Physiochemical properties of catalysts
To explore the effect of the catalyst's properties on the lignin catalytic hydrogenolysis reaction performance, a series of characterization measurements such as BET, SEM, EDX, and XPS were performed. First, the N2 adsorption/desorption measurements were performed to evaluate the catalysts' textural properties after preparation. The BET surface area, pore volume, and pore size are presented in Table 1. The Ni/C catalyst exhibited a BET surface area of 633.4 m2/g, a pore volume of 0.076 cm3/g, and a pore size of 5.30 nm, while the Co/C catalyst showcased a BET surface area of 584.1 m2/g, a pore volume of 0.109 cm3/g, and a pore size of 6.81 nm. The porous structure of activated carbon facilitates the dispersion of metal sites. In comparison, the Ni-Co/C catalyst demonstrated a slightly reduced BET surface area of 556.6 m2/g, akin to Ni/C and Co/C. A similar trend was observed in pore size (3.53 nm). However, upon loading both metals, pore blockage ensued, leading to a significant decrease in both BET surface area and pore size. In addition, the Ni-Cu/C catalyst has the lowest BET surface area and pore volume of 537.6 m2/g and 0.063 cm3/g, respectively. Furthermore, all catalysts exhibited isotherms of N2 adsorption/desorption that were comparable (Fig. S1). The type-IV isotherms demonstrated that all catalysts were mesoporous.
Table 1 Surface properties of the prepared catalysts.
Secondly, SEM and EDX measurements were conducted to reveal the distribution of the loaded metals (Co and Ni) in the Ni-Co/C catalyst (Fig. 1). The SEM and EDX elemental mapping images of Co and Ni revealed that both Co (white dots) and Ni (pink dots) are present on the surface of the activated carbon support and are uniformly dispersed within the tested region. Similarly, Fig. 2 indicated the even dispersion of Cu and Ni elements on the surface of the support. According to reported literature, the interactions between Ni and Co enhanced the metal dispersion and provided more active sites, which typically enhance activity (Zhu et al., 2019a). Fig. S2, Fig. S3, and Fig. S4 represent the SEM and EDX elemental mapping images of the Ni/C, Co/C, and Cu/C catalysts respectively. EDX elemental analysis indicated that the elemental compositions of nickel and cobalt on the surface of Ni-Co/C were 17.76 wt. % and 12.90 wt. %, respectively. Similarly, the elemental compositions of nickel and copper on the surface of Ni-Cu/C were 26.37 wt. % and 6.17 wt. %. The results of EDX elemental analysis revealed that the elements (Ni and Co) coexist on the support.
Fig. 1 SEM images of Ni-Co/C catalyst at a magnification of (a) 3 kX and (b) 30 kX with corresponding EDX mapping of (c) Ni and (d) Co.
Fig. 2 SEM images of Ni-Cu/C catalyst at a magnification of (a) 10 kX and (b) 30 kX with corresponding EDX mapping of (c) Ni and (d) Co.
Furthermore, XPS measurement was performed to further explore the electronic interactions between Ni and Co and to reveal their chemical states. As shown in Fig. 3. The XPS spectra and the associated deconvolutions of Ni 2p on Ni/C and Ni-Co/C catalysts. Fig. 3(a) and (c), reveal that the distinctive peaks between 861.2 and 860.7 eV were assigned to Ni2+, whereas the distinctive peaks between 854.3 and 853.7 eV were assigned to metallic Ni0. Revealing the presence of two different forms of Ni species in the prepared Ni/C and Ni-Co/C catalysts (Zhou et al., 2017). This is due to the possibility of environmental oxygen oxidizing the active Ni metal surface (Li et al., 2020). Ni0 and Ni2+ binding energies exhibited a slight drop which reveals the change in the chemical state of the bimetallic catalysts and the transfer of electrons from Co to Ni species.
Similarly, Fig. 3 (b and d), shows the Co 2p spectra of the Co/C and Ni-Co/C catalysts. The Co3+ and Co2+ are attributable to the peaks at 785.9 eV and 795.7 eV, respectively. Whereas the peak at 802.8 eV corresponded to the Co satellite. Both the Co/C and the Ni-Co/C catalysts exhibit mixed valence states of Co. Conversely, for the Ni-Co/C catalyst, peaks corresponding to Co3+, Co2+, and Co satellite were slightly shifted to higher binding energy positions at 786.4 eV, 796.9 eV, and 804.1 eV. This reveals that the intense interactions between nickel and cobalt result in the transfer of electrons from cobalt to nickel (Hu and Lu, 2007). Additionally, two distinctive peaks in the Co 2p spectra of Co/C and Ni-Co/C catalysts, which appeared at 780.2 eV and 781.7 eV can be attributable to metallic Co0. In summary, the shift observed in the Co 2p and Ni 2p spectrum reveals that the strong electron interactions between nickel and cobalt enhanced the transfer of electrons and the formation of multiple electron-rich Ni sites.
Fig. 3 The deconvolutions of Ni 2p and Co 2p XPS spectra of the (a) Ni/C, (b) Co/C, (c) Ni-Co/C, and (d) Ni-Cu/C catalysts.
Catalytic hydrogenolysis of lignin
The real motives behind the selection of cobalt and copper species as nickel promoters was to improve catalytic activity, stability, and product selectivity. The performance of monometallic catalysts (Cu/C, Co/C, Ni/C) and Ni-based bimetallic catalysts (Ni-Cu/C, Ni-Co/C) on the catalytic hydrogenolysis of lignin was tested in isopropanol solvent at 250 ℃, 4 MPa and 700 rpm, in 4 h. The results of the finding, which includes the percentage of lignin conversion and product distributions, are presented in Table 2 and Fig. S5. Low activity on lignin hydrogenolysis was exhibited by the Cu/C catalyst. It afforded only 16.2 wt. % of the total monomer product at 72.9 % lignin conversion. Also, the introduction of a Co/C catalyst afforded 17.5 wt.% of the total monomer products at a lignin conversion of 78.1 %. Cu/C and Co/C catalysts were not as effective as Ni/C catalysts, with 18.2 wt. % of total monomer product yield and the lignin conversion reaching 78.6 %. Zhu et al. reported catalytic depolymerization of lignin over Pt-Re/TiO2 catalyst under isopropanol/water solvent at 240 in 12 h. Their study achieved 18.7 wt. % of the total yield of monomeric products (Zhu et al., 2019b).
The introduction of Ni-based bimetallic catalysts (Ni-Cu/C) significantly improved the lignin conversion rate and total monomer product yield to 84.7 % and 33.0 wt. % respectively. Because of their hydrogenation abilities, Ni-based catalysts have the potential to increase catalytic performance and the efficiency of reaction by suppressing the formation of char (Korányi et al., 2017). Interestingly, Ni-Co/C catalyst exhibited a superior catalytic performance achieving the highest yield of monomer product of 53.1 wt. % and excellent lignin conversion of 94.2 %. According to Tian et al. (2022), in catalytic hydrogenolysis reaction of lignin model compound, over Ni-Co/SiO2-ZrO2 catalyst at 240 °C and 1 MPa, in 4 h, 48.1 wt. % of the total yield of monomer products and 100 % conversion of guaiacol were achieved. In terms of product distribution, the selectivity of all the products was increased over the Ni-Co/C catalyst. The phenol selectivity was 2.9 wt. %, guaiacol selectivity was 43.8 wt. % and other aromatic products had a selectivity of 5.3 wt. %. The results of this finding revealed that the promotional effect of cobalt and copper in the bimetallic catalysts enhanced catalytic activity, stability, and product selectivity more than their monometallic counterparts.
Table 2. Hydrogenolysis performance of lignin with monometallic and bimetallic catalysts.
Analysis of volatile products
To identify and quantify the volatile products from the lignin hydrogenolysis reaction, GC-MS and GC-FID analyses were performed. The results of the investigation as presented in Table 3 revealed that the monomer products were classified into phenols, guaiacols, other aromatics, and hydrogenated products. Under optimized reaction conditions, phenols afforded 2.9 wt. %, guaiacols afforded 43.8 wt. %, other aromatics afforded 5.3 wt. % and hydrogenated products afforded 1.2 wt. %, respectively. The product distribution reveals that the hydrogenolysis reaction converted all the lignin's main building blocks (H, G, and S units). The results of the quantitative analysis indicated that the maximum yield of 43.8 wt. % was guaiacols, which typically make up the coniferyl (G) structural unit of lignin. Similarly, the yields from sinapyl (S) and p-coumaryl (H) units from the product distribution are relatively low. This revealed that the G-unit is more reactive than the corresponding S and H-units.
Table S1 provides the summary of GC-MS analysis of the product distributions obtained from lignin hydrogenolysis reactions. A total of 20 monomer products were successfully identified by the GC-MS analysis. Their various retention time, chromatogram areas, and chemical structures were also identified.
Table 3 The major components of monomeric product yields
Analysis of nonvolatile products
Raw lignin and the nonvolatile products resulting from lignin depolymerization over both monometallic and Ni-based bimetallic catalysts underwent analysis via FT-IR and GPC characterizations. Initially, the raw lignin exhibited an average molecular weight of 1,717 g/mol with a polydispersity of 2.3, indicating a relatively high molecular weight distribution. However, following depolymerization, GPC measurements demonstrated a substantial decrease in molecular weight to 610 g/mol for the nonvolatile products across various catalysts. This reduction highlights the efficacy of the catalysts in facilitating lignin depolymerization, leading to a significant decrease in molecular weight, thereby offering valuable insights into the catalytic mechanisms involved in the process (Table 4). The results of this investigation revealed that the nonvolatile products of lignin hydrogenolysis in the presence of Cu/C had a relatively high molecular weight (Mw) of 1,693 g/mol and the highest polydispersity (PID) of 2.2. With Ni/C and Co/C catalysts the molecular weight and polydispersity of the obtained nonvolatile products decreased to 929 g/mol, 807 g/mol, and 2.0, 1.8, respectively. Interestingly, the introduction of Ni-based bimetallic catalysts, specifically Ni-Cu/C and Ni-Co/C significantly decreased the molecular weight of the nonvolatile products to 612 g/mol and 610 g/mol respectively.
The results obtained revealed the promotional effect of copper and cobalt in the bimetallic catalyst system enhanced lignin depolymerization reaction. The results were in agreement with the yields of monomer products obtained over monometallic and Ni-based bimetallic catalysts (Table 2). Ni-based bimetallic catalysts also exhibited significantly lower values of polydispersity (1.5), revealing that the nonvolatile products had a concentrated distribution and the excellent hydrogenolysis efficiency of Ni-based bimetallic catalysts.
Table 4 GPC analysis results of Mn, Mw, and PID for raw lignin and nonvolatile products.
In addition, the FT-IR measurements of the nonvolatile products were performed and compared with the FT-IR measurements of raw lignin. As shown in Fig. S6 and Fig. S7, the FT-IR spectra of nonvolatile products obtained from the catalytic hydrogenolysis reaction over Ni-based bimetallic catalysts were comparable. The medium peaks at 3,331 cm-1 and 1,379 cm-1 corresponded to the O–H bending vibration of the phenolic compounds (hydroxyl groups). It was observed that during the hydrogenolysis process, the absorption peaks changed significantly, revealing a transformation of the lignin structure. The O–H peak, for example, increased from 3,285cm-1 to 3,331 cm-1, revealing that the hydrogenolysis process did occur and resulted in an increase in phenolic compounds.
The characteristic absorption peaks at 2,961 cm-1, 2,897 cm-1, and 1,445 cm-1 were attributed to C–H stretching and bending vibration in the alkane or methoxy functional groups. Furthermore, the yield of guaiacols was revealed by an increase in the methoxy functional groups from 2,945 cm-1 to 2,961 cm-1. Similarly, the strong absorption peaks at 1,734 cm−1 and 1,643 cm−1 were attributable to the C=O double bond stretching and bending vibration in the carboxyl group and benzene rings (Shu et al., 2016). The medium and strong peaks at 1,306 cm-1 and 1,128.36 cm-1 were equally attributed to the aromatic rings.
Similarly, the increase in the aromatic ring peak from 1,113 cm-1 to 1,306 cm-1 reveals that the hydrogenation of the aromatic ring did occur, resulting in the formation of hydroxylated intermediate products. The C=C double bond bending vibration in alkenes can be assigned to three strong and medium peaks at 947 cm−1, 808 cm−1, and 675 cm-1, respectively. Also, the appearance of the C=C double bond bending vibration in alkenes after the lignin catalytic hydrogenolysis reaction confirmed that lignin depolymerization exposed functional groups. The Ni-based bimetallic catalysts exhibited better catalytic activity in the lignin hydrogenolysis process through the absorption peak intensities.
Effects of reaction temperature on lignin conversion and yield of monomer products
Reaction parameters such as temperature, pressure, stirring speed, and catalyst dosage have significant effects on the performance of the hydrogenolysis reaction (Shao et al., 2018). Based on this fact, their effects on both the monomer product yield and lignin conversion rate were investigated. The effect of reaction temperature on the yield of monomer products and the lignin conversion rate were investigated. Fig. 4 shows the effect of reaction temperatures between 220 ℃ to 250 ℃ on the yield of monomer products and lignin conversion rate over Ni-Co/C catalyst. At an initial temperature of 220 ℃, 66.8 % of lignin was converted affording 17.5 wt. % monomer products. The lignin conversion rate increased to 78.6 % while the monomer product yield reached 26.3 wt. % as the temperature was raised to 230 ℃. When the reaction temperature was further increased to 240 ℃, the lignin conversion rate also increased to 82.7 % affording 36.9 wt. % monomer products. At an elevated reaction temperature of 250 ℃, 94.2 % of lignin was converted with an impressive yield of 53.1 wt. % of monomer products.
Ambursa et al. (2016) reported 96 % conversion of lignin model compound (dibenzofuran) and 45 wt. % of hydrocarbon selectivity at 250 °C and 10 MPa over Ni-Cu/TiO2 bimetallic catalysts. They observed that excessively high temperatures could cause the repolymerization of the monomer product and the non-volatile liquid products in the reaction to form solid residues (Joffres et al., 2016). Moderate reaction temperature promotes the scission of carbon-oxygen and carbon-carbon bonds by overcoming the energy barrier in the lignin hydrogenolysis reaction (Zhao et al., 2019). Regarding the distribution of products, guaiacol was the highest afforded 43.8 %, and other aromatic products afforded 5.3 wt. %, phenol afforded 2.9 wt. % and the least was the hydrogenated products afforded just 1.1 wt. %. 250 °C was the optimal reaction temperature that achieved the highest monomer product yield of 53.1 wt. % over Ni-Co/C catalyst.
Fig. 4 Effects of reaction temperature on the hydrogenolysis performance of Ni-Co/C catalyst. (Reaction conditions: Lignin = 0.5 g, catalyst = 0.3 g, isopropanol =16 mL, n-decane = 0.5 mL, 4 h, 4 MPa, 700 rpm).
Effects of reaction pressure on lignin conversion and yield of monomer products
Reaction pressure is another important parameter that has a significant effect on the performance of the hydrogenolysis reaction. Fig. 5 shows the results of the investigation conducted on lignin hydrogenolysis reactions over Ni-Co/C catalyst at varying reaction pressures between 2 MPa to 5 MPa. At an initial reaction pressure of 2 MPa, only 70.3 % of lignin was converted, affording 18.4 wt. % monomer products. When the reaction pressure reached 3 MPa, the rate of lignin conversion increased to 77.6 % affording 29.9 wt. % monomer products. Zhai and his group reported the yield of monomer products of 39.5 wt. % over Ni-Fe/C at a reaction pressure and temperature of 2 MPa and 225 ℃ in 6 h where propylsyringol and propylguaiacol were the major products (Zhai et al., 2017).
Both the lignin conversion and the monomer product yield significantly increased to 94.2 % and 53.1 wt. % respectively when the reaction pressure was increased to 4 MPa. Conversely, as the reaction pressure increased further to 5 MPa, both the lignin conversion rate and the yield of monomer products significantly decreased to 86.1 % and 38.8 wt. %, respectively. The yield of guaiacol and phenol monomers decreased significantly at high reaction pressure, most likely as a result of the formation of more volatile and saturated products (Wanmolee et al., 2018). The results of this investigation revealed that the reaction pressure of 4 MPa is the optimal pressure. And the promotion of catalytic hydrogenolysis of lignin at high pressure of 5 MPa is negligible.
Fig. 5 Effects of reaction pressure on the hydrogenolysis performance of Ni-Co/C catalyst. (Reaction conditions: Lignin = 0.5 g, catalyst = 0.3 g, isopropanol =16 mL, n-decane = 0.5 mL, 4 h, 250 ℃, 700 rpm).
Effects of stirring speed on lignin conversion and yield of monomer products
The rate of lignin conversion and yield of monomer products are greatly influenced by stirring speed in the lignin hydrogenolysis reaction. In our previous study, stirring speed promotes the transfer of mass and heat, selectivity of product, and hydrogenolysis reaction kinetics (Ewuzie et al., 2023). It enhances the catalyst-lignin interactions within the reactor and enables effective reactions and uniform product compositions. To maximize the rate of reaction, the gradient of reactant concentration can be maintained under an optimal stirring speed. To run the reactions under kinetic control, the stirring speed was carefully varied from 500 rpm to 800 rpm. Fig. 6 shows the effects of stirring speed on the lignin conversion and yield of monomer products. When the reaction was run at 500 rpm, a relatively low monomer product yield of 19.0 wt. % was achieved at a 71.3 % lignin conversion rate. This would suggest that the reaction was likely limited by mass transfer processes, probably as a result of insufficient mixing and dispersion of the reactants.
However, when the stirring speed was increased to 600 rpm, there was a sharp increase in the yield of the monomer product reaching 32.2 wt. %, while the lignin conversion rate increased to 80.5 %. Interestingly, the lignin conversion rate significantly improved reaching 94.2 %, and affording 53.1 wt. % monomer products as the stirring speed was increased to 700 rpm. This significant improvement suggests that the increased stirring speed provided better contact between the active components and the lignin substrate, enabling the active components to perform more efficiently thereby increasing the lignin conversion rate and monomer product yield. Surprisingly, when the stirring speed was further increased to 800 rpm, both the lignin conversion rate and monomer product yield significantly dropped to 88.4 % and 44.6 wt. %, respectively. These findings revealed that beyond 700 rpm, further increase in stirring speed hurts lignin conversion rate and monomer product yield.
Fig. 6 Effects of stirring speed on the hydrogenolysis performance of Ni-Co/C catalyst. (Reaction conditions: Lignin = 0.5 g, catalyst = 0.3 g, isopropanol =16 mL, n-decane = 0.5 mL, 4 h, 250 ℃, and 4 MPa).
Effects of Ni-Co/C catalyst dosage on the lignin hydrogenolysis reaction
Catalyst dosage can also influence the rate of reaction, lignin conversion rate, monomer product yield, and in commercial applications, the process's economics. An in-depth investigation on the influence of Ni-Co/C catalyst dosage was conducted, and results are displayed in Fig. 7. It can be observed that the monomer product yield and lignin conversion rate exhibited an increasing trend with decreasing catalyst-to-lignin ratios. Under optimal reaction conditions, 0.05 g, 0.1 g, 0.2 g, and 0.3 g of catalyst were used, resulting in lignin-to-catalyst ratios of 10, 5, 2.5, and 1.7. When the hydrogenolysis experiment was conducted with 0.05 g of Ni-Co/C catalyst only 68.8 % of lignin was converted affording 16.5 wt. % monomer products. As for the 0.1 g experiment, the lignin conversion rate and monomer product yield gradually increased to 76.5 % and 27.9 %, respectively.
As the catalyst dosage was further increased to 0.2 g, the lignin conversion rate and monomer product yield equally increased reaching 84.7 % and 45.1 wt. %. Moreover, the standard run with 0.3 g of Ni-Co/C catalyst achieved the highest lignin conversion of 94.2 % and the maximum monomer product yield of 53.1 wt. %. Because of the improved lignin depolymerization that occurs when there are more catalytic sites present, the conversion of lignin and monomer product yield increases progressively with higher catalyst dosage (Zhao et al., 2019).
Fig. 7 Effects of Ni-Co/C catalyst dosage on the lignin hydrogenolysis reaction. (Reaction conditions: Lignin = 0.5 g, isopropanol = 16 mL, n-decane = 0.5 mL, 250 ℃, 4 h, 4 MPa and 700 rpm).
Kinetic study results
To further confirm the improved catalytic performance and promotional effects of Ni-based bimetallic catalysts on lignin hydrogenolysis. A kinetic study on the rate of conversion of lignin with time over Ni-based bimetallic and monometallic catalysts was conducted. Fig. 8 presents the summary of TOF and Ea for the rate of conversion of lignin over Ni-Co/C and Ni/C catalysts. The rate of conversion of lignin exhibited a linear increase from 29.8 % to 94.2 % over the Ni-Co/C catalyst and from 5.3 % to 78.6 % over the Ni/C catalyst between 40 min to 240 min (Fig. 8(a)). Under the same reaction conditions, the TOF over Ni/C catalyst was determined to be 4.4 h−1. This value was calculated from the starting point of the linear kinetic plot. And that of the Ni-Co/C catalyst was calculated to be 36.1 h-1, which is about eight times higher than the value with catalyst Ni/C.
In addition, the kinetic study on lignin catalytic hydrogenolysis over Ni/C and Ni-Co/C catalysts at different temperatures was also conducted. As seen in Fig. 8(b) (Arrhenius plot), apparent activation energy (Ea) was calculated from the slope of the fitted linear plot [In(K) vs 1/T] and the Ea for lignin depolymerization over Ni/C was found to be 50.0 kJ/mol. Lui and co-workers reported that an Ea value of 48.0 kJ/mol was calculated from the kinetic study on a lignin model compound over Pd/C at 240 °C and 3.4 MPa in 4 h (Liu et al., 2017). In this study, the calculated Ea value over the Ni-Co/C catalyst was 36.1 kJ/mol. Which is much lower compared to the Ea value of the Ni/C catalyst. The results of the kinetic study coupled with higher TOF and lower Ea of Ni-based bimetallic catalysts have further revealed the promotional effect of Co and Ni catalysts. Therefore, the addition of Co as a nickel promoter significantly enhanced the catalytic performance in the lignin hydrogenolysis reaction process.
Fig. 8 Kinetic study on catalytic hydrogenolysis of lignin over Ni/C and Ni-Co/C catalysts. (Reaction conditions: Lignin = 0.5 g, catalyst = 0.3 g, isopropanol, = 16 mL, n-decane = 0.5 mL, 240 oC, 40 min, 700 rpm, and 4 MPa N2).
Possible reaction mechanism for catalytic depolymerization of lignin
Based on the excellent results achieved, the possible reaction mechanism for catalytic depolymerization of lignin over Ni-based bimetallic catalyst (Ni-Co/C) under optimum reaction conditions is shown in Fig. 9. Initially, lignin was chemically adsorbed on the surface of the catalyst and interacted with catalytic active sites. As a result, the aromatic ring was hydrogenated and formed hydroxylated intermediate products (phenol, 2-methoxyphenol, and 4-vinylpyridine) under isopropanol solvent. The hydroxylated intermediate products serve as the main building blocks for the catalytic reactions involving the breaking of bonds and fragmentation to produce phenol, guaiacols, and aromatic monomers. The interaction between the hydroxylated intermediate products and the catalytic active components (Ni and Co) facilitated the cleavage of β-O-4, β-5, and C-C ether bond linkages in the lignin structure. According to the report, H-radicals can be transferred to catalytic active sites such as Ni, Fe, Co, and Cu (Zhu et al., 2019a).
Isopropanol as a hydrogen donor solvent, supplied all the required hydrogen for the bond cleavages and hydrogenation processes. In our previous study, Smaller aromatic fragments were formed due to the depolymerization of C-C, β-O-4, and β-5 ether bond linkages in the aromatic rings (Ewuzie et al., 2023). Furthermore, benzene, phenyl alcohol, and guaiiacyl aldehyde (aromatic fragments) reacted with active H-radical (hydrogenolysis reaction) producing phenols, guaiacols, and aromatic monomers. Hydrogen is activated on cobalt sites to produce active H-radicals, due to cobalt's ability to activate hydrogen (Lin et al., 2021). The two main products are phenol and aromatic monomers; hence it has been proposed that the most crucial step in producing the targeted products is the scission of the aromatic ring in lignin. A detailed chemical reaction route that breaks the C-O and C-C bonds in lignin to produce 2.9 wt. % phenols, 43.8 wt. % guaiacols, and 5.3 wt. % aromatic monomers is shown in Table 2. Because of the strong chemical interactions between the metal's d-states and aromatic ring, lignin was chemisorbed onto nickel's active sites. Then, the active H-radicals spill onto Ni sites and attack the C-C and C-O bonds in lignin. After which, the C-C and C-O bonds in lignin were broken, resulting in a 94.2 % lignin conversion.
Fig. 9 Possible reaction mechanism for the catalytic depolymerization of lignin.