Review on Catalytic Cleavage of C–C Inter-unit Linkages in Lignin Model Compounds: Towards Lignin Depolymerisation

Catalytic oxidation is one of the most widely used methods for the breaking of the inter-unit linkages in lignin, especially C–C linkages because of the abundance of hydroxyl groups in lignin [54]. Behling et al. has reported an overview on the recent advances in the oxidative depolymerisation of lignin including some oxidative C–C cleavages [55]. Paper and pulping industries use some of the most advanced oxidative routes for the depolymerisation and eventual removal of traces of lignin present in cellulosic materials [50, 55, 56]. Oxidative depolymerisation of lignins typically result in monomeric oxygenates like carbonyl compounds and carboxylic acids. Hanson et al. reported a general correlation between the oxidative breaking of specific linkage and the resultant product for the model compound 7 represented in Fig. 10 [57]. For example, C_α–C_β cleavage results in aromatic aldehydes (or corresponding carboxylic acid), while the C_α–H cleavage yields corresponding ketones, the breaking of C_β–O bond yields ketones and finally the C_Ph–C_α cleavage results in acrolein and quinone derivatives [58]. This correlation is useful to rationally design catalytic systems for the targeted cleavage of specific bonds to get specific compound in high yield.

Crestini et al. reported a catalytic, chlorine-free, oxidative cleavage of inter-unit linkages in an array of monomeric and dimeric, phenolic and non-phenolic lignin model compounds using a homogeneous methyltrioxorhenium(VII) (MeReO₃) (MTO) catalyst and H₂O₂ as the oxidant [59]. To study the C–C bond cleavage, they used different substituted lignin model compounds having β–O–4 units 9 and diphenylmethane units 10 as substrates (Fig. 11). The model compound 9, used in this study, is a highly functionalised version of the simple model compound 1 (Fig. 6). During the catalytic reaction using phenolic model compound 9, (substrate-1 in Fig. 11) > 98% of the substrate was converted to products. The products mixture includes carboxylic acid on the C_α position (4-hydroxy-3-methoxybenzoicacid), carbonyl group on the C_β position (hydroxyl-ketone), 2,6-dimethoxyphenol and muconolactone. Among these products, 4-hydroxy-3-methoxy benzoic acid was formed by the cleavage of C_α–C_β bond; however, the yield of this product was only 16%. Similarly, when substrates 2 and 3 (Fig. 11) were oxidised, evidence for the cleavage of the C_α–C_β bond was also observed [59]. The C_α–C₁ linkage present in 10 is not found in native lignins, however it is formed because of the condensation reactions during the delignification processes, hence it is prevalent in technical lignins [60]. In an effort to break this C_α–C₁ linkage, two versions of model compound 10 was oxidised using MTO and H₂O₂. For substrate 4 (phenolic model compound in Fig. 11) higher proportion of C_α–C₁ cleavage was observed resulting in aromatic carboxylic acids. For the substrate 5 (non-phenolic model compound in Fig. 11) only trace amount of C_α–C₁ cleavage was observed. They have concluded that C_α–C₁ cleavage is easier in phenolic model compounds, compared to non-phenolic compounds.

Encouraged by these results, Crestini et al. depolymerised technical lignins such as hydrolytic sugar cane lignin (SCL), red spruce kraft lignin (RSL) and hardwood organosolvent lignin (OSL) using MTO and H₂O₂ [59]. This catalytic oxidation reaction resulted in a decrease in the content of aliphatic OH groups (43, 14 and 67% reduction in SCL, OSL and RSL, respectively) and resulting in the formation of more soluble lignin fragments and higher yields monomeric carboxylic acids. Because of this additional C–C cleavage for this catalytic system, it is effective for the depolymerisation of complex technical lignins.

Hanson et al. reported the cleavage of C_Ph–C_α, C_α–C_β and β–1 inter-unit linkages during the aerobic oxidation of different lignin model compounds 7 and 8 (Fig. 8) using different vanadium metal complexes [61, 62]. During the aerobic oxidation of derivatives of model compound 8 with β–1 linkage using (HQ)₂V^V(O)(OⁱPr) (HQ = 8-oxyquinolinate) catalyst, they observed substantial C_Ph–C_α and β–1 cleavage [61]. Again, phenolic model compounds resulted in substantially higher C–C cleavage compared to non-phenolic model compounds and the solvents altered the product distribution. In DMSO solvent benzaldehyde and methanol were the major products, while in pyridine solvent the main products obtained were benzoic acid and methyl benzoate (Fig. 12) [61].

More recently, Ma et al. reported the selective oxidative C–C cleavage in model 1 using VO(acac)₂ catalyst with molecular oxygen as the oxidant. They further show the effect of solvents on the selectivity with acetic acid being the most desired solvent for C–C cleavage [63]. Amadio et al. reported the oxidative cleavage of model compound 7 (Fig. 6, phenolic X, Z = OCH₃, Y = OH) using V3 in Fig. 9, where they have found the effect of solvent on the selectivity of C_Ph–C_α cleavage. The yield of the products as a result of the breaking of C_Ph–C_α bond follows the order ethylacetate > 2-methyl THF > pyridine > THF. When non-phenolic version of the model compound 7 (X = H, Z = MeO, Y = EtO) was used, C–C cleavage was not observed at all [58].

TEMPO (2,2,6,6-tetramethyl-1-piperidin-1-yl-oxyl in Fig. 9) is used as it is or in combination with metal catalysts for the breaking of inter-unit linkages in lignin. Sedai and Baker reported an effective combined catalytic system containing CuCl and TEMPO for the oxidation of 1,2-diphenyl-2-methoxyethanol (model 8, W = OCH₃, Fig. 8), having β–1 linkage, using O₂ as the oxidant [64]. After 48 h reaction at 100 °C, they achieved more than 80% of β–1 cleavage. However, under similar condition, when (dipicolinate)V^V(O)(OⁱPr) V1 (Fig. 9) was used as catalyst, the oxidation of secondary alcohol to ketone was followed by β–1 cleavage in a two step process. However when the intermediate ketone was oxidized by V1, > 90% of cleavage was observed. The CuCl + TEMPO catalytic system is more effective and better than the vanadium catalyst in breaking the β–1 bond in one step. The same group reported Cu(OTf)/2,6-lutidine/TEMPO catalyst system for the aerobic oxidation of model compound 8 having β–1 linkage [65]. In comparison to vanadium complexes, generally this Cu catalyst is superior in breaking C–C linkages for non-phenolic models [65, 66]. For phenolic β–1 model compounds, catalytic amounts of TEMPO were not effective in breaking any C–C linkages, however when stoichiometric amounts of TEMPO were used substantial amount of C_Ph–C_α cleavage was observed [65]. When non-phenolic β–1 model compounds were tested, even with catalytic amounts of TEMPO substantial amount of C_α–C_β bond is broken. From these results we can conclude that for effective cleavage of C–C bonds, CuOTf + TEMPO (stoichiometric) system is more suitable for phenolic model compounds, whereas catalytic amount CuOTf + TEMPO is preferred for non-phenolic model compounds. When a 1:1 mixture of non-phenolic β–1 and β–O–4 model compounds (7 and 8) was used for the oxidation reaction using this catalytic system, substantial amount of C–C cleavage was observed and β–1 model compound got converted more readily compared to the β–O–4 model compound [65]. Rahimi et al. used catalytic amount of 4-acetamido TEMPO, without any metal, for the oxidation of β–O–4 model compound 9 using O₂ (Fig. 13). In this reaction, C_α–C_β cleavage has been clearly observed. This oxidation methodology has been extended to the depolymerisation of real lignin (Aspen lignin). Through detailed analysis of the product mixture, they propose C–C inter-unit cleavage [67].

Díaz-Urrutia et al. compared the catalytic activities of a few vanadium complexes for the oxidative depolymerisation of organosolv lignin and studied the mechanism of the catalytic oxidative cleavage using model compounds. Among all the tested catalysts, only bis(8-oxyquinoline) oxovanadium (V3) (Fig. 9) resulted in C–C cleavage under basic condition. However, under their reaction conditions, CuOTf + TEMPO and TEMPO did not result in C–C cleavage [68, 69]. Another interesting method for the cleavage of C_α–C_β linkage was reported by Patil et al. [70, 71]. Using simple model compounds, in the first step they have oxidised the OH group in C_α in β–O–4 model compound to form ketone using TEMPO/O₂ system. In the second step they converted the ketone to ester using Baeyer–Villiger oxidation (i.e. introducing O in between C_α and C_β), which is then hydrolysed in situ to form carboxylic acid, aldehyde and phenol (Fig. 14). Though this indirect method of breaking C_α–C_β is interesting, it will be less applicable for the depolymerisation of pure lignin [70].

Another two step strategy was proposed by Wang et al. for the cleavage of C_α–C_β bond in β–O–4 lignin model compound 1 (Fig. 6). In the first step, the secondary OH group is oxidised to ketone using VOSO₄/TEMPO catalyst and O₂ as oxidant. In the second step, the ketone is converted to monomeric phenols and carboxylic acids through the cleavage of C_α–C_β bond using Cu/1,10-phenanthroline catalyst and O₂ as oxidant [72]. The bond energy of the C_α–C_β bond decreases from 307.7 kJ mol^− 1 for the alcohol to 205.5 kJ mol^− 1 for the ketone, making the ketone an easier substrate for C–C cleavage [72]. More recently, the same group developed Cu(OAc)₂/BF₃·OEt₂ catalyst for the cleavage of C_α–C_β bond in β–O–4 model compound 1 to produce esters and phenols [73]. Napoly et al. reported Fe (TAML) Li (Fe tetraamido macrocyclic complex) catalyst for the oxidative cleavage of C_α–C_β bond in β–O–4 model compound 1 using (diacetoxyiodo) benzene (DAIB) as the oxidant at 25 °C (Fig. 15) [74].

They futher report that by increasing the water content in the reaction mixture from 5% to a 20% the extent of C_α–C_β bond cleavage increased from 45 to 95%. Though the exact role of water in increasing the selectivity of C_α–C_β bond cleavage is not clearly understood. They have extended this methodology for the cleavage of β–1 linkage as well in a lignin model compound similar to 8 [74].

Luo et al. developed a transition-metal free protocol for the selective oxidative C–C cleavage in lignin model compounds with sodium persulfate as the oxidant [75]. They tested this system for the oxidative cleavage of different inter-unit linkages in many model compounds. Relevant to this review, using sodium persulfate, Luo et al. were able to break β–1 linkage in model compound 3 into benzaldehydes in fairly good yields (ca. 60%). However, they could not translate this methodology for the oxidative depolymerisation of real lignins because of their poor solubility. Cobal salen [Co(salen)] complexes have also been used as homogeneous catalysts for the oxidative cleavage of phenolic and non-phenolic phenylcoumaranes (lignin model compound 4), resulting in the cleavage of β–5′ inter-unit linkages to form benzoquinone derivatives, alkylphenyl ketones, benzoic acid derivatives and densely functionalized phenoxyacrylaldehydes. Some quantities of benzofuran (with β-5′intact) has also been found [76]. Biannic and Bozell used Co-Salen complexes for the selective cleavage of the C_Ph–C_α bond cleavage in β–O–4 model compound instead the typically weaker β-aryl ether linkage (C–O linkage) (Fig. 16). This oxidative cleavage reaction was performed at a milder reaction condition compared to other reported examples. This is one of the few examples where C–C bond is broken selectively compared to the C–O bond [77].

Mottweiler et al. reported the catalytic oxidative depolymerisation of organosolv beech and kraft lignins using transition-metal-containing hydrotalcites or combinations of vanadium and copper species using V(acac)₃ and Cu(NO₃)₂·3H₂O as catalysts using O₂ [78]. Significant reduction in the molecular weight was observed due to the effective cleavage of β–O–4 and other inter-unit linkages. The structure of the modified (depolymerised) lignin post reaction could not be fully understood by NMR, however, based on the resinol structure, they have confirmed that the cleavage of β–β inter-unit linkage. This is again one of the very few examples where β–β inter-unit linkage is broken.

In the case of Ru, specially Ruthenimun-triphos-based complexes such as Ru(H)₂(CO)(PPh₃)(xantphos) have been found to be effective in breaking the C–O bond in β–O–4 lignin model compound 7 [79]. However, it was not effective in breaking the C–C bond until vom Stein et al. reported a two-step redox neutral process for the cleavage of C–C linkages in lignin model compound 7 (Fig. 8). This redox neutral process involves a dehydrogenation-initiated retro-aldol reaction for the C_α–C_β bond clevage in different substituted lignin model compounds containing β–O–4 linkage (Fig. 17) [80].

Another catalytic route for the cleavage of C–C interunit linkage in lignin model compounds is the reductive pathway. Reports on the oxidative cleavage of lignin model compounds are dominated by homogeneous catalysts, whereas for the reductive cleavage of C–C bonds (hydrogenolysis) heterogeneous catalysts are more commonly used. Surprisingly, the first report on the hydrogenolysis of Aspen lignin over copper–chromium catalyst under relatively harsh conditions (220 atm of H₂ pressure at 260 °C) dates back to 1938 by Harris et al. [81]. Later on, the hydrogenolysis of technical lignins such as organosolv lignin was inspired by the hydrocracking processes developed in the petroleum refineries [82, 83]. During the lignin hydrocracking process mainly β–O–4 and C_Ph–C_α linkages are broken [82]. Oxidative routes are preferred to produce monomeric compounds used in the synthesis of bulk and fine chemicals since the resultant monomeric products are mostly oxygenates. However, hydrogenolysis is preferred to produce alkanes to be used for fuel applications. Huber and Corma also investigated the catalytic cracking of lignin using biomass-derived feedstocks mixed with petroleum-derived feedstocks and triglyceride-based feedstocks as substrate [84]. Catalytic hydrogenolysis of C–O linkages are known in the literature, however reports on the hydrogenolysis of C–C linkages in lignin model compounds are not common.

The catalytic hydrodeoxygenation (HDO) process is an interesting route to produce biofuels from biomass feedstock by removing oxygen. Jongerius et al. reported the HDO of lignin model compounds (derivatives of models 4, 6 and 7) using CoMo/Al₂O₃ catalyst. During this HDO reaction, they report the breaking of C–C (β–5) linkage along with the breaking of C–O (β–O–4) linkage. However the 5–5′ linkage in the phenolic model 6 could not be broken under this condition (Fig. 18) [85].

The same group used 1% Pt/Al₂O₃ for the aqueous-phase reforming (APR) of technical lignin and found many monomeric compounds such as syringol or guiaicol in the products mixture and proposed the cleavage of 5–5′ linkage (Fig. 19). Though this has not been confirmed by any studies using model compounds. To our knowledge, this is one of the first reports on the cleavage of 5–5′ linkage in the phenolic lignin model compound 6, proved by the formation of methylguaiacol (7%) and guaiacol (12%) [86].

Yamaguchi et al. reported the cleavage of C–O and C–C bonds in β–O–4 model compound 7 and β–1 model compound 3 using many supported metal catalysts (Pd/C, Pt/C, Rh/C, or Ru/C) in supercritical water without any H₂. They report that the β–1 linkage is broken in supercritical water at 673 K using Rh/C catalyst, forming monomers such as toluene and benzene, which are valuable base chemicals [53].

Microwave irradiation is an alternative route that has been reported to be promising for lignin depolymerisation [6]. This process was studied on simple lignin models and on organosolv lignin using different supported metal (Ni, Pd, Pt and Ru) nanoparticles on mesoporous Al-SBA-15. A 10 wt% nickel supported on mesoporous Al-SBA-15 was found to be the most active catalyst yielding 30% of bio-oil from organosolv lignin after half an hour of reaction using formic acid as hydrogen donor [87, 88]. Zhu et al. utilised the microwave assisted cleavage of inter-unit linkages in lignin model compounds using ferric sulphate Fe₂(SO₄)₃ and HZSM-5 as catalysts to produce aldehydes, secondary alcohol or ketone compounds [89]. Then, they extended this methodology for the catalytic depolymeristion of organosolv lignin to produce aromatic monomers (vanillin, syringaldehyde, methyl vanillate, and methyl syringate). They further report that these products are formed by the selective cleavage of C_α–C_β bond (promoted by the Fe³⁺) over the C–O bond (Fig. 20) [90]. Compared to conventional heating, microwave heating resulted in a 48.9% increase in the cleavage of C_α–C_β bonds for Sigma lignin and this increase is more substantial for organosolv lignin (62.3%) [90]. This microwave assisted cleavage of inter-unit linkages in technical lignin appears to be an interesting approach, however, more work has to be done to further understand why microwave heating is more effective in breaking C–C bonds in lignin.

Biocatalytic valorisation of low molecular weight lignin has been studied extensively and Abdelaziz et al. has reviewed this subject recently [91]. Tien and Kirk reported the first enzymatic degradation of spruce and birch lignins using phanerochaete chrysosporium and H₂O₂. In this they have reported the cleavage of C–C bonds in lignin model compounds [92]. This report inspired many research groups to work in the area of enzymatic degradation of lignin and lignin model compounds [93]. Through careful mechanistic investigation, Schoemaker et al. proposed a mechanism for this enzyme catalysed oxidative cleavage of C_α–C_β bond in β–1 lignin model compound involving single-electron transfer (SET) between aromatic rings resulting in a cationic radical of the substrate is formed [94, 95]. Inspired by these developments, Shimada et al. developed a tetraphenylporphyrinatoiron (III) chloride complex and tertbutyl hydrogenperoxide system for an unprecedented C–C bond cleavage in a non-phenolic β–1 lignin model dimer [94, 96]. They used 1,2-bis(4-ethoxy-3-methoxyphenyl) propane-1,3-diol, which is a combination of 2 β–O–4 linkages (model 1) with a β–1 linkages (model 3) as a model compound for this study. After these reports, enzyme catalysed oxidative depolymerisation has been widely studied, especially using peroxidases [97] and laccases [76]. An example of it is the use of peroxidase pleurotous ostreaus for the depolymerisation of lignosulfonate in the presence of a H₂O₂. Here they report the production of 2,6-dimethoxy-1,4-benzoquinone, benzoic acid, butyl phthalate, and bis(2-ethylhexyl) phthalate at ambient conditions [98]. These products are formed by the cleavage of many inter-unit linkages including C_α–C_β linkage [88, 98]. Cho et al. combined enzyme catalysis with photocatalysis for a highly regioselective C–C bond cleavage in β–1 and β–O–4 model compounds through cationic radicals generated through SET process [99, 100]. In this report, cation radicals were generated by using SET-sensitized photochemical and Ce(IV) and lignin peroxidase promoted oxidative processes. Photocatalytic methodologies have been used for the successful C–C bond cleavage recently. Mitchell and Moody reported the successful photocalytic cleavage of C–C linkages in a range of 1,2-diols and β–O–4 model compounds under visible light (solar) irradiation using 1,4-hydroquinone with a copper supported on aluminum oxyhydroxide catalyst (Cu/AlO(OH)) with oxygen as the oxidant [101]. In this system, the catalyst function as the electron transfer mediator (ETM) through the Cu(I)/Cu(II) redox couple.