Mild and controlled lignin methylation with trimethylphosphate: towards a precise control of lignins functionality

The methylation of lignin is a key reaction for many different purposes, such as increasing thermal stability and controlling lignin functionality for polymer applications, or increasing the yield of aromatic hydrocarbons during lignin pyrolysis. Methylation most often requires the use of toxic reagents, such as dimethyl sulfate. We developed an alternative protocol based on trimethyl phosphate as safer and milder methylating reagent. The reaction proceeds without solvent in only one hour to ensure full and selective methylation of phenolic hydroxyl groups. This specific protocol was successfully applied to several lignins from different botanical origins (solftwood, hardwood, annual plants) and fractionation processes (Kraft, soda, organosolv). This approach further allows to precisely control the methylation in order to prepare lignins with tunable contents in phenolic hydroxyl groups for a large range of potential applications.


Introduction
Lignins are renewable polyphenols abundantly found in all vascular plants, which constitute the main feedstock of biobased aromatic structures. They are now more and more seen as a potential source of aromatic building blocks for chemistry and polymer science. New processes have been developed in the past decades to extract lignins from Kraft 1,2 or sulfur-free processes. Lignocellulosic biorefineries have the potential to bring high amounts of lignins to the market with a variety of architectures depending on the resources and fractionation processes.
The use of lignins in polymer or material science mainly requires preliminary chemical modifications. 3,4 Methylation is probably one of the oldest chemical reactions that was employed by lignin chemists. It was already described in the early stages of the study of lignin chemical structure by Brauns in 1939. 5 Methylation served then as a chemical tool to characterize the structure of lignin, mainly by comparing samples before and after methylation, for instance to quantify phenolic hydroxyl (Ph-OH) and carboxylic (COOH) groups. 6,7 It was also part of the process of the characterization of lignins by the degradative permanganate oxidation method. 8,9 Since then, these characterization techniques have been replaced by more modern techniques, such as NMR spectroscopy, but methylation reactions still found utility for analytical purposes, for instance to increase the solubility in solvents for size-exclusion chromatography (SEC) 10 or gas-chromatography coupled with mass spectrometry (GC-MS). 11 Following the evolution towards the use of lignins in materials and polymer science, methylation reactions later found a renewed interest. Methylated lignins were shown to form miscible blends with aliphatic polyesters or polyethers, [12][13][14] leading to improved mechanical properties. Recently, it was shown that methylation could enhance the compatibility of lignin with natural rubber, making it a potential alternative to carbon black. 15 The methylation of lignins has also been shown to increase their thermal stability by limiting condensation crosslinking reactions, thus potentially facilitating their thermomechanical processing for incorporation in polymer materials. 16 In the context of depletion of some fractions from fossil oils, lignin appears as a potential source of aromatic hydrocarbons such as benzene, toluene and xylene (BTX). However, during the pyrolysis of lignin catalyzed by zeolites, the phenolic OH groups can cause the deactivation of the catalyst, and contribute to increase the coke formation by condensation reactions. 17 The methylation of phenolic OH groups can prevent these phenomena and lead to significant increase in the yields of BTX and polycyclic aromatic hydrocarbons. 18 Similarly, methylation can increase the yield of aromatics obtained during reductive depolymerization of lignin, 19 and increase the storage stability of bio-oils obtained by solvent liquefaction of lignin. 20 Historically, the first protocol describing the methylation of lignin's phenolic OH groups used diazomethane as methylating agent. Diazomethane is a highly toxic and explosive gas, which requires extreme caution during handling. It has progressively been replaced by other methylating agents, the most common being dimethyl sulfate (DMS), even though it is also highly toxic and carcinogenic. 21 The importance of the methylation reaction has thus led researchers to find safer and greener protocols. Recently, the use of dimethyl carbonate (DMC) appeared as a seducing approach. 22 However, DMC methylation requires long reaction times in pressurized reactors with DMSO as solvent. The aliphatic OH groups of lignin are also masked, and the sulfur content of the lignin increases as a result of a side reaction with DMSO, 20 which can be a serious drawback for some applications.
Trialkyl phosphates have been recognized as potential alkylating agents for phenols since the 1930s. 23 They have commonly been used to esterify COOH groups and to methylate aromatic amines or N-heterocyclic compounds, 24 but only scarcely for the methylation of phenols. Nelson described in a 1984 patent the methylation of various phenolic compounds, such as vanillin and syiringaldehyde, with trimethyl phosphate (TMP). 25 Some naphtols and phenols were also reported to be successfully methylated by TMP using BF3 as catalyst under microwave activation. 26 More recently, TMP was used to methylate aliphatic polyols in melt-phase reactions with iron triflate as catalyst. 27

Scheme 1. Methylation of Kraft lignin with trimethyl phosphate (TMP).
In this paper, we describe for the first time the use of TMP to methylate various lignins in a precisely controlled fashion, leaving aliphatic OH (Al-OH) groups largely unaffected (Scheme 1). We first determined the best conditions to obtain fully methylated lignin in agreement with several green chemistry principles, i.e. in a solvent-free protocol, under ambient pressure and for a short reaction time. The applicability of the process to several industrial sulfur-based (Kraft) and sulfur-free lignins (soda and organosolv) was then assessed. A protocol was finally established to obtain lignins with tunable content in phenolic OH (Ph-OH) groups by controlled partial methylations.

Materials
Softwood Kraft lignin (SW-and soda lignin from mixed grasses (Wheat Straw and Sarkanda grass, WS-SL) were obtained from Mead Westvaco (Indulin AT) and Green Value SA (Protobind 1000), respectively. Hardwood organosolv lignins from beech (HW-OL1) and birch (HW-OL2) were produced at Fraunhofer CBP pilot plant (Leuna, Germany). They were extracted by the acetone-based FABIOLA™ process, [28][29][30] and recovered by precipitation of the lignin after dilution with water. Some characteristics of the different lignins are provided in Table 1.
Trimethylphosphate (TMP, 99%) was purchased from Acros Organics and used as received.
where mi and mf are the initial and final lignin masses, [OH + COOH] is the sum of the content in Ph-OH and COOH groups (in mol g -1 ) and 14 g mol -1 corresponds to the increase in molar mass caused by the grafting of the methyl group.
For kinetic studies, the amount of lignin was increased to 0.5 g. Aliquots of the reaction mixture (about 300 µL) were then taken at given reaction times and treated as previously described.

Acetylation
Acetylation of lignins was performed according to a protocol previously described. 32  Fourier transform infrared (FTIR) spectra were recorded on a Nicolet iS10 spectrometer (Thermo Scientific) in attenuated total reflectance (ATR) mode. 32 scans were collected between 500 and 4000 cm -1 at 4 cm -1 resolution.
Dynamic scanning calorimetry (DSC) was performed on a TA Q200 calorimeter. The samples were first maintained at 105 °C for 15 min to erase the thermal history, then cooled to 0 °C at 10 °C min -1 and heated to 200 °C at 10 °C min -1 . The glass transition temperature (Tg) was taken as the midpoint of the change in slope during the heating run.
Tetrahydrofuran (THF, HPLC grade, Fisher Scientific) was used as the eluent at a flow rate of 0.6 mL min -1 . Detection was performed with an Acquity refractive index (RI) detector and an Acquity tunable UV (TUV) detector operating at 280 nm. The samples were dissolved in THF at 5 mg mL −1 and filtered through 0.2 μm PTFE syringe filters prior to injection. When required to ensure full solubility in THF, the samples were first acetylated as described above. The average molar masses and the dispersity were calculated from a calibration with polystyrene standards.

Determination of optimum reaction conditions on SW-KL
Lignin chemistry is complicated by their low solubility in common organic solvents. To solubilize lignin samples, polar aprotic solvents such as DMSO or DMF are usually needed. In order to avoid such solvents, we have recently developed several reactions with cyclic carbonates, which can act as both solvent and reagent to modify lignins. 31,32,34 In these conditions, it is possible to quantitatively derivatize lignins without the need for any organic solvents. Following these results, we applied a similar strategy to methylate lignins with TMP as mild methylating agent. The reaction conditions were first tested on SW-KL.
Using 10 eq. TMP with respect to the acidic groups in lignins (i.e. the sum of Ph-OH and COOH groups), it is possible to obtain a homogeneous solution without the need of an additional solvent, confirming that TMP is a powerful polar aprotic solvent. 24 K2CO3 was chosen as catalyst, and preliminary trials were performed at 80 °C. The corresponding results are presented in Table 2.
Using 0.1 eq. K2CO3, the conversion of Ph-OH groups was limited to 22 -23% and increasing the reaction time didn't lead to any amelioration. Increasing the catalyst content to 0.5 eq. allowed to reach up to 82% conversion, whereas almost 90% conversion was obtained using a stoichiometric amount of K2CO3. Using 1 eq. K2CO3 as catalyst, methylation reactions were then performed at different temperatures (80, 100 and 120 °C), and the evolution of the conversion with the reaction time was followed ( Figure   1). At 80 °C, the maximum conversion achieved was 88%. Quantitative conversions could be achieved at 100 °C in 6 h and at 120 °C in only 1 h. Fully methylated SW-KL were soluble in THF and could thus be directly analyzed by SEC. Figure 2a shows Prolonged reaction times at 120 °C however caused gradual changes in the SEC distributions ( Figure   2b). There is a shift toward higher molar mass when the reaction time is increased (ESI, Figure S1).
This pleads for a strict control of the reaction time at 120 °C to avoid the occurrence of side reactions.

Methylation of lignins from diverse botanical origin and extraction processes
The conditions determined previously were then applied to the methylation of various lignins, originating from different fractionation processes (Kraft, soda and organosolv) and botanical origins (softwood, hardwood and annual plants, Table 1). They present significant differences in terms of  All lignins were successfully methylated under the predetermined conditions (   Table 3). 31 P spectra show the complete disappearance of signals from Ph-OH and COOH groups, which are converted into methoxyls and methyl esters, respectively ( Figure 3). It seems that there is no difference in reactivity between the various types of phenol groups present in the different lignin samples (Scheme 2). The content in Al-OH groups is also reduced, but the decrease was limited to about 22% for SW-KL, and 32 to 36% for WS-SL and HW-OLs, respectively. Part of this decrease is directly caused by the increase in mass related to the grafting of methyl groups onto the phenols. If we consider a sample of initial mass mi containing nAl-OH moles of Al-OH groups, its initial content in Al-OH groups as measured by 31 P NMR is given by Equation (2):

Scheme 2. Chemical structures of the different types of phenolic groups in lignins
After the methylation of all phenol and carboxylic groups, its final mass mf has increased (Equation (3)): with nPh-OH and nCOOH the initial contents in phenols and carboxylic acids.
The final content in Al-OH measured by 31 P NMR is then obtained with Equation (4): It thus appears than 7 to 9 % of the decrease in Al-OH groups is in fact caused by the increase in mass, showing that only about 12% of Al-OH groups in SW-KL and 23 to 29 % in WS-SL and HW-OLs, respectively, have been converted into methoxyls. The proposed methylation protocol thus leaves Al-OH groups largely unaffected, and produces purely aliphatic polyols from lignins, which are recovered in high yields (   Table 3). FTIR spectra also confirm the successful methylation ( Figure 4). The large OH band, initially centered between 3220 and 3360 cm -1 depending on the lignin, is shifted toward higher wavenumbers (3430 -3500 cm -1 ) as a result of the masking of Ph-OH groups. Indeed, signals in the 3220 -3240 cm -1 region has been assigned to intermolecular hydrogen bonds in phenols. 36 Similarly, the modification of polyphenols by oxypropylation, which leaves only Al-OH groups, has been shown to induce similar shift of the OH absorption band in FTIR. 34 The C=O stretch band in COOH, initially located for SW-KL and WS-SL below 1700 cm -1 , is also shifted to higher wavenumbers (1720 -1730 cm -1 ), confirming the conversion of COOH into the corresponding methyl esters. 1 H NMR spectra (available in the ESI, Figure S2), confirm the disappearance of phenolic protons, but are unable to give further information, since the protons of the new methoxyl groups overlap with those initially present in the lignin. SEC of HW-OL1 and HW-OL2 after methylation gives results close to those of their acetylated counterparts (ESI, Figure S3 and Table S1), indicating the absence of significant side reactions. The solubility of WS-SL in THF was enhanced by the methylation, but it was however not fully soluble and could thus not be analyzed by SEC. The reproducibility of the process was evaluated by performing 3 distinct methylations of SW-KL with similar reaction conditions. As seen in Table 3, the reproducibility is excellent regarding the conversion calculated by 31 P NMR. 1 H NMR and FTIR spectra as well as SEC traces also confirm the optimal reproducibility (data available in ESI, Figures S4 to S6).

Insights into the reaction mechanism
As TMP possesses 3 methyl groups, a single molecule would potentially be able to methylate up to 3 phenol groups. However, in their seminal work on alkyl phosphates as methylating agents, Noller and Dutton used 1 eq. of alkyl phosphate per phenol and recovered only moderate yields of phenol ethers. 23 Saidi and Rajabi used a slight excess of TMP (1.1 eq.) to methylate various phenols with BF3 as catalyst under microwave activation. 26 In a recent work on the methylation of aliphatic polyols with TMP, Duclos et al. also reported the need for 1 eq. of alkyl phosphate. 27 They proposed a reaction mechanism under acid catalysis that leads to the formation of a dialkyl phosphate, and further showed that the dialkyl phosphate is unable to alkylate aliphatic alcohols.
To gain insights into the reaction mechanism, aliquots of the crude reaction mixture of the methylation of SW-KL with TMP were taken at regular intervals and 1 H and 31 P NMR spectra were recorded ( Figure 5). 1   A tentative reaction mechanism is shown on Scheme 3, by analogy with the mechanism of the basecatalyzed methylation of phenols with dimethyl sulfate. It involves first the deprotonation of Ph-OH groups by K2CO3, followed by the nucleophilic attack of the phenolate anions on the methyl group of TMP. Since K2CO3 has a relatively low pKa (10.3), it is unable to deprotonate Al-OH groups, which can thus not be activated and remain largely unaffected under these reaction conditions.
Al-OH groups can potentially be methylated by TMP, but it involves the use of strong Brønsted (sulfuric or trifluoromethanesulfonic acid) or Lewis acids (transition metals or lanthanide triflates) as catalyst. 27 Their methylation under neutral conditions has also been reported, but only at high temperature (T > 160 °C) and for extended reaction times. 38 In our case, the use of relatively low temperature (120 °C) and short reaction time (1 h) prevent the reaction of most of the Al-OH groups, leading to a good selectivity towards the methylation of Ph-OH groups.

Scheme 3. Proposed reaction mechanism for the base-catalyzed methylation of lignin Ph-OH group with TMP: deprotonation of Ph-OH groups (a) and nucleophilic attack of the phenolate anion onto the methyl group of TMP (b)
1 H and 31 P spectra of the isolated methylated SW-KL do not show signals for TMP or DMP ( Figure 5), indicating that the workup procedure allows the recovery of a pure product. No other signals are detected on the 31 P NMR spectrum of methylated SW-KL, confirming that no side reaction, such as the formation of phosphate esters, occurs on lignin. 1 H and 31 P NMR spectra taken at regular intervals allow the quantification of DMP formed during the reaction (spectra are available in the ESI, Figure S8). Figure 6 shows the evolution of the DMP content in the reaction mixture, as measured by 1 H NMR. The experiment was replicated three times and the results are presented as average values with the corresponding standard deviations. At the end of the reaction, the content in DMP is around 15 mol%, which is slightly higher than expected considering that all Ph-OH and COOH groups and part of the Al-OH have reacted. A blank reaction was run in the absence of lignin, and did not show any degradation of TMP within 1 h at 120 °C in the presence of K2CO3 ( Figure S10).   Detailed study of 31 P NMR spectra show that the reactivity of the different kinds of Ph-OH groups (Htype, G-type and 5-substituted, Scheme 2) is very similar (data available in the ESI, Figure S11). It confirms that this protocol is applicable to all kind of lignins for the preparation of lignin samples with tailored content in Ph-OH groups. The possibility to control the Ph-OH groups content is also a key to control the thermal properties of lignins. The Tg depends on the Ph-OH groups content, as depicted on Figure 8. Methylation reduces the hydrogen bonding ability and increases the free volume, leading to a gradual reduction in Tg.

Comparison of the method with other methylation procedures
Methylation of lignin is a very common reaction. Its importance to modify lignin in a material science perspective has emerged more recently, leading researchers to focus much more on the process (Table 4). It has been particularly studied in the past few years by Argyropoulos' group. 22,39 They first compared methylations with dimethyl sulfate (DMS) and methyl iodide (MeI), two common reagents for the methylation of phenols. 39 The methylation with DMS affords good results: complete and selective methylation of phenol groups is achieved within 2.5 h at room temperature in water.
However, continuous addition of 0.7 M NaOH was mandatory to maintain the pH above the pKa of the phenols during the reaction. The main drawback of this process is the very high toxicity and the carcinogenicity of DMS (group 2A carcinogen according to IARC classification). 21 The replacement of DMS by MeI is not completely satisfactory. Only 61% of phenol groups of a softwood Kraft lignin were methylated despite an excess of MeI (3 eq), and the selectivity was quite poor, since 38% of the Al-OH groups were also affected. 39 Furthermore, the reaction was performed in DMF, and MeI also has acute toxicity and is suspected to be carcinogenic (group 3 in the IARC classification). 40 The replacement of these toxic chemicals by dimethyl carbonate (DMC) appears then as a valuable option. The ability of DMC to methylate phenols has been largely studied in organic chemistry, 41,42 but it was applied to lignin only recently. Sen et al. 22 showed that softwood Kraft lignin could be fully methylated using 10 eq DMC in DMSO. However, the reaction requires an excess of inorganic base as catalyst (2 eq NaOH) and has to be performed in a pressurized reactor at high temperature (150 °C) for long reaction times (15 h). In addition, the methylation with DMC lacks selectivity, since 75% of the Al-OH groups also react, either via carboxymethylation with DMC or via a side reaction with DMSO. 22 The latter leads to an increase in the sulfur content of the lignin, 20 which can be a strong issue for several applications.
The use of TMP as methylating agent solves several issues encountered with the other protocols for lignin methylation (Table 4). TMP presents much less acute toxicity than DMS and MeI. 43 It was also reported to have considerably less mutagenic activity than DMS. 25 In accordance with green chemistry principles, the reaction is performed without organic solvents, and only water is used for the workup and purification. Full methylations can be achieved with a wide variety of lignin substrates in only 1 h at 120 °C under atmospheric pressure, making it the quickest and simplest methylation procedure described for lignins, and considerably improving the energy efficiency as compared to DMC methylations. Furthermore, its selectivity is better than with DMC since only 12% of the Al-OH groups of a softwood KL are converted into methoxyls.  The final conversion of the Ph-OH groups can be easily tuned by lowering the catalyst content. This way, it is possible to reduce precisely the content in Ph-OH groups of lignin samples, leading to an unprecedented control of their functionality. It is a key for their application in high value-added applications in polymer science.
The developed methylation procedure solves several issues encountered with the other common protocols. TMP is less toxic than dimethyl sulfate or methyl iodide, and the reaction is fast, does not require any organic solvent for the reaction or workup, uses only cheap and benign catalyst, and has a good selectivity towards Ph-OH groups.
Further works will now be focused on the scale up of the reaction as well as on the recovery of the excess of TMP used during the reaction, and in the use of the modified lignins to develop new macromolecular architectures.