Expanding lignin thermal property space by fractionation and covalent modification

To fully exploit kraft lignin's potential in material applications, we need to achieve tight control over those key physicochemical lignin parameters that ultimately determine, and serve as proxy for, the properties of lignin-derived materials. Here, we show that fractionation combined with systematic (incremental) modification provides a powerful strategy to expand and controllably tailor lignin property space. In particular, the glass transition temperature (Tg) of a typical kraft lignin could be tuned over a remarkable and unprecedented 213 °C. Remarkably, for all fractions the Tg proved to be highly linearly correlated with the degree of derivatisation by allylation, offering such tight control over the Tg of the lignin and ultimately the ability to ‘dial-in’ this key property. Importantly, such control over this proxy parameter indeed translated well to lignin-based thiol–ene thermosetting films, whose Tgs thus covered a range from 2–124 °C. This proof of concept suggests this approach to be a powerful and generalisable one, allowing a biorefinery or downstream operation to consciously and reliably tailor lignins to predictable specifications which fit their desired application.


Sample Description
For naming of samples, suffixes were appended to fraction names to distinguish between allylated samples, where non-allylated is denoted with 0, and sequential numbers indicate higher allyl bromide loading (1-6); e.g., the non-allylated methanol fraction sample is "FMeOH-0", and the highest loading FMeOH is "FMeOH-6".

Equipment
All NMR measurements were recorded at 25 o C on either a Varian VNMRS 400 with a PFG probe, or an Agilent MRF400 spectrometer equipped with a OneNMR probe and Optima Tune system at 400 MHz. NMR samples were measured in standard 5 mm-OD NMR tubes and chemical shifts (δ) are reported in ppm. Acquired data was analysed and processed using

Gel Permeation Chromatography (GPC)
Lignin samples were acylated prior to GPC measurements to ensure proper dissolution in the eluent. Lignin (50 mg) was dissolved in pyridine/acetic anhydride (1:1, 1 mL) and stirred overnight at room temperature. The acylated sample was recovered azeotropically, first by addition of toluene (500 μL x 5), then with EtOH (500 μL x 5) to yield a dry brown powder. Some samples required further addition of toluene (3 times) and EtOH (3 times) to fully remove trace impurities. The acylated samples were then dissolved in the tetrahydrofuran (THF) eluent spiked with acetic acid (1.5 mg/mL, 0.1% AcOH/THF v/v) and filtered with a 45 μm PTFE syringe filter before GPC analysis. UV detection was performed at a wavelength of 280 nm and molecular weight determinations were based on calibrations with polystyrene standards (Mn = 104, 208, 312, 416, 520, 625, 729, 833, 937, 1930, 2900, 3790, 6180, 10110, 16500, 24600, 38640 gmol -1 ). As polystyrene evidently doesn't have the same structure as lignin, its use as proxy calibration standard means that reported molecular weights need to be interpreted with some caution in relation to a sample's absolute molecular weight. The use of PS is nevertheless common and allows for benchmarking with existing literature on kraft lignins and their derivative fractions. [1,2] Modulated Differential Scanning Calorimetry (MDSC) Lignin (5 mg) was weighed into a Tzero low-mass aluminium pan, fitted with a Tzero lid which was drilled to produce a venting hole (0.8 mm). Once the pan was loaded into the DSC cell, the sample was equilibrated to 20 °C. Two heating cycles were then performed. For the first cycle, the sample was heated to 105°C at 10 °C/min, was held isothermal for 20 min, cooled to -50 °C at 10 °C/min and held isothermal for 1 min to anneal the sample, remove any residual solvents and to erase the sample's thermal history. Data acquisition was performed on the second heating cycle, whereby the temperature was ramped at 3 °C/min from -50 o C to 190/250 °C with a sinusoidal temperature modulation of 0.66 °C per 50 s. Each sample was measured in duplicate, and the average value was taken.

Attenuated Total Reflectance Fourier Transform Infra-red Spectroscopy (ATR-FTIR)
Before each measurement, a background measurement was taken to remove residual unnecessary signals.
A small quantity of sample was placed onto the diamond aperture, ensuring total coverage. The toner arm was lowered to press the sample onto the diamond and the spectrum was measured with a spectral range of 600-4000 cm -1 with 4 co-added scans. The resulting spectra were then background corrected within the PerkinElmer Spectrum IR software.

Synthetic Methods
Fractionation [3] 100 g of Stora Enso Kraft lignin, ground by pestle and mortar and sieved (450 μm), was suspended in 1000 mL of ethyl acetate in a round bottom flask and stirred for 2 h at room temperature. The insoluble fraction was then separated from the solution by vacuum filtration. The insoluble residue was then air-dried, re-suspended, stirred and extracted under the same conditions with 1000 mL of the subsequent solvent. The soluble fractions were isolated by removal of solvent in vacuo. To minimise residual traces of organic solvent, each fraction was wetted with water to form a paste and dried through rotary evaporation, followed by drying under a gentle stream of compressed air overnight, followed by 24 h in a vacuum oven at 65 o C. The chosen solvent order was ethyl acetate, ethanol, methanol, and acetone, yielding four soluble fractions and an insoluble residual fraction.
Allylation of Lignin [4] Lignin samples were hand ground with a pestle and mortar until a visually homogeneous particle size was achieved. The ground lignin (1 g) was dissolved in an acetone/NaOH solution (1:3, 40 mL). Different volumes of allyl bromide (0.11−3.29 equiv. vs total lignin [OH]) were carefully added through a septum, to the magnetically stirring mixture. The reaction was held at 40 °C for 5 h under a nitrogen atmosphere. The reaction mixture was then reduced in vacuo to remove the acetone and the crude mixture was acidified to pH 2 by adding 35% HCl dropwise. The resulting slurry was purified by centrifuging at 8000 rpm for 5 minutes, decanting the supernatant, followed by washing with demineralised water (10 mL), and centrifuging again. The washing steps were repeated two more times and afterwards the sample was air-dried under a gentle stream of compressed air overnight, followed by 24 h in a vacuum oven at 60 o C.
Thiol-ene Cross-coupling of Lignin [5] Trimethylolpropane tris(3-mercaptoprionate) (0.333 equiv. vs allyl content of lignin) was added to approximately 60 mg of allylated lignin samples in a vial containing 135 μL EtOAc and stirred until homogenous. The mixture was then transferred via pipette to a microscopy glass viewing plate and the solvent left to evaporate for 30 minutes, before curing at 120 o C for 24 h.
Synthesis of 3,3'-dimethoxy-5,5'-dipropyl-[1,1'-biphenyl]-2,2'-diol [6] An aqueous sodium acetate solution was prepared (18.0 g, 100 mL), after which 4-propylguaiacol (16.8 g, 0.10 mol) was added at room temperature. To this vigorously stirring suspension, an aqueous solution of K3[Fe(CN)6] (35.0 g, 0.11 mol, 300 mL) was added dropwise. After total addition, the reaction was left to stir for 24 h, after which the mixture was extracted with DCM (4 x 200 mL). The organic layers were combined, dried with magnesium sulphate, filtered, and concentrated in vacuo to yield a white crystalline solid, which showed resonances in 13 C & 1 H NMR which corresponded to shifts, multiplicities, and peak integrations reported in the literature. [6] Allylation of model compounds to form (1) and (2) Modified from a generalised protocol. [7] To a stirred solution of phenolic compound (605 μmol) in acetone (8 mL) was added K2CO3 (0.166 g, 1.210 mmol), followed by allyl bromide (1.45 mmol, 125 μL) at room temperature. Stirring was then continued for 2 h under reflux, after which the reaction mixture was left to cool. The suspension was then filtered over a silica plug and reduced in vacuo. Following this, the solution was redissolved in DCM (8 mL) and extracted with H2O (3 x 8 mL). Finally the conversion was checked by thin layer chromatography (TLC) (hexane:EtOAc, 9:1, rf = 0.29) and if necessary, the product was purified by flash chromatography (hexane:EtOAc, 9:1).  * Peak overlaps with d6-DMSO peak, 1 H NMR was re-run in CDCl3 for integration of this resonance.

NMR Methods
Quantitative 31 P NMR [8] The peak assignments and integration ranges were based on the work of Argyropoulos [8] and measurements were performed with the internal standard cyclohexanol, as reported previously by our group. [1,9,10] A dry solvent mixture of deuterated chloroform/pyridine (1:1.6, v/v) was prepared and stored over 3 Å molecular sieves. Separate stock solutions of the internal standard (cyclohexanol, 19 mg/mL) and the paramagnetic relaxation agent (chromium (III) acetylacetonate, 5 mg/mL) were prepared using the solvent mixture. 500 μL of the solvent mixture was added to approximately 40 mg of accurately weighed lignin, followed by internal standard and relaxation agent solutions (200 μL & 50 μL, respectively) at room temperature. The sample was stirred for at least an hour and up to overnight to achieve total dissolution. Immediately prior to measurement, the sample was phosphitylated by addition of TMDP (100 μL) and the mixture was allowed to stir for 1 minute before transfer to an NMR tube. A standard phosphorous pulse program with a pulse angle of 45 o , relaxation delay of 10 s and 512 co-added scans was used. Peak shifts were referenced by the sharp peak arising from reaction of residual water with TMDP at 132.2 ppm. Spectra were processed by application of a 3 rd order polynomial baseline correction, followed by automatic phase adjustment, which provided adequate internal standard peak shape.

Heteronuclear Single Quantum Coherence (HSQC) NMR
Lignin (200 mg) was dissolved in 750 μL d6-DMSO overnight. Spectra were recorded using a "gc2hsqcse" pulse sequence and optimised for 145 Hz 1 J with multiplicity editing turned off. The spectral width was -2 to 14 ppm in f2 ( 1 H) and -10 to 190 in f1 ( 13 C) with 256 scans in f2 and 64 scans in f1. For all spectra, a 3 rd order polynomial baseline correction in both dimensions and automatic phase correction was performed. The spectra were apodised in f2 (LB = 0.3 Hz, GB = 0.1) and a sine squared function set to 90 o was applied in both f1 and f2. Spectral shifts were referenced to the shift of the solvent peak for HSQC measurements. Peak assignments were based on existing literature. [9] Fractionation, NMR & DSC Data  [11] and [b] Duval et. al. [3] Yield (%) Fraction