Skip to main content
NCBI home page
RSC Advances logoLink to RSC Advances
. 2026 Jul 7;16(36):38541–38553. doi: 10.1039/d6ra03462d

Comparative study of the impacts of three common fractionation methods on the physicochemical features of lignin

You Wang a, Kendhl Seabright a, Haoxin Zhu a, Hongyu Zhou b, David P Harper a, Mi Li a,✉
PMCID: PMC13340313  PMID: 42416731

Abstract

The intrinsic structural heterogeneity and complexity of lignin arising from diverse botanical sources and isolation processes limit its efficient utilization. Molecular weight strongly governs lignin's physicochemical properties and valorization potential, making fractionation an essential technique to reduce dispersity and tailor desired structures. Here, we systematically compared three common fractionation methods, solvent exchange (SE), acid precipitation (AP), and membrane ultrafiltration (MU), to tailor Indulin AT kraft lignin into low-, moderate-, and high-molecular-weight fractions. The physicochemical features of fractions were comprehensively characterized, including molecular weight distribution, functionality, surface charge, interunit linkages, morphology, and thermostability. The SE method demonstrated clear advantages in producing well-separated fractions with narrow molecular weight distribution and redistributed hydroxyl groups, thereby achieving superior reactivity for polymer synthesis. The AP method was conducive to large-scale lignin fractionation but exhibited poor molecular separation, as revealed by the closely overlapped molecular weights in the fractions. Moreover, acid-induced condensation likely reduced hydroxyl content yet produced more condensed lignin structures and higher thermal stability. The MU method showed large discrepancies in fractionation behavior. The low-molecular-weight fractions displayed low dispersity, with intermediate thermal properties and hydroxyl content compared with AP and SE fractions. However, its separation effectiveness among moderate- and high-molecular-weight lignin heavily relied on the type of membranes employed. This study establishes direct links between common fractionation techniques and key physicochemical properties of lignin fractions, providing important guidance for tailoring the characteristics and structures of lignin toward effective and desired valorization.

The structural heterogeneity and complexity of lignin arising from diverse botanical sources and isolation processes limit its efficient utilization. This can be well tailored using readily available fractionation methods.graphic file with name d6ra03462d-ga.jpg

1. Introduction

Lignin, a by-product of the pulping and biorefining industry and the most abundant renewable aromatic resource derived from terrestrial vascular plants, has significant potential as a sustainable alternative to fossil-based resources for biofuels, chemicals, and materials.1 Currently, the global pulp and paper industry generates approximately 70 million tons of lignin annually.2 Among the various pulping methods, the kraft process dominates, accounting for nearly 90% of chemical pulping, which produces a substantial amount of kraft lignin as a result.3 However, less than 2% of generated lignin is currently utilized for the synthesis of value-added products, such as dispersants, adhesives, and surfactants, while the vast majority is instead burned as a low-value fuel for electricity and heat generation.4

The underutilization of lignin is further exacerbated by its structural heterogeneity and complexity, originating from both its biosynthetic pathways in plants and isolation processes in industries and laboratories. Lignin is primarily composed of three phenylpropanoid monomers—guaiacyl (G), syringyl (S), and p-hydroxyphenyl (H), varying among botanic sources or the tissues of the plants. These monomers are interconnected through a variety of ether (e.g., β-O-4′, α-O-4′ aryl ether, diphenyl ether), carbon–carbon (e.g., β-5′, β–β′, β-1′, biphenyl structures) bonds, and other inter-unit connections.5 Meanwhile, the lignin extraction process in which the native linkages of cellulose and hemicellulose were majorly cleaved, markedly increases its structural complexity and alters the distribution of key functional groups, such as hydroxyl, methoxy, and carboxyl groups.6 The highly complex structure results in pronounced heterogeneity among lignin macromolecules, leading to considerable variability in the composition and performance of lignin-derived materials and chemicals.7,8

Recent studies highlight that variations in molecular size, functionality, and reactivity are closely associated with structural features, which in turn, strongly influence the performance and suitability of lignin for various downstream applications.9 These structure–property variations largely originate from the intrinsic heterogeneity and complexity of lignin, thereby motivating the development of fractionation strategies to tailor lignin structures for targeted valorization pathways.10 High-molecular-weight lignin from hardwood, with condensed lignin structures, promotes graphitic-like crystallization, leading to improved mechanical strength in the resulting lignin-based ultrafine carbon fibers.11 Low-molecular-weight lignin (1860 g mol−1) with improved water solubility, bioactivity, and functional groups from rice straw significantly enhanced maize seedling growth as a biostimulant, whereas the high-molecular-weight lignin counterpart (6840 g mol−1) exhibited growth inhibition due to its condensed hydrophobic structure.12 The study of the relationship between the lignin molecular weight and biological applications showed that the low-molecular-weight lignin (2540 g mol−1) exhibited fewer interunit linkages, higher S/G ratios, and more phenolic hydroxyl groups, enhancing their solubility and antioxidant capacity. In contrast, the moderate-molecular-weight lignin (3940 g mol−1) with balanced functional groups provided an optimized biocompatibility, making it suitable for osteoarthritis treatment. The high-molecular-weight lignin with condensed structures and fewer reactive sites, showed low bioactivity but high biocompatibility, suiting structural application.13 Taken together, developing techniques that enables the tunning of lignin structure is crucial to strengthen the valorization potential of lignin.

Lignin fractionation is an effective method that divides heterogeneous lignin with broad molar mass dispersity (ĐM) into multiple fractions, each with reduced dispersity compared with the original lignin, and these fractions display distinct molecular weight distributions from one another.14,15 Solvent exchange (SE), acid precipitation (AP), and membrane ultrafiltration (MU) emerge as the three most commonly used approaches with distinct underlying separation principles, although other methods such as enzyme-mediated fractionation systems have also been widely explored.16 Based on the different solubility of lignin in a selected organic solvent, SE approaches separate lignin into soluble and insoluble fractions, yielding a certain molecular weight either through selective precipitation or dissolution of a lignin solution or a solid lignin, respectively.16,17 Successive solvent extraction with different solvents, multiple lignin fractions with well-defined molecular weight distributions from a few hundred up to more than ten thousand can be separated.18,19 Alternatively, SE uses binary solvent mixtures by adding an anti-solvent to a good solvent with better dissolving capacity for lignin, such as glycerol/ethanol,20 ethyl acetate/petroleum ether,21 and acetone/hexanes,22 thereby modulating the solubility of the mixture and precipitating several fractions with different molecular weights.16 Furthermore, replacing one of the two solvents with water as an anti-solvent, solvent fractionation can be conducted cost-effectively and environmentally friendly with only one organic solvent, like tetrahydrofuran,23 ethanol,24 acetone,25 acetic acid,26 and γ-valerolactone.27 The average molar mass of the fraction typically decreases along with the addition of water.28 However, there are substantial discrepancies in molecular weight size and homogeneity among these methods, depending on the lignin sources, selected solvents, and the recipes employed.

AP is commonly used for lignin fractionation from black liquor or alkali solutions.29 In basic conditions, lignin exists as a stable colloid due to the negative surface charges from deprotonated phenolic and carboxylic groups. Upon acidification, protonation neutralizes these charges, leading to colloidal aggregation and lignin precipitation. High-molecular-weight lignin typically precipitates first at higher pH values, whereas low-molecular-weight lignin bearing more ionizable groups remains in solution until lower pH levels, allowing for stepwise molecular-weight-based separation when a gradient acidification approach is applied.16,30 The molecular weight and yield of the resulting fractions are largely determined by the pH intervals.31–33 In addition to its simple operation and scalability, AP is compatible with industrial pulping liquors and requires no organic solvents, making it attractive for large-scale lignin recovery.34 However, the selectivity of this technique is limited, as co-precipitation of low- and high-molecular-weight lignin often occurs at overlapping pH ranges, leading to fractions with relatively broad molecular-weight distributions and approximate average molecular weight.33,35 Furthermore, the presence of inorganic salts (e.g., sodium sulfate) formed during neutralization can affect the purity and downstream processing of the recovered lignin.30

MU separates dissolved lignin into permeate and retentate according to the desired molecular weight based on the pore size of the membranes. The process offers continuous operation, low energy demand, reagent-free conditions, and easy integration with other separation technologies.32,36–38 By employing membranes with different molecular weight cut-offs between 1–50 kDa, lignin can be fractionated into high-, medium-, and low-molecular-weight components with relatively narrow dispersity.16,38–40 The separation efficiency and stability of the process depend strongly on membrane material and operating conditions. Polymeric membranes such as polyethersulfone and polysulfone are widely used for aqueous lignin systems, while titanium dioxide- and zirconium dioxide-coated ceramic membranes show superior chemical and thermal resistance, making them more suitable for extreme pH or temperature environments.41 However, the poor solubility of technical lignin in water and the tendency of macromolecular lignin to cause membrane fouling remain a noteworthy challenge.41

To address the discrepancies arising from these three commonly used fractionation techniques, comparative research for one lignin source in one lab with minimized deviation can offer a more direct way to assess the major differences and nuances. For instance, the comparison of AP and MU for the same black liquor indicated that the AP resulted in residual hemicelluloses contained in the lignin fractions, whereas the MU produced purer lignin fractions with controlled molecular weights and lower dispersity.42 Another comparative study focused on the successive extraction & precipitation methods versus MU for lignin fractionation heavily hinged on a large number of low boiling point solvents, despite the successive extraction and precipitation approaches enabled the most distinct molecular weight separation with markedly reduced molar mass dispersity (ĐM).43 Nevertheless, these studies provided limited insight into the deeper physicochemical consequences of fractionation techniques, which critically determine lignin fractions' chemical reactivity, dispersibility, stability, and compatibility in a material matrix. The relationship between fractionation strategies and the properties of fractions (e.g., interunit linkages, compositional unit ratios, morphology, functional group distribution, surface charge, and thermal stability) remains underexplored. Moreover, a comprehensive comparison of SE, AP, and MU regarding their impacts on lignin physicochemical properties under minimized experimental variation has not yet been examined.

In this work, we studied the fractionation of kraft lignin employing the three widely applied but mechanistically distinct fractionation approaches, SE, AP, and MU, for their effects on the physicochemical properties of the lignin fractions. By integrating comprehensive characterizations, including molecular weight distribution, hydroxyl group content, surface charge, chemical structure, morphological features, and thermostability, we proposed potential connections between fractionation strategy and lignin's structure–property relationships. Such comparative insights provide a mechanistic foundation for selecting appropriate fractionation methods based on targeted applications, thereby guiding the rational utilization and valorization of technical lignins for sustainable material development.

2. Materials and methods

2.1. Materials

Industrial Indulin AT softwood kraft lignin (KL) was acquired from MeadWestvaco (Richmond, VA) and utilized as received. All chemicals in this study were purchased from Fisher Scientific, VWR or MilliporeSigma and used as received without further purification. The ultrafiltration devices (200 mL) and regenerated cellulose membrane (filter diameter: 63.5 mm, molecular weight cut-offs of 30 kDa and 10 kDa) were purchased from Amicon® and Ultracel®.

2.2. Lignin fractionation

2.2.1. SE fractionation

Dry KL (10.0 g) was dissolved in 100 mL 75% aqueous ethanol (relatively cheap and benign solvent system) by slowly adding the KL into the solution while stirring for 1 h. Next, the resulting suspension was centrifuged at 5000 × g for 10 min to separate the insoluble lignin as fraction 1 (SE-F1). The supernatant (∼100 mL) was diluted by adding 50 mL water (equivalent of 50% aqueous ethanol) under stirring for 1 h to precipitate lignin again. The new 150 mL suspension was centrifuged at 5000 × g for 10 min to yield the residue as fraction 2 (SE-F2). The supernatant after centrifugation was transferred and concentrated under reduced pressure to remove most ethanol to precipitate lignin. The ethanol-removed suspension was then centrifuged at 5000 × g for 10 min to lead to lignin residue fraction 3 (SE-F3) and the supernatant was decanted. Finally, all three fractions were washed with 50 mL water on filter paper (Fisher Scientific, P8), freeze-dried, and weighed quantitatively.44

2.2.2. AP fractionation

Dry KL (∼85.0 g) was slowly added to 900 mL of a 1 wt% NaOH solution (pH 13.4) while stirring until the pH value became stable (pH ∼10). The obtained suspension was centrifuged at 5000 × g for 10 min to separate the insoluble lignin residue, designated as fraction 1 (AP-F1). The solubilized lignin in the supernatant was further precipitated by drop-wise addition of 6 M H2SO4 under stirring until pH of 7 and pH of 2 to recover lignin precipitated fraction 2 (AP-F2) and fraction 3 (AP-F3), respectively, using the same centrifugation method. All lignin fractions were acidified to pH 2 by drop-wise addition of 6 M H2SO4 and washed with 50 mL water on filter paper (Fisher Scientific, P8) before freeze-drying.31

2.2.3. MU fractionation

Dry KL (6.0 g) was first dissolved completely in 200 mL of 1 wt% NaOH based on the solubility of KL in NaOH solution.45 The lignin solution was separated into three portions using Amicon® Stirred Cell when it was sequentially filtrated through membranes with different cut-off molecular weights at a pressure of 55 psi. The first liquid retentate was collected from the filtration with the 30 kDa membrane employed, representing a fraction denoted as fraction 1 (MU-F1). Then the permeate from the 30 kDa membrane was subjected to filtration using the 10 kDa membrane under the same conditions. The retentate and the permeate were then collected as fraction 2 between 30 kDa and 10 kDa (MU-F2), and fraction 3 < 10 kDa (MU-F3), respectively. Finally, the three liquid fractions were acidified to pH 2 by adding 2 M H2SO4 to precipitate the dissolved lignin. The solid lignin fractions were collected by centrifugating at 5000 × g for 10 min to remove the liquid and washed with 25 mL of water on filter paper (Fisher Scientific, P8) before freeze-drying.37

2.3. Chemical structure characterization

2.3.1. Gel permeation chromatography (GPC)

GPC was used to determine the mass-average molecular weight (Mw) and number-average molecular weight (Mn) of lignin fractions after acetylation.46,47 The molar mass dispersity (ĐM) is defined as the ratio of the Mw to the Mn. The fractionated lignins were derivatized by dissolving ∼3 mg of lignin in a 1 : 1 v/v mixture of anhydrous pyridine and acetic anhydride and keeping the reaction at room temperature in dark with continuous magnetic stirring for 24 h. Pyridine and the remaining acetic anhydride were removed by co-evaporation with ethanol at 45 °C using a rotary evaporator; this process was carried out several times until the acetylated lignins were completely dry. Subsequently, the acetylated lignins were dissolved in tetrahydrofuran overnight and filtered through a 0.45 µm membrane filter before being analyzed using a TOSOH HLC-8320GPC EcoSEC system (Tosoh Corp., Tokyo, Japan) with a refractive index detector and a TSKgel SuperAWM-H column. Tetrahydrofuran was used as the mobile phase, with a flow rate of 0.35 mL min−1. The calibration curve was established using polystyrene standards from 580–6 085 000 g mol−1. The reported values represent the average of duplicate measurements with standard deviation.

2.3.2. Fourier-transform infrared spectroscopy (FTIR) analysis

The FTIR spectra of lignin fractions were recorded using an FTIR spectrometer (Spectrum Two FTIR , PerkinElmer, Shelton, CT, USA) with a universal attenuated total reflection accessory. The spectra were acquired from 4000 to 600 cm−1 with 32 scans and a resolution of 4 cm−1.33

2.3.3. Nuclear magnetic resonance (NMR) analysis

NMR spectra were recorded on a Varian 400 MHz spectrometer (Varian, Inc., Palo Alto, CA, USA). 31P NMR was used to quantify the hydroxyl group contents in lignin fractions.46,47 Dried lignin samples (∼20 mg) were accurately measured and phosphitylated with 2-chloro-4,4,5,5-tetramethyl-1,3,2-dioxaphospholane (TMDP) in an anhydrous solvent mixture of pyridine/CDCl3 (1.6 : 1.0 v/v), and N-hydroxy-5-norbornene-2,3-dicarboximide (NHND) was added as an internal standard. After adding TMDP, the 31P NMR spectra were acquired quickly (within <10 min) at a 45° pulse angle, 1 s acquisition time, and 5 s D1 relaxation delay with 512 scans. The spectra clearly identified and resolved several hydroxyl groups, including aliphatic OH (149.6–145.5 ppm), C5 substituted (144.1–143.1 ppm, 141.6–140.7 ppm), guaiacyl (140.3–138.8 ppm), p-hydroxyphenyl group (138.4–137.6 ppm), and carboxylic hydroxyl groups (135.8–134.2 ppm). The product of TMDP reacted with water at 132.2 ppm was used as the reference peak to report all chemical shifts (Fig. S3). The contents of hydroxyl groups were quantified based on the amount of the added NHND internal standard.

For the structural analysis, ∼50 mg of each lignin fraction was dissolved in DMSO-d6 and transferred into a 5 mL NMR tube. The proton and two-dimensional (1H–13C) heteronuclear single quantum coherence (HSQC) NMR spectra were collected using a phase-sensitive gradient-edited HSQC program (gHSQCAD), following the default acquisition parameters except for using 96 t1 increments, at 4 scans per t1 increment, and a scan range of 0–190 ppm in the F1 (13C) dimension.46,47 The acquired spectra were processed using MestReNova 14.2 (Mestrelab Research S. L., Spain) software where first they were subjected to a 3rd order polynomial baseline correction in all dimensions, then automatic phase correction, Gaussian (90 Hz in F1 and 10 Hz in F2) and exponential (−10 Hz in F1 and −0.5 Hz in F2) apodization, t1 noise reduction, and finally adjusted to a spectrum size of 1024 (in both dimensions). HSQC contour integrals were interpreted as relative values for comparative evaluation among samples rather than absolute quantification. The DMSO peak at δC/δH – 39.5/2.49 ppm was used as internal reference.

2.3.4. Zeta potential measurement

The surface charge of lignin fractions was evaluated using a Zetasizer Nano ZS-90 (Malvern Instruments, UK). Each lignin sample (wet sample before freeze-drying) was redispersed in deionized water with a pH of ∼6. Zeta potential values were obtained at 25 °C using folded capillary zeta cells, and each sample was measured in triplicate.32 The reported values represent the mean and standard deviation of three measurements.

2.4. Thermal and morphology characterization

2.4.1. Differential scanning calorimetry (DSC) analysis

The thermal behavior and stability of lignin fractions were evaluated using differential scanning calorimetry (DSC).48 DSC measurements were performed on a DSC 8500 (PerkinElmer, Shelton, CT, USA). Approximately 5 mg of each dried sample was sealed in aluminum pans and subjected to a heating–cooling–heating program under a nitrogen atmosphere (20 mL min−1). The samples were first heated from 25 to 150 °C at 100 °C min−1 to erase thermal history, cooled to 25 °C at the same rate, and then reheated to 185 °C. After cooling again to 25 °C, a third heating cycle was conducted up to 200 °C. The glass transition temperature (Tg) was determined from the midpoint of the heat capacity change.

2.4.2. Thermogravimetric analysis (TGA)

TGA was carried out on a ThemysOne+ instrument using a TG-DSC rod (Setaram, Caluire-et-Cuire, France) under air flow (30 mL min−1). Approximately 10 mg of each sample was heated from 30 °C to 800 °C at a rate of 10 °C min−1.48 The onset degradation temperatures (Tonset) were determined from TGA curves using the tangent-intersection method.

2.4.3. Scanning electron microscope (SEM) analysis

The morphology of lignin fractions was monitored using a scanning electron microscope (SEM, Auriga Crossbeam, Carl Zeiss, Oberkochen, Germany). Prior to imaging, each lignin powder was mounted on aluminum stubs with conductive carbon tape and an additional silicon wafer support, and sputter-coated with a thin layer of gold to minimize charging.49 Micrographs were collected in high-vacuum mode at an accelerating voltage of 3 kV.

3. Results and discussion

3.1. Molecular weight distribution and fractionation efficiency

The unfractionated KL exhibited a brownish color with a particle size ranging from 10 to 50 µm. After fractionation, the three fractionation techniques yielded physically distinguishable lignin fractions (Fig. 1 and micrographs in Fig. S1). The micro-graphical images showed that the SE fractions displayed progressively smaller particle sizes from F1 (20–50 µm), F2 (3–20 µm) to F3 (100–500 nm), accompanied by a noticeable lighter color, consistent with the higher light absorption and scattering of finer particles. In contrast, the AP fractions exhibited increasing particle aggregation from F1 (5–20 µm), F2 (10–30 µm) to F3 (30–100 µm), resulting in darker appearances in the order. The three MU fractions exhibited similar appearances and particle sizes (20–30 µm), corresponding to the same precipitation step applied during ultrafiltration. In addition, the three methods also produced distinct yield distributions among their fractions (Fig. 1). SE generated relatively balanced yields across F1 to F3 (38.16%, 12.66%, 30.11%), whereas AP showed the highest total lignin recovery (92.4%) with the F3 fraction showing the highest yield of 48.34%. In contrast, MU yielded more than half of its mass in the F1 fraction of 47.89%, with considerably smaller proportions in F2 and F3 (16.84% and 21.40%). These visual, morphological and yield differences indicate that the three methods generate distinctly different types of lignin particular fractions.

Fig. 1. The appearance and yields of kraft lignin fractions obtained by SE, AP, and MU methods.

Fig. 1

Open in a new tab

The unfractionated KL exhibited a weight-average molecular weight (Mw) of 6620 g mol−1 and a broad dispersity (ĐM = 5.8) (Table 1), aligning with reported ranges for KL (Mw 4500–7600 g mol−1 and ĐM up to 9.9).50,51 All three fractionation methods reduced the intrinsic heterogeneity of the KL and produced fractions with distinct molecular-weight characteristics. However, the extent of molecular-weight separation into high-, moderate-, and low-molecular-weight fractions varied substantially among the methods. The SE produced three sharply differentiated fractions with clear molecular-weight gradients (Table 1). The SE-F1 was enriched in high-molecular-weight lignin (Mw = 20 307 g mol−1, ĐM = 10.0), while the SE-F2 (Mw = 4529 g mol−1, ĐM = 2.8) and the SE-F3 (Mw = 1754 g mol−1, ĐM = 2.0) represented well-defined moderate- and low-molecular-weight fractions. The insoluble fraction, SE-F1, contains most of the KL's high-molecular-weight components together with a small portion of moderate- and low-molecular-weight lignin not fully dissolved in the ethanol-solvent system, resulting in its broad molecular-weight distribution (Fig. S2). During the solvent-exchange step, the gradual addition of water progressively decreases lignin solubility. As solubility drops, moderate-molecular-weight lignin precipitates first, whereas low-molecular-weight species remain dissolved because their higher polarity and stronger affinity toward aqueous ethanol mixtures enable them to stay in solution until later stages.28,44 This molecular-weight-dependent solubility trend has been consistently reported for ethanol–water, acetone–water, and other binary solvent systems.24,25,27 However, the sharpness of fractionation depends on the intrinsic heterogeneity of the original lignin, solvent composition, and mixing protocol. The reported value for SE fractions from similar KL (Mw = 6200 g mol−1, ĐM = 4.4) using ethanol–water system are (F1: Mw = 24 000 g mol−1, ĐM = 5.4), (F2: Mw = 6900 g mol−1, ĐM = 2.4) and (F3: Mw = 3900 g mol−1, ĐM = 3.4).44 Our SE fractions exhibit similar molecular Mw (20 307, 4529, 1754 g mol−1), comparable dispersity (ĐM = 2.8) in SE-F2 and narrower dispersity (ĐM = 2.0) in SE-F3 (Table 1). Notably, SE-F2 and SE-F3 together accounted for more than 40% of the total mass yield, indicating that SE can generate substantial amounts of low-heterogeneity lignin in the moderate- and low-molecular weight that are highly relevant for chemical conversion because lower molecular weight and narrower dispersity enhance solubility, accessibility of reactive sites, and overall reactivity in polymer synthesis, like polyurethane.52,53

Table 1. Molecular weight characteristics of kraft lignin (KL) and its fractions obtained by SE, AP, and MU methods. Mw: mass-average molecular weight; Mn: number-average molecular weight; ĐM: molar-mass dispersity (ĐM = Mw/Mn).

Lignin M n (g mol−1) M w (g mol−1) Đ M
KL 1151 ± 24 6620 ± 48 5.8
SE-F1 2037 ± 37 20 307 ± 422 10.0
SE-F2 1626 ± 24 4529 ± 176 2.8
SE-F3 898 ± 8 1754 ± 20 2.0
AP-F1 1242 ± 26 7328 ± 234 5.9
AP-F2 1251 ± 5 5455 ± 309 4.4
AP-F3 1182 ± 6 4317 ± 167 3.7
MU-F1 1138 ± 211 12 443 ± 3427 10.9
MU-F2 651 ± 55 2938 ± 432 4.5
MU-F3 531 ± 26 1600 ± 207 3.0
Open in a new tab

The AP method generated fractions with a molecular-weight distribution that differed markedly from that of the SE method. Although the Mw decreased from AP-F1 (7328 g mol−1) to AP-F2 (5455 g mol−1) and AP-F3 (4317 g mol−1), the separation was limited: all the Mn values remained clustered around ∼1200 g mol−1, nearly identical to the parent KL (Table 1). Dispersity values (ĐM = 3.7–5.9) showed only modest improvement to the KL of 5.8, and the GPC curves exhibited extensive overlap among all three AP fractions, confirming the weak separation of the AP method in the employed conditions (Fig. S2). This limited molecular size resolution likely stems from the intrinsic mechanism of the acid precipitation. As H+ is introduced, lignin colloids undergo protonation-driven charge neutralization, leading to rapid aggregation and precipitation. The neutralized lignin molecules precipitate at nearly the same pH regardless of their molecular weight, causing extensive co-precipitation of all high-, moderate-, and low-molecular weight lignin and thereby preventing sharp molecular-weight separation.16 Prior studies have consistently reported similar outcomes, in which AP fractions exhibit broad molecular-weight distributions and limited separation efficiency, with reported ĐM typically around 6 and up to 16.30,33,34,42 Despite its limited fractionation sharpness, the offered highest processing capacity and lowest solvent consumption among the three methods as well as the greatest total lignin recovery yield (Fig. 1), enabling AP for efficient bulk lignin recovery from kraft black liquor.

The MU method was capable of producing low-molecular-weight lignin in lower molar-mass dispersity, but with limited separation efficiency of moderate- and high-molecular-weight species. Both MU-F2 (Mw = 2938 g mol−1, ĐM = 4.5) and MU-F3 (Mw = 1600 g mol−1, ĐM = 3.0) were composed mainly of low-molecular-weight lignin and narrower distribution compared with the original KL but relatively low yield (Table 1 and Fig. 1, S2). In contrast, the MU-F1, exhibited the highest Mw (12 443 g mol−1) and broad dispersity (ĐM = 10.9), which accounted for more than half of the total yield, indicating that both high- and moderate-molecular-weight lignin and a substantial portion of low-molecular-weight molecules were retained in the retentate (Table 1 and Fig. 1, S2). This molecular-weight pattern aligns with previous observations in ultrafiltration studies, which reported that macromolecular lignin fragments tend to aggregate and deposit on the membrane surface, rapidly initiating membrane fouling owing to their poor solubility.41 As fouling progresses, the effective pore size narrows, and mass-transfer resistance increases, causing a strong decline in permeate flux and sharply reducing the permeability of moderate- and high-molecular-weight lignin.41 Consequently, only low-molecular-weight fragments are able to pass through the fouled membrane, while both moderate- and high molecular weight lignin are largely retained in the retentate. Similar results have been reported in studies using membranes (polysulfone, polyethersulfone, or ceramic) that progressively retained lignin species far below their nominal molecular weight cut-offs despite the use of 10–30 kDa membranes, resulting in the permeates with substantially lower molecular weight lignin than expected.32,37,39,54 Meanwhile, compared with these polymeric and ceramic membranes, regenerated cellulose membranes exhibit even lower lignin permeability, often yielding retentates and permeates with greater molecular-weight contrasts, e.g., 33 500 Da vs. 4900 Da.55

To further validate this membrane-dependent difference, we conducted an experiment using polyethersulfone membranes with the same nominal molecular weight cut-offs as the regenerated cellulose membrane. The polyethersulfone membranes showed noticeably reduced fouling and improved permeation behavior (Fig. S8 and Table S1). The yields of fraction 2 (21.31%) and fraction 3 (21.39%) increased substantially, and the Mw values of PES-F1, F2, and F3 were 8174 g mol−1, 6968 g mol−1, and 3633 g mol−1, respectively, indicating that moderate-molecular-weight lignin successfully permeated the polyethersulfone membrane. Although the dispersity (ĐM) of these PES-derived fractions remained a large dispersity that can be improved by choosing appropriate molecular cut-offs and tuning the filtration parameters (e.g., concentration of lignin solution, pressure, and time), the results clearly demonstrate that membrane material and compatibility strongly influence lignin retention/permeation behavior and the sharpness of molecular-weight separation.

3.2. Hydroxyl group content and surface charge

The hydroxyl group content of lignin is decisive for its functionalization and reactivity, particularly in transformations that depend on nucleophilic sites such as polyol–isocyanate reactions, esterification, and etherification. The unfractionated KL contained 3.85 mmol per g aliphatic OH, 3.94 mmol per g phenolic OH, and 0.59 mmol per g COOH groups, giving a total OH content of 8.38 mmol g−1 (Table 2), consistent with reported values for softwood kraft lignin.43,44 Fractionation generated substantial shifts in the profile of hydroxyl group (e.g., the relative proportions of aliphatic, phenolic, and carboxylic OH) in the individual fraction due to the changes in the makeup of lignin molecules in each fraction.

Table 2. Hydroxyl group distributions of kraft lignin (KL) and its fractions obtained by SE, AP, and MU methods, as determined by quantitative 31P NMR.

Lignin Aliphatic OH (mmol per g lignin) Phenolic OH (mmol per g lignin) COOH (mmol per g lignin) Total (mmol per g lignin)
C5 substituted Guaiacyl p-Hydroxylphenyl Total
KL 3.85 1.10 2.52 0.32 3.94 0.59 8.38
SE-F1 4.10 0.89 1.65 0.18 2.72 0.34 7.16
SE-F2 3.70 1.51 2.50 0.44 4.46 0.53 8.69
SE-F3 3.84 1.58 3.76 0.48 5.81 0.74 10.35
AP-F1 4.24 1.57 2.95 0.32 4.84 0.51 9.60
AP-F2 3.76 1.12 2.77 0.26 4.14 0.67 8.57
AP-F3 3.48 0.86 2.54 0.26 3.65 0.60 7.74
MU-F1 3.46 1.18 2.02 0.18 3.37 0.53 7.37
MU-F2 3.10 1.16 2.46 0.24 3.86 0.62 7.58
MU-F3 2.88 1.11 3.19 0.34 4.64 0.81 8.33
Open in a new tab

For the SE fractions, a trend in the hydroxyl distribution was observed (Table 2). The SE-F1 contained higher aliphatic OH content (4.10 mmol g−1) in comparison to the SE-F2 (3.7 mmol g−1) and SE-F3 (3.84 mmol g−1). In contrast, the phenolic OH and carboxylic OH are in an increasing order from SE-F1, SE-F2, and SE-F3, such as the total phenolic OH is of 2.72, 4.46, and 5.81 mmol g−1, and the carboxylic OH is of 0.35, 0.53, and 0.74 mmol g−1, respectively. This increasing trend is maintained in the total OH from 7.16, 8.69, 10.35 mmol g−1 for the SE-F1, SE-F2, and SE-F3, respectively. This pattern is consistent with literature on solvent fractionation using ethanol–water, tetrahydrofuran–water and γ-valerolactone–water systems.23,27,43,44 It is well-known that carboxylic OH and phenolic OH are more readily ionizable than the aliphatic OH. With less overall polarity, the lignin molecules containing more aliphatic hydroxyls and less non-ionizable hydroxyls are prone to dissolve in the solvent versus the antisolvent water in this case. Thus, the SE-F1 undissolved in the 75% ethanol tends to have less polar lignin with more aliphatic OH. Meanwhile, low- and moderate-molecular-weight lignins contain more phenolic chain ends and more oxidized side groups such as COOH, which increase polarity and strengthen affinity toward aqueous ethanol and dissolution.43,44 Upon addition of more water, certain lignin molecules precipitated as SE-F2, showing higher ionizable OHs than the SE-F1, and the well-dissolved lignin molecules were finally recovered as SE-F3, possessing the largest quantity of phenolic OH and carboxyl groups. Combining the lower molecular weight and narrower molar mass disparities (ĐM) of the SE-F2 and SE-F3, they are likely appropriate for lignin-derived polymer applications and materials development, including lignin-based polyurethanes,52,53 esterified lignin derivatives,56 phenolic resin adhesives,57 and lignin-based epoxy systems,58 requiring high hydroxyl availability and reactivity.

The AP and MU methods produced much weaker differentiation in the distributions of the OH group across their fractions compared with SE (Table 2). Both AP and MU showed a gradual decrease in aliphatic OH from F1 to F3, but phenolic OH and COOH groups did not exhibit trending profiles observed in SE, reflecting their limited capability for OH group redistribution. These trends were matched with the reported values that aliphatic OH content decreased from 1.67 (F1) to 1.43 g mol−1 (F3) in acid precipitation, while the phenolic OH values of three lignin fractions were 1.64, 1.49 and 1.52 g mol−1, respectively.35 Likewise, in ultrafiltration, aliphatic OH decreased from 2.62 to 1.25 mmol g−1 from F1 to F6, while phenolic and carboxyl OH fluctuate within 2.81–4.11 g mol−1 and 0.35–0.52 mmol g−1 among the six fractions.55 In our study, the total hydroxyl content of AP and MU fractions was generally lower than the original KL, indicating that both processes caused partial loss of hydroxyl functionality during fractionation. For AP, fractions displayed a stepwise decline in both aliphatic and phenolic OH, resulting in the total OH decreasing from 9.60 mmol g−1 (AP-F1) to 7.74 mmol g−1 (AP-F3), consistent with the chemically harsh acidic conditions. High proton concentration promotes protonation of benzylic hydroxyls, followed by dehydration and formation of benzylic carbocations, especially for AP-F3 precipitated at pH 2, which readily undergo C–C condensation reactions and may also react with nearby phenolic hydroxyls, thereby consuming free OH groups and limiting their redistribution.59,60 The limited molecular-weight-separation efficiency in AP further restricts the possibility of isolating fractions with distinct hydroxyl compositions (Table 1). For MU, aliphatic OH contents decreased from MU-F1 to MU-F3, while phenolic OH varied only slightly, resulting in total OH values lower than KL (7.37–8.33 mmol g−1). Because MU is driven by size-based permeation rather than solubility differences, it inherently provides weaker structural sorting. Moreover, the dissolution of kraft lignin prior to ultrafiltration and its subsequent post-treatment requires strong alkaline (pH 13.4) and acidic (pH 2) conditions. Such environments can trigger base- or acid-catalyzed structural rearrangements and mild condensation reactions,60 contributing to the loss of hydroxyl groups observed in MU fractions. However, when lignin is dissolved in milder, lignin-compatible organic solvents or weakly acidic media during ultrafiltration, the redistribution of hydroxyl groups in MU can be significantly improved, as demonstrated in previous studies.40,43 To further visualize the relationships between molecular weight and hydroxyl-group distribution among different fractionation methods, correlation plots between Mw and total, aliphatic, and phenolic OH contents are provided in Fig. S9.

As a result, the fractions differ in the number of ionizable OH groups available on their macromolecular surfaces, which in turn governs their surface charge behavior in aqueous media. These distinctions form the basis for evaluating each fraction's suitability in hydroxyl-driven chemical transformations and in applications for colloidal formulations. The differences in hydroxyl-group distribution among the fractions were also reflected in their surface-charge measured in zeta potential (Table 3). The unfractionated KL exhibited a negative zeta potential (−43.4 mV), consistent with reported values (−30 to −40 mV).61 The pKa of phenolic and carboxyl OH groups is typically at 7–10.5 and 3–5, respectively.62 So, they can be partially deprotonated at neutral pH, contributing to lignin's negative surface charge.32 Accordingly, the SE-F2 and SE-F3 showed substantially more negative zeta potentials (−54.7 mV and −58.7 mV), agreeing with their higher phenolic and carboxyl OH contents (Table 2). The highly negative zeta potential enhances electrostatic repulsion between particles and greatly improves colloidal stability. Combining their nanoscale particle sizes (Fig. S1), such strongly charged, well-dispersible fractions are particularly attractive for applications that require stable aqueous lignin systems, including lignin nanoparticles63,64 and emulsifiers.65 In addition, the high negative charge favors electrostatic interactions with cationic polymers (e.g., chitosan) or ions, expanding their potential use in lignin-based composite materials formulations.66 In contrast, AP and MU fractions exhibited only moderately negative zeta potentials at about −40 mV (Table 3), consistent with their lower phenolic and carboxyl OH contents than SE-F2 and SE-F3 (Table 2). The lower density of deprotonatable groups on the lignin molecules results in a weaker surface charge, which diminishes electrostatic repulsion and yields lower colloidal stability than SE-F2 and SE-F3. As a consequence, AP and MU fractions are less suitable for dispersion-based or nanoparticle applications compared with SE-derived fractions. Moreover, the similar zeta potential values among AP fractions further corroborate the observed co-precipitation phenomenon that low- moderate- and high-molecular-weight lignin precipitate at overlapping pH ranges, leading to three fractions with approximate molecular weight distribution (Table 1 and Fig. S2) and surface charge.

Table 3. Zeta potential values of kraft lignin (KL) and its fractions obtained by SE, AP, and MU methods.

Lignin Zeta potential (mV) Lignin Zeta potential (mV) Lignin Zeta potential (mV)
KL −43.4 ± 1.18
SE-F1 −45.1 ± 0.97 AP-F1 −38.6 ± 0.80 MU-F1 −41.8 ± 0.99
SE-F2 −54.7 ± 0.56 AP-F2 −37.8 ± 0.76 MU-F2 −42.2 ± 1.03
SE-F3 −58.7 ± 0.66 AP-F3 −39.2 ± 1.15 MU-F3 −34.3 ± 0.84
Open in a new tab

3.3. Chemical structure analysis

The HSQC spectra (Fig. 2 and 3) provided detailed insight into the interunit-linkage patterns and compositional-unit ratios of the fractionated lignins. Compared with the unfractionated KL (β-O-4′: 66.7%, β-5′: 16.0%, β–β′: 17.3%; Fig. S5), the SE fractions still maintained the high portion of β-O-4′ ether linkages, consistently exceeding 60% across SE-F1, SE-F2, and SE-F3. Correspondingly, the proportions of condensed C–C linkages (β-5′ and β–β′) remained comparatively low, at 14–25% in all SE fractions (Fig. 2), indicating that solvent-exchange fractionation preserves native lignin structures while redistributing molecules into fractions according to their molecular weight and solubility characteristics rather than chemically modifying the linkages. Meanwhile, SE-F1 contained a higher proportion of β-O-4′ linkages than SE-F2 and SE-F3. During kraft pulping, β-O-4′ ether linkages are susceptible to cleavage, whereas condensed C–C linkages are more resistant. As a result, higher-molecular-weight lignin macromolecules retain a greater abundance of surviving β-O-4′ linkages, while lower-molecular-weight fragments generated during pulping display elevated levels of β-5′ and β–β′ condensed linkages.21,27,47

Fig. 2. Aliphatic region of HSQC spectra of lignin fractions obtained by three fractionation methods—SE (a–c), AP (d–f), and MU (g–i). The interunit linkages, such as β-O-4′, β-5′, and β–β′ are colored corresponding to their contours in the spectra.

Fig. 2

Open in a new tab

Fig. 3. Aromatic region of HSQC spectra of lignin fractions obtained by three fractionation methods—SE (a–c), AP (d–f), and MU (g–i). The compositional units, such as G and H, are colored corresponding to their contours in the spectra.

Fig. 3

Open in a new tab

In contrast, the AP and MU fractions displayed a clear increase in condensed interunit linkages compared with the KL. The β-5′ and β–β′ contents in AP fractions increased to 24.9–27.8% and 16.6–20.9%, while these in MU fractions reached up to 26.3% (β-5) and 24.2% (β–β), respectively (Fig. 2). This higher proportion of C–C linkages relative to β-O-4′ aryl ether linkages likely arise from the harsh alkaline or acidic environments intrinsic to both AP and MU processes. Under such conditions, β-O-4′ ether linkages readily undergo acid- or base-catalyzed cleavage, generating reactive intermediates such as benzyl carbocations, enol ethers, or phenylacetaldehyde species, which subsequently promote electrophilic aromatic substitution or radical-mediated coupling pathways. These reactions facilitate the formation of new C–C bonds and thereby increase the abundance of condensed structures in the resulting lignin fractions.67,68

The compositional aromatic units of these fractions were largely unchanged, primarily consisting of guaiacyl (G) units. Across the SE, AP, and MU fractions, the G-unit contents were in the range of 94.8–98.7% in comparison to 94.6% in the unfractionated KL (Fig. S5). Meanwhile, these fractions were found to have 1–5% H-unit depending on the fractionation methods (Fig. 3), consistent with the softwood origin of the lignin. As shown in Fig. S4, FTIR spectra also demonstrate consistent absorption peaks across all fractions and the unfractionated KL (Fig. S4a–c). These absorption peaks correspond to the characteristic lignin functional groups and are assigned individually according to literature.25,31–33 For instance, the broad band around 3448 cm−1 corresponds to the O–H stretching of phenolic and aliphatic hydroxyl groups, and the peaks near 2938 and 2837 cm−1 arise from C–H stretching in methyl and methylene groups. Characteristic aromatic skeletal vibrations are observed at 1595 and 1514 cm−1, while a weak band near 1720 cm−1 indicates minor unconjugated C Created by potrace 1.16, written by Peter Selinger 2001-2019 O stretching. The prominent band at 1268 cm−1, without a corresponding peak at 1339 cm−1, confirms the dominance of G units, consistent with the softwood origin. The remaining absorbance below 1200 cm−1 correspond to C–O and C–H vibrations of alcohols and ethers, typical of kraft lignin.33 Taken together, the NMR and FTIR spectroscopic analysis of these lignins revealed that fractionation method did not overly alter the fundamental functional groups presented in the unfractionated KL, despite there are minor changes in the interunit linkages of lignin that are potentially caused by the extreme acid/alkaline conditions.

3.4. Thermal stability

Thermal stability is a critical property for evaluating lignin's suitability in downstream material applications, particularly in biocomposites and carbon-rich materials, where the resistance to thermal deformation directly determines the processing temperature windows and the performance of the final products. Both the glass transition temperature (Tg) and the onset degradation temperature (Tonset) provide insights into the chain mobility, structural rigidity, and decomposition resistance of a polymer. Across all samples, the Tonset values of all the SE, AP, and MU fractions showed only minor variations, generally falling within 350–400 °C (Fig. S7), which agrees with the reported 300–400 °C.18,47 This similarity is consistent with their majorly unaltered structural feature, such as functional-group profiles and compositional units confirmed by FTIR and NMR.

The glass-transition behavior, by contrast, differed among the three fractionation approaches. All samples exhibited a broad endothermic transition characteristic of amorphous lignin, and their Tg values (Table 4 and Fig. S6) clearly reflected the structural and molecular differences induced by the three fractionation methods. SE fractions exhibited a clear, systematic decrease in Tg from SE-F1 (150.2 °C), SE-F2 (149.9 °C) to SE-F3 (139.7 °C), paralleling their progressive reduction in molecular weight (Table 1). The higher Tg of SE-F1 is related to its large molecular weight and extensive chain entanglement, indicating restricted segmental motion and structural rigidity of this fraction. In contrast, the much lower molecular weight of SE-F3 presumably has easier chain mobility, leading to a lower Tg. This monotonic decrease demonstrates that SE effectively modulates lignin's thermal transition behavior through precise molecular-weight control.69 For the AP and MU fractions, the Tg values (162.7–178.2 °C, 155.4–157.4 °C, respectively) were higher than the SE fractions at 139.7–150.2 °C. The elevated Tg, especially for AP fractions, is likely related with the increase in condensed C–C linkages (Fig. 2), reflecting a more compact and rigid macromolecular structure formed during acid-induced condensation. The reduced amount of β-O-4 linkages and increased C–C crosslinks restrict segmental mobility, imparting a certain degree of enhanced thermal stability to the AP fractions. The relationship between molecular weight and glass transition temperature (Tg) for different lignin fractions is further illustrated in Fig. S10. Such characteristics suggest that AP fractions may be suitable for applications requiring improved thermal resistance, such as fillers, carbon-rich materials, or potential flame-retardant systems.

Table 4. Glass transition temperature of kraft lignin (KL) and its fractions obtained by SE, AP, and MU methods.

Lignin T g (°C) Lignin T g (°C) Lignin T g (°C)
KL 154.9
SE-F1 150.2 AP-F1 162.7 MU-F1 155.4
SE-F2 149.9 AP-F2 178.2 MU-F2 156.7
SE-F3 139.7 AP-F3 165.2 MU-F3 157.4
Open in a new tab

4. Conclusion

This study systematically compared three commonly employed lignin fractionation methods, solvent exchange (SE), acid precipitation (AP), and membrane ultrafiltration (MU), for fractionating industrial Indulin AT kraft lignin, which is originally recovered via an acid precipitation process during pulping liquor treatment. Overall, the results demonstrate that fractionation not only reduces lignin heterogeneity but also strategically tailors its molecular weight, hydroxyl functionality, linkage distribution, morphology, and thermal stability. Although derived from the same feedstock, these methods produced fractions with markedly different physicochemical characteristics of each fraction that potentially affect their downstream valorization. The SE method proved the most effective in generating sharply separated fractions with narrow molecular weight distributions and high hydroxyl contents, resulting in enhanced chemical reactivity and excellent colloidal stability. These properties make SE fractions especially suitable for reactivity-driven applications such as polyurethane precursors, epoxy systems, and lignin nanoparticles. In contrast, the AP method benefits large-scale lignin recovery with minimal solvent consumption but achieves limited molecular-weight separation among the obtained fractions. Its acidic environment appears to cause lignin condensation, which reduced hydroxyl functionality and increased Tg. These characteristics may be advantageous in applications such as in fillers or rubber composites. The MU method can generate low-molecular-weight fractions with low dispersity that might be useful for functional composites and tailored polymer systems. We also found that the separation results highly depend on the selection of the membrane and solution concentration. The comprehensive characterization of the physicochemical features of lignin fractions provides practical guidance for selecting suitable methods to meet the requirements of targeted lignin valorization pathways. However, molecular overlapping among the obtained fractions suggests that fractionation techniques, particularly in a scalable and sustainable manner, that effectively narrow down the molar mass are crucially needed to advance lignin valorization and biorefinery. Future studies should extend this comparative approach to lignins derived from different botanical sources and isolation techniques to further evaluate the universality and limitations of these fractionation strategies.

Author contributions

Y. W. designed the study, performed the experiments, analyzed and interpreted the data, and wrote the original draft of the manuscript. K. S. and H. Z.a assisted with lignin characterization, data interpretation, and result analysis. H. Z.b and D. P. H. provided technical expertise and data discussion. M. L. supervised and managed the project, contributed to conceptualization and methodology, provided resources and funding acquisition, and reviewed and revised the manuscript. All authors discussed the results, revised the manuscript, and approved the final version of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Supplementary Material

RA-016-D6RA03462D-s001

Acknowledgments

This study is funded by the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, Buildings Energy Efficiency Frontiers & Innovation Technologies program under Award Number DE-EE-0010922. Neither the U.S. Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights.

Data availability

The data supporting the findings of this study are available within the article and its supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6ra03462d.

References

  1. Bajwa D. S. Pourhashem G. Ullah A. H. Bajwa S. G. Ind. Crops Prod. 2019;139:111526. doi: 10.1016/j.indcrop.2019.111526. [DOI] [Google Scholar]
  2. Mateo S. Fabbrizi G. Moya A. J. Polymers. 2025;17:952. doi: 10.3390/polym17070952. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Argyropoulos D. D. Crestini C. Dahlstrand C. Furusjö E. Gioia C. Jedvert K. Henriksson G. Hulteberg C. Lawoko M. Pierrou C. Samec J. S. Subbotina E. Wallmo H. Wimby M. ChemSusChem. 2023;16:e202300492. doi: 10.1002/cssc.202300492. [DOI] [PubMed] [Google Scholar]
  4. Souza R. E. Gomes F. J. B. Brito E. O. Costa Lelis R. C. Journal of Applied Biotechnology & Bioengineering. 2020;7:100–105. [Google Scholar]
  5. Katahira R., Elder T. J. and Beckham G. T., in Lignin Valorization: Emerging Approaches, RSC, 2018, ch. 1, pp. 1–20 [Google Scholar]
  6. Saadan R. Hachimi Alaoui C. Ihammi A. Chigr M. Fatimi A. Environmental and Earth Sciences Proceedings. 2024;31:3. [Google Scholar]
  7. Margarida Martins M. Carvalheiro F. Gírio F. Biomass Convers. Biorefin. 2024;14:3183–3207. doi: 10.1007/s13399-022-02701-z. [DOI] [Google Scholar]
  8. Zhang M. Jia L. Li M. Yoo C. G. Peng H. Arvelli S. Zhao J. Adv. Sustainable Syst. 2025;9:2401054. doi: 10.1002/adsu.202401054. [DOI] [Google Scholar]
  9. Gigli M. Crestini C. Green Chem. 2020;22:4722–4746. doi: 10.1039/D0GC01606C. [DOI] [Google Scholar]
  10. Schutyser W. Renders T. Van den Bosch S. Koelewijn S.-F. Beckham G. T. Sels B. F. Chem. Soc. Rev. 2018;47:852–908. doi: 10.1039/C7CS00566K. [DOI] [PubMed] [Google Scholar]
  11. Bai J. Wang S. Li Y. Wang Z. Tang J. Int. J. Biol. Macromol. 2021;187:594–602. doi: 10.1016/j.ijbiomac.2021.07.149. [DOI] [PubMed] [Google Scholar]
  12. Wu D. Wang Y. Qi S. Yuan Y. Guo J. Chen G. Ahmad M. Jiang B. Jin Y. Biomacromolecules. 2023;24:1377–1387. doi: 10.1021/acs.biomac.2c01392. [DOI] [PubMed] [Google Scholar]
  13. Zheng L. Lu G. Pei W. Yan W. Li Y. Zhang L. Huang C. Jiang Q. Int. J. Biol. Macromol. 2021;190:291–300. doi: 10.1016/j.ijbiomac.2021.08.168. [DOI] [PubMed] [Google Scholar]
  14. Sadeghifar H. Ragauskas A. J. ACS Sustainable Chem. Eng. 2020;8:8086–8101. doi: 10.1021/acssuschemeng.0c01348. [DOI] [Google Scholar]
  15. Rodrigues J. S. Lima V. Araujo L. C. Botaro V. R. Ind. Eng. Chem. Res. 2021;60:10863–10881. doi: 10.1021/acs.iecr.1c01704. [DOI] [Google Scholar]
  16. Pang T. Wang G. Sun H. Sui W. Si C. Ind. Crops Prod. 2021;165:113442. doi: 10.1016/j.indcrop.2021.113442. [DOI] [Google Scholar]
  17. Li H. McDonald A. G. Ind. Crops Prod. 2014;62:67–76. doi: 10.1016/j.indcrop.2014.08.013. [DOI] [Google Scholar]
  18. Park S. Y. Kim J. Y. Youn H. J. Choi J. W. Int. J. Biol. Macromol. 2018;106:793–802. doi: 10.1016/j.ijbiomac.2017.08.069. [DOI] [PubMed] [Google Scholar]
  19. Duval A. Vilaplana F. Crestini C. Lawoko M. Holzforschung. 2016;70:11–20. doi: 10.1515/hf-2014-0346. [DOI] [Google Scholar]
  20. Liu C. Si C. Wang G. Jia H. Ma L. Ind. Crops Prod. 2018;111:201–211. doi: 10.1016/j.indcrop.2017.10.005. [DOI] [Google Scholar]
  21. Jiang X. Savithri D. Du X. Pawar S. Jameel H. Chang H. M. Zhou X. ACS Sustainable Chem. Eng. 2017;5:835–842. doi: 10.1021/acssuschemeng.6b02174. [DOI] [Google Scholar]
  22. Cui C. Sun R. Argyropoulos D. S. ACS Sustainable Chem. Eng. 2014;2:959–968. doi: 10.1021/sc400545d. [DOI] [Google Scholar]
  23. Meng X. Parikh A. Seemala B. Kumar R. Pu Y. Wyman C. E. Cai C. M. Ragauskas A. J. Bioresour. Technol. 2019;272:202–208. doi: 10.1016/j.biortech.2018.09.130. [DOI] [PubMed] [Google Scholar]
  24. Goldmann W. M. Ahola J. Mikola M. Tanskanen J. Sep. Purif. Technol. 2019;209:826–832. doi: 10.1016/j.seppur.2018.06.054. [DOI] [Google Scholar]
  25. Boeriu C. G. Fiţigău F. I. Gosselink R. J. Frissen A. E. Stoutjesdijk J. Peter F. Ind. Crops Prod. 2014;62:481–490. doi: 10.1016/j.indcrop.2014.09.019. [DOI] [Google Scholar]
  26. Klett A. S. Chappell P. V. Thies M. C. Commun. Chem. 2015;51:12855–12858. doi: 10.1039/C5CC05341B. [DOI] [PubMed] [Google Scholar]
  27. Wang G. Liu X. Yang B. Si C. Parvez A. M. Jang J. Ni Y. ACS Sustainable Chem. Eng. 2019;7:10112–10120. doi: 10.1021/acssuschemeng.9b01641. [DOI] [Google Scholar]
  28. Jääskeläinen A. S. Liitiä T. Mikkelson A. Tamminen T. Ind. Crops Prod. 2017;103:51–58. doi: 10.1016/j.indcrop.2017.03.039. [DOI] [Google Scholar]
  29. Velez J. Thies M. C. Bioresour. Technol. 2013;148:586–590. doi: 10.1016/j.biortech.2013.08.097. [DOI] [PubMed] [Google Scholar]
  30. Helander M. Theliander H. Lawoko M. Henriksson G. Zhang L. Lindström M. E. BioResources. 2013;8:2270–2282. doi: 10.15376/biores.8.2.2270-2282. [DOI] [Google Scholar]
  31. Alekhina M. Ershova O. Ebert A. Heikkinen S. Sixta H. Ind. Crops Prod. 2015;66:220–228. doi: 10.1016/j.indcrop.2014.12.021. [DOI] [Google Scholar]
  32. Wang G. Chen H. Sep. Purif. Technol. 2013;105:98–105. doi: 10.1016/j.seppur.2012.12.009. [DOI] [Google Scholar]
  33. Lourençon T. V. Hansel F. A. da Silva T. A. Ramos L. P. de Muniz G. I. Magalhães W. L. Sep. Purif. Technol. 2015;154:82–88. doi: 10.1016/j.seppur.2015.09.015. [DOI] [Google Scholar]
  34. dos Santos P. S. Erdocia X. Gatto D. A. Labidi J. Ind. Crops Prod. 2014;55:149–154. doi: 10.1016/j.indcrop.2014.01.023. [DOI] [Google Scholar]
  35. Li X. H. Wu S. B. BioResources. 2014;9:6277–6289. [Google Scholar]
  36. Weinwurm F. Drljo A. Waldmüller W. Fiala B. Niedermayer J. Friedl A. J. Cleaner Prod. 2016;136:62–71. doi: 10.1016/j.jclepro.2016.04.048. [DOI] [Google Scholar]
  37. Shao L. Zhang X. Chen F. Xu F. J. Anal. Appl. Pyrolysis. 2017;128:27–34. doi: 10.1016/j.jaap.2017.11.003. [DOI] [Google Scholar]
  38. Zhu W. Westman G. Theliander H. Nord. Pulp Pap. Res. J. 2016;31:270–278. doi: 10.3183/npprj-2016-31-02-p270-278. [DOI] [Google Scholar]
  39. Toledano A. García A. Mondragon I. Labidi J. Sep. Purif. Technol. 2010;71:38–43. doi: 10.1016/j.seppur.2009.10.024. [DOI] [Google Scholar]
  40. Huang C. He J. Narron R. Wang Y. Yong Q. ACS Sustainable Chem. Eng. 2017;5:11770–11779. doi: 10.1021/acssuschemeng.7b03415. [DOI] [Google Scholar]
  41. Humpert D. Ebrahimi M. Czermak P. Membr. 2016;6:42. doi: 10.3390/membranes6030042. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Toledano A. Serrano L. Garcia A. Mondragon I. Labidi J. Chem. Eng. J. 2010;157:93–99. doi: 10.1016/j.cej.2009.10.056. [DOI] [Google Scholar]
  43. Rohde V. Böringer S. Tübke B. Adam C. Dahmen N. Schmiedl D. GCB Bioenergy. 2019;11:206–217. doi: 10.1111/gcbb.12546. [DOI] [Google Scholar]
  44. Beaudoin D. Langis-Barsetti S. Gagné A. Palus E. Inwood J. Konduri M. K. J. Wood Chem. Technol. 2024;44:133–144. doi: 10.1080/02773813.2024.2314454. [DOI] [Google Scholar]
  45. Melro E. Filipe A. Sousa D. Valente A. J. Romano A. Antunes F. E. Medronho B. Int. J. Biol. Macromol. 2020;148:688–695. doi: 10.1016/j.ijbiomac.2020.01.153. [DOI] [PubMed] [Google Scholar]
  46. Li M. Pu Y. Chen F. Ragauskas A. J. New Biotechnol. 2021;60:189–199. doi: 10.1016/j.nbt.2020.10.005. [DOI] [PubMed] [Google Scholar]
  47. Wang Y. Y. Li M. Wyman C. E. Cai C. M. Ragauskas A. J. ACS Sustainable Chem. Eng. 2018;6:6064–6072. doi: 10.1021/acssuschemeng.7b04546. [DOI] [Google Scholar]
  48. Yu L. Seabright K. Bajaj I. Keffer D. J. Alonso D. M. Hsieh C. T. Li M. Chen H. Dai S. Gandomi Y. A. Maravelias C. T. Harper D. P. Front. Energy Res. 2022;10:849949. doi: 10.3389/fenrg.2022.849949. [DOI] [Google Scholar]
  49. Mishra P. K. Ekielski A. Colloids Interfaces. 2019;3:52. doi: 10.3390/colloids3020052. [DOI] [Google Scholar]
  50. Hu Z. Du X. Liu J. Chang H. M. Jameel H. J. Wood Chem. Technol. 2016;36:432–446. doi: 10.1080/02773813.2016.1214732. [DOI] [Google Scholar]
  51. Haridevan H. Evans D. A. Ragauskas A. J. Martin D. J. Annamalai P. K. Green Chem. 2021;23:8725–8753. doi: 10.1039/D1GC02744A. [DOI] [Google Scholar]
  52. Li C. Jin H. Hou M. Guo X. Xiao T. Cao X. Jia W. Fatehi P. Shi H. Int. J. Biol. Macromol. 2024;256:128290. doi: 10.1016/j.ijbiomac.2023.128290. [DOI] [PubMed] [Google Scholar]
  53. Guo W. Sun S. Wang P. Chen H. Zheng J. Lin X. Qin Y. Qiu X. Int. J. Biol. Macromol. 2022;221:913–922. doi: 10.1016/j.ijbiomac.2022.09.074. [DOI] [PubMed] [Google Scholar]
  54. Allegretti C. Fontanay S. Krauke Y. Luebbert M. Strini A. Troquet J. Turri S. Griffini G. D'Arrigo P. ACS Sustainable Chem. Eng. 2018;6:9056–9064. doi: 10.1021/acssuschemeng.8b01410. [DOI] [Google Scholar]
  55. Sevastyanova O. Helander M. Chowdhury S. Lange H. Wedin H. Zhang L. Ek M. Kadla J. F. Crestini C. Lindström M. E. J. Appl. Polym. Sci. 2014;131:18. doi: 10.1002/app.40799. [DOI] [Google Scholar]
  56. Liu L. Y. Hua Q. Renneckar S. Green Chem. 2019;21:3682–3692. doi: 10.1039/C9GC00844F. [DOI] [Google Scholar]
  57. Gong X. Meng Y. Lu J. Tao Y. Cheng Y. Wang H. Macromol. Chem. Phys. 2022;223:2100434. doi: 10.1002/macp.202100434. [DOI] [Google Scholar]
  58. Lu X. Gu X. Int. J. Biol. Macromol. 2023;229:778–790. doi: 10.1016/j.ijbiomac.2022.12.322. [DOI] [PubMed] [Google Scholar]
  59. Hata T. Nonaka H. Biomass Convers. Biorefin. 2018;8:339–343. doi: 10.1007/s13399-017-0282-6. [DOI] [Google Scholar]
  60. Gong Z. Shuai L. Trends Chem. 2023;5:163–166. doi: 10.1016/j.trechm.2022.12.005. [DOI] [Google Scholar]
  61. Beisl S. Loidolt P. Miltner A. Harasek M. Friedl A. Molecules. 2018;23:633. doi: 10.3390/molecules23030633. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Hubbe M. A. Alén R. Paleologou M. Kannangara M. Kihlman J. BioResources. 2019;14:1–21. [Google Scholar]
  63. Pylypchuk I. V. Riazanova A. Lindström M. E. Sevastyanova O. Green Chem. 2021;23:3061–3072. doi: 10.1039/D0GC04058D. [DOI] [Google Scholar]
  64. Schneider W. D. H. Dillon A. J. P. Camassola M. Biotechnol. Adv. 2021;47:107685. doi: 10.1016/j.biotechadv.2020.107685. [DOI] [PubMed] [Google Scholar]
  65. Brenelli L. B. Mariutti L. R. B. Portugal R. V. de Farias M. A. Bragagnolo N. Mercadante A. Z. Franco T. T. Rabelo S. C. Squina F. M. Ind. Crops Prod. 2021;167:113532. doi: 10.1016/j.indcrop.2021.113532. [DOI] [Google Scholar]
  66. Quezada G. López F. Romero P. González G. ACS Omega. 2025;10:16138–16146. doi: 10.1021/acsomega.4c09697. [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Lan W. Luterbacher J. S. Chimia. 2019;73:591. doi: 10.2533/chimia.2019.591. [DOI] [PubMed] [Google Scholar]
  68. Ye Q. Yokoyama T. J. Wood Sci. 2020;66:80. doi: 10.1186/s10086-020-01928-6. [DOI] [Google Scholar]
  69. Saito T. Perkins J. H. Vautard F. Meyer H. M. Messman J. M. Tolnai B. Naskar A. K. ChemSusChem. 2014;7:221–228. doi: 10.1002/cssc.201300509. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

RA-016-D6RA03462D-s001
RA-016-D6RA03462D-s001.pdf (1.7MB, pdf)

Data Availability Statement

The data supporting the findings of this study are available within the article and its supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6ra03462d.

Articles from RSC Advances are provided here courtesy of Royal Society of Chemistry

RESOURCES