Metal-alkali catalytic valorization of lignocellulose towards aromatics and small molecular alcohols and acids in a holistic approach

In this work, we developed an approach of one-pot complete catalytic conversion of woody biomass into two value product streams: lignin-derived aromatics (monomer yield of 68.54% and oligomer yield of 29.65% based on lignin mass) and (semi-)cellulose-derived small molecular alcohols (about 59.60% of biomass mass). These could be afforded by conducting lignocellulose depolymerization over metal-alkaline catalysts in a mixed n-butanol/H2O solvent system at 250 °C and 30 bar H2. In the valorization process, the homogenous mixture of n-butanol-H2O solvents extract and depolymerize both lignin and hemicellulose, while the catalysts and H2 are essential to cleave the inter-/intramolecular linkage of lignocellulose into target products. After the reaction, the phase separation of n-butanol and H2O takes place when system temperature drops below 125 °C, providing a mild and effective strategy to isolate lignin-derived aromatics (n-butanol phase) from small molecular alcohols/acids (aqueous phase). Ru/C and alkali catalysts are collected by filtration from n-butanol phase and H2O phase, respectively. Meanwhile, the effect of metal-alkali coupled catalysts facilitates the cleavage of β-O-4 linkage of lignin and increases the attainability of (semi-)cellulose-derived oligomers and the small molecular alcohols. This catalytic system provides a versatile valorization approach for biomass catalysis to bio-based chemicals.


Introduction
The decline of crude oil attainability for fuel and chemicals feedstock (Bastin et al. 2019), the proportion of energy and chemicals provided by renewable Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/ s10570- 021-04156-3. resources (such as biomass) is gradually increasing and becoming more important for the sustainable production of fuels and chemicals (Li et al. 2015;Lin and Huber 2009). Combined with people's concern and pursuit of healthier and environmentally benign products, it has become a common trend to use natural ingredients in various products to replace non-renewable raw materials with renewable resources and minimize the impact of industrial waste on the environment (Ferrini and Rinaldi 2014;Sun et al. 2018b;Borrero Lopez et al. 2018;Liao et al. 2020). Therefore, a shift of producing of selected high-value chemicals from fossil to renewable carbon resources can ease the energy crisis, open up a fine chemical industry chain and reduce CO 2 emissions (Binder and Raines 2009;Shen et al. 2015).
Among diverse renewable resources, biomass has the advantage of being abundant, clean and cheap with huge potential of utilization, and has been set to become one of the most important energy sources in the future Liu et al. 2019c). Biomass serves as a natural renewable chemical resources that can be converted to high added-value products (Sun et al. 2018a;Liu et al. 2019b;Kuna et al. 2017). For example, cellulose and hemicellulose produce high value-added liquid fuels and fine chemicals (liquid fuel oil, food additives and important intermediates such as ethanol, ethylene glycol, hydroxyl ketenes, propylene glycol, glycerol, butanediol, hexadiol, polyol, acetone alcohol, acetone, lactic acid, etc.) Kuna et al. 2017;Liu et al. 2019aLiu et al. , 2019cSong et al. 2019;Huber et al. 2006;Ji et al. 2008;Zhang et al. 2010;Li et al. 2019). Lignin, the only renewable aromatic polymer, is showing its significant utilization potential for manufacture of flavors, perfumes, and spices Wang et al. 2020), synthesizing value-added chemicals (dimethyl sulfoxide, alkyl phenol and lubricating oil, etc.) (Zakzeski et al. 2010;Schutyser et al. 2018;Gandini 2011;Liu et al. 2019b), liquid fuel (aromatic/olefin/paraffin) (Ragauskas et al. 2014;Li et al. 2015;Zhao et al. 2009) and polymers (polycaprolactone, polyester, polyurethane, phenolic resin, matrix ink and enhanced plant oil-based polyurethane composite materials, etc.) (Ha et al. 2012;Yuhe Liao* 2020;Gandini 2011;Van den Bosch et al. 2015a;Wang et al. 2020).
Although biomass is an attractive resource, biomass utilization at present has low utilization rate of raw materials and high energy consumption. Thus, it is necessary to optimize biomass utilization for competition with fossil-based processes based on the cost and environmental impact (Alonso et al. 2017;Liao et al. 2020). To address the issues and realize multicomponent full conversion of lignocellulose, it is badly in need of developing the overall biomass refining technologies with low energy requirements and high carbon (and mass) efficiency.
In recent years, the one-pot method has been adopted to convert lignocellulose into multiple chemicals by some research teams. For instance, Zhang et al. selectively converted biomass into two sets of small molecular compounds over 4% Ni-30%W 2 C/ AC catalyst in water at 235°C and 6 MPa H 2 . The lignin in biomass was converted into phenolic compounds, while the cellulose and hemicellulose were selectively degraded to ethylene glycol (EG) and 1,2propylene glycol (1,2-PG). Especially, 36.3% of monomer phenol yield and 76.1% of glycol yield were afforded when raw material was birch wood (Li et al. 2012). Wang et al. reported that corn stalk was efficiently transformed in water with a partially oxidized commercial Ru/C catalyst at 200°C and 3 MPa H 2 . Lignin was selectively converted into cycloalkanes, while the cellulose and semi-cellulose fractions were converted into polyols, such as sugar alcohols and diols . In our previous work, LiTaMoO 6 combined with Ru/C were applied to catalyze the transformation of real biomass in the dilute phosphoric acid aqueous solution system at 230°C and 6 MPa H 2 . Compared to the above reports, the hemicellulose and cellulose fractions in real biomass in this previous work were converted to C5-C6 gasoline alkanes with a yield of 82.4%, and lignin was partially converted into monomer phenols (Liu et al. 2015). Recently, an integrated biorefinery strategy of lignocellulose was reported in Science by Bert Sels' group and 78 wt % of birch was converted into the xylochemicals. Carbohydrate pulp and lignin oil were separated in one-pot by reductive catalytic fractionation (RCF) of lignocellulose. After that, lignin oil was extracted to obtain phenolic monomers that were converted to 20 wt% of phenol and 9 wt% of propylene (on lignin basis) by gas-phase hydroprocessing/dealkylation, and the residual phenolic oligomers (30 wt%) were used in printing ink. Whereas the carbohydrate pulp was used to produce bioethanol (40.2 g/L -1 ) by biological fermentation .
Due to the great difference of structures among lignin, hemicellulose and cellulose, the final products derived one-pot conversion of biomass are too complex and quite different. It is necessary to separate the products after depolymerization reaction for the products further upgrading or utilization. However, the product separation is difficult to operate and increase production costs. In other words, although one-pot catalytic conversion of lignocellulose to hydrocarbon fuel has a certain prospect of application, there are still some shortcomings in the preparation of high value-added chemicals. Therefore, the strategy of valorization lignocellulose into chemicals needs to be further improved.
Herein, we developed and investigated a catalytic valorization route, targeting the one-pot conversion of woody biomass and the separation of components into two sets of value products: (i) lignin derived aromatics, (ii) (semi-)cellulose-derived small molecular (C2 * C5) alcohols and acids (Fig. 1). Intending to provide an effective depolymerization and separation approach of the solubilized fractions, woody biomass was complete depolymerized over metal-alkali catalysts in biphasic n-butanol/H 2 O solvent system. After the reaction, lignin-derived phenolic compounds and (semi-)cellulose-derived products were automatically extracted and separated by n-butanol phase (upper) and aqueous phase, respectively. This approach reduced the difficulty of the product separation and production cost in one-pot complete dissolution polymerization system.

Materials
All used chemical reagents and materials in this work are shown in the supporting information.
Typical process for lignocellulose valorization Depolymerization experiments were conducted in a 100 ml autoclave (Hastelloy alloy, made by Anhui Kemi Machinery Technology Co., Ltd.) equipped with a mechanical stirring. In a typical depolymerization reaction, 2.0 g pre-extracted powder was put into the reactor with the catalyst (0.2 g Ru/C or / and 0.4 g MgO) and the n-butanol/water solvent mixture (40 mL, v/v = 1:1). The reactor was sealed, the air was replaced in the reactor by N 2 four times and then by H 2 three times. A total of 30 bar of H 2 was pressurized at room temperature. The reaction mixture was heated to 200°C (*10°C min -1 ) and then kept for 2 h with continuous stirring at 700 rpm. After that, the reactor was cooled and the gaseous products were collected at room temperature. Fig. 1 The scheme of eucalyptus powder depolymerization over Ru/C-MgO in a mixture of n-butanol/H 2 O solvents process targeting (i) lignin-derived aromatics, (ii) small molecular (C2 * C5) acids and alcohols, (iii) retrieved Ru/C and Mg(OH) 2

Products separation
The mixture of liquid and solid were taken out and the reactor was washed with 15 mL water and 15 mL n-butanol. The obtained mixture was filtered to separate the solid and liquid mixture. The mixture of solid (Ru/C, Mg (OH) 2 (MgO converted to Mg (OH) 2 in hydrothermal condition) and depolymerization residue) was separated and collected through liquidliquid extraction method with n-butanol and H 2 O. Ru/ C primarily resided (due to its relatively polar) in n-butanol phase, while Mg (OH) 2 and the residue of eucalyptus powder (if any) located at aqueous phase. After liquid-liquid extraction, portion of Ru/C was extracted by n-butanol phase and then was retrieved, the remainder Ru/C blended with Mg(OH) 2 and the residue of eucalyptus powder (if any) in H 2 O phase was extracted again with fresh n-butanol. This step was repeated until n-butanol phase remained relatively clear. All of the collected Ru/C was washed with ethanol to remove n-butanol and other adsorbents until alcoholic filtrate relative clear. The mixture (the residue of eucalyptus powder and Mg(OH) 2 ) was calcinated to retrieve MgO in air under 550°C for 4 h. On the other hand, the liquid products were transferred to a separating funnel and then automatically separated into n-butanol phase (upper phase) and aqueous phase (bottom phase), as depicted in Fig. 1. After phase separation, a 20 mL fresh n-butanol was added to aqueous phase for extraction lignin-derived fragments, repeating this step three times. In this way, the total n-butanol fraction measured calculating about 80 mL and the total aqueous fraction was about 43 mL.

Determine the components of lignocellulose and residue
The bits of eucalyptus wood was smashed and sieved to obtain the wood powder (40 * 60 mesh). Then the extractions were removed by a Soxhlet extractor. This wood powder, hereinafter marked as ''extracted powder'', was used for composition analysis and depolymerization. The composition is summarized in Table S1. The composition of eucalyptus powder and depolymerization residue were analyzed by NREL according to NREL Laboratory Analytical Procedures (Sluiter A. 2008).
Eucalyptus powder and residue before and after depolymerization were characterized by XRD and FT-IR analysis.

Analysis and measurement of gaseous and liquid products
The products in n-butanol phase were analyzed by GC-FID and GC-MS. Acetophenone was added as an internal standard for quantification. Aqueous phase products were analyzed by HPLC. The gaseous products were analyzed by GC. The distribution of the molar weight of the lignin products was investigated using gel permeation chromatography (GPC). The structure characteristic of lignin fragments were analyzed by HSQC analysis. The detail information was provided in the supporting information.

Result and discussion
Ru/C cooperated with MgO for catalytic valorization lignocellulose Eucalyptus powder, composition summarized in Table S1, was completely catalytic depolymerized to liquid products over Ru/C-MgO catalysts in biphasic solvent (n-butanol/H 2 O). The eucalyptus powder valorization process is displayed in Fig. 1. Here, 2.0 g pre-extracted eucalyptus powder, 0.2 g of Ru/C and 0.4 g MgO catalysts, and 20 mL H 2 O, 20 mL n-butanol were put in a 100 mL autoclave equipped with a mechanical stirring. The autoclave was sealed, purified with N 2 and pressurized with H 2 (30 bar), and the mixture was stirred (700 rpm) and heated to a given temperature of 250°C for 6 h. After cooling to room temperature, the gaseous and liquid fractions were collected and processed. The solids (mostly Ru/C and Mg (OH) 2 ) were washed with fresh H 2 O and n-butanol. The filtrate instantly separated into n-butanol (containing lignin-derived products) and aqueous phase (including (semi-)cellulose derived products) (Fig. 1). Thereby this catalytic depolymerization system provides an effective approach of lignocellulose depolymerization and the soluble products separation (Scheme S1, supporting information for experimental and separated procedures). Evaporation of n-butanol yielded a viscous oil with 23.85wt% of the initial biomass weight. The weight of oil in the n-butanol phase is close to the total lignin content of the eucalyptus sawdust (25.97 wt%), indicating that the extensive lignin in extracted wood powder was converted.
The lignin oil in n-butanol phase contains a large number of aromatic monomers, corresponding to a yield of 68.54wt% based on total lignin mass of the eucalyptus powder. The monomer aromatics mainly include propanol-substituted guaiacyl/syringol (label as G3/S3-OH), propyl-substituted guaiacyl/syringol (label as G3/S3), methyl-substituted guaiacyl/syringol (G1/S1), and a small amount of guaiacol/2, 6-dimethoxyphenol (label as G0/S0) and ethyl-substituted guaiacyl/syringol (label as G2/S2) (Fig. 2a). Compared to the monomer aromatics (G3/S3-OH, G3/ S3 and G2/S2) derived from Ru/C catalyzed eucalyptus depolymerization in n-butanol/H 2 O solvents (Renders et al. 2016a, another 14.32 wt% of G0/S0 and G2/S2 (on lignin oil basis) were obtained in Ru/C-MgO case due to the presence of MgO. The molecular weight distribution of the lignin-derived non-volatile products in n-butanol phase was investigated by GPC analysis. The result in Fig. 2c displays that n-butanol phase consists of monomers, dimers, trimmers and oligomers. The D p (from 14 * 17.5 min) of oligomers is 1.25. The largest oligomers elute at 15 min, which corresponds to about a M w of 1600 gmol -1 . The result of HSQC NMR indicates that the lignin interunit ether bonds almost disappear (Fig. 2e), and no lignin typical absorption peak in FT-IR curve of residue further evidences the absence of lignin in residue (Fig. 2d). Hence, we infer that lignin is degraded efficiently over Ru/C-MgO catalysts in the biphasic n-butanol/water solvent system. Products in aqueous phase were analyzed by GC-MS and quantified with HPLC (Fig. 2b). The products comprise acetic acid (29.65 mgÁg biomass -1 ), EG (96.85 mgÁg biomass -1 ), 1,2-PG (88.78 mgÁg biomass -1 ), ethanol (79.61 mgÁg biomass -1 ), 1-hydroxy-2-butanone (69.46 mgÁg biomass -1 ), tetrahydrofurfuryl alcohol (94.03 mgÁg biomass -1 ) and oligomers (22.2 mgÁg biomass -1 ), a little C5 * C6 polyol components (glucose, sorbitol and xylitol). FT-IR ( Fig. 2d) result shows the cellulose and hemicellulose were almost converted since no typical peak was observed. Therefore, we conclude that cellulose and hemicellulose pretty much depolymerized effectively to small molecular alcohols/ acids over Ru/C-MgO in this system. Additionally, except for cellobiose, the other oligomer fragments are hard to qualitatively analyze ( Fig. S9-6 h) and the mass of oligomer was calculated via Eq. (9) in supporting information. Besides H 2 , very few other gaseous generated and the mass was not calculated (Table S2). A general mass flowchart of the main lignocellulose constituents is presented in Fig. 3.
The fractions from liquid and solid phases (not included gas phase due to a little amount of gas), in relation to the composition of the eucalyptus (Table S1). All numbers are expressed as wt % relative to the mass of eucalyptus powder. Compared to Fig. 2a andb, acetic acid, EG and 1,2-PG in n-butanol phase were calculated into n-butanol phase products in Fig. 3, similarly, the monomer aromatics in aqueous phase were calculated in aqueous phase products in Fig. 3. Reaction conditions: 20 mL n-butanol, 20 mL water, 2.0 g extracted eucalyptus powder, 0.2 g Ru/C and 0.4 g MgO, 30 bar H 2 , 250°C, 6 h.

Effect of solvent on lignocellulose valorization
To illuminate the necessity of biphasic solvents, valorization of eucalyptus with Ru/C or/and MgO was performed in pure n-butanol and pure water. After reaction, a similar product separation procedure (supporting information) was conducted for qualitative and quantitative analysis.
When eucalyptus powder was depolymerized in pure n-butanol solvent system over Ru/C, MgO and Ru/C-MgO catalysts, lignin oil yield was 9.13 wt%, 10.28 wt% and 13.65 wt%, respectively. Apparently, pure n-butanol has poor effective for the depolymerization process, which is also reflected by the low lignin monomer yield (10.1 wt %, 20.2wt % and 23.6 wt %, respectively, on lignin basis, Fig. 4a). Additionally, a small amount of aqueous phase product ( Fig. 4b and Fig. S3B) also confirmed that n-butanol had very poor effect on the cleavage of linkage in cellulose and hemicellulose.
Compared to depolymerizing in pure n-butanol system, the valorization process in pure water system provided higher aromatic monomer yield. For instance, the monomer yield was increased to 28.56 wt% and 26.05 wt% in Ru/C and Ru/C-MgO cases, respectively (on lignin basis, Fig. 4a). Despite this, product yields are lower than those of the biphasic solvent system, confirmed by the weaker peak intensity response of GPC curves (Fig. S1). The structure of lignin depolymerization products obviously changed in pure water system, and the products A * F were the cyclohexanol derivatives ( Fig. 4a) derived from the hydrogenation of benzene ring and disproportionation of side chain alkyl group (Verboekend et al. 2016;Liao et al. 2020). The hydrogenation of benzene ring probably ascribed to the sorption behavior of phenols on the catalyst surface. Bert Sels and coworkers also considered that the phenol incompatibility with water led to the phenols adsorbing on the catalyst surface and facilitate benzene ring hydrogenation He et al. 2014aHe et al. , 2014b. In contrast, the presence of alcohols, as well as Ru/C catalyst, phenolic compounds are dissolved and protected by the alcohol solvent from being subjected to benzene ring hydrogenation and subsequent demethoxylation He et al. 2014a). The products from the valorization process in n-butanol/H 2 O also confirm this conclusion (Fig. 4a), i.e., benzene ring and side chain of methoxy group are reserved in n-butanol/H 2 O.
Apparently, the biphasic n-butanol/H 2 O solvent system provided the higher lignin oil and monomer aromatic yields compared to either pure n-butanol or pure H 2 O system, the result is consistent with the lignin depolymerization in ethyl acetate/H 2 O (Lv et al. , 2017, methanol-H 2 O (Chen et al. 2016;Renders et al. 2016b) and ethanol-H 2 O (Renders et al. 2016b;Lv et al. 2017). On the other hand, the molecular weight of aqueous phase product from pure water system is lower than that of from biphasic solvent system (Fig. 4b). The main course of this result is the polarity of n-butanol-H 2 O lower than pure H 2 O (n-butanol/water mixture is monophasic when the temperature is above 125°C (n-butanol/H 2 O upper critical solution temperature)) Maczynski et al. 2007). The low polarity of n-butanol-H 2 O prevented the oligomers degrading in hydrolysis process, which was also evidenced by more oligomers in aqueous phase (HPLC, Fig. S3a and c). However, it cannot be ignored that butanol/ H 2 O has a dual behavior , when the temperature is higher than 125°C, they becomes a homogeneous system (homogeneous phase solvent system is considered more appropriate for b Fig. 2 Analysis of n-butanol phase, aqueous phase and residue obtained from eucalyptus valorization over Ru/C combined with MgO catalyst. Reaction conditions: 20 mL n-butanol, 20 mL water, 2.0 g extracted eucalyptus powder, 0.2 g Ru/C and 0.4 g MgO, 30 bar H 2 , 250°C, 6 h. The yield of products were calculated according to the calculation formulas in supporting information. The n-butanol phase products were qualitatively and quantitatively analyzed by GC-MS and GC spectrometers, and aqueous phase products were qualitatively and quantitatively analyzed with GC-MS and HPLC spectrometers. a Lignin depolymerization product yield, expressed as mg products per g eucalyptus . Acetic acid, EG and 1, 2-PG in n-butanol phase were not included in Fig. 2 a. b The yield of depolymerization products from cellulose and hemicellulose in aqueous phase, expressed as mg per g biomass, while the monomer aromatics in aqueous phase were not contained in Fig. 2 b. c GPC analysis of non-volatile fractions in n-butanol phase, inset a corresponding M w -d w /d logM graph. d FT-IR analysis of material and depolymerization residue. e HSQC NMR of non-volatile fractions of n-butanol phase , which avoids the inherent complexity of the biphasic phase solvent system. When the temperature is lower than 125°C, the solvent system gets back to two phases and separates simultaneously products. Therefore, n-butanol/H 2 O mixture is homogeneous phase under valorization temperatures (160 * 280°C) and more appropriate for biomass valorization (Fig. S11-12).
As for the non-volatile compounds in lignin oil, it can be found that just a few non-volatile compounds were dissolved by the n-butanol/H 2 O solvents without catalyst (GPC, Fig. S1a), which is consistent with the strong lignin characteristic absorption peaks in FT-IR curve of the residue (Fig. S2). Although the typical absorption peaks of lignin in the residue significantly reduced over MgO catalyst, the non-volatile compounds still distributed in lignin oil in the form of macromolecules (Fig. S1a). In the Ru/C and Ru/C-MgO cases, a large amount of lignin fragments (nonvolatile compounds) with M w = 155 * 897 were afforded after depolymerization. These indicated that both of Ru/C and Ru/C-MgO demonstrated the good degradation ability of lignin in n-butanol/H 2 O system. The weak lignin characteristic absorption peaks in FT-IR curves of the residue (Fig. S2) also confirmed the results (Fig. S1c). In addition, as shown in Fig. 5, most of the alkyl side chains in non-volatile compounds were retained over all catalysts, while the residual intermolecular linkages in lignin fragments were different over different catalysts. For instance, without catalyst, many intermolecular linkages were observed in non-volatile compounds, such as Observably, the selectivity of aqueous phase product differ strongly. A large amount of soluble oligomer, some acetic acid, acetol, formic acid and glycerine were afforded without catalyst ( Fig. 4b and S3a), implying the solvothermal action of n-butanol/ H 2 O has the effect on the cleavage of intermolecular linkage of (semi-)cellulose. Compared to without catalyst, the weight of aqueous phase products was slightly increased in MgO case, the effect of MgO led to decrease the yield of soluble oligomer and increase the acetic acid, formic acid, lactic acid (LA) and glycerol yield ( Fig. 4b and S3a). This is attributed to convert the cellulose and hemicellulose into small molecular acids under alkaline conditions (Yan et al. 2010;Jin et al. 2008;Jin and Enomoto 2011;Zhang et al. 2012). It was reported that alkaline catalyst (MgO) promoted monosaccharides (/glucose/fructose/ erythritol/aldoses) to convert into lactic acid because the coordination of oxygen-Mg 2? (Mg 2? coordinated to the ring oxygen or hydroxyl of monosaccharides) decreased the activation energy of breaking sugar rings or/and C-C bonds (Debruijn et al. 1987;Yang and Montgomery 1996;Jin and Enomoto 2011). Meanwhile, in the presence of Mg(OH) 2 , lactic acid also was produced from the direct decomposition of glycerin, and the decomposition of lactic acid could lead to produce formic and acetic acids (Jin and Enomoto 2011;Ding et al. 2018). Some of acetic acid and formic acid might also come from the depolymerization of lignin (Ma 2020;Lv et al. 2017).
Differently, the yield of 1-hydroxy-2-butanone and oligomers obviously increased when reaction was conducted over Ru/C combined with MgO ( Fig. 4b and S3a). The generation of 1-hydroxy-2-butanone might be ascribed to glucose (from cellulose hydrolysis) isomerize to fructose by alkaline catalyst, the C-C bond of fructose was cleaved by inverse aldol reaction and then cleaved the C-O bond to obtain 1-hydroxy-2-butanone (Deng and Liu 2014;Wang et al. 2019b). The increase of oligomer yield benefited from the synergetic effect between Ru/C and MgO that promoted the cleavage of the hydrogen linkages among lignin, hemicellulose and cellulose. However, the synergistic effect between Ru/C and MgO inhibited the conversion of glucose to C5/C6 alcohols ( Fig. 4b and S3a).
Thus, the synergistic effect between Ru/C and MgO catalysts promoted to cleave b-O-4 (A c ) linkages and inhibited to break b-b (B c ) linkages of lignin. This interaction also increased the attainability of aqueous phase oligomers and the small molecular alcohols (such as EG, 1, 2-PG, 1-hydroxy-2-butanone) rather than C5/C6 alcohols and sugar.

Influence of hydrogen pressure and reaction network
In addition to the effect of catalyst interactions and solvents, hydrogen pressure influenced the outcome of lignocellulose valorization in biphasic n-butanol/H 2 O solvent. To investigate its influence on the catalytic depolymerization process, the depolymerization process was performed under different hydrogen pressures (0 * 40 bar). Without external pressure (0 bar H 2 ), the G1/S1, G2/S2 and G3/S3 = monomers were selectively obtained over Ru/C-MgO catalysts (Fig. 6a). Despite the high monomer selectivity, the total monomer yield was low (about 17 wt%), indicating inefficient stabilization of reactive intermediates. This is confirmed by GPC analysis, just obtaining a small number of oligomers with large molecular weight (Fig. 6c). Increasing the H 2 pressure from 0 to 10 bar, the yield of aromatic monomers strongly enhanced depending on the increase of G3/ S3-OH yield. The aromatic monomers mainly comprised G3/S3 and G3/S3-OH when the H 2 pressure increased from 10 to 40 bar, the selectivity of G3/S3-OH was slightly increase. The monomer yield was the highest when the H 2 pressure was 30 bar. G3/S3-OH, G1/S1and G2/S2 were the predominant monomers if the H 2 pressure in the range of 10 * 30 bar. Above 30 bar H 2 , the monomer yield didn't increase with a slightly increase selectivity of G2/S2 and G1/S1, which was ascribed to quinone methides was converted to ethyl/methyl-substituted G/S through hydrolysis and hydrogenation/ decarbonylation under high H 2 pressure over alkali catalysis ). Meanwhile, a 3.90% yield of homosyringic acid (3, 5-dimethoxy-4-hydroxyphenylacetic acid, S2 = OOH) was also obtained, it may be derived from aromatic aldehyde oxidation (Rinesch et al. 2017;Zhu et al. 2020).
The trends deduced from Fig. 6a contribute to constructing the reaction network for native lignin catalytic conversion (Fig. 1). Coniferyl and sinapyl alcohol were identified as key intermediates in recent reporters Kumaniaev et al. 2017;Van den Bosch et al. 2017). The unsaturated compounds (G3/S3 = and G1/S1 = O) can even generate by solvothermal effect and are unstable at elevated temperature (Scheme 1, pathway P1, P3 and P8) in absence of catalyst (Fig. 4a) (Kumaniaev et al. 2017). Coniferyl and sinapyl alcohol are the key substances in Scheme 1, they can either undergo hydrogenation of the C a = C b bond to propanol-substituted G/S (pathway P2); repolymerization, producing soluble lignin oligomers (pathway P4); hydrogenolysis of the hydroxy group on C c to propenyl-substituted G/S (pathway P3); or retrograde aldol of the C = C bond of side chain in propenyl-substituted G/S to formoxylsubstituted G/S (pathway P8) or/and non-substituted G/S (pathway P9).
In the present of catalysts, the alkali depolymerization and hydrogenation/hydrogenolysis are combined to depolymerize lignin. In this couple system, not only the coniferyl and sinapyl alcohol were the key intermediates (from the hydrogenation/hydrogenolysis processes), but also the quinone methide (3) was also the pivotal intermediates (from alkali depolymerization reaction, pathway M3). In absence of hydrogen, the hydrogenolysis (pathway P3, M7), hydrolysis (pathway M6), hydrogenation (pathway M8) and decarbonylation (pathway M9 /decarboxylation (pathway M11) routes prevailed, as indicated by the high selectivity towards G3/S3 = , G2/S2 and G1/S1 monomers (Fig. 6a). The solvents or/and solubilized carbohydrates probably acted as the reducing agents (Galkin et al. 2016;Song et al. 2013). However, the rather low monomer yield obtained suggested the repolymerization reaction occurred due to the unsaturated sidechains (pathway M5 and/or P4 and/or P7) (Chen et al. 2016;Renders et al. 2016a). Repolymerization can be prevented by hydrogenating these unsaturated bonds (Chen et al. 2016;Renders et al. 2016a), thereby produce stable phenolic products. Under a relative low H 2 pressure (10 bar), hydrogenation of coniferyl/ sinapyl alcohol (pathway P2) and hydrogenolysis of propanol-substituted G/S (pathway P5) were the major reaction pathways, resulting in generating G3/S3-OH and G3/S3 monomers (Fig. 6a). Meanwhile, 2.21% of G3/S3 = monomers yield (Fig. 6a) indicated that hydrogenolysis of coniferyl/sinapyl alcohol (pathway P3) and subsequent hydrogenation of propenyl-Scheme 1 Proposed reaction network of lignin valorization to aromatic monomers over Ru/C combined with MgO substituted G/S (pathway P6) also constituted the reaction pathways. With the H 2 pressure further increasing, direct hydrogenation of coniferyl/sinapyl alcohol (pathway P2) were predominant and more G3/ S3-OH monomers were selectively produced. Hydrogenolysis of propanol-substituted G/S (pathway P5) was not continued when the H 2 pressure increased from 20 to 40 bar with a very slightly change of G3/S3 monomers yield. This change of selectivity can be explained by the fact that hydrogenolysis reactions are the negative order in a relative high hydrogen pressure, in contrast to hydrogenation reactions (Bernas et al. 2008). Meanwhile, quinone methide took off methoxy group (pathway M4) to enol ether and subsequently hydrolyzed (pathway M6) or/and conducted hydrogenolysis (pathway M7) to ethenyl/ ethanoyl-substituted G/S, and then ethenyl-substituted G/S hydrogenated to ethyl-substituted G/S (pathway M8); or conducted decarbonylation of ethanoyl-substituted G/S to methyl-substituted G/S (pathway M9) (Li et al. 2015;Schutyser et al. 2018). It is thereby inferred that high H 2 pressure is contributed to the increase of G2/S2 and G1/S1 monomers. It is worth noting that acetoxy-substituted G/S was detected when hydrogen pressure was higher than 30 bar, which was attributed to oxidize ethanoyl-substituted G/S to acetoxy-substituted G/S (pathway M10), and the oxygen likely come from lignocellulose depolymerization process.
In addition, the effect of hydrogen pressure on the products in aqueous phase is illustrated in Fig. 6b. In absence of external hydrogen, the aqueous phase products yield was a little low than those of under other pressurized hydrogen, while the monomers yield was the highest. Although the mass of oligomers and the component of main monomers were similar with the aqueous phase products over MgO catalyst under n-butanol/H 2 O (Fig. 4b), MgO had great effect on the selectivity of the monomer products (Ryu et al. 2019;Swatloski et al. 2002). DMF was obtained under 0 and 10 bar H 2 , which was ascribed to the hydrogenation and subsequent hydrogenolysis of 5-hydroxymethylfurfural (HMF) that produced from catalytic dehydration of cellulose under low hydrogen pressure. Much soluble aqueous phase products increased with the mass of oligomers increasing and monomers decreasing when the hydrogen pressure increased from 10 to 40 bar, indicating that high H 2 pressure facilitated the cleavage of intermolecular linkages between lignin and cellulose/hemicellulose. The mass of acetic acid and methanol just slightly changed with the H 2 pressure increasing ( Fig. 6b and d), indicating that they were obtained from lignin depolymerization instead of (semi-)cellulose depolymerization (Huang et al. 2017b;Lv et al. 2017). With increasing H 2 pressure, the mass of aqueous phase products gradually increased and reached a plateau at 30 bar. The mass of oligomers followed a similar trend, as HPLC profile shown (Fig. 6d). Thus it could be seen that depolymerization (i.e. hydrogenation) of the soluble cellulose and hemicellulose fragments required a higher hydrogen pressure than lignin valorization. At the relative low hydrogen pressure, lignin depolymerized effectively through the hydrogenation-dehydration-hydrogenation (pathway P2 ? P3 ? P6, in Scheme 1). Overall, 30 bar H 2 is the optimal pressure range to acquire high yields of both aromatics and the mass of aqueous phase products.

Effect of reaction temperature
Subsequently, we investigated the effect of reaction temperature, since it is known that temperature greatly promotes hemicellulose and cellulose hydrolysis and delignification during depolymerization. Similar experiments were therefore conducted at 160 * 280°C. Results of these experiments are summarized in Fig. 7. As shown in Fig. 7a, the yield of lignin oil and lignin monomer increased with the temperature increasing, accompanying the yield of aromatic monomers increased slowly when the temperature above 220°C. These results were also verified by GPC analysis (Fig. S5). The GPC chromatogram of the non-volatiles obtained from 160 * 280°C displays the response peaks of oligomers (eluted before 17.5 min) and monomers (eluted after 17.5 min) become large and strong when the temperature increases from 160°C to 280°C, and the changes of the monomer aromatics peak areas are also full accordance with the slowly increasing monomer aromatics yield at 220 * 280°C (Fig. S5, Fig. 7a). In Fig. 7a, a small number of the benzene ring hydrogenation products were detected (the benzene ring hydrogenation products were counted into the other (Fig. 7a)) when the depolymerization reaction was conducted at 160 or 180°C. These results indicated that the lignin was incomplete depolymerized at 160 and 180°C and the remaining hydrogenation activity of Ru/C was applied for the hydrogenation of benzene rings. Obviously, the change of reaction temperature shown less effect on the aromatic components (G3/S3-OH, G3/S3, G2/ S2 = OOH, G2/S2, G1/S1 and G0/S0), while the yield of lignin oil and monomers increased with the temperature increasing (Fig. 7a). Both the increase of the total monomers yield and G3/S3 yield indicated that the oligomers were converted to G3/S3-OH and subsequent hydrogenolysis of G3/S3-OH (pathway P3) and hydrogenation of G3/S3 = to G3/S3 were conducted (pathway P6). With temperature elevating, the yield of G3/S3-OH slightly changed because the oligomer was converted to G3/S3-OH. Meanwhile, it is seemed that the yield of G2/S2 = OOH and G2/S2 were not affected under elevated temperature, implying the M8 and M10 reaction pathways were insensitive to temperature. Interestingly, G1/S1 and G0/S0 derived from decarbonylation reaction of M10 and P9 pathways, respectively. Their yields changed slightly at 200 * 250°C, while G1/S1 yield increased and G0/ S0 yield decreased obviously at 280°C, indicating that the pathway M10 was facilitated and pathway P9 was inhibited when reaction temperature was above 250°C. The decrease of G0/S0 was ascribed to the generation rate of G1/S1 = O could not match the rate of G1/S1 = O decarbonylation (Lv et al. 2019), resulting in the condensation of surplus G1/S1 = O.
Moreover, the reaction temperature shows great effect on hemicellulose and cellulose degradation process. For instance, below 200°C, just some hemicellulose degraded to form soluble oligomers, while the aqueous phase product mass was rapidly increased to 230 mgÁg biomass -1 at 200°C (Fig. 7b). Correspondingly, the cellulose typical XRD diffraction peaks (2h = 15.2°(1-10), 22.8°(200), 34.5°(004)) in residue were weakening at 160 * 250°C, and disappeared at 280°C (Fig. S7). From XRD analysis of depolymerization residue, the diffraction peaks at 15.2°(1-10) and 34.5°(004) were completely removed from cellulose matrix to dissolve in aqueous phase at 220°C, while the diffraction peak at 22.8°(200) disappeared above 250°C.The increasing temperature led to an increased dissolution of hemicellulose and cellulose, from 7.1wt% up to 100 wt%. The mass of oligomers gradually increased from 41 mgÁg biomass -1 to 250 mgÁg biomass -1 with increasing temperature (Fig. 7b). Meanwhile, a large amount of solubilization oligomers were converted to small molecular products with the temperature increasing (Fig. 7b). The mass of acetic acid changed slightly, being similar with the trend in different H 2 pressure cases (similar mass of 55 mgÁg biomass -1 ), indicating that the change of reaction temperature and hydrogen pressure had less effect on the generation of acetic acid. Moreover, when the temperature elevated from 200 to 280°C, the selectivity towards EG and 1,2-PG increased, from 5 mg g biomass -1 up to 68 mg g biomass -1 and 16 mg g biomass -1 up to the plateau 58 mg g biomass -1 , respectively. It can be explained by the fact that high temperature promotes glucose's C-C bonds to cleave by hydrogenation/hydrogenolysis. The mass of 1-hydroxy-2- Fig. 7 a Lignin monomers, b carbohydrate products obtained from catalytic depolymerization in n-butanol/water under different temperature. Reaction conditions: 20 mL n-butanol, 20 mL water, 2.0 g extracted eucalyptus powder sawdust, 0.2 g catalyst, 0.4 g MgO, 160 * 280°C, 30 bar H 2 , 2 h butanone almost didn't change at 200 * 280°C. It is worthwhile to notice that tetrahydrofurfuryl alcohol, DMF and ethanol were detected at high H 2 pressure and temperature (30 bar and 250/280°C) or in low H 2 pressure cases (0 and 10 bar), as shown in HPLC curves ( Fig. 6d and Fig. S6), the reasons for these results are not clear.

Effect of contact time
Increase contact time may enhance high conversion or/and yield but reduce the product selectivity because of multi-step reactions. Long contact time have indeed been reported for lignin valorization but not in the view of co-depolymerizing cellulose. The effect of contact time on the depolymerization was therefore studied under the optimal reaction conditions (20 mL n-butanol, 20 mL water, 2.0 g extracted eucalyptus powder sawdust, 0.2 g Ru/C, 0.4 g MgO, 250°C, 30 bar H 2 .).
As shown in Fig. 8a, increasing the contact time (from 2 to 24 h) exhibited a little effect on lignin oil yield but did change the yield of total monomer aromatics and the selectivity of monomer aromatics. The yield of total monomer aromatics increased from 38.89% for 1 h, reached to the top of 68.54% for 6 h, and then decreased to 51.21% for 24 h. This change can be explained that lignin and oligomers were degraded and further converted to monomer aromatics in the contact time rage of 1 * 6 h, and the decrease yield of monomer aromatics were ascribed to repolymerization when the contact time longer than 8 h. These results also confirmed by GPC analysis: a small amount of high M W oligomer was detected at the start of the reaction and the high M W fragments were obtained after the reactions under too long contact time (Fig. S8). The similar trend was also evidenced in previous reports (Huang et al. 2017a). The formation of high M W fragments were ascribed to the condensation of unsaturated double bond or/and carbocation on benzene ring (or on Ca in side-chain) under the long contact time . Thus, longer time did not enhance lignin oil yield. Meanwhile, with reaction time increasing, more coniferyl/sinapyl alcohols released from lignin matrix were converted to G3/ S3-OH (pathway P2), G3/S3 (pathway P3 ? P6) and G0/S0 (pathway P3 ? P8 ? P9) (Scheme 1). The yield of G3/S3-OH was increased from 22.12% to the highest 35.84% at 6 h and then decreased. The yield of G0/S0 monomers did not obviously affected when the contact time was C 8 h (Fig. 8a), attributing to the repolymerization of G1/S1 = O or/and coniferyl/sinapyl alcohols (Deuss et al. 2015). These were also confirmed by the GPC analysis (produced new high M W fragments at 8 h, 16 h and 24 h, Fig. S8). Moreover, no obvious change of G3/S3 yield during 2 * 24 h indicated that the formation of G3/S3 (propyl-substituted monomers) primarily derived from the pathway P3 ? P6. Whereas the yield of G1/S1 increased with the contact time increasing, Fig. 8 a Lignin-derived monomer aromatics, b carbohydrate products obtained from catalytic depolymerization in n-butanol/ H 2 O under different contact time. Reaction conditions: 20 mL n-butanol, 20 mL water, 2.0 g non-extracted eucalyptus powder sawdust, 0.2 g catalyst, 0.4 g MgO, 250°C, 30 bar H 2 , 1 * 24 h which is responsible for the decarbonylation of G2/ S2 = O over alkali-catalyzed (Lv et al. 2019).
The reaction time also affected the mass of aqueous phase products. The aqueous phase product mass was about 33.30 wt % (based on eucalyptus mass) when the contact time was 1 h, and then increased to the highest mass of 63.35 wt % with the contact time increasing to 8 h, but the mass of aqueous phase products decreased when the contact time was longer than 8 h. Meanwhile, the mass of oligomer firstly increased and subsequently decreased to complete dissolve and then increased to a plateau. The mentioned above variation trends are likely correlated with the magnesium species during depolymerization process, that is, MgO was converted successively to Mg(OH) 2 , Mg 5 (-CO 2 ) 3 (OH) 2 Á4H 2 O and MgCO 3 with the increasing contact time (Fig. S10). These changing magnesium species were also reported and exhibited similar effect on the cellulose depolymerization products (Zhang et al. 2010;Nakagawa et al. 2014). The main aqueous phase products are soluble oligomers, tetrahydrofurfuryl alcohol, ethanol, 1-hydroxy-2-butanone, 1, 2-PG, EG, glycerin and LA ( Fig. 8b and Fig. S9). Expect for methanol and acetol, the mass of main monomers increased when the contact time increased from 1 to 8 h. As the contact time continued to prolong, the mass of total aqueous phase products and most monomers decreased but oligomers appeared again, which were attributed to the change of magnesium species.
The stability of the solvent and catalyst recycle To investigate the effect of depolymerization process on the stability of n-butanol solvent, the gaseous and aqueous products were analyzed after conducting the valorization of eucalyptus or without feedstock under 250°Cin n-butanol/H 2 O for 6 h. From the result of GC and GC-MS, a portion of n-butanol was converted whether the feedstock was added or not (Fig. 9, Table S2). Without feedstock, the main components in gaseous phase were H 2 , CH 4 , C 2 H 6 and C 3 H 8 , suggesting that the activity of catalyst was contributed to cleave C a -C b bond in n-butanol for generation a large amount of CH 4 (25.85%) and C 3 H 8 (15.18%, not include propane of n-butanol phase (Fig. 9)) (  Renders et al. 2018). A similar reaction of propanol was converted to methane and ethane over hydrogenation catalyst (Lan et al. 2018). Butyl butyrate and propane were detected in n-butanol phase (Fig. 9), ascribing to esterification between n-butanol and butyric acid (originated from butaldehyde oxidation) over MgO catalyst. Therefore, a portion of n-butanol solvent was converted to propane, methane and butyl butyrate in n-butanol/H 2 O solvents over Ru/C-MgO catalysts. However, just a few gaseous products (H 2 , CO 2 , CO, CH 4 , C 3 H 8 and C 4 H 10 ) were obtained when feedstock was added (Table S2), CO 2 , CO came from decarbonylation and decarboxylation reactions. Butyl butyrate and propane were also decreased in n-butanol phase (Fig. 9), suggesting that changing the catalyst performance and reducing the catalyst dosage would be reduced the side-reaction of the solvents .
In this work, MgO was converted to Mg(OH) 2 during the hydrothermal process and also confirmed by the XRD analysis of the solid mixture (depolymerization residue and catalyst) (Fig. 11). After reaction, we conducted the recycle experiment of catalysts. Mg(OH) 2 , located at the bottom of aqueous phase, was calcined at 550°C(6 h) to obtain MgO for reuse. Meanwhile, the used Ru/C located in n-butanol phase was washed with ethanol and dried at 105°C (10 h) for reuse with regenerated MgO. In this way, a black powder could be recycled about 97.4 wt% of the initial Ru/C (0.2 g), the white powder was afforded about 92.5wt% (viz. 0.37 g) of the original MgO (0.4 g, Fig. S10). To investigate whether the decrease of recovered catalyst (MgO and Ru/C) mass is related to the metal leaching off, the metal ion concentration in liquid products was detected by ICP-AES analysis. As Table S6 shown, the Ru and Mg metal leaching was not detected in liquid products after the reaction, or the Ru and Mg leaching was below the limit of detection (were \ 0.1 and \ 0.1 mgÁL -1 , respectively) (Table S6), indicating almost no obvious Ru and Mg species was dissolved by catalytic depolymerization and intermediate organic compounds in the process. Therefore, the decrease mass of recovered catalyst were related to operation. Subsequently, the eucalyptus powder depolymerization experiment (Fig. 1) was carried out over the recycled catalysts under the given conditions (fresh catalysts were added to make up for the loss of retrieved catalysts). As Fig. 10 shown, the recycled catalysts afforded a decrease lignin oil (24.36 wt% vs. 25.50 wt%, basis of feedstock) and aromatic monomer yield (56.58 wt% vs. 68.54wt%, basis of lignin) compared to fresh Ru/C-MgO, with a little higher selectivity towards G3/H3-OH monomers (Fig. 10a). Similarly, the yield of aqueous product was also slightly decreased (596.03 vs. 556.27 mg g -1 , basis biomass). The decreasing of lignin oil, monomer yield and aqueous phase products probably because of the altered property of catalyst, especially MgO. The reused catalysts relevant characterization was analyzed by XRD (Fig. 11). The X-ray diffraction patterns displayed no Ru diffraction peak was detected as fresh catalyst, implying Ru particles were not obviously aggregated (the TEM image of Fig. 12  (a * b) and (e * f) also demonstrated this result). While obvious difference between fresh and used MgO are exhibited as the diffraction peaks shown (PDF # 32-0671, Fig. 11). For instance, the MgO diffraction peaks (PDF # 45-0964, periclase) were obviously became weak due to the MgO particles became small and a portion of MgO species changed the crystal form. Although the TEM image of  n-butanol, 20 mL water, 2.0 g extracted eucalyptus powder sawdust, 0.2 g Ru/C, 0.4 g MgO, 250°C, 30 bar H 2 , 6 h (c * d) shown no obvious aggregation particle was observed on the surface of spent MgO, the Fig. 12 (g * h) also confirmed that a portion of MgO exhibited another crystal form (Fig. 12h, MgO crystal structure with lattice fringes spacing 0.31 nm). To conclude, this recycle experiment confirms that the used Ru/C can be reused with a good catalytic activity and spent MgO was changed with a decline activity.

Conclusions
A metal-alkaline catalytic biomass valorization approach was developed, the one-pot complete conversion and separation of woody biomass into two value product streams: lignin derived aromatics and (semi-)cellulose derived small molecular (C2 * C5) alcohols and acids. Lignocellulose depolymerization was conducted in a mixture of n-butanol/H 2 O (homogeneous phase) at 250°C in metal-alkaline coupling catalytic system. The mixed solvents extracted and depolymerized of both lignin and hemicellulose, while the metal-alkali catalysts and H 2 were necessary to break inter-/intramolecular linkages of lignocellulose into target products (aromatics, alcohols and acids). After reaction, the liquid products were separated into n-butanol phase (rich of aromatics) and aqueous phase (rich of small molecular Fig. 11 The XRD patterns of fresh and used Ru/C and MgO Fig. 12 The TEM images of fresh and spent Ru/C and MgO catalysts: a * b fresh Ru/C; c * d fresh MgO; e * f spent Ru/C; g * h spent MgO alcohols and acids) when the reaction system temperature below 125°C. Furthermore, the used Ru/C and alkali catalysts were recovered in n-butanol phase and H 2 O phase, respectively. This work also exhibits that simultaneous depolymerization all components of lignocellulose in one-pot catalytic refining process is feasible, but one should take into consideration that the high yield and selectivity of products can hardly obtain without the appropriate reaction conditions, such as the catalyst, hydrogen pressure, solvent composition and temperature had different consequences for valorization the main component of biomass.