Catalytic depolymerization of kraft lignin for liquid fuels and phenolic monomers over molybdenum-based catalysts: The effect of supports

WANG Yi-shuang ， CHEN Ming-qiang,,* ， SHI Jing-jing ， ZHANG Jin-hui ， LI Chang ， WANG Jun

（1.School of Chemical Engineering, Anhui University of Science and Technology, Huainan 232001, China；2.Analytical and Testing Center, Anhui University of Science and Technology, Huainan 232001, China）

Abstract: Catalytic lignin depolymerization (CCLD) for liquid fuels and phenolic monomers was investigated over various supports including clays (e.g., sepiolite (SEP), attapulgite (ATP), and montmorillonite (MTM)), and oxides (e.g., Al2O3 and SiO2) as well as their supported Mo-based catalysts under supercritical ethanol.The characterization results demonstrated that different supports with diverse structural properties could affect the textural structures, surface Mo5+ content, and acid sites distribution.Clay-based supports had more strong acid sites as compared with Al2O3 and SiO2, which went against the production of lignin oil (LO) and led to form more solid products during CLD experiments.Meanwhile, the obtained petroleum ether-soluble product (PEsp) in LO catalyzed by sole supports was mainly alkyl/alkoxy substituted phenols.Additionally, Mo species (especially Mo5+) significantly increased the yields of LO and PEsp.Mo/SiO2 had the highest surface Mo5+ species, showing the highest LO yield of 85.2%, in which the produced alkyl/alkoxy substituted phenols reached 450.3 mg/glignin.Among the clay-supported Mo catalysts, Mo/SEP presented superior LO (82.3%) and PEsp (70.8%) yields and the generated substituted phenols reached 398.8 mg/glignin.This paper systematically reported the application of green and environmentally friendly clay-based materials in lignin conversion, which provides some key information for the development of clay catalysts for biomass conversion.

Key words: kraft lignin；catalytic depolymerization；lignin oil；Mo-based catalysts；clay materials；phenolic monomers

Efficient utilization of lignocellulosic biomass resources is of great significance for achieving the goal of carbon peak and carbon neutral in China.Cellulose,hemicellulose and lignin are the three major components of lignocellulosic biomass.Compared with the mature industrialized application technology of cellulose and hemicellulose, efficient lignin valorization still remains in the research and development stage.At present, lignin is mainly utilized through direct combustion technology, which leads to a serious waste of resources and emission of toxic contaminants[1-3].Lignin, as one of the main components of lignocellulose (accounting for 15%-30% of its weight), is an amorphous and threedimensional network polymer containing a large number of aromatic structures, which is composed of guaiacyl, syringyl, andp-hydroxyphenyl phenylpropane structural units interwoven by ether linkages and C-C bonds[4].For the time being, catalytic conversion of lignin into liquid fuels and high valueadded platform compounds (e.g., phenolic monomers)is considered to be one of the promising utilization technologies[2,5,6].

On account of extremely complicated natural macromolecular structure, lignin presents high inertia and lower contact probability with the catalytic active sites during catalytic depolymerization.Additionally,the formed highly reactive intermediates (e.g., benzene free radicals) will be rapidly converted into large molecules through condensation/polymerization during the depolymerization process, and the obtained macromolecular polymers are difficult to decompose and more prone to form a large amount of coke[7-10].Therefore, precisely designing catalysts and catalytic reaction processes to achieve efficient conversion of lignin under mild conditions is considered as a promising strategy for solving above mentioned technical bottleneck in lignin conversion.

The catalytic hydrogenation of lignin is one of the research hotspots in recent years.This is attributed tothe fact that the formed active hydrogen species (H*)can effectively react with the highly reactive intermediates that are generated in the depolymerization process, then inhibiting the polymerization.Furthermore, active hydrogen species also can attack the C=C, C=O, and other unsaturated bonds, weakening/breaking the C-C and C-O bonds between structural units and transforming lignin into small-molecule products[11-13].Although molecular hydrogen as a hydrogen source has been extensively studied in the lignin hydrodepolymerization, the activation of molecular hydrogen and the excessive hydrogenation of benzene ring increase the energy consumption and reduce the selectivity of the target product.Xiao et al.found that MoOx/CNT facilitated directional breakage of β-O-4 bond in lignin and achieved 47% of monophenols in the H2atmosphere,but the destruction of the aromatic structure leading to the generation of cycloalkanes was unavoidable[14].

A large number of studies have demonstrated that ethanol is widely used in the catalytic lignin depolymerization (CLD), in which ethanol acting as a chelating agent andin-situhydrogen source could effectively inhibit the polymerization of intermediates and promote the hydrogenation process[15-19].In addition, supercritical ethanol could fully dissolve lignin fragments and increase their contact chances with the catalytic sites[18-20].The group of Hensen studied the catalytic depolymerization behavior of CuMgAl oxide in a supercritical ethanol system and demonstrated that the supercritical ethanol system was conducive to the directed conversion of lignin into phenolic monomers[17-19].Furthermore, active hydrogen species produced by ethanol dehydrogenation could improve the hydrogenolysis of the chemical bonds of lignin and benefit the deoxygenation of monomeric and oligomeric intermediates[18].

In recent years, transition metal catalysts (e.g., Cu,Ni, and Mo) have been extensively studied in the CLD process under supercritical ethanol[16-25].Huang et al[21]found that CuMgAlOxcatalyst could achieve one-step conversion of lignin in supercritical ethanol, obtaining 23% of monomer yield without char formation.Jeong and his colleagues[22]investigated the CLD overMOMgAlOycatalysts (M=Co, Ni and Cu) in supercritical ethanol and demonstrated that Cu exhibited the highest monoaromatic yield of 18.4%.In addition, they also studied catalytic depolymerization of protobind lignin over ZSM-5 supported metal catalysts in supercritical ethanol and revealed the effect of metal types and Cu loadings on the distribution of monoaromatic compounds[23].Recently, our group explored CLD employing Mo/sepiolite, in which sepiolite (SEP) as a clay material was creatively introduced into the CLD field, and the influences of Mo content, calcined temperature, and reduction treatment for Mo/SEP catalyst on the yields of liquid fuel and phenols were clearly presented[16,20,24].In addition, it was also confirmed that Mo-based catalysts presented excellent activity and selectivity in cracking β-O-4 bonds of lignin[26,27]and MoO3phase could selectively cleavage C-O bonds of lignin, thus significantly increasing the yield of aromatics and phenolic compounds in other reaction media[28,29].

The researchers found that the support composition and structure also played an important role in determining catalytic reaction, adjusting surface acidity/alkalinity, and optimizing active metal dispersion and valence state[30,31].For example, the hydrodeoxygenation and dearomatization of anisole were directly affected by Br?nsted and Lewis acid sites on the surface of ASPO-11, SBA-15 and γ-Al2O3[30].In addition, CoMo loaded on different supports (e.g.,Al2O3, ZrO2and TiO2) with different textural and acidic properties could regulate the reducibility, acidity, and metal-carrier interaction of the CoMo catalyst and then control the hydrogenation and deoxygenation performance of anisole[31].Recently, natural clay materials such as SEP, attapulgite (ATP), and montmorillonite (MTM) with environment-friendly feature, wide availability, low cost, and superior adsorption have been widely investigated in various catalytic processes[4,25,32-36].However, comprehensive research focused on the effect of the kinds of supports on the catalytic performance of Mo-based catalysts for CLD under supercritical ethanol without supplying additional hydrogen has not been reported.

Considering to above-mentioned situations, in this paper, the Mo-based catalysts that used various supports including SEP, ATP, MTM, Al2O3and SiO2loaded Mo active species were prepared by an impregnation method.The structural properties of the as-prepared catalysts were characterized by XRD, FTIR, N2adsorption-desorption, NH3-TPD and XPS and their catalytic performance was determined by executing CLD experiment under supercritical ethanol.

1 Experimental

1.1 Reagents and apparatus

The utilization of lignin was kraft lignin (Indulin AT?) purchased from MeadWestvaco (Shanghai China).The results of the ultimate and proximate analysis of lignin were listed in Table 1.Al2O3and SiO2oxides were purchased from Sinopharm Chemical Reagent Co., Ltd.The natural clays such as SEP, ATP,and MTM were collected from Gansu, Jiangsu and Shandong in China, respectively.The reagents such as nitric acid, anhydrous ethanol (AR), petroleum ether(AR) and ammonium molybdate tetrahydrate[(NH4)·6Mo7O24·4H2O (AR)]were purchased from Sinopharmaceutical Chemical Reagents Co., Ltd.The CLD test was carried out in a high-pressure reaction vessel that was designed and produced by Shanghai Yanzheng Experimental Instrument Co., Ltd.

Table 1 Ultimate and proximate analysis of lignin

1.2 Catalyst preparation

Initially, Al2O3and SiO2oxides were calcined at 500 °C in the air for 4 h and the natural clays (SEP,ATP, and MTM) underwent infiltration processing using 5 mol/L nitric acid for 2 h at 60 °C to remove impurities, and then were filtrated, dried and ground to standby application.

Subsequently, SEP, ATP, MTM, Al2O3and SiO2supported Mo catalysts were prepared by the impregnation method, in which the Mo content was fixed at 40%.Firstly, 4.0900 g of (NH4)·6Mo7O24·4H2O was added into 70 mL of deionized water in 250 mL beaker to fully dissolve at 60 °C under continuous stirring.Next, 5.0000 g of support (SEP,ATP, MTM, Al2O3or SiO2) was added into the above Mo precursor aqueous solution and then further stirred for 3 h to fully impregnate.Subsequently, the temperature was raised to 90 °C to evaporate the moisture and then obtain a solid product.Lastly, the solid product was dried overnight at 105 °C, ground and screened for obtaining precursor powder of the catalyst.The precursor powder was calcined at 500 °C in the air for 4 h in a tubular furnace with a heating rate of 4 °C/min to prepare the described Mo-based catalysts such as Mo/SEP, Mo/ATP, Mo/MTM,Mo/Al2O3, and Mo/SiO2.

1.3 Catalytic polymerization test

The CLD experiment was carried out in a 100 mL high-pressure reaction vessel.Test operation procedure was presented in our previous report[20].Specifically,1.0000 g of lignin, 30 mL of anhydrous ethanol, and 0.5 g of Mo-based catalyst were together added into the vessel.Then, the vessel was sealed by flange construction and the residue air was replaced with N2three times and then the vessel was filled 0.5 MPa N2.Subsequently, the reaction system was heated to the target temperature and kept for a certain amount of reaction time to implement the CLD reaction.After the reaction finished, the high-pressure reaction vessel was quickly cooled to room temperature with cold water,and the yielded mixture was filtered through the sand core funnel to separate the solid and liquid products.Then, the residual ethanol solvent in the liquid product was rotatably evaporated at 44 °C to completely remove ethanol, so the enriched liquid product was named lignin oil (LO) and weighed.Following, the LO was dissolved in petroleum ether (PE) under ultrasonic oscillation and then processed filtration using an organic filter head.The obtained filtrate was rotatably evaporated at 38 °C to remove PE to produce a PE-soluble product (PEsp) and weighed.The separated solid was dried at 105 °C oven for 12 h and weighed.Therefore, the weight of the solid products (possibly including trace amounts of unreactive lignin, oligomer,and coke) was obtained by subtracting the amount of catalyst.In order to clearly present the effect of supports on Mo-based performance for CLD, the quantitative indicators such as liquefaction rate of lignin and yields of LO, PEspand solid product were calculated based on the following formulas (1)-(3).

1.4 Catalyst characterizations and lignin oil analyses

The elemental contents of C, H, O and N in kraft lignin are measured employing an elemental analyzer(model: Vario EL Ⅲ, Germany).

The crystal structure of the catalyst was analyzed by the SmartLab SE X-ray spectrometer (XRD) made by Rigaku Company, Japan.

Fourier transform-infrared spectroscopy (FT-IR)of the catalyst was measured using Nicolet IS50 produced by Thermo Fisher Scientific in the wavelength range of 4000-400 cm-1using a KBr tablet.

The N2adsorption-desorption isotherm of the catalyst was measured on the ASAP2020 adsorptor manufactured by Micromeritics, USA.The specific surface area (SBET) of the catalyst was calculated by the BET method, and the average pore volume (v) and pore size (d) of the catalyst were calculated by the BJH method.

The NH3-TPD curves of the catalyst were measured on the TP-5080 Adsorbometer manufactured by Tianjin Xian Quan Instrument Co., Ltd.(China).Generally, the sample was firstly experienced pretreatment process at 200 °C for 1 h in the 30 mL/min He flow, then the temperature was cooled down to 50 °C and adsorbed NH3using 50 mL/min of 5%NH3/He flow for 1 h to realize sample saturation adsorption.Subsequently, the sample was heated to 800 °C with 10 °C/min speed in a 40 mL/min He flow and the desorbed signals of NH3were recorded.

The X-ray photoelectron spectroscopy (XPS) of the catalyst was measured on a Thermo Fisher Scientific ESCALAB250Xi (UK) with a monochrome AlKα source (1486.68 eV) under the conditions of 150 W and 500 mm beam spots.XPS spectra were further fitted and analyzed using the Thermo Avantage v5.52 software (Thermo Fisher Scientific, Micro Focus Ltd).

The monomer components in LO were distinguished on the GC-MS instrument (Thermo Scientific ISQ 7000 MS detector) equipped with a TG-5MS capillary column (30 m × 0.25 mm × 0.25 μm)and quantified by the internal standard method using acetophenone as internal standard.The specific heating ramp was as follows: the oven temperature program increased from 50 °C (held for 3 min) to 280 °C at a rate of 10 °C/min and holding at 280 °C for 10 min[20].

2 Results and discussion

2.1 Characterizations of supports and Mo-based catalysts

2.1.1 XRD characterization

The XRD patterns of supports and Mo-based catalysts are shown in Figure 1.As can be seen from Figure 1, the characteristic peaks of SEP were observed at 2θ= 8.7°, 17.6°, 19.7°, 20.8°, 25.2°, 26.5°, 34.6° and 45.0° (Figure 1(a)), the specific peaks of ATP were found at about around 8.5°, 13.7°, 19.8°, 31.1° and 34.8°(Figure 1(b)) and the diagnostic peaks of MTM were detected at about 2θof 6.1°, 19.8°, 29.3°, 35.1°and 61.9° (Figure 1(c)).Furthermore, the typical peaks of Al2O3at 2θ= 32.3°, 37.5°, 45.7° and 67.0° and the wide diffraction peak of SiO2at 20°-25° were all determined by XRD.The broad peak of SiO2support suggested that silicon oxide was amorphous.For ATP support, some peaks of quartz (SiO2) at 20.7° and 26.4°were also found.These results were consistent with previous reports[20,32-36].For all supported-Mo catalysts,apart from some peaks of supports, some sharp diffraction peaks were detected at 12.8°, 23.3°, 25.7°,27.3°, 33.7°, 39.0°, 45.7° and 49.2°, which were attributed to the crystalline phase of MoO3[20].This result demonstrated that active Mo species in the Mobased catalyst mainly existed as MoO3structure and the support did not significantly affect the crystalline structure of active Mo species.It was also found that the characteristic peaks of support substance decreased or even disappeared, attributing to the dilution effect of the Mo species signal for support signal.

2.1.2 FT-IR analysis

Figure 2 showed the FT-IR spectra of various supports and as-prepared catalysts.For clay supports(SEP, ATP and MTM), the stretching vibration peaks in the range of 3700-3600 cm-1were attributed to the vibration of the hydroxyl groups that existed in the interior of their skeleton and surface structures and the peaks at 3430 and 1636 cm-1were assigned to the stretching vibrations of physically adsorbed H2O[20].Meanwhile, the peak at 1032 cm-1was attributed to the vibration of the Si-O bond in the [SiO4]tetrahedral sheets of clay supports.The peaks at 513-535 cm-1and 472 cm-1came from the signal of Si-O-Si bond vibration[20].In addition, the peak at 1449 cm-1in ATP corresponded to the breathing vibrations of Si-O-Mg/Al bonds[37].For Al2O3, only a broad peak at3700-3500 cm-1was found, which confirmed the existence of hydroxyl groups.The stretching vibrations of hydroxyl groups and the SiO2framework were also observed in the spectrum of SiO2.After the introduction of Mo active species, the vibrations of hydroxyl groups of supports presented varying degrees of decline and even disappearance.Additionally, some fresh vibration peaks at 990 cm-1, 875 cm-1/818 cm-1,and 606 cm-1belonged to the vibration characteristics of Mo=O, Mo-O-Mo, and Mo-O bond of MoO3species, respectively[20].These results demonstrated that the Mo metal precursor interacted with the structure hydroxyl groups of supports to form MoO3species during the catalyst preparation process, and the support promoted the distribution of active Mo species.

Figure 1 XRD patterns of supports and Mo-based catalysts

Figure 2 FT-IR spectra of supports and Mo-based catalysts

2.1.3 N2 adsorption-desorption test

The textural characteristics of supports and Mobased catalysts were characterized by N2adsorptiondesorption and the obtained results were presented in Figure 3 and Table 2.

Table 2 Textural properties of supports and Mo-based catalysts

Figure 3 N2 adsorption-desorption isotherms and pore size distribution profiles of the supports and Mo-based catalysts

It can be seen that the supports all presented the Ⅲ-type adsorption-desorption isotherms,suggesting that all supports had definitely mesoporous structure[20,35,37].Furthermore, the isotherms of SEP,ATP and SiO2had an H3 hysteresis loop, indicating that the holes in these materials mainly included plate slits, cracks, and wedges structures, and those of MTM and Al2O3exhibited an H2hysteresis loop, illustrating that the pore structures in these supports were complex,which might include typical “ink-bottle” holes, tubular holes with uneven pore size distribution and densely packed spherical particle interstitial holes.Meanwhile,their corresponded pore size distribution profiles displayed in the right of Figure 3 also demonstrated the fine differences of pore size among the five supports.The overwhelming majority of pores was the mesoporous structure in the five supports, but there was a small number of macropores in SiO2.After loading Mo species, the shapes of N2adsorptiondesorption isotherms of all Mo-based catalysts remained unchanged but the strengths were significantly reduced.The data in Table 2 also demonstrated the obvious reduction in terms of BET specific surface area (SBET) and BJH desorption cumulative volume of pores (v).In addition, the pore size distributions of Mo-based catalysts presented similar changes with N2adsorption-desorption isotherms.The BJH desorption average pore diameter(d) of support was lower than that of corresponding Mo catalysts except for the SiO2and Mo/SiO2.This result indicated that the loaded Mo species blocked the smaller holes of the carrier materials.From the data in Table 2, it is observed that theSBETof the Mo-based catalysts decreased in the order of Mo/SiO2＞Mo/Al2O3＞ Mo/MTM ＞ Mo/ATP ＞ Mo/SEP, whileddecreased in the order of Mo/SiO2＞ Mo/SEP ＞Mo/ATP ＞ Mo/MTM ＞ Mo/Al2O3.

2.1.4 XPS characterization

Figure 4 exhibits the XPS spectra of Mo 3din Mo-based catalysts.It was found that all Mo 3dXPS spectra presented two major peaks at binding energies about 233 and 236 eV, which corresponded to two split spin orbitals such as 3d5/2and 3d3/2of Mo 3dlevel,respectively, according to the previous reports[38,39].After deconvolution, Mo 3dXPS spectra of all catalysts could be severally divided into two pairs of peaks.A pair of peaks were located at 231.0 and 233.1 eV assigned to the Mo5+, while the coupled peaks at 232.3 and 235.5 eV were ascribed to the Mo6+.The presence of Mo5+species confirmed the existence of the interactions between Mo species and support framework in Mo-based catalysts[25].It has been reported that Mo5+species were the actual active sites for lignin conversion under supercritical ethanol[40].Therefore, the surface relative content of Mo6+and Mo5+species was also determined by XPS technology and the result is presented in Figure 4.It can be seen that Mo/SiO2catalyst had the highest Mo5+content,reaching 17%, while Mo/MTM catalyst exhibited the lowest Mo5+content, in only 4%.The contents of Mo5+in Mo/Al2O3and Mo/SEP were 9%, and that in Mo/ATP was 7%.This result demonstrated that the structures and properties of support materials could affect the valence state distributions of active Mo species.

Figure 4 XPS spectra of Mo-based catalysts

2.1.5 Surface acidity analysis

The surface acidity of solid catalyst is commonly determined using NH3temperature-programmed desorption (NH3-TPD) Technology.Figure 5 displays the NH3-TPD profiles of the supports and as-prepared Mo-based catalysts.It is well accepted that the location and area of NH3desorption peak generally indicate the strength and amount of acidic sites over the catalyst surface.Popularly known, the NH3desorption peak below 250 °C is ascribed to the weak acid sites, that between 250 and 450 °C is attributed to the medium acid sites, and that above 450 °C is assigned to strong acid sites[4].Consequently, the clay-based supports(SEP, ATP, MTM) mainly contained the strong acid sites, except for MTM had a handful of weak acid sites,as demonstrated by the NH3-TPD profiles in Figure 5.The weak acid sites were attributed to the Br?nsted acid sites (surface hydroxyl groups) originated from the Al/Mg constituent in clay[41].The strong acid sites were mainly ascribed to low coordination oxygen species[42],which strongly adsorbed HN3molecules and reacted with them to form a NOxdesorption peak.This could enhance the adsorption of intermediate reactive species during lignin depolymerization to produce coke through the dehydration reaction.In addition, there were mainly weak and medium acid sites over the Al2O3surface while only weak acid sites were presented in SiO2surface.Furthermore, it was clearly found that the intensities of NH3-TPD profiles of five supports were discrepant, and the quantitative NH3monolayer uptake (μmol NH3/g sample) was presented in Table 2, which could be identified as the amount of acid sites.One obviously observed that the surface content of acid sites showed the following order: ATP＞ MTM ＞ SEP ＞ SiO2＞ Al2O3.

Figure 5 NH3-TPD profiles of supports and Mo-based catalysts

After the introduction of Mo species into the above supports, the NH3desorption peaks displayed significant changes as compared to their corresponding supports.Overall, the intensities of NH3desorptionpeaks of Mo-based catalysts significantly reduced.Additionally, the NH3desorption peak of Mo/SEP moved to a higher temperature and only weak and medium acid sites appeared in Mo/ATP.Mo/MTM exhibited a similar acid site distribution to MTM.The NH3desorption peak of Mo/Al2O3became more homogeneous as compared with Al2O3, implying the acid sites were more uniform.For Mo/SiO2, the distinct strong acid sites were detected except for the presence of weak sites.The total surface content of acid sites over Mo-based catalysts exhibited the following order:Mo/MTM ＞ Mo/SEP ＞ Mo/ATP ＞ Mo/SiO2＞Mo/Al2O3and the data are listed in Table 2.

2.2 Lignin catalytic depolymerization test

2.2.1 Effects of supports and catalysts on CLD

The effect of supports on the catalytic performance of Mo-based catalysts for CLD was detailed investigated under a high-pressure reaction vessel at 280 °C for 4 h with the following conditions:1.0 g of lignin, 0.5 g of support or catalyst, 30 mL of ethanol and 0.5 MPa initial N2.The yields of LO, PEsp and solid product were presented in Figure 6.It can be seen that the yielded LO over pure support was very low and mainly ranged from 23.5% (over SEP) to 32.2% (over Al2O3).Furthermore, the clay-based supports displayed relatively lower LO yields while SiO2and Al2O3presented higher LO.This confirmed that the acid sites on the support surface were the major active sites for promoting CLD, while strong acid sites were not conducive to lignin depolymerization.The obtained higher solid product over clay-based supports could also demonstrate this point.It was worth noting that the PEsp (containing monomer and dimer products) yields of five supports were similar, implying that the acid site strength had no effect on monomer and dimer yields for CLD.Obviously, all Mo-based catalysts showed significant increases in terms of LO and PESPyields, demonstrating that molybdenum oxide species had a unique performance in CLD.Additionally, the LO yields over Mo-based catalysts showed significant differences.Among them, Mo/SiO2exhibited the highest LO (85.2%) and the Mo/ATP and Mo/MTM presented fewer LO (51.2% and 43.6%respectively).Mo/SEP and Mo/Al2O3had analogous LO yields, in 82.3% and 83.1%, respectively.Furthermore, the PEspyields over Mo-based catalysts revealed a similar variation trend with LO but the yielded solid products showed a contrary tendency.Previous studies had demonstrated that synergistic effect between Mo5+species and acid sites could promote the cracking of the C-O bonds of lignin[16,20]and Mo5+species involved in the transformation of lignin depolymerization intermediates to give small molecule products[43,44].In the XPS analysis, the support materials were confirmed to affect the content of Mo5+species, and the surface content showed the following order: Mo/SiO2＞ Mo/Al2O3= Mo/SEP ＞ Mo/ATP ＞Mo/MTM.Meanwhile, the strong acid sites inhibited the yield of LO.It can be found that Mo/MTM had superior strong acid sites and the lowest content of Mo5+species, so it showed the lowest LO yields.Therefore, it was very appropriate that Mo/SiO2contained the highest content of Mo5+species, and suitable acid sites (most of them were weak acid sites)endowed it with the highest LO and SPPEyields.In addition, Mo/SiO2had a larger pore size (27.6 nm),which improved the heat and mass transfer and contact rate of active sites with reactants, thus promoting the conversion of lignin to LO and monomers.Among the clay-based supports loaded Mo catalysts, Mo/SEP presented the unique CLD performance.This study could further provide useful information on developing clay materials for biomass conversion.

2.2.2 Monomer identification in lignin oil

For convenience, the PEspobtained using SEP,ATP, MTM, Al2O3, SiO2, Mo/SEP, Mo/ATP,Mo/MTM, Mo/Al2O3and Mo/SiO2were named as PEsp-a, PEsp-b, PEsp-c, PEsp-d, PEsp-e, PEsp-A, PEsp-B,PEsp-C, PEsp-D and PEsp-E, respectively.The monomers in PEspwere identified using GC-MS and their corresponding yields were obtained by the internal standard method.The obtained GC-MS profiles of all PEspare shown in Figure 7.From the GC-MS profiles, it can be found that the monomers were similar over various supports and mainly contained aliphatic alcohols, esters, phenols, and benzene.Furthermore, the mass spectrum of PEsp-d was more complex than those of other supports.This is mainly due to the effect of Lewis acid sites in Al2O3(it was not the focus of this study).In addition, the PEspproduced by Mo-based catalysts clearly had different constituents compared with supports produced PEsp, in which the esters and phenols clearly increased.In order to further reveal the effects of supports and Mo species on the distribution of monomers in PEsp, all identified monomers were divided into three types, such as aliphatic oxygenates including alcohols and esters(containing the serial number (SN)∶1, 3, 5, 6, 7, 8, 10,11 and 17 as shown in Figure 7), alkyl/oxyalkylated benzenes (containing the SN: 2, 16, 28 and 31) andalkyl/alkoxy substituted phenols (containing all the remained SN).Furthermore, the monomer yields are listed in Table 3 and Table 4.It was found that the yielded aliphatic oxygenates and substituted phenols over supports were significantly lower than those produced over Mo-based catalysts.This might be due to the fact that the pyrolysis reaction of lignin occurred mainly over only supports, where the support materials had limited ability to crack the C-O bond in the structure of lignin.It had been confirmed that Mo active species did not only promote the cracking of the C-O bonds of lignin but also could interact with supercritical ethanol to form the homogeneous Mo5+ethoxide species (Mo(OEt)5)[16,20], therefore enhancing the yields of aliphatic oxygenates and substituted phenols.Additionally, the content of alkyl/oxyalkylated phenols decreased in the order of PEsp-D ＞ PEsp-E ＞PEsp-A ＞ PEsp-B ＞ PEsp-C, which was different from the variation trends of Mo5+content and LO yields.

Figure 6 Results of CLD over supports and Mo-based catalysts

Figure 7 GC-MS results of all obtained PEsp

Generally, lignin was pyrolyzed to form alkoxycontaining fragments during CLD and then C-O bonds of lignin fragments were cleaved by Mo5+to generate small molecular active intermediates.These active intermediates were then alkylated with ethanol radicals to form alkylated phenols.On the other hand, ethanol radicals could also combine with the chain fragments formed by the cleavage of lignin fragments to produce aliphatic monomers in LO.However, the formation of various monomers was very easily affected by the amount and types of catalyst acid sites and the pore and interface structures of catalysts.So, the illumination of CLD reaction mechanism was still the challenge of biomass conversion.The present study provided some key information about using green and economical clay-based materials for lignin stabilization and expanded the research field of biomass conversion.

In order to further illustrate the performance of Mo-based catalysts in this paper, we compared the experimental data with previous reports.The group of Li prepared various complex and exquisite catalysts including ZnCoOxnanoplate, hollow Ni-Fe (HNiFe2O4) and bifunctional Ni/DeAl-beta catalysts, and these catalysts could achieve more than 80% of LO and 70% of PESPyields under the mixture of 1, 4-dioxane/methanol solvent and 2 MPa H2through adjusting reaction conditions[7,8,10].Compared with them, although the as-prepared Mo-based catalysts in this paper showed inferior performance, the economical efficiency in terms of manufacture of the catalysts and reaction medium (without extra hydrogen molecules) enhanced the competitiveness of the prepared Mo-based catalysts.In addition, considering the reports of Hensen et al.that had reported a series of noteworthy results about depolymerization of lignin in supercritical ethanol over CuMgAlOxcatalysts[17-19].Among them, 36% of monomers was yielded without formation of char after reaction at 340 °C for 4 h[18].In Table 4, it can be found that the only monomers of phenols of Mo/SEP beyond the value (398.8 mg/glignin)are 39.8% based on the calculation method in ref.[18].Therefore, the Mo-based catalysts reported in this paper have a good research prospect from both economy and yield/selectivity.

Table 3 Yields of identified monomer products in PEsp obtained over various supports

Table 4 Yields of identified monomer products in PEsp obtained over various Mo-based catalysts

3 Conclusions

In this paper, a series of green and environmentally friendly clay-based materials (SEP,ATP, MTM) and supported Mo-based catalysts were firstly applied in catalytic depolymerization of kraft lignin, and their catalytic performance was also compared with oxides (Al2O3and SiO2) and oxidessupported Mo catalysts.Various analyses and characterizations confirmed that the supports influenced the surface pore structure, acidity and the content of Mo5+species in Mo-based catalysts.Due to the existence of strong acid sites, clay-based supports presented lower LO yields comparing with oxide supports.Mo/SiO2exhibited the largest specific surface area and average pore size and the highest Mo5+species, and it consequently produced the highest LO yield of 85.2%, where the yield of the substituted phenols reached 450.3 mg/glignin.It was worth mentioning that Mo/SEP also presented promising LO(82.3%) and PEsp(70.8%) yields and the generated substituted phenols reached 398.8 mg/glignin.