Shan
Jiang
a,
Riyang
Shu
*b,
Anqi
Wang
a,
Zhuoli
Deng
a,
Yuhong
Xiao
a,
Jiajin
Li
a,
Qingwei
Meng
ac and
Qian
Zhang
*ac
aSchool of Chemical Engineering and Light Industry, Guangdong University of Technology, Guangzhou 510006, China. E-mail: zhangqian@gdut.edu.cn
bGuangdong Provincial Key Laboratory of Functional Soft Condensed Matter, School of Materials and Energy, Guangdong University of Technology, Guangzhou 510006, PR China. E-mail: shuriyang@gdut.edu.cn
cGuangdong Provincial Laboratory of Chemistry and Fine Chemical Engineering Jieyang Center, Jieyang 515200, China
First published on 17th July 2024
High-quality liquid biofuels can be produced from renewable lignin-derived phenolic compounds through an efficient hydrodeoxygenation (HDO) process in which the traditional catalysts usually include metal sites and acid sites that catalyze the hydrogenation and deoxygenation procedures respectively. This work presents a novel acid-free NixMoyN/C catalyst from the perspective of green chemistry providing a new pathway for HDO of lignin-derived phenolic compounds that involves hydrogenation deoxygenation and hydrogenolysis at the same time. A series of NixMoyN/C catalysts were prepared by varying the Ni:
Mo molar ratio among which the Ni1Mo3N/C catalyst showed the best HDO performance. Guaiacol could be completely converted at 260 °C after 4 h with 95.8% cyclohexane selectivity. In addition a small amount of benzene could be obtained as a valuable fuel additive by-product by altering the conventional HDO reaction path. By shortening the reaction time benzene could be obtained as an intermediate product with a relative high selectivity. Based on the characterizations using XRD BET SEM TEM XPS H2-TPD and EPR, the results demonstrate that the multiple active components of the Ni1Mo3N/C catalyst allow it to efficiently catalyze the hydrogen activation and C–O bond cleavage even under acid-free conditions. The existence of the active phases of Ni Ni2Mo3N and β-Mo2C as well as the interaction between Ni and Mo metals together contributed toward efficient HDO performance. Not only for the various phenolic model compounds the feasibility of Ni1Mo3N/C catalysts for upgrading raw lignin oil was also demonstrated with the hydrocarbon content increasing from 5.7% to 88.4%. Notably arenes accounted for 18.2% of the hydrocarbon products which confirmed the occurrence of hydrogenolysis in the catalytic process. This work provides a novel route for the conversion of lignin-derived phenolic compounds to produce high-quality hydrocarbon liquid biofuels especially the direct production of arene components.
Hydrodeoxygenation (HDO) is a commonly used pathway to improve the quality of lignin-derived phenolic compounds and it mainly involves hydrogenation and deoxygenation procedures.6 Catalysts are conducive to completing HDO reactions by lowering the reaction activation energy significantly. Metal sites catalyze the hydrogenation procedures while acid sites catalyze the deoxygenation procedure.7 Heterogeneous catalysts are commonly employed in this process due to their good recovery characteristic which is quite compatible with green chemistry principles.8 The metal components of the catalyst are the key catalysis sites to accomplish the hydrogenation process. Many different types of metal-based catalysts have been studied including noble metals (like Pt Pd and Ru)9,10 transition metals (like Ni Cu and Co)11 and metal carbides (like Mo carbide).12 Among these the high cost of noble metals prevents their widespread usage although they can show outstanding HDO catalytic performance under moderate reaction conditions. Transition metals possess low cost and capable catalytic activity compared to noble metals which has seen them attract much attention especially based on green chemistry and sustainability considerations.11 In addition most of the reported studies have presented the formation of saturated hydrocarbons because metal-based catalysts usually have a good hydrogenation ability.9,10 However with a view to practical application the usage of unsaturated aromatic additives in liquid fuels especially arenes can enhance the octane number of automotive fuels reduce the volatility and play a positive role in engine efficiency and its life span.
Transition metal catalysts with a single metal component generally lead to an inadequate HDO activity and efficiency.13 To improve the original sites electronically and geometrically a second active component can be added. This strategy enables changing the surface state of the catalysts and can increase the catalytic activity product selectivity and stability.14–16 Given the strong synergistic catalytic effects a number of bimetallic component catalysts have been studied for the HDO reaction.17,18 For instance Wang et al. studied the catalytic performance of the bimetallic NiS2/MoS2 catalyst in the HDO conversion of guaiacol. Their results showed that the bimetallic component catalyst was more efficient than that of a single component under the same circumstances.19 Also our previous study explored bimetallic Ru-based catalysts in the HDO of lignin-derived phenolic compounds and showed that the addition of Ni and Co metal to Ru/SiO2–ZrO2 catalysts can significantly improve their HDO activities promoted by the synergetic effect of the bimetal species.20,21 Cheng et al. synthesized a series of Fe–Co/SiO2 catalysts via an impregnation process. The bi-component metal-based catalyst system achieved a significantly higher hydrocarbon yield (22.4%) compared to the single-component catalyst.22 Therefore the bimetal catalytic strategy provides an effective method to enhance the HDO performance.
The usage of metal–acid bifunctional catalysts can significantly enhance the deoxygenation of phenolic compounds.23 However catalyst deactivation can also occur due to the fact that excessive acid sites induce a higher rate of coke formation.24 Laurent et al. investigated the catalytic effect of CoMo/Al2O3 on the HDO of guaiacol. Due to the strong acidity of Al2O3 the substrate could easily adsorb on the support and condense to form char. They also found a low carbon number balance which indicated the formation of deposited carbon in the reaction.25 Echeandia et al. discovered that less coke was formed during phenol HDO over an activated-carbon-loaded Ni–W catalyst than an alumina-loaded one.26 To mitigate the negative effects of these polymerization products in the reaction system the use of an acid-free catalyst may be a solution. Also the preparation process of acid-free catalysts avoids potential harm to the environment and has benefits from both economic and green catalysis considerations due to eliminating the acid usage.
In this paper a series of NiMo-based bimetallic catalysts on a carbon support were synthesized without the addition of an acid component in line with a green chemistry perspective. The active species composition oxygen vacancy properties and NiMo intermetallic interactions could be modulated by varying the Ni:
Mo ratio. The catalysts were then employed for the HDO reaction of lignin-derived phenolic compounds. A series of catalyst characterizations were carried out and the HDO performance of the different catalysts was compared in order to reveal the relationship between the catalysts’ physicochemical properties and the reaction performance. Finally a potential catalytic reaction mechanism was proposed based on the characterization and experiment results.
![]() | (1) |
![]() | (2) |
As for the HDO reaction of raw lignin oil the substrate included a certain amount of phenolic compounds (about 0.15 g) which was extracted from 3 g raw lignin oil with 20 mL octane. Then the substrate and 0.05 g catalyst were added into the reactor and the reaction process was the same as that for the above HDO process of phenolic compounds. Considering the complexity of the lignin oil only semiquantitative measurements were conducted by GC-MS before and after the HDO reaction. The product content was calculated based on the proportions of the peak areas.
![]() | ||
Fig. 1 (a) XRD patterns of the different NixMoyN/C catalysts. (b) XRD patterns of the Ni1/C Mo3/C and Ni1Mo3/C catalysts. |
The specific surface area and pore structure of a catalyst have a significant impact on the exposure of active sites. The results from the N2 porosimetry tests of the different NixMoyN/C catalysts are shown in Table 1 and Fig. S1.† Generally all the catalysts showed typical type IV isotherms and hysteretic rings in a high range of P/P0 (Fig. S1†). Combined with the pore volume data in Table 1 mixed structures with mesopores and micropores were demonstrated. Among these catalysts the Ni1Mo3N/C catalyst had the largest mesopore volume value due to the enhanced formation of β-Mo2C during calcination in which the Mo2C phase is reported to have an abundant mesopore structure.30,31 This result was also supported by the XRD analysis (Fig. 1). Moreover with the addition of excess Mo metals in the Ni1Mo4N/C catalyst both the specific surface area and the pore volume decreased including the mesopore volume. Overall the Ni1Mo3N/C catalysts had the largest mesopore volume value which is favorable for the adsorption and diffusion of large molecular substrates. The element content of the different prepared NixMoyN/C catalysts was also investigated. Based on the measurements by ICP-OES analysis and organic elemental analysis the results for the catalyst composition are listed in Table 1. It could be seen that the actual element content values were close to the theoretical ones and the Ni1Mo3N/C catalysts presented a Ni/Mo molecular ratio of 0.43.
Catalysts | Surface areaa (m2 g−1) | Pore diameterb (nm) | Pore volumec (cm3 g−1) | Element content (wt%) | Ni/Mog | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|
V tot | V mic | V meso | Nid | Mod | Ce | Ne | Of | ||||
a MultiPoint Brunauer–Emmett–Teller (BET) method. b Barrett–Joyner–Halenda (BJH) method. c V tot: total pore volume determined using density functional theory (DFT) and Boehm's titration. Vmic: micropore volume calculated using the t-plot method. Vmeso: mesopore volume determined by the difference between Vtot and Vmic. d Ni and Mo contents in the catalysts determined by ICP-OES analysis. e C and N contents in the catalysts determined by organic elemental analysis. f Elemental oxygen content based on the difference between the other elements and the total. g Ni/Mo atomic ratios calculated from the ICP-OES results. | |||||||||||
Ni1Mo1N/C | 113 | 1.9 | 0.053 | 0.040 | 0.013 | 11.9 | 15.9 | 38.0 | 2.1 | 32.1 | 1.2 |
Ni1Mo2N/C | 138 | 2.4 | 0.083 | 0.045 | 0.038 | 15.0 | 39.5 | 20.8 | 2.9 | 21.8 | 0.62 |
Ni1Mo3N/C | 88 | 3.2 | 0.072 | 0.020 | 0.052 | 15.2 | 57.7 | 13.8 | 1.5 | 11.8 | 0.43 |
Ni1Mo4N/C | 60 | 3.6 | 0.054 | 0.011 | 0.043 | 12.3 | 61.4 | 9.2 | 2.4 | 14.7 | 0.33 |
The measurements for the catalyst morphology were conducted by SEM and TEM characterizations. The SEM images of the different catalysts are presented in Fig. 2 and Fig. S2† in which a large amount of granular material could be observed on the catalyst surface. Based on the XRD analysis these particles were probably metal Ni and Ni2Mo3N species. There was a tendency for the size of these particles to decrease with the increasing addition of the Mo metal element which may be related to the occurrence of a transition from Ni to Ni2Mo3N. Comparatively the Ni1Mo3N/C catalyst showed the smoothest surface and the smallest metal particle size. SEM-based EDS elemental analysis of the different NixMoyN/C catalysts was also carried out (Fig. S3–S6†). Each type of element was uniformly distributed on the catalyst surface indicating the effectiveness of the preparation method. To further understand the internal structure of these catalysts TEM characterization analysis was also performed (Fig. 3a–d). The dark region in these images is representative of the metal particles and the light color region is representative of the C support. The results show that with the increase in Mo addition the particle size became smaller gradually. The Ni1Mo3N/C catalyst showed a small metal particle size as well as high dispersion. Although the Ni1Mo4N/C catalyst seemed to have the smallest metal particle size the particle dispersion was not uniform which probably had a negative effect on the HDO catalytic performance. The TEM-based EDS examination of Ni1Mo3N/C catalysts (Fig. 3e–i) revealed that Ni elements were concentrated inside the particle form whereas the Mo C and N components were evenly dispersed. Furthermore the lattice characteristics of the specific metal phase were examined by HR-TEM. The images in Fig. 3j–m show the lattice characteristics of the (110) (111) and (210) crystal planes of Ni2Mo3N and the (111) crystal plane of the metal Ni. Besides lattice fringes belonging to β-Mo2C were also found in the Ni1Mo3N/C catalyst (Fig. 3n and o). These results were in good agreement with the XRD results (Fig. 1).
![]() | ||
Fig. 3 TEM images of (a) Ni1Mo1N/C (b) Ni1Mo2N/C (c) Ni1Mo3N/C and (d) Ni1Mo4N/C catalysts (e–i) EDS elemental mappings and (j–o) HR-TEM images of the Ni1Mo3N/C catalyst. |
The valence states of the catalyst components were explored in detail by XPS. The whole spectra of the NixMoyN/C catalysts are shown in Fig. S7† and the specific element spectra are shown in Fig. 4. As for the Mo 3d spectrum in Fig. 4a the distance between the Mo 3d3/2 and Mo 3d5/2 bimodal peaks due to spin–orbit splitting was 3.1 ± 0.3 eV.32 The signals of the resolved peaks at 229.9 and 232.9 eV were related to Mo4+ and Mo6+ which were also attributed to the MoO2 and MoO3 phases respectively. The fluctuating trend in the content of the above oxide components in these NixMoyN/C catalysts (Table 2) was consistent with the changes in the oxygen content given in Table 1 which suggests that an increase in the Mo component would result in a greater tendency to produce nitrides or carbides in the NixMoyN/C catalysts. Furthermore the signal near 228.5 eV in the Ni1Mo3N/C and Ni1Mo4N/C catalysts was particularly enhanced which points to the presence of Mo2+ in the form of Mo–C bonds for the Mo2C phase.33 This result was also consistent with the XRD analysis showing the presence of Mo2C crystals. The binding energy at 228.9 eV was attributed to the Moδ+ (0 < δ < 4) species which were mainly present in the Ni2Mo3N phase.34,35 In addition a 232.3 eV signal was observed corresponding to the MoOxCy phase and assigned to the Mo5+ valence state formed by the insertion of oxygen atoms into the carbide lattice of Mo. The undercoordinated Mo5+ facilitates the formation of oxygen vacancies and promotes the HDO reaction.36 Moreover the positions of the Mo deconvolution peaks were also discriminated and the contents of Mo species with different valence states were calculated (Table 2). It was found that the binding energy values of Mo2+ Moδ+ and Mo5+ in the Ni1Mo3N/C catalysts were shifted to a lower direction compared to the other catalysts. This suggests that the electronic structure of Mo atoms is altered in nitrides and carbides and these negative shifts are related to the down-shift in the d-band center and the high charge density around the Mo species.37 These results clearly demonstrate the formation of the Ni2Mo3N and β-Mo2C active phase.
![]() | ||
Fig. 4 XPS spectra of NixMoyN/C catalysts: (a) Mo 3d spectra (b) Ni 2p spectra (c) C 1s spectra (d) N 1s spectra (e) O 1s spectra. |
Samples | Mo | Ni | |||||
---|---|---|---|---|---|---|---|
Mo2+ (Mo2C) | Mo4+ (MoO2) | Moδ+ (Ni2Mo3N) | Mo5+ (MoOxCy) | Mo6+ (MoO3) | Ni0 | Ni2+ (NiO & Ni2Mo3N) | |
Reaction conditions: 20 mL octane solvent 0.1 g guaiacol 0.05 g catalyst 1 MPa H2 reaction time was 4 h. | |||||||
Ni1Mo1N/C | 228.6 eV (20.3%) | 229.9 eV (25.8%) | 229.1 eV (7.4%) | 232.2 eV (7.7%) | 232.8 eV (38.7%) | 852.9 eV (17.3%) | 856.4 eV (82.7%) |
Ni1Mo2N/C | 228.5 eV (3.2%) | 229.8 eV (20.8%) | 228.7 eV (41.2%) | 232.0 eV (8.3%) | 232.7 eV (26.6%) | 853.1 eV (20.3%) | 856.4 eV (79.7%) |
Ni1Mo3N/C | 228.4 eV (41.8%) | 230.0 eV (14.2%) | 228.6 eV (10.4%) | 232.0 eV (25.5%) | 233.5 eV (8.1%) | 853.2 eV (10.6%) | 856.5 eV (89.4%) |
Ni1Mo4N/C | 228.4 eV (36.2%) | 230.1 eV (24.4%) | 228.7 eV (10.2%) | 232.1 eV (15.9%) | 232.5 eV (13.4%) | 853.2 eV (30.1%) | 856.4 eV (69.9%) |
Deconvolutional analysis of the Ni 2p spectrum (Fig. 4b) and calculation of the atomic content (Table 2) were also conducted. The results show that Ni0 and Ni2+ valence states were predominantly present in the different NixMoyN/C catalysts corresponding to binding energies of 852.9 eV and 856.4 eV respectively.38 The unavoidable Ni surface oxidation resulted in the existence of Ni2+ valence states due to the easy oxidation property of Ni metal species. Moreover most the Ni2+ component in the Ni1Mo3N/C catalyst resulted from the involvement of Ni species in the formation of the Ni2Mo3N phase which allowed the transformation of Ni0 to Ni2+.39 Ni0 species in Ni1Mo3N/C catalyst were observed to have a higher binding energy compared to that in the other catalysts. This shift corresponds to the change in Mo species and led to the moving to a lower binding energy which demonstrates the charge transfer of Ni toward Mo and produces a strong interaction between Ni and Mo. The C 1s peaks of the different NixMoyN/C catalysts in Fig. 4c could be deconvoluted into three peaks corresponding to C–C (284.8 eV) C–O (286.1 eV) CO (288.5 eV) respectively.38 In addition signals attributed to C–Mo bonds were found at 283.7 eV in the cases of the Ni1Mo3N/C and Ni1Mo4N/C catalysts and indicated the presence of Mo carbides.33 The XPS spectrum of the N 1s peaks in Fig. 4d showed the presence of pyrrolic N (399.3 eV) and pyridinic N (380.0 eV)35 indicating that N had been introduced successfully. A peak at a binding energy of 397.2 eV assigned to the N–Mo bond could be observed for all the catalysts while the strongest N–Mo interactions were obtained in the Ni1Mo3N/C catalyst.40 Moreover the strong peak observed near 394.2 eV belonged to the Mo 3p spectrum.34
To further investigate the relationship between the undercoordinated Mo5+ and oxygen vacancies in the MoOxCy phase the O 1s orbital was also analyzed by XPS (Fig. 4e). The deconvolution spectrum at 530.8 eV belonged to lattice oxygen (OL) which was predominantly present in the oxide phase state of Mo metal. The peak of 532.4 eV belonged to the oxygen vacancies (OV) and was related to the insertion of lattice carbon into the oxide lattice of Mo to form Mo5+ which was capable of a rapid interconversion with Mo6+. Such a redox cycle facilitates the activation of the C–O bond and avoids the hydrogenation of the CC bond.37,41,42 Moreover the surface OH groups (OC) and chemisorbed dissociated oxygen species were associated with the peaks at 533.0–534.2 eV. The content of oxygen vacancies could be estimated by the OV/OL ratio. These NixMoyN/C catalysts had OV/OL ratios of 0.28 0.33 0.48 and 0.35 respectively with specified Ni–Mo ratios x
:
y of 1
:
1 1
:
2 1
:
3 and 1
:
4. These results were consistent with the fluctuation of the Mo5+ species. Overall the Ni1Mo3N/C catalyst possessed the most oxygen vacancies which would probably facilitate the substrate adsorption and promote the HDO efficiency.
To further characterize the oxygen vacancy property EPR characterization was employed to capture information on the unpaired electrons in different NixMoyN/C catalysts. As shown in Fig. S8† all the NixMoyN/C samples had a symmetrical pair of peaks at g = 2.006 which is a typical diagnostic signal for oxygen vacancies. The intensity of the EPR signals was also used to determine the relative content of the oxygen vacancies.43 Here the Ni1Mo3N/C catalyst exhibited the strongest symmetry signal with most electrons trapped in oxygen vacancy sites which was consistent with the XPS results.
In order to explore the hydrogen-adsorption capacity of the catalysts H2-TPD characterization of the different NixMoyN/C catalysts was performed. As shown in Fig. 5 compared to the other catalysts the reversible desorption signals of the Ni1Mo3N/C and Ni1Mo4N/C catalysts were compressed to a narrow range of desorption temperatures of 650–700 °C and 690–740 °C respectively. This suggests that the active species for the main chemisorption of H2 had changed. With the increasing Mo the hydrogen-activation species was transformed from metal nickel to bimetallic nitrides and the formed Mo carbide phase also possessed a certain hydrogen-adsorption ability. Besides the doping of N elements into the NiMo alloy lattice resulted in a continuous energy gap around the Fermi level in Ni2Mo3N which was associated with a rich metal–ligand structure.44 Such inherent metallic properties help to accelerate the charge-transfer process and produce adsorption and dissociation effects on hydrogen. In addition Mo2C was formed as a hexagonal closed packed crystal by the insertion of C atoms while the electron transfer from Mo to C also lowered the electron density surrounding the Mo active site resulting in weak hydrogen bonding and an increased susceptibility to interference with the hydrogen-adsorption process.45,46 Moreover it has been reported that Mo2C possesses some Brønsted acidic sites on its surface that have the ability to break C–O bonds in a dehydration process.47 These complementarities indicate the possibility of a synergy effect between Ni2Mo3N and Mo2C. In addition the H2 desorption peak of the Ni1Mo3N/C catalyst was shifted toward lower temperature compared to that of the Ni1Mo4N/C catalyst which demonstrates that desorption of the hydrogen can occur at low temperature and can promote the HDO reaction.
Catalysts | Conversion (%) | Temperature (°C) | Product selectivity (%) | |||||
---|---|---|---|---|---|---|---|---|
Ni1Mo1N/C | 7.9 | 260 | 8.8 | 8.1 | 3.9 | — | 79.2 | — |
Ni1Mo2N/C | 87.1 | 260 | 16.2 | 1.7 | 56.2 | 25.2 | 0.8 | — |
Ni1Mo3N/C | 99.9 | 260 | 95.8 | 4.2 | — | — | — | — |
Ni1Mo4N/C | 93.0 | 260 | 33.8 | 13.1 | 27.6 | — | 17.6 | 8.0 |
Ni1/C | 91.2 | 260 | 11.5 | 1.3 | 48.0 | 38.6 | 0.6 | — |
Mo3/C | 24.3 | 260 | — | — | — | — | 99.9 | — |
Ni1Mo3/C | 99.9 | 260 | 17.7 | — | 57.6 | 24.7 | — | — |
Ni1Mo3N/C | 53.2 | 240 | 14.6 | 4.0 | 58.6 | 15.6 | 7.2 | — |
Ni1Mo3N/C | 38.4 | 220 | 7.0 | — | 47.3 | 44.0 | 1.6 | — |
Combined with the XPS and EPR analysis it could be found that the variation of the Mo5+ species content in these NixMoyN/C catalysts was consistent with the variation in their oxygen vacancy content as well as their HDO performances. Mo5+ species are easily oxidized by carbon-containing substrates to the higher valence Mo6+ phase thus easily adsorbing the surrounding oxygen-containing substrates with reducing properties.37 In addition the stronger interactions between Ni and Mo in the Ni1Mo3N/C catalyst resulted in a higher charge density around the Mo atoms and the stronger metallicity of Mo species also contributed to a higher efficiency of the catalyst.38 Besides there was also a small amount of Ni0 which also plays a role in the H2-activation procedure. However this is different from the conventional metal–acid site HDO catalytic process. In the absence of significant acid sites the Ni1Mo3N/C catalyst still exhibited good HDO performance indicating that Ni1Mo3N/C does not completely rely on acid sites to catalyze the conversion of guaiacol to cyclohexane. The combination of hydrogenation deoxygenation and hydrogenolysis finally resulted in a high HDO efficiency of guaiacol. Overall the addition of Mo to the Ni1Mo3N/C catalyst contributed to the destruction of the initial carbonaceous structure leading to a decrease in the specific surface area and pore size. While the increased contact between Ni and Mo causes Mo oxides to change into carbide and bimetallic nitride phases. This formation enhances the interaction between Ni and Mo and increases the content of Mo5+ simultaneously thus achieving a high HDO activity. Eventually there is a synergistic catalytic effect between the Ni2Mo3N phase and β-Mo2C phase with remarkable lattice features each of which plays a specific catalytic role for a particular reaction step. Due to these factors the Ni1Mo3N/C catalyst could display high HDO activity even in the absence of acid conditions.
Given that the HDO performance of the catalysts is highly related to the reaction temperature the influence of the reaction temperature on guaiacol HDO was examined with the Ni1Mo3N/C catalyst. As can be seen in the Table 3 relatively few cyclohexane products were produced at a reaction temperature of 220 °C and the conversion of guaiacol was only 38.4%. The conversion rose to 53.2% when the reaction temperature was increased to 240 °C although the selectivity for methoxy cyclohexanol dramatically decreased. This result suggests a considerable elimination of the oxygen-containing functional groups. Also both the conversion and the selectivity of the cyclohexane product rose to about 100% when the reaction temperature was raised to 260 °C. This indicates that increasing the reaction temperature can result in the complete deoxygenation products and that the reaction temperature plays a significant role in the HDO process.
A comparison between the Ni1Mo3N/C catalyst in this work and the reported catalysts with acid-free supports was also conducted. Non-precious metals Ni and Co are frequently combined with other active species (such as phosphorus Fe metal and molybdenum carbide) to create HDO catalysts for lignin-derived phenolic compounds. Also the catalytic hydrogenation is commonly carried out at temperatures above 300 °C to achieve high conversion rates (Table 4 entries 2–5). The requirement for high temperature puts high demands on the reactor equipment. The metal Ni has also been used by researchers to couple with molybdenum carbide to design catalysts for the HDO reaction. This kind of catalyst typically needs a higher initial hydrogen pressure to promote CC saturation which results in a wastage of energy and incurs potential risks (Table 4 entries 6–9). In the studies of the HDO of non-precious metal Co the production of fuel-grade products was difficult to achieve due to its lack of deoxygenation capacity (Table 4 entries 10 and 11) but it may play a role in partial HDO for the production of oxygenated chemicals. Carbides and sulfides of Mo metal as well as carbides of W have also been used to conduct the HDO of lignin-derived phenolic compounds. Molybdenum carbide species have a certain hydrogenolysis capacity but due to the insufficiency of their hydrogen activation the combination of them with components with a strong hydrogen-activation capacity may be an effective solution to improve their HDO capacity (Table 4 entries 12–15). Mo sulfide species are traditionally used as catalysts for fossil energy improvements but it is difficult to maintain their stability due to the gradual loss of their S elements during the reaction (Table 4 entries 16 and 17). Also the gradual popularization of S-free renewable fuels today signals resistance to the widespread usage of S-containing catalysts. W2C has also been used in HDO for the upgrading of biomass fractions to fuels due to the fact that W has a configuration of an extra-nuclear electron similar to that of Mo (Table 4 entry 18). However its performance seems to be inferior to that of the active Mo-based catalysts even under more adequate reaction conditions. Metal oxides have also been used as HDO catalysts for the upgrading of phenolic compounds (Table 4 entries 19 and 20) but their performances are limited and need further development. Overall in comparison to other reported works using different Ni-based catalysts and Mo-based catalysts without acid addition the results in this work were far superior in terms of the mild reaction conditions and high hydrocarbon product selectivity.
Entry | Catalyst | Substrate | T (°C) | Atmosphere condition | Conversion (%) | Hydrocarbon product selectivity (%) | Hydrocarbon product yield (%) |
---|---|---|---|---|---|---|---|
1 | Ni1Mo3N/C (this work) | Guaiacol | 260 | 1 MPa H2 | 99.9 | 99.9 | — |
2 | Ni/ZrP48 | Guaiacol | 300 | 4 MPa H2 | 100.0 | — | 80.0 |
3 | Ni–Fe/CNT49 | Guaiacol | 300 | 3 MPa H2 | <99.0. | <90.0 | — |
4 | NiMoS2/CMK-350 | Guaiacol | 300 | 5 MPa H2 | 100.0 | <60.0 | — |
5 | Ni–P-5-30051 | p-Cresol | 350 | 4 MPa H2 | 85.0 | 65.8 | — |
6 | Ni/β-Mo2C52 | Dihydroeugenol | 260 | 2 MPa H2 | 100.0 | 55.0 | — |
7 | Ni1/β-Mo2C52 | 25-Dimethoxyphenol | 260 | 4 MPa H2 | 100.0 | <99.6 | — |
8 | Ni1/β-Mo2C52 | Catechol | 260 | 4 MPa H2 | 100.0 | <99.9 | — |
9 | Ni–MoC-80053 | Guaiacol | 320 | 4 MPa H2 | 73.2 | 2.6 | — |
10 | Co/ZrP48 | Guaiacol | 300 | 7 MPa H2 | 100.0 | — | 76.0 |
11 | Co–Mo-0.5-20054 | p-Cresol | 275 | 4 MPa H2 | 100.0 | 6.3 | — |
12 | Mo2C/CNF55 | Guaiacol | 300 | 2 Mpa H2 | 67.0 | 11.1 | — |
13 | Mo2C/CNF56 | Guaiacol | 300 | 2 MPa H2 | 79.7 | — | 60.0 |
14 | Mo2C@C57 | Guaiacol | 340 | 2.8 MPa H2 | 76.3 | — | 0 |
15 | Mo2C/CNF58 | Guaiacol | 300 | 2 MPa H2 | <99.0 | — | <42.7 |
16 | MoS2/AC59 | Eugenol | 300 | 3 MPa H2 | 99.9 | — | 4.2 |
17 | MoS2/C60 | Guaiacol | 300 | 5 MPa H2 | 86.0 | — | <10.0 |
18 | W2C/CNF61 | Guaiacol | 350 | 5.5 MPa H2 | 66.0 | <2.0 | — |
19 | RuO2–ZM62 | Guaiacol | 300 | 2 MPa H2 | <70.0 | — | <10.0 |
20 | MoO3/AC59 | Eugenol | 300 | 3 MPa H2 | 74.6 | — | 0 |
To reveal the HDO reaction pathway of guaiacol the effect of the reaction time on the real-time product distribution was also investigated. As shown in Fig. 6 the conversion of guaiacol could reach 89.9% after 2 h of reaction and the selectivities for cyclohexane benzene methoxycyclohexanol and cyclohexanol were 23.7% 12.0% 33.4% and 22.5% respectively. The selectivity of cyclohexane increased to 95.8% when the reaction time was increased to 4 h. Remarkably the products from the reaction time of 0.5 h showed a distribution of 42.1% phenol and 7.9% anisole. This implies that a portion of the guaiacol is first dehydroxylated or methoxylated prior to hydrogenation during the reaction. This pathway is different from common studies which typically take into account hydrogenation of the benzene ring of lignin-derived phenolic compounds at the metallicity site into a saturated product as a favorable condition for the deoxygenation reaction during the operation of a bifunctional metal–acid multiphase catalyst.15 Moreover the selectivity for methoxy cyclohexane and methoxy cyclohexanol products remains low most likely due to their roles as chemical intermediates that may be quickly converted to cyclohexane and cyclohexanol respectively. Based on the real-time product distributions the possible reaction pathways were identified. As shown in Fig. 7 there are two different reaction pathways. (1) Guaiacol is first successively deoxygenated at the β-Mo2C site to produce an unsaturated benzene ring without oxygen followed by hydrogenation at the Ni2Mo3N site and Ni site to produce a cyclohexane product; (2) guaiacol first saturates the benzene ring with activated hydrogen at the Ni2Mo3N site and Ni site and then dehydrates at the β-Mo2C site to form a fully deoxygenated cyclohexane. These two reaction pathways display a competitive relation thus resulting in the formation of benzene and methoxycyclohexanol as the intermediate products respectively.
![]() | ||
Fig. 6 Results of varying the reaction time on the HDO of guaiacol. Conditions: 0.1 g guaiacol 0.05 g Ni1Mo3N/C 20 mL octane 260 °C 1 MPa H2 4 h. |
![]() | ||
Fig. 7 Schematic illustration showing the catalytic role of the multiple active components in the catalysts and the possible HDO pathways of guaiacol over Ni1Mo3N/C catalyst. |
To further investigate the reusability of the Ni1Mo3N/C catalyst recycling tests were also conducted. The recycled catalyst was washed with deionized water and ethanol after each reaction and the subsequent recycle reaction was then carried out under the same parameters as previously mentioned. As can be seen in Fig. 8 the catalyst exhibited high recycle stability even after six consecutive operations maintaining 99.9% conversion and 99.9% hydrocarbons selectivity (benzene and cyclohexane). These results demonstrate the high stability of the Ni1Mo3N/C catalyst which presents it with great potential for practical application in the HDO upgrade of lignin-derived phenolic compounds.
![]() | ||
Fig. 8 Results for the catalyst recycling times on the HDO of guaiacol. Conditions: 0.1 g guaiacol 0.05 g Ni1Mo3N/C 20 mL octane 260 °C 1 MPa H2 4 h. |
Owing to the high concentration of unsaturated bonds and oxygen-containing groups raw lignin oil is rarely directly used as a liquid fuel until it has been refined and upgraded by an HDO approach. The Ni1Mo3N/C catalyst was also investigated to examine its HDO characteristics for the lignin oil composite fractions. Table 6 and Fig. S9† display the precise composition of the lignin oil both before and after HDO upgrading. Analysis of the raw lignin oil indicates that it consisted of 5.7% hydrocarbons 19.0% alkylphenols and 74.4% guaiacols and syringols and 0.9% other oxygenated compounds. The stability and calorific value of lignin oil were significantly increased by HDO modification at 260 °C. The guaiacols and syringols components completely disappeared with the hydrocarbon content increasing to 88.4%. The hydrocarbon products included a certain amount of arenes and their content reached 18.2% which was similar to the results for the HDO of other model compounds. This result is quite different from other reported studies since the removal of oxygen from lignin oil is often accompanied by a process of benzene ring saturation due to the consumption of large amounts of hydrogen and the existence of strong acid sites.64,65 Overall the novel Ni1Mo3N/C catalyst in this work has the potential to provide new pathways to convert lignin oil into high-quality liquid biofuels.
Raw lignin oil | Upgraded lignin oil | ||||
---|---|---|---|---|---|
RT (min) | Component | Content (%) | RT (min) | Component | Content (%) |
Conditions: 0.05 g Ni1Mo3N/C 3 g raw lignin oil extracted by 20 mL octane 1 MPa H2 260 °C 4 h. The content was calculated by the peak areas. Components listed are those represented for more than 0.4% of the content. RT: retention time. | |||||
Hydrocarbons | 5.7 | Hydrocarbons | 88.4 | ||
22.92 | 3-Methylphenylacetylene | 1.5 | 1.72 | Cyclohexane | 22.9 |
26.05 | Naphthalene | 3.2 | 2.51 | Cyclohexane 13-dimethyl- | 14.0 |
27.30 | Naphthalene 2-methyl- | 0.9 | 2.84 | Cyclohexane ethyl- | 9.0 |
Alkylphenols | 19.0 | 3.59 | Benzene | 0.7 | |
27.38 | Phenol 26-dimethyl- | 1.0 | 4.37 | Cyclohexane propyl- | 19.4 |
28.04 | Phenol 3-methyl- | 8.7 | 6.16 | Toluene | 3.1 |
28.58 | p-Cresol | 8.4 | 7.51 | Cyclohexane butyl- | 1.2 |
29.24 | Phenol 3-ethyl- | 1.0 | 9.49 | Ethylbenzene | 3.1 |
Guaiacols and syringols | 74.4 | 13.06 | Benzene propyl- | 2.4 | |
27.01 | Phenol 2-methoxy- | 19.8 | 19.99 | Indane | 1.8 |
27.09 | Phenol 2-methoxy-5-methyl- | 1.7 | 22.73 | Benzene 1-methyl-4-(2-propenyl)- | 0.9 |
28.79 | Phenol 2-methoxy-4-propyl- | 5.1 | 23.17 | Benzene 2-ethenyl-14-dimethyl- | 1.1 |
29.11 | Phenol 2-ethyl-6-methyl- | 1.3 | 23.56 | Naphthalene 1234-tetrahydro- | 2.7 |
29.19 | Eugenol | 9.1 | 24.92 | Naphthalene 1234-tetrahydro-6-methyl- | 1.1 |
29.50 | 26-Dimethoxytoluene | 1.3 | 25.30 | Benzene cyclohexyl- | 1.0 |
29.73 | Phenol 2-methoxy-4-(1-propenyl)- | 6.6 | Alkylphenols | 9.8 | |
29.78 | Phenol 26-dimethoxy- | 5.1 | 24.66 | Phenol 3-methyl-6-propyl- | 1.7 |
30.30 | Phenol 2-methoxy-4-(1-propenyl)-(Z)- | 18.4 | 27.38 | Phenol 26-dimethyl- | 0.7 |
30.62 | Benzene 123-trimethoxy-5-methyl- | 2.7 | 28.04 | Phenol 3-methyl- | 1.9 |
31.43 | Phenol 26-dimethoxy-4-(2-propenyl)- | 2.0 | 28.50 | Phenol 3-ethyl- | 0.4 |
32.00 | (E)-26-Dimethoxy-4-(prop-1-en-1-yl)phenol | 1.2 | 28.57 | Phenol 23-dimethyl- | 2.4 |
Other oxygenated compounds | 0.9 | 29.63 | 2-Methyl-6-propylphenol | 0.7 | |
25.16 | Acetophenone | 0.9 | 29.76 | Phenol 2-propyl- | 1.1 |
Guaiacols and syringols | 0 | ||||
Other oxygenated compounds | 1.8 | ||||
22.27 | Benzene 1-methoxy-3-methyl- | 0.9 | |||
23.65 | 3-Ethylphenol methyl ether | 0.9 |
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4gc02298j |
This journal is © The Royal Society of Chemistry 2024 |