|Home | About | Journals | Submit | Contact Us | Français|
The depolymerization of lignin to bioaromatics usually requires a hydrodeoxygenation (HDO) step to lower the oxygen content. A mixed Cu–Mg–Al oxide (CuMgAlOx) is an effective catalyst for the depolymerization of lignin in supercritical ethanol. We explored the use of Ni-based cocatalysts, i.e. Ni/SiO2, Ni2P/SiO2, and Ni/ASA (ASA = amorphous silica alumina), with the aim of combining lignin depolymerization and HDO in a single reaction step. While the silica-supported catalysts were themselves hardly active in lignin upgrading, Ni/ASA displayed comparable lignin monomer yield as CuMgAlOx. A drawback of using an acidic support is extensive dehydration of the ethanol solvent. Instead, combining CuMgAlOx with Ni/SiO2 and especially Ni2P/SiO2 proved to be effective in increasing the lignin monomer yield, while at the same time reducing the oxygen content of the products. With Ni2P/SiO2, the lignin monomer yield was 53 wt %, leading to nearly complete deoxygenation of the aromatic products.
Biomass is currently mainly used to generate heat and power, satisfying about 10% of the global primary energy demand.1 The use of lignocellulosic biomass is expected to grow as it constitutes a renewable source of fuels and chemicals. Lignocellulosic biomass, which consists of cellulose, hemicellulose and lignin, is the most abundant, fastest growing and cheapest form of land-based biomass.2 Although most attention has been focused on the valorization of the cellulosic and hemicellulosic parts of biomass, creating value from lignin is a necessity to make lignocellulosic (2nd generation) biorefineries profitable. Although lignin is usually burned to recover its heating value, it is likely that not all this energy is needed in a biorefinery setting. Chemical conversion processes can add more value to lignin by depolymerizing the three-dimensional network of monomethoxylated (guaiacyl, G), dimethoxylated (syringyl, S) and nonmethoxylated (p-hydroxyphenyl, H) phenylpropanoid units.3,4 Lignin is considered as a potential future source of biorenewable aromatics.5
Several approaches have been explored to depolymerize lignin, involving pyrolysis, hydrocracking, hydrogenolysis, oxidation, and hydrolysis, leading to monomeric units such as cyclic hydrocarbons and aromatics.3,4,6,7 Usually, the yield of aromatic monomers is higher when hydrogen or hydrogen-donating solvents such as methanol,8,9 ethanol,9,10 and iso-propanol9,11 are employed in the hydrogenolysis of lignin.6 Lignin can, for instance, be converted into monomeric cyclohexyl derivatives in supercritical methanol at 300 °C.8 Using a carbon-supported Ni catalyst, a range of useful hydrocarbon products was obtained in methanol at 200 °C in the presence of molecular H2.9 Supercritical ethanol is an effective solvent for the catalytic conversion of Kraft lignin at 280 °C using a carbon-supported Mo-carbide catalyst.10 In this approach, no external hydrogen was needed. Iso-propanol has also been mentioned as a promising solvent and hydrogen transfer agent in the depolymerization of organosolv lignin at 300 °C using Raney Ni catalysts.11
We have earlier reported a method to convert lignin into aromatics, making use of a mixed CuMgAl-oxide catalyst.12 This one-step approach comprises thermal cracking of lignin in supercritical ethanol and protection of the resulting fragments by reactions with ethanol; it yields predominantly alkylated aromatics in large amounts (>50 wt %) with no or little char formation. This approach can be employed to convert different types of lignin including Kraft lignin. Detailed investigations have demonstrated that the use of ethanol serves three purposes: it acts as a capping agent to protect intermediate phenolic fragments from repolymerizing (through C- and O-alkylation), it scavenges formaldehyde obtained by demethoxylation of lignin’s methoxy groups, and it provides hydrogen for metal-catalyzed hydrogenolysis reactions.13 Recently, we optimized the preparation of the mixed CuMgAl oxide catalyst and the reaction conditions toward high aromatics yield in the depolymerization of a soda lignin derived from wheat straw.14
Usually, hydrodeoxygenation (HDO) is considered as a valuable step to upgrade the phenolic compounds in oils derived from lignin depolymerization, i.e., to lower the oxygen content.15 Alumina-supported CoMo and NiMo sulfides, which are widely used as catalysts for hydrotreating operations in oil refineries, can be used for this purpose.16,17 Although these catalysts display high HDO activity in their sulfide form, the sulfur content of lignin-derived bio-oils is too low to maintain the catalyst in the active sulfided state during HDO reactions.18 Accordingly, transition metal phosphides have been considered as catalysts to catalyze HDO reactions.18,19 The preparation of metal phosphides is relatively straightforward and, similar to conventional hydrotreating catalysts, base metals such as Ni and Mo make excellent metal phosphide HDO catalysts.18 Transition metal phosphides are usually prepared by reduction of a metal salt or oxide precursor in the presence of a phosphorus compound.19 In the classical phosphate reduction approach, the support is impregnated with a metal nitrate and ammonium phosphate followed by reduction at high temperatures in the 500–800 °C range.20 As support, silica is preferred over alumina (c.q., aluminum phosphates form, hindering phosphidation of the metal component).21 The use of more reactive phosphorus compounds such as phosphite, hypophosphite, and phosphine has also been explored.19,22 The benefit of these precursors is that the reduction can be carried out at lower temperatures, which leads to higher dispersion of the final active metal phosphide phase.19
Some of us were the first to report that Ni2P is an active catalyst for hydrodenitrogenation reactions of oil feedstock.23 This catalyst also displays high catalytic activity in hydrodesulfurization reactions.24 More recently, the promise of Ni2P/SiO2 for the HDO of guaiacol,25−27 dibenzofuran,28 and methyl-tetrahydrofuran29 has been demonstrated. In another study, a bio-oil obtained by pyrolysis of lignocellulosic biomass was upgraded using Ni2P/SiO2 as a HDO catalyst.26 Usually, it is found that the HDO activity of nickel phosphides is higher than that of metallic Ni catalysts.30 In this context, it is interesting to mention that metallic Ni catalysts have been employed for the depolymerization of lignin in the presence of molecular hydrogen yielding substituted alicyclic and aromatic hydrocarbons.31,32 A Ni catalyst supported on Al-SBA-15 catalyst was able to convert lignin to cycloalkanes.33 Despite the promise of this approach, the high rate of aromatic ring hydrogenation is problematic in view of the process economics.34 Ni-based heterogeneous catalysts for energy and environmental applications have been recently reviewed.35
In the present study, we report for the first time about the substantial synergy between the earlier described mixed CuMgAl-oxide catalyst and nickel (phosphide) catalysts in the depolymerization of lignin in supercritical ethanol. Under optimized conditions, using Ni2P/SiO2 as the HDO catalyst, a temperature of 340 °C and a reaction time of 4 h the monomers yield from soda lignin was 53 wt %. This result compares favorably to the monomer yields obtained in experiments with either of the two catalysts. Also Ni/SiO2 and Ni/ASA (ASA = amorphous silica alumina) catalysts were found to perform in a favorable manner with the mixed CuMgAlOx catalyst. A detailed comparison was made of lignin conversion under optimized conditions using these catalysts. Our aim was to clarify the influence of the different active phases and supports on the monomer yield and degree of deoxygenation.
Protobind 1000 alkali lignin was purchased from GreenValue. It was produced by soda pulping of wheat straw (sulfur-free lignin with less than 4 wt % carbohydrates and less than 2 wt % ash). Detailed experimental information (chemicals, detailed catalyst preparation and characterization, lignin residue analysis) is given in the Supporting Information (SI) and in our preceding publications.12−14
Mixed CuMgAl-oxides were prepared by calcination of a hydrotalcite precursor. The Cu loading was 20 wt % and the (Cu + Mg)/Al ratio 4. This catalyst is denoted as CuMgAlOx or in short as “Cu”. Supported Ni catalysts were prepared by pore volume impregnation using an aqueous Ni(NO3)2·6H2O solution. In all cases, the Ni loading was 10 wt % (Table 1). As supports, silica (SiO2, pore volume 1.3 mL/g, surface area 210 m2/g), and amorphous silica alumina (ASA, pore volume 0.7 mL/g, surface area 350 m2/g, 55 wt % nominal Al2O3 content) were used. After impregnation, the catalyst precursors were dried and calcined in air. The catalysts are denoted as NiO/SiO2 and NiO/ASA. Part of the NiO/SiO2 catalyst was impregnated with an aqueous solution of (NH4)2HPO4 (P/Ni = 1 initial precursor ratio). This phosphate-impregnated catalyst, denoted as NiO-P/SiO2, as well as NiO/SiO2 and NiO/ASA were reduced in H2 flow at 620 °C for 3 h. The resulting reduced catalysts are named Ni2P/SiO2, Ni/SiO2, and Ni/ASA. A more detailed description of the preparation of these materials can be found in the SI.
Soda lignin was converted in a 100 mL stainless-steel high-pressure Parr autoclave. Typically, the autoclave was charged with a suspension of 0.5 g catalyst and 1.0 g lignin in 40 mL ethanol. When two catalysts were combined, the autoclave was charged with 0.5 g CuMgAlOx and 0.5 g Ni-containing catalysts. When the ratio of CuMgAlOx and Ni2P/SiO2 catalysts was optimized, their mass was varied at 0.25, 0.5, and 0.75 g, but the sum of catalyst masses remained 1 g. The reactor was then sealed and purged with nitrogen several times to remove air. After leak testing, the pressure was increased to 10 bar with nitrogen or 30 bar with hydrogen and the reaction mixture was heated to 340 °C under continuous stirring at 500 rpm within 1 h. After a reaction time of 4 h (or in one case 8 h), the heating oven was removed, and the reactor was allowed to cool to room temperature. A workup procedure was developed to distinguish light (THF-soluble) lignin residue (THF = tetrahydrofuran), heavy (THF-insoluble) lignin residue and char. A detailed description of this workup procedure can be found in our previous report,13 and a shortened version is presented in the SI.
The liquid phase product mixture was analyzed by a Shimadzu 2010 GC-MS system equipped with a RTX-1701 column (60 m × 0.25 mm × 0.25 μm) and a flame ionization detector (FID) together with a mass spectrometer (MS) detector. Identification of products was based on a search of the MS spectra with the NIST11 and NIST11s MS libraries. The GC peaks with the same molecular weight (Mw) were unified and presented by the structure determined by GC-MS. These products were further divided into four groups, namely hydrogenated cyclics (−O (oxygen-free)) [HC–O], hydrogenated cyclics (+O (oxygen-containing)) [HC+O], aromatics (−O) [Ar–O], and aromatics (+O) [Ar+O], according to the nature of the ring structure and functional groups. The FID response factors were calculated using the Effective Carbon Number (ECN) method36 to determine the relative response factors corrected by the molecular weight of the compounds relative to n-dodecane, which served as the internal standard. The lignin monomers and ethanol product yields, the aromatic hydrogenation (HYD) and hydrodeoxygenation (HDO) degrees, and the yields of lignin residue (LR) and char were determined as follows:
The composition of the gas phase was analyzed by sampling a small amount of gas from the autoclave on an Interscience Compact GC system, equipped with Molsieve 5 Å and Porabond Q columns, each fitted with a thermal conductivity detector (TCD) and an Al2O3/KCl column fitted with a flame ionization detector (FID).
GPC analyses were performed on a Shimadzu apparatus equipped with two columns connected in series (Mixed-C and Mixed-D, polymer Laboratories) and a UV–vis detector at 254 nm. The column was calibrated with polystyrene standards. Analyses were carried out at 25 °C using THF as eluent with a flow rate of 1 mL/min. For the lignin residue analysis, the sample was prepared at a concentration of 2 mg/mL in THF. All the samples were filtered using 0.45 μm filter membrane prior to injection.
Figure Figure11 shows the XRD patterns of the reduced Ni catalysts. Reduction of NiO-P/SiO2 leads to phase-pure Ni2P/SiO2, while metallic Ni catalysts were obtained by reducing NiO/SiO2 and NiO/ASA. The nominal Ni loading of these catalysts is around 10 wt %. The textural properties of the Ni-based and the CuMgAlOx catalyst are collected in Table 1. These catalysts were used in different combinations to convert soda lignin in ethanol under supercritical conditions. The reaction temperature was 340 °C, and the reaction time was 4 h. Hydrogen was not added to those reaction experiments that involved the use of CuMgAlOx, as this catalyst produced in situ hydrogen by reforming of ethanol.13,37 When only supported Ni was used as the catalyst, the autoclave was pressurized by 30 bar H2 atmosphere. Another experiment involved the use of Ni2P/SiO2 in 10 bar N2 atmosphere. The results of these eight reaction experiments are shown in Table 2.
The monomer yields obtained upon reaction of soda lignin using only Ni2P/SiO2 and Ni/SiO2 catalysts were around 10 wt % (entries 2, 3, and 7 in Table 2). These reactions were started in 30 bar H2 gas (except entry 2 in Table 2), which is equivalent with 75 mmol of hydrogen. Very little (9 mmol for Ni2P/SiO2 in N2) or a nearly similar amount of hydrogen (60–70 mmol for Ni2P/SiO2 in H2 and Ni/SiO2 in H2) remained after reaction over these catalysts (entries 2, 3, and 7 in Table S1), which implies that hydrogen was not produced (Ni2P/SiO2) or was not consumed. From this, it can be concluded that lignin depolymerization toward aromatic monomers was very limited when only silica-supported Ni and Ni-phosphides were used as catalysts.
The use of CuMgAlOx and Ni/ASA catalysts gave similar lignin monomer yields (27–28 wt %, entries 1 and 5 in Table 2). The product distributions of these two reaction experiments were however very different (Figures Figures22a and S1): the mixed oxide catalyst yields nearly the same amounts of oxygen-free and oxygen-containing aromatics, whereas the product mixture derived from using Ni/ASA contains predominantly oxygen-containing aromatics. This difference indicates that the Ni catalyst is much less active in the HDO of primary oxygen-containing aromatics reaction products. Lercher et al. reported that 18 wt % lignin monomers (mainly alkanes and naphthalenes, 74% selectivity) were obtained by depolymerizing lignin over a 30 wt % Ni/ASA catalyst in dodecane at 250 °C in 40 bar hydrogen.31,32 In the present study higher lignin monomer yield and lower aromatic ring hydrogenation were obtained at higher reaction temperature (28% and 16%, respectively, at 340 °C, entry 5 in Table 2). The reason for this can be that these authors used dodecane, while we use ethanol as the solvent. It is known that the solvent has significant influence on product selectivity for hydrogenolysis/hydrogenation processes.11 There is strong evidence that alcoholic solvents such as methanol and ethanol tend to strongly adsorb on the Ni metal catalyst surface by forming alkoxy species. These alkoxy species partially block the active sites for hydrogenation reactions which explains the much lower rate of aromatic ring hydrogenation activity compared to the reactions in nonpolar alkane solvents.11 Besides, extensive alkylation of the aromatic rings by ethanol might also prevent their hydrogenation.
In order to improve the monomer yield and HDO degree obtained using CuMgAlOx, we employed combinations of this catalyst with different Ni-based catalysts. Lignin monomer yields using these combined catalyst systems were quite similar, but much higher than with either the CuMgAlOx or Ni-based catalysts (entries 4, 6, and 8 in Table 2). The product distributions for the combined-catalyst experiments are summarized in Table 2 and presented in more detail in Figure Figure22. While oxygenated and oxygen-free aromatics were the predominant products in nearly similar amount with CuMgAlOx as the catalyst, using an additional Ni-based HDO catalyst resulted in a shift of the product mixture toward oxygen-free aromatics and cyclohexenes. Table 2 shows that the total lignin monomer yield is higher for the combined catalyst systems than for the separate experiments, except for the combination of CuMgAlOx and Ni/ASA. The HDO degree is strongly increased by the combined-catalyst experiments from about 50 to 75 wt % and higher. The aromatics hydrogenation degree is also increased but to a lesser extent, i.e., from about 30 wt % of the lignin-derived monomers were hydrogenated when CuMgAlOx was the only catalyst to 35–47 wt % when an additional HDO catalyst was present. The slightly lower aromatics hydrogenation degree in the experiment with Ni/ASA is consistent with the lower amount of H2 formed as compared with Ni2P/SiO2 and Ni/SiO2 (entries 6, 4, and 8 in Table S1).
The present results are very promising in comparison to HDO of phenolic compounds, for instance obtained by pyrolysis of biomass, which usually results in extensive hydrogenation of the aromatic rings. Typically, noble metals or Ni are used as catalysts to lower the oxygen content of the phenolic compounds as reviewed by Daud and co-workers.15 In comparison with these studies, the aromatics hydrogenation degree in the present study is much lower at very high HDO selectivity. Another aspect worth discussing is the formation of reaction products from ethanol. Cu and basic sites also catalyze Guerbet38,39 and esterification reactions, producing higher alcohols and esters from ethanol (entry 1 in Table S2).14 Adding Ni2P/SiO2 or Ni/SiO2 to the CuMgAlOx only slightly increased the conversion of the solvent. This is beneficial, as the unconverted ethanol can be recycled. On the other hand, the use of Ni/ASA resulted in much higher ethanol conversion, mainly due to formation of diethyl ether (see Tables S2 and S3). Dehydration of ethanol is catalyzed by acid sites present on alumina and amorphous silica alumina.40 Besides diethyl ether, also hydrocarbons are produced using Ni/ASA, which is likely due to a combination of acid-catalyzed dehydration and hydrogenation. This reaction consumes hydrogen and explains the lower aromatic hydrogenation degree. These reactions will not take place on the nonacidic silica-supported Ni catalysts.41 While there is clear synergy between the two catalysts with respect to lignin depolymerization, especially with Ni2P/SiO2 and Ni/SiO2, the effect of combining two catalysts appears to be additive with respect to ethanol yield.
We next discuss the lignin residue and char as the byproducts obtained during lignin conversion. The results in terms of lignin monomers yields, THF-soluble and THF-insoluble lignin residues (LR), and char are presented in Table 3. The high yields of THF-soluble lignin residue and the higher than 100% total yields are due to the significant alkylation of lignin-derived products using the CuMgAlOx catalyst.14 The THF-soluble lignin residue represents fragments of lignin that have been depolymerized to lower molecular weight.12 Guerbet and esterification reactions catalyzed by CuMgAlOx have been found to be more important in suppressing repolymerization of lignin fragments than alkylation reactions by ethanol.14 GPC analysis reveals the molecular weight distribution of these fragments in THF-soluble LR (Figures Figures33 and S2). Accordingly the weight-averaged molecular weight (Mw) of fragments from the single CuMgAlOx catalyst experiment is the highest (1027 g/mol). The molecular weights of experiments carried out in the presence of Ni/ASA (711 g/mol) and Ni2P/SiO2 (649 g/mol) were significantly lower (Figure Figure33 and Table 3). The elemental composition (C, H, and O content) of these lignin residues are shown in Figure Figure44. Accordingly, the oxygen content of the residues obtained using Ni/ASA (12%), CuMgAlOx and Ni/ASA (11%), and CuMgAlOx and Ni2P/SiO2 (11%) were much lower than that of the parent lignin (32%) (Figure Figure44 and Table 3). The oxygen content of the CuMgAlOx (14%) and CuMgAlOx and Ni/SiO2 (15%) were slightly higher, suggesting that the combined catalytic systems with Ni2P and Ni/ASA were effective in reducing the oxygen content already of the light (THF-soluble) lignin residue (Table 3).
Two-dimensional heteronuclear single quantum coherence (HSQC) NMR is frequently employed for the characterization of the starting lignin and solid lignin residues.42 We proved in this way that significant alkylation reactions occur on the aromatic rings and phenolic groups of the solid fraction during deconstruction with CuMgAlOx as the catalyst.14 These C-alkylated and oxygenated (O-alkylated) groups are marked in the HSQC NMR spectra shown in Figure Figure55. Comparing the different spectra, it is seen that the amount of C-alkylated groups in the light residue is higher for the CuMgAlOx and combined CuMgAlOx and Ni2P/SiO2 catalyzed experiments than for Ni2P/SiO2 catalyzed one, while the intensity of aromatic H groups decreases (Figure Figure55). Such changes were not observed for the residue obtained using the combination of CuMgAlOx and Ni/SiO2 and CuMgAlOx and Ni/ASA catalysts (Figure Figure55). Thus, we infer that alkylation is enhanced slightly by Ni2P, whereas reduced Ni catalysts are not active in alkylation.
Heavy (THF-insoluble) lignin residue and char represent the repolymerized lignin fragments.12 They originate from condensation reactions between lignin fragments and they are typically adsorbed to the solid catalyst.14 The basic sites of CuMgAlOx catalyst help to reduce char formation (Table 3).14 Very few heavy residue and char were formed on the single Ni-based catalysts, except when the Ni2P/SiO2 catalyst was used without added hydrogen (entry 2 in Table 3). This result shows that a hydrogenation functionality can prevent repolymerization reactions during lignin upgrading and it most likely relates to the saturation of reactive carbon–carbon double bonds. Comparatively, more heavy lignin residue and char were obtained when CuMgAlOx was combined with Ni/ASA (entry 6 in Table 3). This most likely correlates with the lower contribution of Guerbet-type reactions (entry 6 in Table S2), which are known to suppress repolymerization and char formation.14 Altogether, the sum of heavy lignin residue and char were similar in the CuMgAlOx and combined CuMgAlOx and Ni(2P)/SiO2 catalyst systems (entries 1, 4, and 8 in Table 3), which means that the degree of repolymerization is also similar in these reactions.
We also varied the ratio of the CuMgAlOx and Ni2P/SiO2 catalysts toward optimum lignin monomer yield. To this end, we kept the total catalyst mass at 1 g and varied the amounts in the following ratios 0:1, 0.25:0.75, 0.5:0.5, 0.75:0.25, and 1:0. Lignin monomers yield, aromatic HYD and HDO degrees, and ethanol product yields are presented in Table 4. In all three cases where two catalysts were combined, there was a clear synergy in terms of the lignin monomer yield. Overall the highest lignin monomer yield was obtained when an equivalent ratio of the two catalysts was used. At the same time, the HDO degree was much higher for the optimum catalysts ratio, presumably indicating that the synergy is at least in part related to removal of oxygen species. This should be related to a reduction of phenolic groups that can be involved in repolymerization reactions. It can also be seen that there is only a slight effect on the ethanol-product yield when the CuMgAlOx content is varied. Finally, we employed the optimum catalyst combination in an experiment where the reaction time was extended to 8 h at 340 °C. Compared to the 4 h experiment, the lignin monomer yield was increased to 80 wt % at nearly similar HDO degree of 78 wt %. At the same time, however, the aromatics hydrogenation degree was higher at 66 wt % and also slightly more ethanol-derived products were formed.
The use of Ni-based HDO cocatalysts for the CuMgAlOx-catalyzed depolymerization of soda lignin in supercritical ethanol was investigated. Ni/ASA and Ni/SiO2 catalysts were prepared by incipient wetness impregnation. The Ni2P/SiO2 catalyst was obtained by impregnation of NiO/SiO2 with diammonium phosphate and reduction at 620 °C. While Ni/SiO2 and Ni2P/SiO2 displayed poor performance in the deconstructing lignin without CuMgAlOx, metallic Ni dispersed on the acidic ASA support gave similar yields as the mixed oxide catalyst. The high acidity of the support resulted in high solvent conversion (diethyl-ether formation). The Ni/SiO2 and Ni2P/SiO2 were found to operate synergetically with CuMgAlOx in soda lignin depolymerization. The lignin monomer yield was strongly increased and at the same time the degree of oxygen removal was much higher. As hydrogen is in situ formed due to the action of CuMgAlOx, these experiments did not require the addition of hydrogen. The best overall performance is achieved by combining CuMgAlOx with Ni2P/SiO2, resulting in 53 wt % lignin monomers yield at a HDO degree of 83 wt % HDO. One aspect of the observed synergy is that in situ hydrogen generation by the CuMgAlOx catalyst will not only activate the supported Ni and Ni-phosphide catalysts but also retain them in a highly active state. The other important aspect is that the Ni-based phases are involved in hydrogenation of reactive intermediates released from lignin by the action of the CuMgAlOx catalyst. Such reactive intermediates contain aldehyde and olefinic groups, which are known to be involved in condensation reactions that will decrease the lignin monomer yield.
This research has been performed within the framework of the CatchBio program under project number 053.70.337.
† T.I.K. and X.H. contributed equally to this work.
The authors declare no competing financial interest.