Lignin conversion into C₄ dicarboxylic acids by catalytic wet peroxide oxidation using titanium silicalite-1

Highlights

• Titanium silicalite-1 improves the oxidation of lignin towards C₄ dicarboxylic acids.

• Succinic acid yields improved more than four times for three studied lignins.

• Indulin AT had the best results, over Lignol, E. globulus kraft, and alkali lignins.

• Titanium silicalite-1 is a stable and reusable catalyst for lignin oxidation.

Abstract

Lignin valorisation towards added-value products has become a relevant topic to consolidate a future circular bioeconomy. In this context lignin oxidation to C₄ dicarboxylic acids (C₄-DCA) by catalytic wet peroxide oxidation is emerging as a value-added strategy, supported by the extensive use of these building blocks in several industrial fields. In this work, lignins from different sources and processes (Indulin AT, Lignol, alkali and E. globulus kraft lignins) were oxidised using H₂O₂ and titanium silicalite-1 catalyst (TS-1) under different operating conditions (temperature, pH, time, H₂O_2, and TS-1 load). Indulin AT was the lignin leading to the highest succinic acid yield (11.3 wt%), and TS-1 catalyst enhanced its production four times over the non-catalysed reaction. Malic acid was also produced at high yields, especially for Lignol lignin. The other lignins (E. globulus kraft, and alkali lignins) also produced these C₄ acids but at lower yields. The catalyst remained stable at the used experimental conditions, and showed potential to be reused for several cycles without being deactivated. Overall, the catalytic conversion of lignin to C₄-DCA can help to guide the pathway to renewable chemicals production.

Introduction

Lignin is obtained as a by-product in pulp mills, where it is burned in the recovery boiler to produce energy and heat. However, this step is usually considered a bottleneck in pulp production (Ahmad et al., 2020; Mathias, 1993). Consequently, using part of this residual lignin as a raw material for conversion to added-value products would improve the sustainability of the pulping industry without affecting pulp production and the energy balance of the mill (Junghans et al., 2020; Rodrigues et al., 2018; Silva et al., 2009).

The physical and chemical properties of technical lignins change depending on the feedstock source, the pulping process, and the isolation method (Kienberger et al., 2021). Lignin structure is based on three monomers: p-hydroxyphenyl (H), guaiacyl (G), and syringyl (S), which are present in different ratios and linked differently depending on the biomass origin (Kamm et al., 2008; Upton and Kasko, 2016). Softwood lignin usually has a high G:H ratio and no S units. Units are linked by β-O-4 (43–50 %), 5−5’ (10–25 %), β-5 + α-O-4 (9–12 %) linkages, among others. Hardwood is an S:G:H lignin with a higher S content. Units are linked by β-O-4 (50–65 %), 5−5’ (4–10 %), α-O-5 (4–8 %), and α-O-5′ (6–8 %) linkages, among others (Rodrigues Pinto et al., 2011). These structural differences introduce challenges for the depolymerisation and upgrading of the recalcitrant lignin matrix. Cooking methods focused on high-quality cellulose, such as kraft, sulfite, and alkaline processes, use harsh conditions to achieve their objective, causing strong modifications in the lignin structure (Sun et al., 2018). The alkali process cleaves many of the ester and ether bonds between lignin and carbohydrates, and also some of the lignin C-C bonds (Kim et al., 2016). Kraft process produces a highly modified and partially fragmented lignin, with α-aryl ether and β-aryl ether linkages cleavage, recondensation reactions, and sulphur incorporation (Li et al., 2015; Sun et al., 2018). Other milder extraction processes, such as ionic liquid extraction, organosolv, and milled-wood lignin, cause fewer modifications to the lignin structure. Organosolv uses organic solvents, such as primary alcohols, alone or mixed with water at high temperatures and pressures, and produces lignins with structures closer to the native one, with slight modifications like cleavage of β-O-4 linkages, repolymerizations, and formation of new C-C bonds (Sun et al., 2018).

Depolymerisation is an important route to achieve new products from the lignin. Oxidative lignin depolymerisation has several advantages, leading to the incorporation of new functional groups (e.g., alcohol, aldehydes, and carboxylic groups) and production of low-molecular-weight fractions, giving rise to added-value products such as phenolics, dicarboxylic acids, and others. However, it remains challenging to overcome the low selectivity towards target products and circumvent the over-oxidation to gaseous compounds (Ahmad et al., 2020; Liu et al., 2019). C₄ dicarboxylic acids, currently derived from petrochemical sources and carbohydrates fermentation routes, are used in several fields as final products (e.g. food additives) and as raw materials for pharmaceuticals and polymer industries (Zhang et al., 2020). They are also pointed out as the biomass-derived building blocks for the future development of a greener chemistry (Jong et al., 2011; Werpy and Petersen, 2004). Therefore, achieving C₄ dicarboxylic acids from lignin would allow the attainment of added-value products without competing with food and fossil resources.

Non-catalytic lignin oxidative depolymerisation has been previously reported with succinic and maleic acids yields lower than 3% (Abdelaziz et al., 2019; Demesa et al., 2015; Figueirêdo et al., 2019; Hasegawa et al., 2011; Vega-Aguilar et al., 2021). Better results were obtained with catalytic oxidation, especially in the presence of heterogeneous catalysts, such as perovskite-type oxides, chalcopyrite, and heteropoly acids (i.e., phosphotungstic and phosphomolybdic acids) (Ansaloni et al., 2017; Bi et al., 2018; Cronin et al., 2017; Demesa et al., 2017; Ma et al., 2014). The use of Fenton reagent (Fe²⁺/H₂O₂) gave rise up to 8% maleic acid with phenol (Faisal, 2009), although with other lignin model compounds, the yield was lower than 2% (Kang et al., 2019). When lignin was used, no C₄-DCA were obtained (Zeng et al., 2015), suggesting that this oxidation methodology is not an efficient strategy for lignin depolymerisation towards C₄-DCA. Perovskite-type oxides have also been tested due to their oxidative activity and the presence in their structure of transition metals with two different oxidation states. However, catalysts like LaFeO₃, LaMnO_3, and CeFeO₃ showed no improvement for C₄-DCA production comparatively with non-catalysed experiments (Ansaloni et al., 2017). Chalcopyrite (CuFeS₂) has been used with lignin and lignin model compounds with good results. Namely, catechol oxidation yielded 6% malic acid and 8% succinic acid, and diluted-acid corn stove lignin oxidation yielded 7% succinic acid and 1% malic acid (Ma et al., 2014). Chalcopyrite nanoparticles in acidic pH enhanced succinic acid yield up to 12 % when an industrial lignin was oxidised, accompanied by low fumaric and maleic acids yields (Bi et al., 2018), confirming that acid yields are dependent on the used oxidation reaction conditions. Heteropoly acids (i.e., phosphotungstic acid (H₃PW₁₂O₄₀) and phosphomolybdic acid (H₃PMo₁₂O₄₀)) showed low improvement for succinic acid yield, with disadvantages such as higher costs and recovering difficulties (Demesa et al., 2017). Other catalysts based on V and Mo oxides or pyrophosphates (deposited in materials like Al₂O₃, TiO₂, and HZSM-5), used in lignin gas-phase oxidation gave rise to no more than 2% of succinic acid (Lotfi et al., 2016, 2015). In most of these studies, significant amounts of oxalic, formic, and acetic acids were obtained (as degradation products of lignin over-oxidation). Still, the final yield and C₄-DCA type varied according to the used catalyst and applied reaction conditions. In most of these works, succinic acid yield varied between trace levels up to 2−3 wt% (lignin-basis), while maleic and fumaric acids were reported to be obtained at contents less than 1%.

Different oxidants, such as O₂, H₂O₂, O₃, and peroxy acids, have enough oxidative power to break down most ether linkages and some C-C linkages (Ma et al., 2018a). However, a strong oxidant is needed to cause ring-opening reactions to achieve dicarboxylic acids (Sun and Argyropoulos, 1996). Molecular oxygen oxidation occurs only in alkaline conditions and is usually intended for mild oxidation to achieve aldehydes. Its oxidative power is relatively low to origin ring-opening reactions at high yields (Cabral Almada et al., 2020; Li et al., 2015). Peroxide oxidation is more active than O₂ oxidation by eliminating liquid/gas mass transfer barriers and releasing free radicals, allowing faster oxidations in milder and cleaner conditions (Bhargava et al., 2006; Cheng et al., 2017). Hydrogen peroxide oxidation causes the cleavage of β-O-4 linkages, forming non-condensed structures, especially in softwood lignins. In hardwood lignins, demethoxylation of S units to G units happens, which could be repolymerised, increasing the condensed OH structures (Ahmad et al., 2020). These small fragments can further have the aromatic ring cleaved to dicarboxylic acids. Ozone can also be used for lignin oxidation, leading to ring-opening reactions and generating C₄-DCA (Figueirêdo et al., 2019). However, O₃ handling can be complex since it has to be generated in situ (Ahmad et al., 2020). Recently, electrochemical depolymerisation was used for lignin oxidation, releasing O₂ and H₂O₂ in situ; however, this technique leads to small amounts of C₄-DCA (Di Marino et al., 2019).

Among the catalysts enhancing H₂O₂ oxidative power, titanium silicalite-1 catalyst (TS-1) has been used for lignin model compounds peroxide oxidation to C₄-DCA with good results. Namely, it was reported that guaiacol is oxidised mainly to maleic acid, while vanillic acid is primarily converted to succinic acid, an important feedstock for renewable polymer production (Su et al., 2014; Vega-Aguilar et al., 2020). This catalyst has an MFI zeolite structure, with less than 3% TiO₂ and hydrophobic properties, allowing the oxidation of non-polar compounds in aqueous medium, such as phenolic compounds (Clerici, 2015; Gamba et al., 2009). The active species in the TS-1 catalysed peroxide oxidation is the Ti−OOH group, which shows better reactivity than H₂O₂ (Clerici, 2015).

Under the current state of the art, the catalytic peroxide oxidation of lignin to C₄-DCA using TS-1 catalyst is still lacking. Given the reported advantages of this catalyst towards the production of C₄-DCA, namely the previous studies conducted with model compounds, its use to improve the production yield of target acids (e.g., succinic acid) from lignin is a promising strategy towards a circular economy. Therefore, in this work, lignins from different origins (Indulin AT, Lignol, alkali, and E. globulus kraft lignins) were oxidised using H₂O₂ and TS-1 catalyst focusing C₄-DCA production, especially succinic acid. Moreover, the stability of the TS-1 catalyst to the applied reaction conditions, taking as reference five consecutive utilizations, was also be studied.