPAN-80) was added. Following this, elemental sulfur (2.5 wt.%) was suspended in the reaction mixture. The reactant ratios are shown in Table 3.

The Mo content (wt.%) in the reaction mixture was calculated with the equation:

where m_substrate – mass of aromatic hydrocarbon, g; M – molar mass of catalyst precursors or aromatic hydrocarbon (substrate), g/mol; m _solution – solution mass, g; m_solution  = m_H2O + m_toluene  + m_S + m_substrate  + m((NH₄)₆Mo₇O₂₄∙4H₂O) + m(Ni(NO₃)₂∙6H₂O) + m_surfactant, m_solution  = 5 g.

The autoclave was pressurized with carbon monoxide to 5.0 MPa, placed in a thermostatically controlled oven and heated up to reaction temperature with ramp of 15 °C/min. The stirring speed was kept at 400–500 rpm. We suppose these conditions and stirring rate free the system from diffusion limitations. When the temperature controlled at the bottom of the reactor by thermocouple reached the set point, the reaction was carried out for 4–8 h. After a predetermined time, the autoclave was cooled and disassembled. Under the reaction conditions the catalysts’ precursors, forming through the high-temperature decomposition of water-soluble metal salts, catalyze the WGSR to form hydrogen, followed by its interaction with elemental sulfur. Thus, the resulting hydrogen sulfide acts as a sulfiding agent to produce an active sulfide form of the catalyst. The catalyst was separated by filtration followed by step-by-step washing with toluene and hexane. Then, it was dried in inert atmosphere and evaluated by XRD, XRF, TEM and XPS techniques. The composition of liquid product was analyzed using GC (FID) provided with capillary column Petrocol™ (Supelco 0.25mmх50 m). The column temperature was varied depending on the reactant compound, and temperature programming was used to improve peak resolution. From the GC analysis the conversion of model substances (C, %) and was calculated as follows:

where S _s – peak area of unreacted substrate, %; S _p – sum of peak areas of products, %.

The selectivity (S, %) was determined from the equation:

where S _i – peak area of i-product, %.

Surfactant influence

Nowadays more attention has been paid to bulk or unsupported transition metal catalysts. Unlike alumina or silica–alumina supported sulfide, dispersed catalysts provide contact of substrate molecules with active component excluding diffusion limitations and mass transfer (17) resulting in high activity. Unsupported catalysts due to its uniform dispersion in feeds acts as hydrogen transfer agent. Thus, these catalysts favor the rapid hydrogen consumption deactivating organic radicals and preventing thereby their condensation and polymerization resulting in suppression of coke formation. When hydrogen reacts with sulfur ions in MoS₂ that are located in the corner and edge, the unsaturated sites and sulfur ion vacancies being active sites for hydrogenation are formed. Meanwhile, the Mo–H and S–H moieties are unstable and react rapidly with the free hydrocarbon radicals decreasing the rate of condensation or polymerization reactions [20], [30], [31], [32], [33], [34]. Among the variety of different techniques, the decomposition of water- or oil-soluble metal salts followed by sulfidation is simple and reproducible, allowing for preparation of binary and ternary sulfides with controlled stoichiometry [21], [22]. Generally, oil-soluble precursors are more active than water-soluble ones, but too expensive [23]. The lower activity of catalysts obtained from water-soluble precursors is caused by fast water evaporation leading to particle agglomeration and low active phase dispersion [23].

In this work water-soluble ammonium heptamolybdate tetrahydrate (NH₄)₆Mo₇O₂₄·4H₂O and nickel nitrate hexahydrate Ni(NO₃)₂·6H₂O were used as catalyst precursors and elemental sulfur as a sulfiding agent. The active phase formed in situ at high-temperature during hydrogenation of aromatics in W/O emulsions under CO pressure according to Scheme 1.

Scheme 1:

Formation of Ni–Mo catalysts in W/O emulsion during hydrogenation of aromatics from water-soluble precursors under CO pressure.

Elevating the temperature of the reaction mixture (at the range of 45–420 °С) so as to decompose of ammonium molybdate and nickel nitrate, leads to formation of intermediates as well as metal oxides (Scheme 2) [35], [36], [37], [38], [39], [40], [41]. Interaction between Mo and promoter supposed to occurs spontaneously at ambient conditions [20]. The cationic species of Ni interact with molybdate anions [20] followed by sulfidation and leading to precipitation of dark solids. Thus, Ni²⁺ as well as Mo₄O₁₃ ^2– exists in solution simultaneously up to 200–240 °C that may provide formation of NiMo₄O₁₃ species. Moreover, (NH₄)₆Mo₇O₂₄ *4H₂O precursor also may decomposed to MoO₃*(H₂O)₂ at 300 °C [42] and reacts with NiNO₃ to form hybrids containing NiMoO₄*2H₂O and MoO₃ [43], [44]. When the temperature rises, the NiMo₄O₁₃ as well as NiMoO₄ may acts with hydrogen sulfide to form NiMo_xS_y particles [45]. These mixed-metal oxides are easier to reduce due to it is easier to adsorb and dissociate H₂S on Ni sites than on Mo sites of an oxide with water removal. Meanwhile, decomposition to form metal oxides also takes place.

Scheme 2:

High-temperature decomposition-sulfidation of Ni and Mo salts [35], [36], [37], [38], [39], [40], [41].

To control particle size and high active phase dispersion, the stabilizing of W/O emulsions by surfactant was applied. Sorbitan monooleate (SPAN-80) with a hydrophilic-lipophilic balance (HLB) value of 4–6 favoring the formation of invert emulsions was chosen [46]. The hydrogenation of 1- and 2-methylnaphtalenes was chosen as a model reaction to evaluate the surfactant influence on catalytic properties. The catalyst obtained through the surfactant-assisted synthesis in W/O emulsion was named NiMoS + SPAN, while counterpart formed without stabilizing agent was named NiMoS. According to the experimental data the former is more active in 2-methylnaphtalene (2MN) hydrogenation, about twice as high as the latter (by conversion for 4–6 h, see Fig. 1a). At higher reaction time of 8 h the conversion of 2MN over NiMoS and NiMoS + SPAN becomes comparable exceeding 75%. When the temperature rises, the conversion of 1-methylnaphtalene (1MN) also increases both for NiMoS and NiMoS + SPAN catalysts while the latter being more active (Fig. 1b).

Fig. 1:

Conversion of (a) 2MN (Т = 380 °С) and (b) 1MN (t = 6 h) over Ni–Mo sulfide catalysts depending on surfactant presence. Reaction conditions: ω(SPAN-80) = 4 wt.%, p(CO) = 5 MPa, ω _substrate  = 10 wt.%., ω(H ₂ O) = 20 wt.% , СО:Н ₂ О = 1.8 (molar ratio).

To evaluate the influence of surfactant concentration on catalytic activity the mass content of SPAN-80 was varied from 2 to 8 wt.% (Fig. 2). On rising surfactant content, the catalytic activity increases achieving maximum at 4 wt.% and was unaffected by higher concentration of SPAN-80. Such a significant difference in activity depending on surfactant is caused by its influence on sulfide phase morphological characteristics and metal distribution over Ni and Mo species at the surface of sulfide catalysts.

Fig. 2:

Influence of surfactant concentration on 2MN conversion over Ni–Mo sulfide catalysts. Reaction conditions: Т = 380 °С, p(CO) = 5 MPa, ω(2MN) = 10 wt.%., ω(H ₂ O) = 20 wt.%, t = 4 h, СО:Н ₂ О = 1.8 (molar ratio).

To evaluate the characteristics of active phase species, the catalyst formed during hydrogenation from water-soluble metal precursors was separated by filtration followed by step-by-step washing with toluene and hexane. Then, it was dried in inert atmosphere and evaluated by XRD, XRF, TEM and XPS techniques.

Morphological characteristics of active phase species

The catalysts reveal the presence of a typical multilayer sulfide structure (Fig. 3) with interlayer distance approximately of 0.62–0.63 nm differs from basal plane of MoS₂ crystalline [47]. According to calculation based on statistical evaluation of sulfide particles from TEM images, NiMoS + SPAN has a narrow distribution of slab length (Fig. 3) with average stacking number of 3.8 (Table 4a). For NiMoS the slabs are 6.0–18.0 nm in length within comparable content of sulfide particles having the length 6.0–12.0 nm and 12.0–18.0 nm. Meanwhile the number of stacks per particle is about 4.5. The formation of multilayered particles with stacking number of 6–9 also takes place from the presence of surfactant. Thus, the surfactant-assisted synthesis leads to increase the dispersion of active phase characterized by highly stacked of 3–4 layers in MoS₂ particles with 14.5 nm lengths providing higher activity of this catalyst.

Fig. 3:

TEM microphotographs and distribution of slab length for (a) NiMoS + SPAN and (b) NiMoS catalysts.

XRD pattern of the NiMoS + SPAN catalyst are shown in Fig. 4. The peaks at 2θ = 14.5° (hkl = 002), 33.0° (hkl = 101) and 58.4° (hkl = 110) correspond to MoS₂ [48]. Nevertheless, the XRD pattern shows poor crystallinity with a weak peak at 2θ ≈ 14°, characterizing a small stacking of MoS₂ layers caused by synergistic effect of Ni and Mo atoms. The second reason is the intercalation of organic and inorganic molecules into the interplanar space of MoS₂ phase leading to violation of the coplanarity of crystallographic planes. The (101) and (310) reflections at 2θ = 31.8° and 58.4° respectively are ascribed to molybdenum oxides [49], while peaks at 2θ = 30.0° and 34.4° (hkl = 101) correspond to NiS phase [50]. The reflections at 2θ = 45.5° and 53.3° indicates NiS₂ or Ni_xMoS_y phase [50]. The reflections around 27°, 30°, and 44° attributed to α-NiMoO₄ phase are not quite resolved, nevertheless the nickel molybdate formation should not to exclude [51], [52].

Fig. 4:

XRD pattern of NiMoS + SPAN catalyst.

According to XRF data (Table 4b), the composition of NiMoS is similar to NiMoS + SPAN catalysts. When the surfactant is used the Ni content is slightly lower compared to NiMoS catalyst. These results are in a full accordance with XPS data (Table 4b) revealing the active phase composition and chemical state of Mo, Ni and S.

The Mo3d-S2s, Ni2p and S2p core level spectra are shown in Fig. 5. The peak at 226.0 eV is characteristic for S²⁻ ions of S2s spectra [19], [20]. For both catalysts the most intense photoelectron Mo3d doublet with binding energy of 228.9 and 232.0 eV corresponds to Mo⁴⁺ species of the MoS₂ phase, while two peaks at 232.8 and 235.9 eV is attributed to NiMoO₄ as well as Mo⁶⁺ oxidation state [20], [25], [53], [54]. Quantitative XPS analysis of the Mo 3d_5/2 core level spectra shows that Mo atoms are predominantly in sulfide with relative content of 71.2 and 85.0 for NiMoS and NiMoS + SPAN catalysts respectively (Table 5).

Fig. 5:

Mo3d-S2, Ni 2p and S 2p core level spectra of (a) NiMoS + SPAN and (b) NiMoS catalysts.

For Ni 2p core level spectra a strong difference is observed. The binding energy region between 854.0 and 854.5 eV corresponds to NiMoS phase [19], [20], [25], whose the relative content for NiMoS + SPAN catalyst reaches almost 50%, while Ni 2p core level spectra for NiMoS counterpart indicates the formation of NiMoO₄ (54.0%) predominantly. It is confirmed by doublet peaks at 857.1 and 874.3 eV [53]. Thus, Ni is fully in mixed oxide or sulfide phases. The XPS data also confirm that NiS with characteristic binding energy of 852.9–853.0 eV doesn’t form. Meanwhile peaks at 856.1–856.3 eV is assigned to Ni⁺² species either in oxide or salts form [19], [20]. The peak attributed to NiMoS phase in Ni 2p core level spectrum for NiMoS + SPAN catalyst slightly shifted to lower binding energies at 0.5 eV caused by the residual presence of carbon species. In the S2p_3/2 core level spectra two peaks at 162.0–162.2 and 163.0 eV are assigned to S²⁻ and (S₂)²⁻ sulfur species respectively [19], [20]. The photoelectron peak with binding energies at the range of 168.7–168.9 eV is attributed to sulfur in SO₄ ^2– form [19], [20]. While for NiMoS + SPAN the S percentage in sulfide form is close to 72%, in the case of NiMoS it doesn’t exceed 53%. The latter is characterized by higher content of (S₂)²⁻ sulfur species (27.9 vs. 14.7%) and sulfate state (19.3 vs. 13.0%). The surfactant-assisted synthesis probably provides better dispersion of catalysts precursors leading to formation of nickel molybdate. Besides, surfactant is favorable for sulfidation and presumably provides the organic species that remain in the precipitate and are transformed to carbonaceous matter on further sulfidation [20]. The sulfate species (SO₄ ^2–) may be formed during migration of sulfur atoms from the metal centers onto the O sites during nickel molybdate sulfidation [55].

Based on calculation results, the effective amount of Ni in active phase species for NiMoS is close to 2.5 wt.%, while those obtained for NiMoS + SPAN catalyst is 2.9 wt.%. When the surfactant is used, the effective Mo content in MoS₂ phase reaches up to 45%, which is 6% higher compared to NiMoS counterpart (Table 6). The Ni/Mo ratios in the NiMoS slabs as well as on edges are equal for both catalysts.

Catalytic testing

The hydrogenation of aromatics was chosen as a test reaction to evaluate the behavior of Ni–Mo sulfide nanosized catalysts assembled in situ from water-soluble metal precursors in W/O emulsions stabilized by surfactant. The catalytic evaluation of more active NiMoS + SPAN catalyst was performed in a high-pressure batch reactor at T = 380–420 °С, p(CO) = 5 MPa with water content of 20 wt.% and CO/H₂O molar ratio of 1.8 for 4–8 h. At a temperature above 420–430 °С, the thermal cracking or dehydrogenation of hydrocarbons takes place, so the temperature range was limited to 420 °С. At a temperature below 320 °С the aromatics ring hydrogenation is affected by kinetic factor [56], [57], [58]. This range is also determined by conditions, when the metal precursors are decomposed to oxides followed by sulfidation forming active phase. The lower the temperature, the longer time is needed for active phase formation. Recently we established the reaction induction period where the catalysts assembly depends on temperature and varies from 2 to 3 h at 360–400 °C [25], [26], [27], so the reaction time was not lower than 4 h.

The temperature-dependent activity of Ni–Mo sulfide catalysts in hydrogenation of model substrates is shown in Fig. 6. The anthracene conversion reaches to 97–100% over the whole temperature range, while for 1MN and 2MN it doesn’t exceed 92 and 86% respectively even at 420 °С. This is in a full accordance with data reported [59], where the reactivity of polycyclic aromatics in hydrogenation increases in contrast with that of monocyclic aromatics.

Fig. 6:

Temperature-dependent conversion of model substrates over Ni-Mo sulfide catalysts. Reaction conditions: p(CO) = 5 MPa, ω(substrate) = 10wt.%., ω(H₂O) = 20 wt.% , t = 6 h, ω(SPAN-80) = 4wt.%, CO:H₂O = 1.8 (molar ratio).

Anthracene hydrogenation (Scheme 3) proceeds faster in comparison with methyl-substituted naphthalenes. Thus, the 97% conversion was achieved at 380 °C after 4 h. The reaction products are comprised of dihydroanthracene (2HA), tetrahydroanthracene (4HA), octahydroanthracene (8HN) as well as trace amount of perhydride (14HA). At the temperature above 420 °C thermal cracking occurs.

Scheme 3:

The pathway for anthracene hydrogenation [59].

At 380 °C the anthracene hydrogenation results in 2HA mainly (Fig. 7a). It is due to reactivity of the conjugated double bonds in the anthracene molecule, when the carbon atoms at positions 9 and 10 have high delocalized electron density facilitating the hydrogen consumption [59], [60]. Nevertheless, 2HA content decreases with reaction time accompanied by increase of 4HA and 8HA amounts. On rising temperature to 400 °C, the 4HA becomes the main reaction product with selectivity of 50% at 8 h. The 2HA content decreases while 8HA accumulates in reaction products (Fig. 7b). At higher temperature up to 420 °C, the products distribution is an analog of the one at 380 °C with higher 8HA initial content being time independent (Fig. 7c). Thus, at 400 °С hydrogen consumption is maximal and minimal at 420 °С (Fig. 7d). At higher temperatures and hydrogen pressure deficiency, dehydrogenation may proceed leading to formation of 4HA or 2HA.

Fig. 7:

Product distribution for anthracene hydrogenation at (a) 380 °C, (b) 400 °C and 420 °C over Ni–Mo sulfide catalysts and (d) hydrogen consumption depending on reaction temperature. Reaction conditions: p(CO) = 5 MPa, ω(substrate) = 10 wt.%., ω(H ₂ O) = 20 wt.%, ω(SPAN-80) = 4 wt.%, СО:Н ₂ О = 1.8 (mol).

In the case of methyl-substituted naphthalenes the lower conversion may be caused by steric hindrances during the adsorption of substrate molecules on the active catalytic sites. The time and temperature dependencies of catalytic activity in 2MN and 1MN hydrogenation are presented in Fig. 8. To achieve higher conversions, the reaction temperature should be at least 380 °С and time at least 4 h. On increasing the temperature by 20 °С (from 360 to 380 °С), the 2MN conversion at 6 h increase up to 27%. With the lapse of reaction time to 8 h at 400–420 °С the conversion rises exceeding 90% conversions. Under reaction conditions 1MN and 2MN transform into 5- and 6-methyltetralines (5MT and 6MT), i. e., unsubstituted aromatic rings are hydrogenated probably due to steric effect of methyl group during the sorption. Thus, at 380 °С and 4 h the 6MT:2MT and 5MT:1MT ratios are approximately 1:1. After increasing temperature and reaction time the 6MT and 5MT contents are 1.2–1.5 times higher than 2MT and 1MT respectively. Meanwhile, the selectivity to 6MT and 5MT varies in the range of 50–53%. The higher the temperature and reaction time, the higher the concentration of decalines: 21% for 1MN and 19% for 2MN at 420 °С after 8 h. The Ni–Mo sulfide catalysts were found to be slightly more active in 2MN hydrogenation at least at 380 °С. Despite of comparable 1MN and 2MN conversions, the decalines’ content in hydrogenation products of 2MN is 2–3% lower.

Fig. 8:

Product distribution for hydrogenation reaction of (a) 1MN and (b) 2MN over Ni–Mo sulfide catalysts. Reaction conditions: p(CO) = 5 MPa, ω(substrate) = 10 wt.%., ω(H ₂ O) = 20 wt.%, ω(SPAN-80) = 4 wt.%, СО:Н ₂ О = 1.8 (mol).

The substituent effect in aromatic naphthalene rings on catalytic activity is presented in Fig. 9. The experimental results showed that hydrogenation activity decreases following trend 1,8-dimethylnaphthalene>>1,3-dimethylnaphthalene>2,6-dimethylnaphthalene≈2,3-dimethylnaphthalene>2-ethylnaphthalene.

Fig. 9:

Product distribution for hydrogenation reaction of (a) 1MN and (b) 2MN over Ni–Mo sulfide catalysts. Reaction conditions: p(CO) = 5 MPa, ω(substrate) = 10 wt.%., ω(H ₂ O) = 20 wt.%, ω(SPAN-80) = 4 wt.%, СО:Н ₂ О = 1.8 (mol).

The highest conversion of 45% was achieved for 1,8-dimethylnaphthalene with 100% selectivity to tetralines. For 1,3-dimethyl-substituted substrate the conversion decreases dramatically to 10% being comparable to 2,6-dimethylnaphthalene. The conversion to tetralines not higher than 5% was found for 2,3-dimethyl- and 2-ethyl-naphthalenes. Similar to 1- and 2-methyl-naphthalenes the hydrogenation of asymmetric dimethyl-substituted substrate carries out through the unsubstituted aromatic ring indicating that steric factors probably influence on the sorption mechanism over active metal sites.

Reuse and recyclability

The recycling tests were carried out for 2MN hydrogenation at 380 °С for 6 h with elemental sulfur (the curve «+S» in the Fig. 10) for active phase formation and without it (the curve «-S» in the Fig. 10). At the first cycle, during 2MN hydrogenation, active sulfide particles formed by decomposition of metal precursors and sulfidation of oxides. Then, the reactor was cooled, depressurized and disassembled, reaction products were separated by decantation and in situ assembled catalyst was used without regeneration many times in 2MN hydrogenation. The catalyst was found to be more active in the second cycle without the induction period for the formation of active phase taking place for the first time (Fig. 10). For the next four cycles in the sulfur-assisted process the 2MN conversion slightly decreased, being maintained at 93–95% level. However in the absence of sulfur, the catalyst activity decreases dramatically with cycle number. Thus, the 2MN conversion after six cycles doesn’t exceed 25% that may be caused by the influence of water on active sulfide phase as well as a lack of hydrogen sulfide in the reaction system leading to metal oxides formation. The H₂S generated under reaction conditions exits the system when the reactor is depressurized and disassembled after second reaction cycle thereby sulfiding agent is absent when catalyst is reused.

Fig. 10:

Recycling tests of Ni–Mo sulfide catalysts in 2MN hydrogenation. Reaction conditions: p(CO) = 5 MPa, t = 6 ч, ω(substrate) = 10 wt.%., ω(H ₂ O) = 20 wt.% , ω(SPAN-80) = 4 wt.%, ω(S) = 2.5 wt. % (when the sulfur added) , СО:Н ₂ О = 1.8 (molar ratio).

To evaluate the characteristics of active phase species for the NiMoS + SPAN catalyst after six cycles of 2MN sulfur-assisted hydrogenation, it was separated by filtration followed by step-by-step washing with toluene and hexane. The XPS spectra of reused catalyst revealed the Ni and Mo atomic concentration was close to initial composition (Table 8). The Mo content decreased by 0.3at.% while Ni content was maintained at 2.2at.%. The S/(Ni + Mo) ratio is also comparable (2.0 vs. 1.9). It confirms, that at the first 2–3 h, which is ascribed to the induction period, the oxide catalyst precursors formed slowly by metal salt decomposition, which reveals that it is the rate-determining step. The sulfidation is rather fast based on high catalytic activity data upon recycling. Deconvolution of Mo3d-S2s core level spectra revealed that 75.1% of Mo is in sulfide state (Table 7) with doublet peaks at 228.7 and 231.8 eV (Fig. 11) [19], [20], [61]. The characteristic binding energy of 230.1 eV is ascribed to oxysulfide which contributed 11.4 % and 13.5% of Mo⁶⁺ [13], [19], [20], [61]. Taking into account the same reaction condition from cycle to cycle it is to be assumed that the minor difference in Mo valency state for initial and recycled catalysts is caused by the its contact with air before XPS analysis which resulted in MoS₂ transformation to MoS_xO_y form. It is also confirmed by the decrease of NiMoS phase as well as sulfide to 29.3 and 68.9% respectively. Meanwhile, the NiMoO₄ content is more than 22% higher compared to the initial composition of the catalysts. The effective Mo content in MoS₂ phase decreases to 40.3%, while the ωNiMoS content is 1.8 wt.% and the Ni/Mo ratio in the NiMoS slabs is the same as the initial catalysts (Table 8).

Fig. 11:

Mo3d-S2s, Ni 2p and S 2p core level spectra for NiMoS + SPAN catalyst after six cycles of 2MN sulfur-assisted hydrogenation.