Highly Efficient Semi-Continuous Extraction and In-Line Purification of High β-O-4 Butanosolv Lignin

Innovative biomass fractionation is of major importance for economically competitive biorefineries. Lignin is currently severely underutilized due to the use of high severity fractionation methodologies that yield complex condensed lignin that limits high-value applicability. Mild lignin fractionation conditions can lead to lignin with a more regular C-O bonded structure that has increased potential for higher value applications. Nevertheless, such extraction methodologies typically suffer from inadequate lignin extraction efficiencies and yield. (Semi)-continuous flow extractions are a promising method to achieve improved extraction efficiency of such C-O linked lignin. Here we show that optimized organosolv extraction in a flow-through setup resulted in 93–96% delignification of 40 g walnut shells (40 wt% lignin content) by applying mild organosolv extraction conditions with a 2 g/min flowrate of a 9:1 n-butanol/water mixture with 0.18 M H2SO4 at 120°C in 2.5 h. 85 wt% of the lignin (corrected for alcohol incorporation, moisture content and carbohydrate impurities) was isolated as a powder with a high retention of the β-aryl ether (β-O-4) content of 63 linking motifs per 100 C9 units. Close examination of the isolated lignin showed that the main carbohydrate contamination in the recovered lignin was butyl-xyloside and other butoxylate carbohydrates. The work-up and purification procedure were investigated and improved by the implementation of a caustic soda treatment step and phase separation with a continuous integrated mixer/separator (CINC). This led to a combined 75 wt% yield of the lignin in 3 separate fractions with 3% carbohydrate impurities and a very high β-O-4 content of 67 linking motifs per 100 C9 units. Analysis of all the mass flows showed that 98% of the carbohydrate content was removed with the inline purification step, which is a significant improvement to the 88% carbohydrate removal for the traditional lignin precipitation work-up procedure. Overall we show a convenient method for inline extraction and purification to obtain high β-O-4 butanosolv lignin in excellent yields.


Continuous integrated mixer/separator (CINC)
The CINC, a 3535-B Arrowhead Drive (2002) has a max rpm of 6000, max throughput of 2 L/min, 200 g filling mass and 0-100 o C temperature range and is connected to the extraction liquor output with a 2 mm silicon tube. Synthesis of (2R/S,3R,4S,5R)-2-butoxytetrahydro-2H-pyran-3,4,5-triol (butyl xyloside): 10 g of D-xylose was heated at reflux conditions in a 90% n-BuOH/10% H 2 O mixture containing 0.18 M H 2 SO 4 for 2 hours. The reaction mixture was concentrated in vacuo to give the desired product as a viscous yellow oil which was characterized by NMR-spectroscopy without further purification.

Gel Permeation Chromatography (GPC):
The molecular weight of the isolated butanosolv lignin was determined by gel permeation chromatography (GPC) analysis. The analysis was carried out on a Hewlett Packard 1100 series THF-GPC. The lignin samples (10 mg) were dissolved in THF with toluene as a flow marker. Prior to analysis, the samples were filtered with a syringe filter (0.45 µm, PTFE). Analysis was performed with PSS WinGPC UniChrom.
Biomass analysis: Analysis of the carbohydrate and lignin content of the biomass sources and residual samples were performed following the NREL methodology (Sluiter et al. 2008). A 1,000 g dried raw biomass or residual biomass sample was added to a round bottom flask and 10 mL of a 72% aqueous H 2 SO 4 was added. The resulting solution was incubated at 30 °C for 1 hour. After 1 hour, 280 mL of deionized water was added in order to obtain a 4% aqueous H 2 SO 4 solution. The mixture was heated at reflux conditions for 1 hour and afterwards allowed to cool down to room temperature. Acid insoluble lignin was obtained by vacuum filtration and dried overnight in a vacuum oven at 60 °C. The amount of acid soluble lignin was determined by UV-VIS spectroscopy, using a background of a 4% sulfuric acid in deionized water. The samples were diluted in order to obtain an absorbance in the range of 0.7 -1.0 at wavelength 240 nm.
The carbohydrate composition of the original material and the obtained residues (oven dried) was analyzed with HPAEC-PAD after acid hydrolysis (samples were diluted 20 -50 times before injection). Therefore, 11 mg of each sample was mixed with 0,9 mL 72% (w/w) sulfuric acid and incubated for 1h at 30°C. After that 9.9 mL milliQ water was added, diluting the acid to 1 M. The incubation was continued for 3h at 100°C to fully hydrolyze the carbohydrates into monomers. All monomers were quantified by integrating the peak area of corresponding standards (arabinose, galactose, glucose, glucuronic acid, mannose, rhamnose, and xylose). Total sugar content was calculated as a sum of all neutral sugars and glucuronic acid .
High performance anion exchange chromatography (HPAEC) on a Dionex Ultimate 6000 system (Thermo Scientific, Sunnyvale, CA, USA) equipped with a CarboPac PA-1 column (2 mm x 250 mm ID) in combination with a CarboPac PA-1 guard column (2 mm x 50 mm ID) and PAD detection. System was controlled by the Chromeleon 7.2.9 software (Thermo Scientific, Sunnyvale, CA, USA). Elution of monosaccharides (0.25 mL min-1) was performed with a multi-step-gradient using the following eluents: A: 0.1M NaOH, B: 1M NaOAc in 0.1M NaOH, C: 0.2 M NaOH, and D: milliQ water. All analyzed monosaccharides elute in the first 20 min with 16% A, 84% D. Followed by 5 min with 45%A, 5%B, 50%D and 15 min with 60%A, 40% B. To regenerate the column it was flushed 12 min with 100% C by increasing flowrate in first 2 min to 0.35 mL min-1. Finally, the column is equilibrated for 12 min with 16%A, 84%D by decreasing the flowrate in the first 2 min to 0.25 mL min-1.

Moisture content measurement:
The moisture content of the biomass and the isolated lignin was determined with a PCE-MA 110 moisture meter. An ascension program to 120 °C was used during which the weight of the sample is tracked. At 120 °C the measurement is stopped if between 2 iterations the weight change is less than 0.2%. The obtained data was used to correct all yield calculations.

Sankey diagram:
The Sankey diagram was made based on the following data and assumptions.
The reported values between brackets were determined by the NREL analysis of the biomass.
The other values are based on the determined mass of the isolated solid product (corrected for moisture content, butoxy-incorporation and carbohydrate impurities). It is assumed that no loss of product occurs in the separation steps. The reported value for lignin lost is the difference of the isolated lignin and the determined Klason Lignin content. The online tool SankeyMatic was used to generate the Sankey diagrams.  Table S4: Solvent efficiency of a wide range of lignin extractions, all showing a lower solvent efficiency than our reported extraction procedure in this work. (Bauer et al., 2012;Wildschut et al., 2013;Shuai et al., 2016;Lancefield et al., 2017;Smit and Huijgen, 2017;Wang et al., 2017;Ouyang et al., 2018;Weidener et al., 2018;Dong et al., 2019) 3. Extraction and work-up flow schemes Figure S2: Schematic representation of the influence of the reactor loading on the biomass level. The small cylindrical tube used as reactor causes an increased contact time of the solvent with biomass, resulting in an increased lignin extraction.   Figure S6: Precipitation of the concentrated butanosolv extraction liquor in water in the absence of sodium sulfate (left) and precipitation with sodium sulfate added to the water (right) which clearly enhances lignin flocculation. Figure S7: Solubility test of butyl-xyloside in water, clearly showing the mediocre solubility of modified xylose. During the lignin precipitation step part of the butyl-xyloside will also precipitate and be isolated as a solid product. Figure S8: 2D-HSQC NMR (d6-acetone) overlap of D-Xylose (red) and butyl-xyloside (green).    Table S5: Aromatic -OH and Aliphatic -OH content of butanosolv lignin and prior and after caustic soda treatment (5-substituted -OH was set at 1.00) as determined by 31 P-NMR analysis. Figure S11: Molecular weight distribution of the organic phase lignin (OPL), Cinc residual lignin (CRL) and water soluble lignin (WSL) fraction after biphasic treatment determined by GPC.

Lignin type
Weight average molecular weight (Da) Organic phase lignin 3900 CINC residual lignin 6900 Water Soluble lignin 1900 Table S6: Weight average molecular weight of the different lignin isolated with biphasic extraction. Figure S12: 2D HSQC NMR spectrum (d6 acetone) of the product obtained after butanosolv extraction of pine wood.