Synthesis of lignin model compound containing a β-O-4 linkage

1 Introduction

Current efforts toward developing sustainable resources for energy sector and chemical industry have highlighted the potential of non-edible lignocellulose biomass (LCB) as a renewable source of both fuels and chemicals [1], [2], [3]. Out of the three major components of LCB, i.e. cellulose, hemicellulose, and lignin, extensive efforts have been devoted to catalytic transformation of the former two constituents into fuels and chemicals [4], [5], [6], [7], [8], [9]. However, only limited studies have been dedicated to the conversion of lignin to value-added products despite the fact that it contains almost 30% of the organic carbon on earth and a potential feedstock for renewable aromatic chemicals [10], [11], [12]. Lignin is, hitherto, treated as a waste product in the pulp and paper industry, where it is burnt for heat supply [13], [14].

Lignin, an amorphous biopolymer, contains monomers including p-coumaryl, coniferyl, and sinapyl alcohols linked by various C–O (e.g. β-O-4, α-O-4 and 4-O-5) and C–C (e.g. β-5 and 5-5) bonds as shown in Figure 1. The β-O-4 linkage is the predominant substructure in lignin and accounts for 50%–60% of all its C–O linkages [10], [11], [12]. Selective cleavage of these linkages, especially the β-O-4 linkage, is a key step for the valorisation of lignin to aromatic compounds [15], [16], [17]. Toward this end, an efficient synthesis and large-scale availability of dilignol model compounds containing a β-O-4 linkage is crucial for the development of robust catalytic methods for the transformation of lignin to useful chemicals. While continuing our efforts to utilize synthetic tools for investigating biological problems [18], [19], [20], [21], we are now extending our endeavors towards designing greener and sustainable energy sources and the development of efficient catalysts.

Fig. 1:

Representative structure of lignin biopolymer showing various types of linkages: Blue circles: C–O linkages, and green circles: C–C linkages. The structure of β-O-4 lignin model compound 1 is shown in the box.

Various methods for the synthesis of dilignol model compounds with β-O-4 linkage have been reported [22], [23], [24], [25]. However, most of those methods are either low yielding, laborious, or require special equipment. In this article we report a convenient synthesis of monomeric β-O-4 model compound, 1-(4-hydroxy-3-methoxyphenyl)-2-phenoxypropane-1,3-diol (1), from commercially available starting material vanillin (2). The method is adaptable to synthesize oligomeric lignin model compounds containing β-O-4 linkages.

2 Results and discussion

The initial step involves protection of the phenolic OH of 2 to the benzylated (Bn) derivative 3, which was accomplished by treating 2 with K₂CO₃ as a base in DMF [26]. The generated phenoxide ion was subsequently reacted with benzyl bromide to afford benzyl-protected vanillin 3 in 95% yield. At this point, we also sought to use a benzoyl (Bz) group as an alternative phenol protection. Main incentive behind this replacement was to accomplish the late-stage deprotection and ester reduction by LiAlH₄ in a single step, thus reducing the number of steps from 6 to 5. However, our efforts toward the synthesis of the Bz-protected vanillin derivative were unsuccessful because of poor yield and instability of the protective group. Therefore, we used Bn protection for our synthetic route.

Elaboration of the aliphatic chain was accomplished via a Wittig reaction on substrate 3a by using (2-methoxy-2-oxoethyl)triphenylphosphonium bromide (5) in presence of sodium ethoxide. The phosphonium bromide 5 was synthesized by reacting a solution of 2-bromoacetate (4) with triphenylphosphine in ethyl acetate [26]. A white precipitate, thus formed, was deprotonated by sodium ethoxide in ethanol and the resulting ylide was reacted in situ with the protected aldehyde 3 to obtain the trans cinnamate derivative 6a as the major (70%) product (Scheme 1) after purification by column chromatography. The ¹H NMR spectrum of the compound indicated two doublets, one signal at downfield chemical shift (δ=7.60 ppm) and the other one at δ=6.56 ppm with a coupling constant of J = 16 Hz, indicating a trans configuration of the olefinic protons. A minor product 6b (20%) having a cis configuration (6b) was also formed. This was confirmed by the presence of two olefinic protons resonating at δ=6.91 and 5.87 ppm, respectively, with a mutual coupling constant of J = 12.9 Hz.

Scheme 1:

Synthesis of racemic dilignol (±)-1 from vanillin (2) [(i) K₂CO₃, BnBr or BzCl, DMF, r.t., 5 h; (ii) PPh₃, EtOAc, r.t., 17 h; (iii) 5, NaOC₂H₅, EtOH, r.t., to 55°C, 1.5 h; (iv) NBS, diphenylthiourea, acetonitrile:water (4:1), 0°C, 2 h; (v) KH, phenol, DMF, 0°C for 2 h and then to r.t., 8 h; (vi) LiAlH₄, diethyl ether, r.t., 1 h; (vii) H₂/Pd, ethyl acetate, r.t., 5 h].

To introduce the 1,2-dioxygenated functional group of the target dilignol (1), we envisioned 1,2-halohydrin as a suitable precursor. The correct regiochemistry of the hydroxyl group and bromine atom was accomplished by using the reaction conditions reported by Bar [27]. Briefly, the trans cinnamate (6a) was stirred with diphenylthiourea and NBS in a mixture of CH₃CN: H₂O (4 : 1) at 0°C for 2 h. Water was added to quench the reaction, and then the product was extracted with ethyl acetate. The crude residue was finally loaded on a silica gel column and eluted with ~ 50% ethyl acetate in hexane to afford the desired compound 7 as a brownish gummy material in 94.5% yield. The IR spectrum of the compound showed a broad peak at 3497 cm⁻¹, indicating the presence of the OH group. The structure of the bromohydrin 7 was further confirmed with the help of ¹H NMR and MS spectra. The absence of the two trans olefinic protons in the cinnamate derivative 6a and appearances of a doublet of doublet at δ=4.79 ppm (J=5.3 and 9.6 Hz) and a doublet at δ=4.52 ppm (J=9.6 Hz) indicated the formation of bromohydrin 7 as previously reported [27]. The formation of bromohydrin 7 resulted in the generation of two stereogenic centers in the aliphatic chain, and we expected the formation of 7 as a racemic mixture. However, since only the trans cinnamate (6a) was used in this reaction, bromohydrin 7 was obtained in diastereomerically pure form. The natural β-O-4 linkages of lignin polymer are chiral in nature, we envisioned that the racemic dilignol model compound, prepared from racemic 7, would have no stereochemical effect on the outcome of the valorisation of lignin polymer by the metal-based achiral catalysts, currently being developed in our laboratory.

To avoid the potential intramolecular elimination reaction or the formation of epoxide by intramolecular displacement of Br by the –OH functionality, the phenol was reacted with KH in DMF, subsequently mixed with the DMF solution of bromohydrin 7 to afford the phenyl ether 8 in 92% yield. The yield of the reaction was initially very low owing to the loss of the product during work up using dichloromethane, probably because of its polar nature. The IR spectrum of the product showed a broad OH peak at 3390 cm⁻¹. The ¹H NMR spectrum of the compound showed multiple resonances between δ=7.65–6.74 ppm amounting to 8 H, indicating the presence of an additional benzene ring in the structure. Due to the combined anisotropic effects of the benzene ring and the carbonyl group the C–H protons in the alkyl chain are shifted ~ 0.6 ppm upfield from the corresponding resonances of the C–H protons in the starting bromohydrin 7.

The ester functionality of 8 was reduced to the primary alcohol by using LiAlH₄ in diethyl ether for 30 min at room temperature to furnish 9 as a colorless solid in 85% yield. The reaction mixture was diluted with water and the product extracted several times with ethyl acetate. Presence of a broad hydroxyl peak at 3364 cm⁻¹ and absence of the C=O stretching at 1738 cm⁻¹ indicated the reduction of the ester functionality to a primary alcohol. The ¹H NMR of the product showed two doublets of doublet for the methylene protons, resonating at δ=4.03 and 3.88 ppm along with the –CH₂Ph peaks at δ=5.02 ppm confirming the reduction of the ester group.

The final step in the synthetic route was removal of the benzyl protective group. This was accomplished by stirring 9 in the presence of 5% Pd/C and H₂ gas, affording the final product as a white solid in 92% yield. It is interesting to note that the benzylic hydroxy (–OH) group survived under the milder reducing conditions as reported previously [28], [29], [30], [31]. The ¹H NMR spectrum of the product showed the disappearance of the benzylic methylene protons signaling at δ = 5.02 ppm as well as the benzene protons.

The synthetic strategy proposed in this article can be adapted to accomplish oligomeric lignin models containing only the β-O-4 linkages. Since they mimic the polymeric nature of the lignin molecule, the oligomeric models are becoming more and more practicable than the monomeric or dimeric models in designing and discovering catalytic methods for converting lignin to aromatic compounds. Toward this end, methyl cinnamate 6a can serve as the starting point. Thus, removal of the protective benzyl group from 6a by hydrogenolysis at low temperature can be followed by the bromohydrin step, which in presence of a base could afford the oligomeric model compound containing β-O-4 linkages.

3 Conclusions

In conclusion, we are reporting here a convenient synthesis of dilignol model compound containing a β-O-4 linkage in six steps with overall yield of 50% starting from the commercially available vanillin. The model β-O-4 linkage will be used for the development of catalytic reactions yielding a diverse range of aromatic raw material by the valorisation of lignin biopolymer.

4 Experimental section