Lignin separation and tailoring with phenol-assisted deep eutectic solvents for enhancing alkyl phenols yield in pyrolysis

Material and DES synthesis

Corn stalk, collected from Shaanxi, China, was processed by crushing and subsequently screened to obtain a powder with a particle size of 40-60 mesh before use. Corn stalk was composed of 45.11% cellulose, 20.10% hemicelluloses, and 21.26% lignin. The following chemicals were provided by Shanghai Aladdin Biochemical Technology Co., Ltd., China: monoethanolamine (MEA, 99% purity), ChCl (98%), NH₄Cl (98%), tetra-n-butyl ammonium chloride (TBAC, 98%), ethanol (95%), and anhydrous sodium acetate (≥ 99%). Additionally, acetic acid (analytical reagent, AR grade), sulfuric acid (AR grade), and hydrochloric acid (AR grade) were sourced from Modern Oriental (Beijing) Technology Development Co., Ltd., China. Cellulase [Cellic@CTec3, 150 filter paper units (FPU)·mL^-1] was furnished by Novozymes, located in Beijing, China.

For the preparation of DES, three HBAs, including ChCl, NH₄Cl, and TBAC, were mixed with MEA at varying molar ratios. The mixtures were then heated to 85 °C and stirred for 6 h to produce a uniform and clear solvent. Prior to use, the prepared DES was transferred to a desiccator for storage.

DES pretreatment

First, 3.0 g of eucalyptus was blended with 30 mL of DES, and then heated at varying temperatures (100-160 °C) and durations (2-10 h). Once the reaction was complete, the pressure tube was removed from the reactor and allowed to cool to room temperature. The mixture was then gradually introduced into 90 mL of ethanol while stirring at a rate of 1,000 rpm for 0.5 h. Subsequently, the solvent was subjected to centrifugation to separate out the residue solids and the filtrate. The filtrate was gathered and then concentrated with the aid of a rotary evaporator. Subsequently, the concentrated solution was added dropwise to 100 mL of deionized water to obtain the recovered lignin. The recovered lignin was subjected to a triple washing process using deionized water. Based on HBA types, the regenerated lignin fractions were labeled as ChCl-120-6, tetrabutylammonium chloride (TBAC)-120-6, and betaine (Bet)-120-6, among which TBAC is tetrabutylammonium chloride and Bet is betaine. Based on phenol types, the regenerated lignin fractions were named Ph-120-6, Sy-120-6, and Gu-120-6, among which Ph is phenol, Sy is syringol and Gu is guaiacol. Based on reaction temperature at ChCl/FA/Ph DES, the regenerated lignin fractions were labeled as Ph-100-6, Ph-120-6, Ph-140-6, and Ph-160-6. Based on the reaction duration at ChCl/Ph/FA DES, the regenerated lignin fractions were labeled as Ph-120-2, Ph-120-4, Ph-120-6, Ph-120-8, and Ph-120-10. Based on the molar ratio of ChCl/FA/Ph DES, the regenerated lignin fractions were labeled as 1/2/0.05, 1/1.5/0.05, 0.5/1/0.05, and 0.8/1/0.05. For comparison purposes, eucalyptus was subjected to a treatment with pure ChCl/FA DES at 120 °C for 6 h, and the resulting lignin fraction was labeled as ChCl/FA.

Once the lignin had been separated, the remaining liquid was gathered and concentrated using a rotary evaporator to remove water, bringing it back to the initial volume of the DES. Taking into account the FA lost during rotary evaporation, the DES solution that was reclaimed was then brought back to the pH of the fresh DES by adding FA, and it was subsequently employed for another round of pretreatment. The lignin fractions obtained from each cycle were labeled as 1C, 2C, 3C, 4C, and 5C, corresponding to their respective cycle numbers.

Lignin pyrolysis

The lignin samples were analyzed using pyrolysis-gas chromatography/mass spectrometry (Py-GC/MS) techniques with a CDS 6150 Series pyrolyzer coupled to a Perkin Elmer Clarus 560 GC/MS system. For each analysis, a precise amount of lignin (0.1 mg) was loaded into a quartz tube and pyrolyzed at 400 °C for 20 s, achieving a heating rate of 20,000 °C/s. The identification of pyrolysis products was carried out using the NIST17.L mass spectrometry database^[29].

NMR analysis

All nuclear magnetic resonance (NMR) spectra of lignin were recorded on a Bruker AVIII 400 MHz (Bruker, Germany) spectrometer at 25 °C. The liquid was dissolved in 0.2 mL of dimethyl sulfoxide (DMSO)-d6. The peak (2.49 ppm) of DMSO-d6 was used as internal standard. The 1H NMR spectra were acquired according to previous papers^[5,6]. For the 2D heteronuclear single quantum coherence (HSQC) NMR, 30 mg of lignin was dissolved in 0.5 mL of DMSO-d6. The data was processed using standard Bruker Topspin-NMR software and the DMSO-d6 was used as an internal standard (δC/δH 39.5/2.49). For 31P NMR, 20 mg of lignin was dissolved in a solvent mixture composed of anhydrous pyridine and deuterated chloroform, with cyclohexanol as an internal standard, chromium (III) acetylacetonate solution as relaxation reagent and 2-chloro-1, 3, 2-dioxaphospholane as the phosphitylating reagent.

Delignification effectiveness at different DES systems

In this study, eucalyptus was pretreated by using FA-based DES with various phenols under different conditions (HBA types, phenol types, molar ratio, temperature, and time). The delignification rate and lignin extraction yield under DES pretreatment are shown in Supplementary Figure 1. For the DESs containing different HBA types, the delignification rate after the ChCl-based DES pretreatment was about three times that after the TBAC-based DES pretreatment and six times that after the Bet-based DES pretreatment [Supplementary Figure 1A]. Moreover, the ChCl-based DES exhibited the highest lignin extraction yield (61.58%). This result indicated that ChCl was the best HBA in the synthesis of DES for achieving efficient delignification. When various phenols were used as an enhancer in DES, the delignification rate and lignin extraction yield after phenol-enhanced DES pretreatment were greater than those of the eucalyptus pretreated with the pure ChCl/FA DES pretreatment (69.97%). Moreover, the Ph-based DES presented superior delignification rate and lignin extraction yield compared to the Sy- and Gu-based DES [Supplementary Figure 1B]. Within the ChCl/Ph/FA DES system, where the molar ratio of Ph remained constant, an increase in FA content and a decrease in ChCl content corresponded to a higher delignification rate and lignin extraction yield [Supplementary Figure 1C]. Moreover, the highest delignification rate and lignin extraction yield were achieved when the molar ratio of ChCl/Ph/FA DES was 1:0.05:2. The delignification rate elevated from 31.75% to 85.48% with increasing temperature [Supplementary Figure 1D]. The previous research revealed that high temperatures are conducive to decreasing the viscosity of DES, thus promoting mass transfer and improving lignin dissolution in the DES system^[16]. Nevertheless, when the temperature exceeded 120 °C, the delignification rate did not increase considerably. Besides, excessive reaction temperature could destroy the lignin structure and increase the costs of DES pretreatment, which could largely restrict the subsequent lignin application and industrial development. Combined with delignification efficiency and lignin structure, the optimal temperature was 120 °C. Furthermore, the impact of the ChCl/Ph/FA DES treatment duration on the delignification rate was evaluated. The delignification rate initially increased with the extension of reaction time before it began to decline, with the highest rate of delignification observed at 6 h [Supplementary Figure 1E]. Therefore, the optimal pretreatment condition was to use ChCl/Ph/FA DES with a 1:0.05:2 molar ratio at 120 °C for 6 h. In addition, the effect of DES cycle times on the delignification rate was investigated. Notably, the recycled DES preserved its ultra-high efficiency in delignification even after undergoing five cycles [Supplementary Figure 1F]. However, the lignin extraction yield exceeded 100% after the twice recycling pretreatment and the yield was enhanced with the increase of the DES cycles. This may be due to the condensation of lignin oligomers from the previous cycle into larger molecules, which, together with the newly extracted lignin, co-precipitated during this cycle. Since the delignification rate in pretreatment is higher than the lignin extraction yield, this phenomenon can explain why the lignin extraction exceeds 100% after the cyclic pretreatment.

Structure analysis of the regenerated lignin

The delignification rate is primarily related to the ChCl/Ph/FA DES treatment conditions, which could reflect the damage degree of DES to the cell wall of biomass. Notably, the structural features of the DES regenerated lignin can markedly determine the final lignin-based product quality. Considering the prospective use of these lignin fractions in pyrolysis, the functional groups of the lignin had a great effect on the selectivities and contents of pyrolysis products. Herein, the structural changes of AL and the regenerated lignin fractions after pretreatment were observed using Fourier transform infrared spectroscopy (FT-IR) analysis [Supplementary Figure 2 and Figure 1A]. The peak at 3,497 cm^-1 was assigned to the stretching vibrations of hydroxyl groups (OH) including the aliphatic and phenolic OH. This peak exhibited higher intensity in pure ChCl/FA lignin than phenols-based DES lignin [Supplementary Figure 2 and Figure 1A]. The higher OH content in pure ChCl/FA lignin was mostly due to the essential fragmentation of β-O-4 linkages^[9]; instead, the incorporation of a phenol compound in DES can decrease the extent of ether linkages fracture via the explicit arylation with phenols^[30]. Besides, the intensity of 3,497 cm^-1 peak in Ph-120-6 was the lowest compared to Sy-120-6 and Gu-120-6, indicating that Ph-based DES showed better ability in preventing the fracture of β-O-4 linkages [Figure 1A]. The principal distinction between AL and the regenerated lignin samples was noted at 1,722 cm^-1 related to the unconjugated C=O stretching in lignin side chains, such as ketones, ester groups. The peak in the regenerated lignin was attributed to the formation of Hibbert ketone during acidic DES treatment, which can be further confirmed by NMR spectra^[27]. Furthermore, the strong absorption peak at 1,722 cm^-1 detected in the regenerated lignin could imply the occurrence of an acylation reaction between the OH group in the lignin side chain and the carboxyl group in FA during DES pretreatment^[31]. The absorption peaks at 1,595, 1,509, and 1,462 cm^-1, which correspond to the aromatic ring structure of lignin, were observed in all lignin samples, proving that the DES pretreatment did not change the main structure of lignin. The peak corresponding to the vibrations of C-C in the syringyl (S) structure was detected at 1,327 cm^-1, and the peaks at 1,240 and 1,124 cm^-1 represented the C=O and C-H stretching vibration in the guaiacol (G) ring. The peak at 1,034 cm^-1 was assigned to the aromatic C-H in-plane deformation and C-O vibration. The evidence showed that the regenerated lignin after DES pretreatment retained S and G units, which represented the typical hardwood lignin.

Figure 1. Structure properties of lignin fractions. (A) FT-IR curves; (B) molecular weight; (C) content of various OH groups detected by ³¹P NMR; (D) TG curves; (E) DTG curves. FT-IR: Fourier transform infrared spectroscopy; ³¹P NMR: phosphorus nuclear magnetic resonance; TG: thermogravimetric; DTG: derivative thermogravimetric.

During DES pretreatment, variation in the molecular weight of lignin inevitably occurred, resulting from both depolymerization and condensation reactions^[32]. The weight average molecular (M_w) and number average molecular (M_n) of the lignin fragments were detected based on acetylated lignin using gel permeation chromatography (GPC) analysis, and dispersity (Đ_M) is equal to M_w divided by M_n [Supplementary Figure 3 and Figure 1B]. As shown in Figure 1B, the M_w of control AL was about twice that of the pure ChCl/FA DES lignin and phenols-based DES lignin. This phenomenon was primarily ascribed to the cracking of interconnections between lignin subunits, especially involving the substantial breakage of β-O-4 linkages. The M_w of the pure ChCl/FA DES lignin and ChCl/FA/phenols DES lignin did not significantly change, while the Đ_M of the lignin (2.83) derived from the pure ChCl/FA DES was roughly twice that of the ChCl/FA/phenols DES lignin (1.33-1.56) [Figure 1B]. The results illustrate that the ChCl/FA/phenols DES pretreatment promoted the formation of small and uniform molecular weights of lignin, which was more beneficial for obtaining valuable lignin pyrolysis products. For the type of HBA, the M_w and Đ_M of the ChCl-based DES lignin were lower than those of the TBAC- and Bet-based DES lignin, suggesting that ChCl-based DES facilitated the depolymerization of lignin [Supplementary Figure 3A]. In addition, the molar ratio and reaction time showed little effect on the molecular weight of lignin [Supplementary Figure 3B and D]. Furthermore, the M_w and Đ_M of the ChCl/FA/Ph lignin at 120 °C were the lowest compared to those of lignin at 100, 140, and 160 °C [Supplementary Figure 3C]. The result indicated that a low temperature made it difficult to destroy the lignin structure, while a high temperature could promote the condensation of lignin^[33].

The differences in OH groups in AL and the regenerated lignin were analyzed by ³¹P NMR. Figure 1C illustrates that the aliphatic OH content in AL far exceeds that of the regenerated lignin. The substantially reduced aliphatic OH content indicated that the dehydration and acylation reaction could occur in the DES system. Besides, phenol types in DES had a minimal impact on the aliphatic OH content. However, elevated reaction temperatures and extended reaction times led to a considerable decrease in the aliphatic hydroxyl group content. The S-OH content in the Sy-120-6 sample was higher than that of the Ph-120-6 and Gu-120-6 samples. Similarly, a higher uncondensed G-OH content in the Gu-120-6 sample and a higher H-OH content in the Ph-120-6 sample were observed. These results implied the introduction of phenols in the DES system into lignin fractions. Besides, the DES pretreatment resulted in more formation of condensed G-OH, suggesting that a condensation reaction takes place during the DES process. The COOH content did not almost change in all lignin samples. The total phenolic OH content in AL, including S-OH, G-OH, and p-hydroxyphenyl (H)-OH, was significantly lower compared to that of the regenerated lignin. The higher total phenolic OH content in the regenerated lignin could result from the substantial breakage of β-O-4 linkages and the introduction of some phenol compounds in the DES system to the lignin structure. The total phenolic OH content increased with the temperature and then remained stable after 140 °C. Furthermore, prolonging the reaction time resulted in more generation of S-OH, condensed G-OH, H-OH, and total phenolic OH groups, likely ascribed to the severe rupture of β-O-4 linkages.

To delve into structure changes of the regenerated lignin during DES pretreatment, the type and content of the internal linkages in the lignin were examined using 2D HSQC NMR, with the principal lignin substructures depicted in Supplementary Figure 4. The signals were primarily divided into the aliphatic and aromatic regions [Figures 2 and 3]. In the aliphatic regions, the dominating cross-signals including methoxyl groups (OCH₃), β-O-4 (A), β-5 (B), and β-β (C) linkages were identified [Figure 2]^[34]. After DES pretreatment, it was noted that the Aα signal in the ChCl/FA sample disappeared completely; on the contrary, the Aα signal was still present in Ph-120-6, Sy-120-6, and Gu-120-6 samples but the signal intensity was significantly reduced compared to that in AL. The result proved that the introduction of phenol compounds in the DES system could prevent the breakage of β-O-4 linkage, consistent with ³¹P NMR result. Nevertheless, after the pretreatment of ChCl/FA/Ph DES, it was found that the Aα signal could be retained at a low temperature (120 °C) and short time (2 and 6 h), but completely disappeared at a high temperature (140 and 160 °C) and long time (10 h). Moreover, the intensity of signals relating to β-5 and β-β linkages progressively diminished as the reaction temperature and duration increased. The result showed that β-O-4 linkages were susceptible to reaction temperature and time. Additionally, two novel signals A’α (4.63/81.0 ppm) and A’γ (4.32/62.0 ppm) were observed in the regenerated lignin. These signals indicated the formation of the acylated β-O-4 substructure resulting from the acylation reaction between carboxyl group in FA and OH groups in the lignin side chain^[35]. Meanwhile, it can be seen that the regenerated lignin shows the signal of Hibbert’s ketone, suggesting the breakage of ether linkages^[36].

Figure 2. Comparison of aliphatic regions of lignin fractions in 2D HSQC spectra. HSQC: Heteronuclear single quantum coherence.

Figure 3. Comparison of aromatic regions of lignin fractions in 2D HSQC spectra. HSQC: Heteronuclear single quantum coherence.

In the aromatic domain, the signals of the S and G units were presented [Figure 3]. Compared to AL, the G signal in the ChCl/FA sample caused severe condensation, but the signal in the Ph-120-6, Sy-120-6, and Gu-120-6 samples was only slightly changed. The result indicated that the introduction of the phenol compounds in the DES system can protect the lignin structure to some degree^[30]. However, in the ChCl/FA/Ph DES system, when the reaction temperature and time exceeded 120 ^oC and 6 h, respectively, the G units underwent condensation. Furthermore, Moreover, the intensity of signals corresponding to S-type lignin units progressively rose, whereas the signals associated with G units became weak, indicating that the condensation of lignin occurred^[37]. Besides, the formyl signal (8.10/161.7 ppm) was observed in all DES regenerated lignin samples, which was caused by FA in the DES system, consistent with the previous literature^[30].

The relative abundances of the internal linkages and S/G ratio in lignin were determined based on the number per 100 aromatic rings (Ar) by integrating the related signals. Figure 3 illustrates that the β-O-4 linkage content in the regenerated lignin is rather lower than that in AL, indicating the substantial cleavage of β-O-4 linkages in the DES. Moreover, the β-O-4 linkage content was remarkably reduced in the ChCl/FA sample compared to that in the Ph-120-6, Sy-120-6, and Gu-120-6 samples, proving that the introduction of a phenol compound to the DES could decrease the degree of ether linkage breakage. Besides, the β-O-4 linkage content of the Ph-120-6 sample (12.33/100 Ar) was slightly higher than that of the Sy-120-6 (6.80/100 Ar) and Gu-120-6 (7.88/100 Ar) samples. The result indicated that the Ph had a better protective effect on lignin structure compared to the Sy and Gu, which could be attributed to the low steric hindrance of Ph. Additionally, it was seen that the increase in the reaction temperature and time resulted in the decrease in the internal bonds including β–O–4, β–5, and β–β. The result implied that a relatively high temperature and long duration during DES processing usually caused significant damage to the lignin structure. When the pretreatment condition was the same, the highest S/G ratio was detected in Sy-120-6 (1.86), followed by ChCl/FA (1.73), Ph-120-6 (1.67), AL (1.46), and Gu-120-6 (1.25). The highest S/G ratio in Sy-120-6 could be due to the grafting of Sy onto the lignin structure, causing the increase of S unit content. Similarly, the lowest S/G ratio in Gu-120-6 was ascribed to the introduction of Gu onto the lignin structure, thus enhancing the content of G units in lignin. During the ChCl/FA/Ph DES pretreatment, the reaction temperature had a greater effect on the S/G ratio than time.

Thermogravimetric (TG) analysis is a usable method to evaluate thermal stability, which will be conducive to predicting the pyrolysis behavior of lignin that follows^[25]. The thermochemical degradation of lignin as a heterogeneous polymer with various functional groups, occurred across a wide temperature range. The thermal degradation characteristics of AL and the regenerated lignin were investigated through TG and derivative thermogravimetric (DTG) analysis [Figure 1D and Supplementary Figure 5]. The AL showed three main degradation stages, while the decomposition process of the regenerated lignin was mainly divided into four stages. All lignin fractions exhibited the first degradation stage at 30-100 °C, which was ascribed to moisture evaporation. The AL and ChCl/FA samples only exhibited one large decomposition at 100-470 °C, and the maximum weight loss rate temperature of the ChCl/FA sample (356 °C) was higher than that of AL (286 °C). This result indicated that the thermal stability of the ChCl/FA lignin was largely enhanced due to the considerable breakage of unstable ether linkages after pretreatment. Comparatively, the phenols-based DES lignin samples appeared with two primary degradation peaks of 100-300 °C and 300-400 °C [Figure 1D]. The former degradation peak at 100-300 °C likely arose from the fracture of unstable ether bonds (β-O-4 linkages) and small molecular lignin, echoing the result of GPC and 2D HSQC analysis in which the regenerated lignin from the phenols-based DES pretreatment possessed lower molecular weight and a few ether linkages, respectively^[36]. The temperature of the later degradation peak of the phenols-based DES regenerated lignin was close to the maximum weight loss rate temperature of the ChCl/FA lignin. The stage at 300-400 °C resulted from the rupture of C–C bonds between lignin units^[38]. At 400-600 °C, the mass loss continued gradually but weakly compared with the previous stages. This stage was mainly due to the degradation or condensation of the aromatic ring^[39]. Further, the char yield of the regenerated lignin after the phenols-based DES pretreatment (31.28%-35.77%) was greatly decreased, following that of the ChCl/FA sample (40.88%) and AL (50.80%) [Figure 1E]. This indicated that the DES pretreatment, especially in the addition of phenols, was beneficial for suppressing the formation of coke from lignin pyrolysis. When changing the molar ratio, reaction temperature, and time, the thermal degradation behaviors of these lignin fractions were similar to those of the different phenols-based DES lignin fractions [Supplementary Figure 5]. Moreover, the reaction temperature showed a greater impact on the char yield compared to the molar ratio and reaction time.

Impact of DES reuse on lignin structure

In this study, the optimal pretreatment parameters (ChCl/FA/Ph DES with a 1:2:0.05 molar ratio, 120 °C, 6 h) were used for the recycling experiment. Supplementary Figure 1F depicts the delignification yield across five cycles of recycling. The DES after the fourth cycle resulted in 67.99% delignification yield, which was slightly reduced compared to the delignification yield of the fresh DES (76.36%). The reduced delignification efficiency was attributable to the diminished acidity of the recycled DES. In addition, the lignin extraction yield was importantly enhanced with the increase of cycle number, which was attributed to the co-precipitation of the lignin fractions left over from the last cycle and the lignin fractions removed this cycle [Supplementary Figure 1F]. To better elucidate structural changes of the recovered lignin during the recycling treatment, comprehensive characterization techniques were used to investigate these lignin fractions. The functional groups of the recovered lignin after each recycled DES pretreatment were detected by FT-IR and compared with those of the AL and ChCl/FA lignin samples [Figure 4A]. The peaks at 1,595, 1,509, and 1,462 cm^-1, representing aromatic ring stretching vibrations, were retained, indicating that the recycled DES pretreatment did not disintegrate the aromatic skeleton of lignin. The spectrum of the regenerated lignin exhibited a novel peak at 1,720 cm^-1. This was likely ascribed to the acylation reaction between OH groups in lignin side chains and FA and the formation of Hibbert’s ketone from the cleavage of β–O–4 bonds.

Figure 4. Structure characterization of AL and lignin fractions after cycle. (A) FT-IR curves; (B) aliphatic domain of 2D HSQC spectrum; (C) aromatic domain of 2D HSQC spectrum; (D) molecular weight; (E) DTG curves; (F) TG curves. AL: Alkaline lignin; FT-IR: Fourier transform infrared spectroscopy; HSQC: heteronuclear single quantum coherence; DTG: derivative thermogravimetry; TG: thermogravimetry.

The 2D HSQC NMR spectrum in the regenerated lignin after the fifth cycle showed that the β-O-4 linkage signal disappeared completely, but the signal of Hibbert’s ketone was observed in the aliphatic domain [Figure 4B]. This proved that the recycled DES solution facilitated the fracture of β–O–4 bonds, echoing the FT-IR results. In the aromatic domain, the condensed signals were observed in the regenerated lignin after the fifth cycle, suggesting that the condensation of lignin took place within the solution of the recycled DES [Figure 4C]. Notably, the intensity of the condensed signals of the 5C sample was stronger than that of Ph-120-6 sample, due to the condensation of some small molecular lignin fragments carried in the previous cycles. Subsequently, the OH group content in the regenerated lignin after the fifth cycle was investigated by ³¹P NMR and compared with the lignin after the fresh DES pretreatment. The contents of the aliphatic OH and total Ph-OH groups in the regenerated lignin after the fifth cycle were higher than those of the lignin fraction after the fresh DES treatment [Supplementary Figure 6]. This could result from the substantial fracture of ether bonds and the deposition of some solubilized lignin fractions from previous cycles.

The GPC results of the regenerated lignin after different cycles are shown in Figure 4D. The M_w and Đ_M of the regenerated lignin after cycles were lower than those of AL but were higher than those of the 1C sample. The result implied that the recycled DES showed an excellent ability to improve the cleavage of the labile linkages. The higher M_w and Đ_M of the regenerated lignin compared to the 1C sample were likely due to the precipitation of small molecular lignin fractions dissolved in the previous cycles onto the lignin structure. The DTG curves showed that the weight loss rate temperatures of the regenerated lignin after cycles were close to those of the 1C sample [Figure 4E]. The TG curves revealed that the char yield of the 5C sample (35.28%) was slightly higher than that of the 1C sample (31.28%) [Figure 4F]. This suggested that the recycled DES was not conducive to suppressing the formation of coke from lignin pyrolysis.

Pyrolysis behaviors of lignin

To achieve the value-added conversion of the regenerated lignin, a fast pyrolysis experiment was conducted at 400 °C by Py-GC/MS. The chromatograms of AL and regenerated lignin after the DES pretreatment are delineated in Figure 5A, and twelve lignin pyrolysis products were identified. Notably, the pyrolysis curves of all Ph-based DES lignin observed a phenol peak, while no phenol peak was detected in the pyrolysis curves of the AL, ChCl/FA, Sy-120-6, and Gu-120-6 samples. This result indicated the successful introduction of the Ph in the ChCl/FA/Ph DES system onto the lignin structure, aligned with the NMR results. The areas and selectivities of different lignin pyrolysis products are shown in Supplementary Tables 1 and 2, respectively.

Figure 5. Pyrolysis products obtained from AL and DES regenerated lignin. (A) Py-GC/MS chromatograms and (B) selectivity (point plot) and area (bar graph). Py: Pyrolysis; GC/MS: gas chromatography/mass spectrometry; AL: alkali lignin; DES: deep eutectic solvent.

For better comparison, the phenolic products after lignin pyrolysis were divided into three major categories: no side-chain phenols, alkyl phenols, and carbonyl phenols according to their side-chain structures in aromatic compounds. The areas and selectivities of the three kinds of phenolic compounds are presented in Figure 5B. Among these phenolic compounds, the area and selectivity of the carbonyl phenols were the lowest. The AL pyrolysis produced the largest area of the no-side-chain phenols, while the ChCl/FA lignin pyrolysis produced the smallest area. Comparatively, the area of the no-side-chain phenols from the regenerated lignin after the phenol-containing DES pretreatment fell between the above two lignin fractions. This could result from the retention of partial β–O–4 bonds and the grafting of phenols in the DES onto lignin, which was confirmed by NMR spectra. The largest alkyl phenols area does not necessarily correspond to the optimal overall yield. The previous research also revealed that more β-O-4 linkages were conducive to facilitating the pyrolysis of lignin to various phenolic compounds^[40]. Besides, the pyrolysis of the phenols-containing DES lignin fractions generated more alkyl phenols than the no- side-chain phenols except for the 5C lignin. Furthermore, the largest area of the alkyl phenols was observed in the Ph-100-6 lignin, following that in the Ph-120-6 lignin. On the contrary, the selectivity of the alkyl phenols in the Ph-120-6 lignin was the highest among the regenerated lignin after the phenols-containing DES pretreatment. Although the pyrolysis of the Ph-100-6 lignin obtained the largest alkyl phenols area, the delignification efficiency was extremely weak at 100 °C during the ChCl/FA/Ph DES pretreatment, which will tremendously limit the final total yield of alkyl phenols. In short, the pyrolysis of the Ph-120-6 lignin would produce the high-yield alkyl phenols with high selectivity. Furthermore, the recycled DES significantly decreased pyrolysis efficiency compared to the fresh DES, likely attributed to the high condensation of lignin in the recycled DES.