Several monomeric lignin model compounds with different alkyl side chains (i.e., methylguaiacol, MG; ethylguaiacol, EG; propylguaiacol, PG) that are typically produced from reductive depolymerization (RD) were oxidized with CuO in an alkaline condition. After the reaction, the products were analyzed by gas chromatography with flame ionization detection or mass spectrometry (GC-FID and GC-MS) via an improved method as described in the METHODS. The results in Figure 2 suggested that alkaline copper oxide oxidation (Ox^CuO) could even break the alkyl side chains (or C-C bonds) of lignin monomers to give oxidation products (e.g., vanillin, vanillic acid, and acetovanillone) at 200°C for 10 min. In contrast, almost no monomeric oxidation products were observed after the Ox^CuO of MG (<5%) under a conventional condition (170°C, 30 min) (Chen, 1992), while the yields of the monomeric oxidation products for MG, EG and PG were all significantly increased about two times under a severe condition (210°C, 30 min). Prolongation of the reaction time further improved the yields of the monomeric oxidation products for MG from 47.8% (30 min) to 55.6% (40 min) and 63.1% (60 min), respectively. These results indicated that an elevated temperature and a long reaction time facilitated the C-C cleavage during Ox^CuO. Similar monomeric oxidation products were also observed after the Ox^CuO of p-methylphenol (MH) and methylsyringol (MS) that had different aromatic nuclei. In addition, an H-type (MH, 25.3%) model gave oxidation products in much lower yield than G-type (MG, 47.8%) and S-type (MS, 51.7%) models (Figure 2). It was in line with the result of nitrobenzene oxidation (Chan et al., 1995), where electron donating groups (i.e., methoxyl groups) on the aromatic nuclei of lignin monomers enhanced the rate of oxidation. Therefore, the structure of the aromatic nucleus could also affect the reactivity of these lignin monomers to give oxidation products apart from the reaction temperature and time, which could be explained by the different ability of the aromatic nuclei to supply electrons to the benzylic carbon atoms (i.e., syringyl > guaiacyl > 4-hydroxyphenyl) (Shuai and Saha., 2017).

Considering the good performance of Ox^CuO on breaking C-C bonds in lignin monomers, we further examined its ability to oxidatively break the robust interunit C-C bonds in the β-β′ and β-5′ moieties of lignin other than β-O-4′ linkages. Three representative dimeric lignin model compounds present in Figure 1 were selected as substrates for oxidative depolymerization (Ox^CuO and Ox^NB).

As showed in Figure 3, the β-O-4′ dimeric compound gave a higher monomer yield (62.9%) for Ox^CuO at 170°C for 30 min than Ox^NB at 170°C for 2.5 h (53.5%). As the reaction condition was intensified, the monomer yield for Ox^CuO of the β-O-4′ dimeric compound increased first and then decreased, and a theoretical monomer yield was achieved at 200°C for 10 min. The high monomer yield suggested that no condensation reaction of the monomeric products occurred in such Ox^CuO system. The decreased monomers under a severe condition (210°C for 30 min) could be due to the further degradation of the monomer products. Nevertheless, such severe condition facilitated the cleavage of C-C bonds in β-5′ and β-β′ dimeric compounds. The highest monomer yields of 44.8% and 50.2% were achieve at 210°C for 30 min via oxidatively breaking β-5′ and β-β′ bonds, respectively. These results suggested that an elevated temperature (210°C) was necessary for β-5′ and β-β′ cleavage, which was in line with the results of oxidation of alkyl side chains in lignin monomers discussed above.

As discussed above, the Ox^CuO system could not only enable the oxidative depolymerization of lignin, but also could reduce the monomer yields via further oxidative degradation of the oxidation products. To investigate the stability of the oxidation products, we performed the Ox^CuO of both G-type and S-type oxidation products under different conditions. The conversion of the substrates and the selectivity of the oxidation products were presented in Figure 4. For example, vanillin could readily be converted to vanillic acid at a relatively low temperature (170°C), guaiacol was only produced from vanillin at an elevated temperature (>200°C). Although the formed vanillic acid showed good stability at 170°C, substantial conversion of vanillic acid (49.8%) and high guaiacol selectivity (94.4%) could be achieved at 200°C for 10 min (Figure 4A). These results suggested that once vanillin was formed in the Ox^CuO system, it could be readily oxidized to vanillic acid, followed by the formation of guaiacol via further decarboxylation at an elevated temperature (>200°C). Whereas only 13.2% of the acetovanillone was further converted to other monomer products (i.e., guaiacol, vanillin, and vanillic acid) with a total selectivity of 62.9% even under a severe condition (210°C for 30 min), confirming that acetovanillone was relatively stable in the Ox^CuO system. As for S-type monomeric oxidation products, a transformation path similar to G-type monomeric oxidation products was observed. However, compared to G-type monomers, S-type monomers showed higher reactivity (or conversion rate) and were more easily degraded (Figure 4B), which was consistent with the result that S-type alkylphenols had the highest oxidation reactivity (Figure 2). The above results proved that the poor stability of the monomeric oxidation products under a severe oxidation condition was the main reason for the loss of monomer yields during oxidative degradation of lignin (or lignin model compounds). Therefore, the optimal temperatures for the oxidative cleavage of different interunit linkages should be carefully evaluated for the high-yield monomer production and more accurate structural analysis.

To examine the performance of Ox^CuO on lignin depolymerization, we conducted the reactions with Eucalyptus wood particles. By following the traditional procedure (Chen, 1992), a monomer yield of only 54.7% was achieved at a 5 wt% biomass loading of 5 wt% (Figure 5). However, the previous model compound studies have shown that the Ox^CuO enabled the cleavage of β-O-4′, β-β′, and β-5′, a theoretical monomer yield (based on the total content of β-O-4′, β-β′, and β-5′ moieties of hardwood lignin, Figure 1) could be around 70% if all of these linkages were selectively cleaved. Such difference (54.7% vs. 70%) motived us to explore the possible factors that caused the lowered oxidation product yield and to provide a more accurate method for lignin structural analysis.

Since O₂ was known to facilitate both product formation and degradation during alkaline aerobic lignin oxidation (Mathias and Rodrigues., 1995; Wu and Heitz., 1995; Schutyser et al., 2018), we speculated that the dissolved O₂ in the reaction liquid and residual O₂ in the reactor might be responsible for the low monomer yield especially at a low biomass loading. To validate the speculation, we inspected the lignin monomer yields for Ox^CuO of Eucalyptus wood particles with various O₂ inputs. As expected, even a 0.2 MPa O₂ input could significantly reduce the monomer yields from 54.7% to 39.6% at a relatively low biomass loading of 5 wt% (Figure 5), suggesting that the existence of O₂ had negative effect on the Ox^CuO process. While further increasing the O₂ input to 0.5 MPa had a slight effect on the monomer yield (36.9%) due to the limited reaction time. Interestingly, an improved monomer yield to 64.8% was observed when the biomass loading was increased to 10 wt%. It was because the high biomass loading mitigated the negative effect of the side reactions (such as overoxidation) on the monomer production in the reactor where a fixed volume of oxygen was present. Under the condition of a limited amount of O₂ input (<0.2 MPa) and a high biomass loading of 10 wt%, a comparable monomer yield (60.4%) to the condition of no extra O₂ input (64.8%) could be obtained. When the reactor was pressurized with up to 0.5 MPa O₂, the monomer yield was dramatically reduced to below 50%. Therefore, the existence of O₂ was an obstacle towards accurate analysis of lignin structure. In order to obtain high yields of monomers, it was preferable to remove O₂ from the reaction liquid and the reactor via ultrasonication combined with N₂ flushing prior to the oxidation reaction (see METHODS); besides, higher biomass loading could mitigate the effect of O₂ on monomer yields. Monomer yields of 71.9% and 69.1% could be achieved via the improved method of Ox^CuO at a biomass loading of 5 wt% and 10 wt%, respectively, which were close to the theoretical monomer yield (around 70%).

The types and ratios of subunits as well as the amount of interunit linkages in lignin vary with the types and sources of plants (Vanholme et al., 2010; Maeda, 2016; Tolbert et al., 2016). Therefore, three representative lignocelluloses (i.e., Eucalyptus, Masson pine, and corn stover) were selected as feedstocks to study the depolymerization of lignin in hardwood, softwood, and grass, respectively. The oxidative degradation of lignin was conducted in NaOH aqueous solution with CuO or nitrobenzene as an oxidant. Classical reductive depolymerization of lignin was conducted in methanol with Ru/C as a catalyst.

As shown in Figure 6A, reductive depolymerization (Red) of lignin enabled monomer yields of 51.9% for Eucalyptus lignin, 28.9% for Masson pine lignin, 47.6% for corn stover lignin via the cleavage of only labile C-O ether linkages (Bosch et al., 2015; Shuai et al., 2016). In contrast, higher yields of lignin monomers could be obtained via oxidative depolymerization due to the additional cleavage of C-C bonds. The total monomer yield of 71.9% for Ox^CuO of Eucalyptus was much higher than that of 59.8% for Ox^NB and 51.9% for Red, demonstrating that the improved Ox^CuO had better performance (shorter reaction time and higher monomer yields) on lignin depolymerization than conventional Red and Ox^NB. Although the similar products could be obtained from the three lignocelluloses (Figure 6B), the least monomers yield were provided by Masson pine using the same oxidant (CuO or nitrobenzene) (Figure 6A). Furthermore, the contents of condensed structures such as biphenyl structures (5–5′ linkages) in softwood lignin (10%–25%, Figure 1) were higher than those in hardwood lignin (4%–10%). Such 5-5′ bonds were unable to be broke by such oxidation system (Tamai et al., 2015), limiting the release of monomers. Therefore, the improved method of Ox^CuO could increase the lignin monomer yields of Eucalyptus more obviously than Masson pine and corn stover.