Making a feint to the east and attacking in the west.

------Huainanzi·Bingluexun

The world relies on nonrenewable fossil resources, which provide various fuels, chemicals and materials. Biomass is another storage form of carbon and solar energy that has attracted great interest, mainly because of its potential as a sustainable source of chemicals and fuels[1, 2]. Lignin is an attractive biomass feedstock. As shown in Fig. 1, three phenylpropane units assemble into lignin with a complex three-dimensional structure and different linkages in plants [3]. Lignin is regarded as an alternative resource to oil for preparing aromatics. Although different lignin has different linkages distributions, β-O-4 linkage is the most abundant in both softwood and hardwood lignin. Therefore, selective cleaving of β-O-4 linkage has become a hot topic in the past several years [1, 2].

In recent years, studies have been directed towards the depolymerization of native lignin residing in wood. These reactions are performed at high temperature in hydrocarbon solvents. The product contains lignin fragments, aromatics and cyclohexane derivatives. In 2013, we reported for the first time that methanol was a suitable solvent for converting native birch lignin over a nickel supported on activated carbon (Ni/AC) catalyst via a fragmentation-hydrogenolysis process [4]. Our further work showed that Ni catalysts could also catalyze the hydrogenolysis of lignin sulfonate with a high content of sulfur to aromatics in ethanediol [5]. Since then, more studies have shown that an alcohol solvent was a preferred choice for native lignin conversion. As well as the outstanding performance of Ni/AC catalyst prepared by carbothermic reduction in an alcohol solvent [4, 5], we have also shown that the unique structure of Ni/AC has an obvious effect on the conversion route of lignin depolymerization [6]. However, the native lignin conversion process is overwhelmingly complex, which obstructs further works to promote the catalytic ability of known catalysts.

Different from the approach of the direct depolymerization of native lignin, another protocol termed bottom-up focuses on the mechanism of cleaving the typical C-O or C-C bonds in lignin model compounds at the molecular level. Although there is a vast difference between studies on model compounds and studies on native lignin, because of the complex lignin structure, the bottom-up approach has presented many effective catalytic systems and strategies in the native lignin conversion.

Selective cleaving of basic C_α-C_β or C_β-OAr bonds in the β-O-4 lignin model can obtain two phenyl units. To achieve this goal, different methods focusing on new reaction routes have been reported in the past 8 years. During this period, our group has performed some studies on the β-O-4 lignin model conversion. Our strategy is to "sell a dummy" on the lignin structure. (Note: As the Huainanzi·Bingluexun quote says, "Sell a dummy" means making a feint on the east and attacking in the west.) Our strategies mainly begin with adjacent functional group modification (AFGM, Fig. 2), which can directly decrease the bond dissociation enthalpy (BDE) of targeted bonds or introduce new "substrate sites" for further conversion. These two ideas combining with appropriate catalytic systems can make the conversion of β-O-4 lignin model efficient and selective.

As shown in Fig. 2, the first oxidation of C_α-OH to C_α=O (β-O-4 alcohol→ I-1) can reduce about 90 kJ/mol BDE of the C_β-OAr bond [7]. Thus, the C_β-OAr bond connected with a C_α=O can be cleaved easily with metal powders in an acid solution via an uncatalyzed process [8]. The reductive species could be the H⁺ and e^- generated from the reaction of metal and acid. To cleave the C_β-OAr bond in a catalytic process, we used TiO₂ with an abundance of Ti³⁺ on the surface to efficiently catalyze the photocleavage of the C_β-OAr bond in β-O-4 lignin ketones models with H⁺ and e^- under 365 nm irradiation [9], during which sodium acetate promotes the generation of H⁺ from the consumption of ethanol (Fig. 3). In addition, the Pd/ZnIn₂S₄ can catalyze the photocatalytic oxidation of C_α-OH in β-O-4 alcohols to C_α=O with 455 nm light.

To achieve the first transformation of C_α-OH to C_α=O and the consequent hydrogenolysis or reduction cleavage of C_β-OAr bond, besides the first selective oxidation [8, 10], the transfer-hydrogenation process could be an atom economical method [11-14]. We have shown that Pd/C catalyzes the transformation of β-O-4 lignin model to phenol and ketone via a dehydrogenation-hydrogenation process, and performed density functional theory calculations to investigate the cleavage mechanism of the C_β-OAr ether bond in 2-phenoxy-1-phenylethanol over the Pd(111) surface [15]. The favorable reaction pathway (Fig. 4, in red) proceeds as follows: the lignol 2a is first dehydrogenated to generate its corresponding ketone 2e; the ketone continues to dehydrogenate on the C_β by a first equilibrated keto-enol tautomerization to its enol form 2f and then -OH dehydrogenation; the C_β-OAr ether bond cleavage happens afterward, leading to surface intermediates (1j and 1d) followed by hydrogenation to yield acetophenone and phenol.

Because there are other ether linkages (α-O-4, 4-O-5) present without an adjacent hydroxyl group to provide an active hydrogen, converting lignin via oxidation-hydrogenation strategy is still a feasible method. However, the uncontrolled hydrogenation system can catalyze the hydrogenation of aromatic rings, and the ketone can hydrogenate back to alcohol before the C_β-OAr cleavage. Thus, the hydrogenation system should prefer to catalyze the cleavage of C_β-OAr rather than hydrogenate the C_α=O bond or aromatic rings. We have shown that NiMo sulfide is an appropriate catalyst for this purpose (Fig. 5) [16]. Because the hydrogenative edge Mo sites of MoS₂ are in a linear arrangement (Fig. 5(D)) and the Mo sites are the active sites to cleave C_β-OAr bonds, there are few Mo sites available as side reaction centers for the hydrogenation of aromatic rings, which leads to a high yield of aromatic rings. Furthermore, if the edge Mo atoms are partly replaced by the less active Ni²⁺, the potential side reaction sites of C_α=O hydrogenation can be reduced, and the yield of aromatic products generated from the cleavage of C_β-OAr bond increases. Besides the transformation of lignin model compounds, the yield of phenolic monomers from birch wood is up to 32% using this two-step strategy.

For the transformation of β-O-4 lignin model, different catalytic routes contain different intermediaries, which can cause variations in the C_β-OAr ether bond cleavage. The dominating factor for the C_β-OAr bond cleavage should contain the target bond BDE of the original substrate and one of the activated intermediates [17]. In an ideal route for cleaving the β-O-4 linkage, an activated intermediate is generated after a series of activation steps with low activation energy, whose target bond has a lower energy and easily breaks in the subsequent transformations. Analyzing from the substrate molecule to an intermediate molecule, the first dehydrogenation or oxidation of β-O-4 alcohols (C_α-OH) to β-O-4 ketones (C_α=O) has provided many efficient strategies in the β-O-4 linkage cleavage. Because the conversion of a lignin linkage model always contains fragments (radical, cation or anion), analyzing from the substrate molecule to a potential intermediate fragment may present more efficient strategies.

Combined with density functional theory calculation, we have shown that the formation of β-O-4 radical after losing a hydroxyl group at C_α can greatly decrease the C_β-OAr BDE from 247.9 to 45.3 kJ/mol, facilitating its cleavage (Fig. 2). In the hydrogenation of lignin β-O-4 model, we proposed an efficient dehydroxylation-hydrogenation strategy for the C_β-OPh bond cleavage over a NiMo sulfide catalyst (Fig. 6) [17], which is not via C_β-OPh ether bond direct cleavage nor C_α-OH dehydrogenation to C_α=O before the subsequent transformation. The β-O-4-A loses a hydroxyl group (C_α-OH) at weak and medium strong acid sites of the catalyst. The generated PhCH^δ+CH₂OPh transforms to PhCH·CH₂OPh by obtaining an electron from the catalyst redox cycle. Because the possible intermediate PhCH·CH₂OPh has a much lower C_β-OPh bond BDE than β-O-4-A, the C_β-OAr bond breaks easily and generates styrene, ethylbenzene, phenol and various ether products under a hydrogenation condition with an alcohol solvent. Toste et al. [18] reported a vanadium-catalyzed non-oxidative mechanism, and Goldman et al. [19] reported an iridium-catalyzed dehydroaryloxylation mechanism for the C_β-OAr cleavage, in which the C_α radical or C_α-M species was the key intermediate. However, this effect of C_α radical or C_α-M species generation on the BDE change in C_β-OAr bonds drew little attention in subsequent studies.

As well as the widely-used methods that focus on reducing the BDE of the target bonds to cleave the lignin linkage, establishing a feasible and controllable reaction route is also a valuable premise to achieve this goal. Modifying the substrate during the pretreatment or reaction process is a common approach to establish the desired reaction route. In some situations, these pre-modifications can cause the increase of the targeted bond BDE, but these methods can introduce new substrate sites on the substrates for further conversion.

As shown in Fig. 2, the first oxidation of C_α-OH to C_α=O (β-O-4 alcohol → Ⅱ-1) can cause the increase of C_α-C_β BDE by 30 kJ/mol, but the H of the new generated C_β-H can be easily abstracted, and the C_β can work as the substrate site to react with active oxygen species or other cleavage reagent, causing the cleavage of C_α-C_β bond. In our recent work, we showed that Cu(OAc)₂/1, 10-phenanthroline[20] and Cu(OAc)₂/BF₃·OEt₂[21] were effective catalysts in cleaving C_α-C_β bonds in β-O-4 ketones, yielding acids, esters and phenols. In-depth studies showed that C_β-H bond activation was the rate determining step, and the substrate site C_β reacted with active oxygen species, generating an C_β-O-Cu₂ intermediate (Fig. 7(A)) or an oxygen-centered radical (Fig. 7(B)) and causing the cleavage of the C_α-C_β bond.

The first oxidation of C_α-OH to C_α=O (Ⅱ-1, Fig. 2) can introduce an active C_α site, which can react with some nucleophilic cleavage reagents, such as H₂O₂, RCOOOH or NH₂OH, and the Baeyer-Villiger oxidation [10] or potential Bechman rearrangement can insert an O or N atom between C_α=O and C_β or C₁ of the aromatic rings. Combined with a subsequent hydrolysis process, the C_α-C_β or Ar-C_α in the original β-O-4 substrate can be cleaved. The acid catalyzed dehydration at -C_αHOHC_βH-to C_α=C_β can also enhance the connection between C_α and C_β, but the C_β-OAr bond in the C_α=C_β-OAr structure (Ⅱ-2, Fig. 2) can be easily cleaved via an acidolysis process, during which the H⁺ species attacks the C_α to generate a C_αH-C_β=O⁺-Ar and weakens the connection between C_β and OAr [22-24]. Given that the oxidation of C_α-OH to C_α=O can reduce the BDE of C_β-OAr and the simultaneous dehydration at -C_βHC_γH-OH can produce a potential C_γ=C_β-OAr for further acidolysis as discussed above, the oxidation-dehydration-acidolysis strategy is more promising than direct acidolysis. Stahl et al. [25] reported that the simple HCOOH/HCOONa system could efficiently catalyze the cleavage of C_β-OAr in oxidized β-O-4 linkages via a redox-neutral process, and the key intermediate (Ⅱ-3, Fig. 2) contains a C_α=O and C_γ=C_β-OAr structure. Furthermore, the only transformation of C_γHOH can promote the cleavage of C_α-C_β via a potential retro-aldol process with an intermediate containing a -C_γHO (Ⅱ-4, Fig. 2) [26].

Although various works have been reported, there are more possibilities to further explore the cleavage strategies for the β-O-4 linkage and other lignin linkages. This perspective briefly highlights the recent progress on this topic. Referring to and summarizing these works, we have revealed the "sell a dummy" strategy for the efficient cleavage of β-O-4 linkage, which focuses on the first modification of adjacent functional group and not the direct cleavage of targeted C_α-C_β or C_β-OAr bond. Two concepts, including reducing the target bond BDE and introducing a new substrate site, can provide new guidance for designing the possible reaction route, which can accelerate catalyst screening or mechanism study of lignin and lignin model conversion. Furthermore, the pre-modification of adjacent functional group can also give some guidance in restraining the lignin condensation [16, 27], which can further promote the conversion of native lignin.