1

Electronic Supplementary Information (ESI)

Promising bulk production of a potentially benign bisphenol A replacement from a hardwood lignin platform

S.-F Koelewijn,a C. Cooreman,b T. Renders,a C. Andecochea Saiz,b S. Van den Bosch,a W. Schutyser,a W. De Leger,c M. Smet,c P. Van Puyvelde,d H. Witters,e B. Van der Bruggen*b and B.F. Sels*a

a. Dept. of Microbial and Molecular Systems (M2S), Centre for Surface Chemistry and Catalysis (COK),

KU Leuven, Celestijnenlaan 200F, 3001 Leuven, Belgium b. Dept. of Chemical Engineering, Process Engineering for Sustainable Systems (ProcESS),

KU Leuven, Celestijnenlaan 200F, 3001 Leuven, Belgium c. Dept. of Chemistry, Polymer Chemistry and Materials,

KU Leuven, Celestijnenlaan 200F, 3001 Leuven, Belgium d. Dept. of Chemical Engineering, Soft Matter, Rheology and Technology (SMaRT),

KU Leuven, Celestijnenlaan 200F, 3001 Leuven, Belgium e. Dept. of Environmental Risk and Health, Applied Bio & molecular Systems (ABS),

Flemish Institute for Technological Research (VITO), Boeretang 200, 2400 Mol, Belgium

*Corresponding authors. E-mail: bart.vanderbruggen@kuleuven.be / bert.sels@kuleuven.be

I. Materials and methods 2

II. Tables 11

III. Figures 14

IV. References 24

Electronic Supplementary Material (ESI) for Green Chemistry.This journal is © The Royal Society of Chemistry 2018

2

I. Materials and methods

Chemicals and materials

All commercial chemicals were analytical reagents and were used without further purification. 5 wt% Ru on carbon, silica gel (pore

size 60 Å, 70-230 mesh, 63-200 µm), formaldehyde solution (37 wt% in H2O), N-methyl-N-(trimethylsilyl) trifluoroacetamide

(MSTFA, ≥98.5%), n-hexane (>95%, 0.659 g/mL at 20oC), anhydrous pyridine (99.8%), dichloromethane (DCM, >99%),

tetrahydrofuran (THF, >99%), toluene (>99%), methanol (>99%), ethanol (>99%), chloroform-d (CDCl3, 99.8 atom% D),

triethylamine (≥99%), 17β-oestradiol (>98%), benzyltriethylammonium chloride (BTEAC, 99%), terephthaloyl chloride (TPC, ≥99%)

and potassium trifluoroacetate (KTFac, 98%) were purchased from Sigma-Aldrich. Concentrated hydrochloric acid (HCl, 37wt%),

sodium hydroxide (NaOH, 99.4%), anhydrous magnesium sulphate (MgSO4, >99%), anhydrous acetone (≥99.5%) and anhydrous

diethyl ether (≥99%) were purchased from Fisher Scientific. Bisphenol A (BPA, 2,2’-bis(4-hydroxyphenyl)propane, >99.0%) was

purchased from TCI Europe. Syringol (S, 2,6-dimethoxyphenol, 99%), n-heptane (99+%) and acetonitrile (ACN, 99.9+%) were

purchased from Acros Organics. Dimethylsulfoxide (DMSO, 99.5%) was purchased from Labscan. 4-Methylsyringol (MS, 2,6-

dimethoxy-4-methylphenol, >97%) and chloroform (HPLC grade, 99.5%) were purchased from Alfa Aesar. 1,1,1,3,3,3-hexafluoro-

2-propanol (HFIP, 99%) was purchased from Fluorochem. 4-n-Propylsyringol (PS, 2,6-dimethoxy-4-n-propylphenol, >98%) was

produced via reductive catalytic fractionation (RCF) of birch wood (see section 1.2). Birch (Betula pendula) was obtained from a

local sawmill (Ecobois, Ghent, Belgium). Water was purified using a Millipore Milli-Q Advantage A10 water purification system to

a resistivity higher than 18 MΩ.cm.

Methods and procedures

Sawdust preparation

Dry birch wood was milled and sieved to obtain a sawdust fraction with a size of 1.5-15 mm. Subsequently, a two-step extraction

procedure was followed using a Soxtex 2055 Avanti apparatus to remove any extractives like fats, waxes, resins and

terpenoids/steroids1 which can interfere with analysis procedures (e.g. determination of the Klason lignin content). Porous

thimbles were filled with ~2.5 g sawdust, and were completely submersed for 15 minutes in 70 mL of a boiling solvent mixture

comprising 33 vol% ethanol in toluene. Next, a standard Soxhlet extraction step was executed in which the thimbles were kept

above the boiling mixture for 3 h. After cooling, samples were washed with ethanol and dried overnight at 80oC. Because

completely dry sawdust is hygroscopic and difficult to handle, the sawdust was stored in an open recipient to equilibrate with air

humidity for minimum 24 h, resulting in a water uptake of circa 4 wt%. This sawdust, hereinafter referred to as ‘pre-extracted

sawdust’, was used for catalytic hydrogenolysis.

Determination of Klason lignin content

Product yields in lignin depolymerisation literature are typically based on the amount of acid insoluble lignin, also called Klason

lignin, in the lignocellulose sample. The determination of the Klason lignin content of birch was based on a procedure from Lin &

Dence.2 Triplicate samples of pre-extracted sawdust (1 g each) were transferred to 50 mL beakers after which 15 mL of a 72 wt%

H2SO4-solution was added. The mixture was left at room temperature for 2 h while continuously stirred with a magnetic rod.

Afterwards the content of each beaker was transferred to a round-bottom-flask which already contained 300 to 400 mL of water.

The beakers were rinsed and additional water was added until a H2SO4 concentration of 3 wt% was reached. The diluted solution

was boiled for 4 h under reflux conditions, to maintain a constant volume and acid concentration. After filtration of the hot

solution, a brown lignin precipitate was retained. The precipitate was washed with hot water to remove any leftover acid and the

obtained residue was dried at 80oC overnight. The Klason lignin content (20.6 wt%) was determined relative to the oven dried

substrate by averaging the measured weight of the residues.

3

Isolation of PS from birch sawdust

For the catalytic hydrogenolysis experiment, pre-extracted birch sawdust (60 g; 20.6 wt% Klason lignin), catalyst powder (9 g Ru/C;

15 wt%) and methanol (300 mL) were loaded into a 600 mL stainless steel batch reactor (Parr Instruments & Co.). The reactor was

sealed, flushed threefold with N2, and pressurised with H2 (10 bar at room temperature). Subsequently, the reaction mixture was

stirred (500-750 rpm) and heated to 235oC (~ 15oC.min-1). When the reaction temperature was reached, the temperature was

kept constant for 4 h after which the reactor was cooled and depressurised at room temperature. Afterwards, the reactor contents

were quantitatively collected by washing extensively with methanol followed by a filtration step to separate the solid residue (i.e.,

pulp and catalyst) from the liquid product mixture.

After reductive processing, the liquid and solid reactor contents were quantitatively collected and separated from each other over

a glass fritted filter. Subsequently, the methanol was evaporated. In this way a light-brown oil was obtained, which was subjected

to a threefold liquid-liquid extraction using dichloromethane (DCM) and water to separate the soluble lignin products from sugar-

derived products or other polar components. Finally, the DCM-extracted phase was dried to obtain the ‘lignin oil’ (11.74 g;

containing mono-, di- and oligomers). More details about the composition of this lignin oil were reported previously.3,4 This lignin

oil was subjected to a threefold reflux extraction with n-hexane (3 x 25 mL) and the extract was distilled in vacuo to obtain a

transparent yellow oil (6.53 g; containing mainly monomers) rich in PS (see Table S4). The product was further purified by column

chromatography (25% v/v acetone in n-heptane) on silica gel, to remove guaiacyl monomers and minor dimeric impurities. This

solvent system was previously reported to be a sustainable alternative to commonly used ethyl acetate-hexane mixtures.5

Yield: 4.24 g (34 % on lignin basis). 1H NMR (300 MHz, CDCl3, 25oC, TMS): δH = 0.94 (t, 3J(H,H)= 7.3 Hz, 3H; -CH2CH3), 1.62 (sex,

3J(H,H)= 7.3 Hz, 2H; -CH2CH2CH3), 2.51 (t, 3J(H,H)= 7.3 Hz, 2H; -ArCH2CH2-), 3.88 (s, 6H; -OCH3), 5.36 (s, 1H; -ArOH) and 6.40 ppm

(s, 2H; -m-ArH); 13C NMR (400 MHz, CDCl3, 25oC, TMS): δC = 146.9 (C2/6), 133.9 (C4), 132.7 (C1), 105.1 (C3/5), 56.3 (Ce), 38.4 (Ca), 25.0

(Cb) and 13.9 ppm (Cc); MS (70 eV, EI): m/z (%): 196 (36) [M+•], 167 (100) [M+•-•CH2CH3]

Early-stage feasibility estimation of distillation as a separation technique

The relative volatility (α) is determined as:

𝛼 = 𝛼𝑖,𝑗 = 𝐾𝑖

𝐾𝑗

in which K is the equilibrium ratio of the vapor concentration to the liquid concentration and component i has a lower boiling

point than component j. K is directly related to the vapor pressure and the activity coefficient, by Raoult’s extended law. The

relative volatility is a direct indicator of the feasibility of distillation6,7 and should be interpreted as follows:

• If α = 1, component i and j are equally volatile, which means they cannot be separated by means of distillation.

• The higher the relative volatility α, the easier the separation of the respective components by distillation. In practice,

industrial scale distillations are rarely undertaken when the relative volatiliy is lower than 1.05.6

The vapor-liquid equilibrium ratios, i.e., the K-values, of PG and PS (the two main components in the birch lignin oil) were derived

from a simple Aspen Plus® simulation. The relative volatility αPG,PS was estimated to be 1.458, which indicates that distillation is

feasible. The K-values were derived at 278.69oC and 1 atm since this is close to the estimated boiling point of the lignin oil.

A critical note should be made: the K-values were estimated by a simulation with Aspen Plus®, which in turn based its calculations

on parameters estimated by the UNIFAC group contribution method. Furthermore, the simulation input is a simplification of the

real composition of the lignin oil. Therefore, the real relative volatility of PG and PS can differ from the calculated value of 1.458

at the operating conditions. Also, the relative volatility will differ depending on the chosen conditions.

Lange (2017)8 recently introduced the distillation resistance (Ω) to assess the potential of distillation as a separation process. The

distillation resistance Ω is defined as:

4

𝛺[𝑜𝐶−1] = 𝛺𝑓𝑒𝑒𝑑 = 100 𝑥 ∑𝑓𝑖

∆𝑇𝑖𝑖

in which fi represents the mass fraction (in wt%) of the total feed intake and ∆Ti the temperature difference (in oC) between the

atmospheric boiling point of component i and the next heavier component (when the components of a multicomponent mixture

are ranked from low to high boiling point). The distillation resistance is based on the assumption that the distillation duty is

determined by the duties needed to vaporise and condense the distillate streams. The distillation resistance should be interpreted

as follows:

• Ω < 1oC-1 : the distillation is simple and a heat duty of approximately 1 GJ per ton of feed is required.

• 3 ≤ Ω ≤ 7oC-1 : the distillation is demanding, requiring higher heat duties, estimated at 3-8 GJ per ton of feed.

Because the distillation cost has to be paid by the sales of the valuable product(s), Lange (2017)8 suggests the normalisation of

the distillation resistance with the weight fraction of valuable products fprod (in wt%), resulting in the following expression:

𝛺𝑝𝑟𝑜𝑑[𝑜𝐶−1] =𝛺𝑓𝑒𝑒𝑑

𝑓𝑝𝑟𝑜𝑑=

100 𝑥 ∑𝑓𝑖

∆𝑇𝑖𝑖

𝑓𝑝𝑟𝑜𝑑

In this article we define fprod as the total weight fraction of 4-n-propyl monomers, being the sum of PS and PG.

The overall distillation cost per ton of product with 25% capital charge and an energy price of $5 per GJ (based on data of 2003)

is expressed as:

𝐷𝑖𝑠𝑡𝑖𝑙𝑙𝑎𝑡𝑖𝑜𝑛 𝑐𝑜𝑠𝑡 [$ 𝑡𝑝𝑟𝑜𝑑−1 2003] = 60 𝑥

𝛺𝑝𝑟𝑜𝑑0.65

𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦 [𝑘𝑡𝑎−1]0.35 + 3 𝑥 𝛺𝑝𝑟𝑜𝑑

This model was tested for several distillation case studies.8 It was concluded that the the cost ceiling lies at $100 per ton of product

for biorefineries. This also corresponds to a ceiling of Ωprod of 15-17 for a production capacity of 200-400 kilotons per annum (kta-

1), which is commonly encountered in biorefineries. Consequently, if Ωprod is lower than this ceiling value, a distillation cost of less

than $100 per ton of product can be estimated for a biorefinery with a capacity of 200-400 kta-1, often promising a plausible

economic outcome.

The distillation resistance Ω is useful as a comparitive parameter for different distillation scenario’s, and for an early-stage

feasibility estimation. However, the parameter is restricted in accuracy and the limitations should be acknowledged:

• An absolute deviation of ± 1 GJ per ton of feed should be taken in account when estimating the heat transfer duty.

• Ω does not incorporate the energy duty needed to bring the feed to the distillation temperature. Lange (2017)8 notes

that this duty is often negligible, unless the distillate represents only a small fraction of the feed.

• Special attention should be paid to non-ideal behaving mixtures such as azeotropic mixtures.

For more details, analysis and examples the reader is kindly referred to the original work of Lange (2017).8

Input for distillation model

» Simplified input model for the monomer-enriched birch lignin oil (in n-hexane)

The lignin oil after extraction with n-hexane is a mixture of identified monomers (see Table S4), together with a small fraction of

dimers. This dimer fraction account for approximately 3.0 wt% of the lignin oil, and consists of a variety of apolar dimeric

compounds in low concentrations.3 Hence the dimeric fraction will be neglected in the simulation model. This simplification is

reasonable for the following reasons:

• The dimeric fraction consists of a variety of compounds in low concentrations. Consequently, a single dimeric

component will not have a significant influence on the simulation results.

5

• The dimeric compounds are significantly heavier than the monomeric compounds. In a distillative separation, the

dimeric compounds stay in the bottom of the column.

• Improving the efficiency of the n-hexane extraction, by for example controlling the contact time, mixing regime, or

temperature, could lower the dimer fraction in the lignin oil even further.

The concentrations and boiling points of the identified monomers are provided in Table S4.

» Thermodynamic model: UNIFAC

Adequate physical properties are essential to obtain an accurate simulation model with Aspen Plus®. Table S4 indicates that some

compounds of the lignin oil are not available in the database of Aspen Plus®, and that due to lack of experimental data, their

physical property parameters are unknown. Consequently, an appropriate thermodynamic model was selected to estimate the

necessary parameters. The following factors were considered for the choice of the model:

• The components in the lignin oil are polar and are non-electrolytes.

• The applied operating pressure in the distillation column is expected to be lower than 10 bar.

• The interaction parameters of the lignin oil components are unknown, since some of the components are not available

in the Aspen Plus® database.

Based on these considerations and the recommendations by Carlson (1996),9 UNIFAC was chosen as the appropriate physical

property method. UNIFAC (UNIQUAC Functional-group Activity Coefficient) is a group contribution method. Each compound is

considered as a collection of functional groups with a specific volume and surface area. Each functional group will influence local

interactions both due to its size and energy.6 UNIFAC can therefore predict local interaction coefficients based on the composition

of the compounds, without the need of experimental data. Since no experimental data are available, UNIFAC and its recent

improvements in Aspen Plus® are the best predictive method currently available.10

Modelling of continuous distillation on industrial scale with Aspen Plus®

The industrial scale distillation unit was simulated as a continuous process in Aspen Plus® (see configuration in Fig. S1). The aim

was to design a distillation train that can separate 100 kmol/hr (i.e., 85 ktfeed.a-1) of monomer-enriched birch lignin oil dissolved

in n-hexane (LIG-OIL) and supplied at 25oC and atmospheric pressure.

The first column (SEP001) for the straightforward n-hexane removal was designed without prior shortcut design calculations. It

was presumed that a nearly ideal split could be achieved and that rigorous simulation would not be too challenging. Therefore, a

simulation column with 10 equilibrium stages (N) and a reflux ratio (R) of 3 was directly erected in Aspen Plus®.

The Fenske-Underwood-Gilliland (FUG) method was used to estimate the design parameters (Nmin, N, Rmin and R) for the second

(vacuum) column (SEP002; see Table S2). As a design goal to initiate the calculations, an ambitious molar split of 99.9% was aimed

for PS and a split of 99% for PG. The same design parameters were also determined for a less ambitious split of 95% for PS and

90% for PG. The second column for the separation of PS (and PG) was designed to have 57 stages and a reflux ratio of about 10.

Next, a sensitivity analysis and design optimisation were conducted with Aspen Plus® to optimise the purity and recovery of PS

(and PG). The primary goal was to optimise the purity of the recovered PS-rich fraction as much as possible. In the sensitivity

analysis the influence of four parameters was investigated: i) the distillate-to-feed (D:F) ratio, ii) the feed stage, iii) the condenser

pressure and iv) the reflux ratio (see Table S3). The following settings and limitations were fixed for the sensitivity analysis and for

the optimisation:

• The recovery of PS is prioritised over the recovery of PG for birch wood lignin oil.

• The pressure drop per tray is fixed at 0.1 psi (±6.9 mbar) per tray, following a common-practice rule of thumb.6,11

• The number of stages of SEP001 was set at N = 10 and the reflux ratio was chosen to be R = 3.

6

• For SEP002, the number of stages was fixed at N = 57 and the reflux ratio was set at R = 10.10, as was calculated for a

99.9% molar split of PS and a 99% split of PG with the FUG-method.

Sizing of the distillation column SEP002

Two critical considerations justify a closer look at the required number of stages for this column:

• At the design stage, it was assumed that the number of equilibrium stages (N) is equal to the actual required number of

plates (Nact). However, the number of plates in a trayed column or the equivalent packing height in a packed distillation

column needs to be adjusted with a certain stage efficiency.6 Therefore, the required number of stages will likely be

higher than the calculated 57 equilibrium stages, adding to the height of the distillation column.

• The infrastructure cost of a distillation column is strongly related to the number of equilibrium stages (N) required for

the separation. Taking maintenance requirements into account, steel trays are often approximately 3 mm thick and are

typically spaced at a minimum distance of 30 cm per tray.12 For a column with 57 stages, this leads to a minimal

estimated column height of 17.27 meters. Reducing the column height by minimizing the number of equilibrium stages,

will most likely reduce capital investment costs.

Therefore, the sizing of the second column SEP002 was investigated with the developed Aspen Plus® model (see Fig. S2). Since

the sensitivity analysis showed a relatively small influence of the condenser pressure and the reflux ratio (R) on the purities of PS

(and PG), these values remained fixed at 0.1 bar and 10.10, respectively. Optimisation of sizing was performed with regard to

distillate-to-feed ratio (D:F) and the the feed stage.

Reaction procedure

Reactions were conducted as previously described.13 In a typical reaction, to alkylsyringol (36 mmol) in a 100 mL Duran® laboratory

bottle with screw cap, 37 wt% formaldehyde (18 mmol) and concentrated HCl (50 mL, 2.5 M in MilliQ H2O) were added, as well

as a magnetic stirring bar (45 mm). At the start of the reaction, the bottle is partially (2/3) submerged in a pre-heated stirred

temperature-controlled oil bath. The bath is held at 100oC under continuous stirring (1000 rpm) for the indicated time. Once at

room temperature, the aqueous acidic phase was decanted and the resulting ‘organic phase’ analysed.

Reaction analysis

Analysing mixtures of structurally resembling monomers, dimers and higher oligomers is not straightforward. Gas

chromatography (GC) and gel permeation/size exclusion chromatography (GPC/SEC) were adapted for this analysis and compared.

» Work-up

The ‘organic phase’ contains dimers, oligomers, unreacted monomers and trace amounts of the aqueous acidic phase (<3 wt%).

To avoid overestimation of ‘total organic phase recovery’, underestimation of product yields, or damaging the column, this

‘organic phase’ was dissolved in acetone or diethyl ether and dried with MgSO4. After filtration, the solvent was removed by rotary

evaporation (150 mbar at 80oC for 1 h) yielding a semi-solid or viscous oil. This ‘dried organic phase’ was analysed by both GC and

GPC/SEC for determining product yields and monomer conversion. Due to the high chemical similarity between reagent and

products, the dimer yield and monomer conversion can only be measured accurately by GC. Due to the limited volatility of

(derivatised) oligomers, the use of GC is excluded for their quantification. To confirm their presence, GPC/SEC measurements

were performed. The ‘total organic phase recovery’ is calculated from the ratio of the weight of the ‘dried organic phase’ and the

weight of the ‘theoretical organic phase at full dimer yield’ (cf. n mol dimer, resulting from the ideal stoichiometrical reaction

between the initial n mol formaldehyde and 2n mol methoxylated phenols without oligomer formation).

7

» Gas Chromatography (GC)

GC analysis was performed on a Hewlett-Packard (HP) 5890 with a CP-SIL 5CB WCOT fused silica column (30 m x 0.32 mm, film

thickness of 1.0 µm), equipped with an FID detector (310oC) and ChemStation software. The injection port and initial oven

temperatures were 300 and 35oC, respectively. This temperature was held for 4 min, increased to 300oC at 10oC.min-1, and held

there for 40 min. Prior to GC analysis, the samples were derivatised via trimethylsilylation with N-methyl-N-(trimethylsilyl)

trifluoroacetamide (MSTFA). In a typical sample preparation, the ‘dried organic phase’ (30 mg), together with the non-alkylated

syringol (S, 10 mg) and bisphenol F (BPF, 10 mg) as external standards were accurately weighed in a glass vial and homogeneously

mixed with 100 μL of anhydrous pyridine and 250 μL of MSTFA. To guarantee a complete derivatisation with MSTFA, the samples

were heated for 15 min at 80oC and subsequently diluted with 1 mL acetonitrile (ACN). One µl of the sample was injected at a split

ratio of 1:100. This GC methodology allows to separate the external standards, unreacted alkylsyringol and m,m’-isomers, and

gives an indication of the presence of higher oligomers (i.e., trimers) which only partially evaporate. Quantification of mono- and

dimers was performed by calibration of the pure (isolated) reagent/product against the applied external standards.

» Gel Permeation/Size-Exclusion Chromatography (GPC/SEC)

To obtain more insight into the presence, type and (relative) amount of higher oligomers, the distribution of the molar mass of

the ‘dried organic phase’ was investigated using GPC/SEC. Therefore, a sample was solubilised in THF (~ 10 mg.mL-1) and

subsequently filtered with a 0.2 µm PTFE membrane to remove any particulate matter to prevent plugging of the columns.

GPC/SEC analysis was performed on a Waters e2695 Separations Module with a pre-column and a Varian M-Gel column (3 µm,

mixed), equipped with a Waters 2988 Photodiode array detector (UV detection at 280 nm), Empower software and using THF as

the mobile phase (1 mL.min-1) at 40oC. Together with a melting point determination and TGA (see Product characterisation), this

GPC/SEC analysis was conducted as a means to validate the purity after precipitation and crystallisation. In contrast to 1H NMR,

which provides a number-average oligomer length, GPC/SEC analysis allows to distinguish all encountered oligomers.

Crystallisation and/or precipitation of bissyringols

» 3,3’-methylenebis(2,6-dimethoxy-4-n-propylphenol) – m,m’-BSF-4P

The crude yellowish-orange dense oil was dissolved in diethyl ether and dried with MgSO4. After filtration of this ‘dried organic

phase’, heptane was added, and macrocrystals were formed from the slow evaporation of this diethyl ether-heptane solution. On

the contrary, addition of excess heptane yielded an emulsion. Highly pure (≥99.0 %) crystals were harvested by filtration and dried

in vacuo (25oC, ~2 mbar) overnight.

Yield: 4.36 g (60 %). M.p. (powder) 102-103oC; m.p. (crystals) 103-104oC; 1H NMR (300 MHz, CDCl3, 25oC, TMS): δH = 0.86 (t,

3J(H,H)= 7.3 Hz, 6H; -CH2CH3), 1.34 (sex, 3J(H,H)= 7.5 Hz, 4H; -CH2CH2CH3), 2.54 (t, 3J(H,H)= 7.9 Hz, 4H; -ArCH2CH2-), 3.51 (s, 6H; -2-

OCH3), 3.85 (s, 6H; -6-OCH3), 3.94 (s, 2H; -ArCH2Ar-), 5.30 (s, 2H; -ArOH) and 6.45 ppm (s, 2H; -m-ArH); 13C NMR (100 MHz, CDCl3,

25oC, TMS): δC = 145.9 (C2), 145.4 (C6), 136.5 (C1), 132.8 (C4), 126.2 (C3), 107.8 (C5), 60.0 (Ce*), 56.1 (Ce), 35.5 (Cb), 24.5 (Cc), 23.7

(Ca) and 14.2 ppm (Cd); MS (70 eV, EI): m/z (%): 405 (10) [M+•+•H], 404 (39) [M+•], 209 (55) [M+•-•C9H11O3], 208 (100) [M+•-•C9H12O3],

195 (14), 179 (24), 167 (14); FTIR (KBr): ṽmax = 1026 (C-O), 1252 (C-O), 1311 (C-O), 1500 (arom. C=C), 2835 (aliph. C-H), 2866 (aliph.

C-H), 2931 (aliph. C-H), 2947 (aliph. C-H), 2962 (aliph. C-H), 3332 (polymeric O-H) and 3545 cm-1 (monomeric O-H)

» 3,3’-methylenebis(2,6-dimethoxy-4-methylphenol) – m,m’-BSF-4M

The crude yellowish-white solid was dissolved in acetone and dried with MgSO4. After filtration of this ‘dried organic phase’,

heptane was added, and macrocrystals were formed from the slow evaporation of this acetone-heptane solution. On the contrary,

addition of excess heptane yielded a precipitate. Highly pure (≥99.0 %) crystals were harvested by filtration and dried in vacuo

(25oC, ~2 mbar) overnight.

8

Yield: 4.70 g (75 %). M.p. (powder) 144-145oC; m.p. (crystals) 145oC; 1H NMR (300 MHz, CDCl3, 25oC, TMS): δH = 2.14 (s, 6H; -

ArCH3), 3.57 (s, 6H; -2-OCH3), 3.84 (s, 6H; -6-OCH3), 3.93 (s, 2H; -ArCH2Ar-), 5.30 (s, 2H; -ArOH) and 6.43 ppm (s, 2H; -m-ArH); 13C

NMR (100 MHz, CDCl3, 25oC, TMS): δC = 146.0 (C2), 145.2 (C6), 136.4 (C1), 128.0 (C4), 125.9 (C3), 108.6 (C5), 60.1 (Ce*), 56.1 (Ce),

24.4 (Ca) and 20.0 ppm (Cb); MS (70 eV, EI): m/z (%): 349 (19) [M+•+•H], 348 (89) [M+•], 181 (80) [M+•-•C9H11O3], 180 (100) [M+•-

•C9H12O3], 167 (18), 165 (25), 151 (27), 137 (21), 121 (14); FTIR (KBr): ṽmax = 1057 (C-O), 1078 (C-O), 1298 (C-O), 1315 (C-O), 1502

(arom. C=C), 2837 (aliph. C-H), 2910 (aliph. C-H), 2935 (aliph. C-H), 2947 (aliph. C-H), 2966 (aliph. C-H), 3296 (polymeric O-H) and

3408 cm-1 (monomeric O-H)

Reasoning behind the bissyringol abbreviations

The abbreviations of the m,m’-bissyringols (BSF-4M and BSF-4P) follow the same logic as the abbreviations of known industrial

bisphenols, such as bisphenol F (BPF). In analogy to BPF, the first letter ‘B’ refers to ‘bis’, and the second letter refers to the

phenolic core type, being either phenolic (P), guaiacylic (G) or syringylic (S). The third letter originates from the reactive carbonyl

species, i.e., formaldehyde (F), used to form the carbon bridge. Next, the aromatic position of the additional alkyl (-xA) chain is

appointed based on the IUPAC nomenclature (vide supra). The (p,p’- or m,m’-) coupling pattern is denoted as the position of the

bridging carbon relative to the aromatic OH-group.

Product characterisation

Several characterisation techniques were applied to asses the molecular structures of bissyringols and their corresponding

polymers. A detailed description is listed below:

» Spectrometric analyses

Idendification of (underivatised) dimeric GC signals was accomplished by gas chromatography-mass spectrometry (GC-MS) on a

Agilent 6890 series with a HP1-MS capillary column equipped with an Agilent 5973 series mass spectroscopy detector (EI, 70 eV

ionisation energy).

» Spectroscopic analyses

Liquid-phase 1H, 13C, 13C DEPT-135o and 1H-13C HMBC nuclear magnetic resonance (NMR) spectra were acquired on Bruker Avance

instruments (300, 400 and 600 MHz) with automated samplers. The chemical shifts (δ) are reported in parts per million (ppm)

referenced to tetramethylsilane (1H) or the internal NMR solvent signals (13C). In a typical sample preparation, a dried sample (~10

mg for 1H and ~50 mg for 13C) is homogeneously dissolved in 650-750 µL of deuterated solvent (CDCl3) and transferred to a NMR

tube. Fourier-transform infrared (FT-IR) spectra of dried KBr pellets, pre-mixed with pure product (1 wt%), were recorded in vacuo

on a Bruker IFS 66v/S instrument. Powder X-ray diffraction (PXRD) patterns were recorded on powdered samples on a STOE Stadi

P Combi diffractometer with an image plate position sensitive detector (IP-PSD) in the region 2θ = 5 to 60° (Δ2θ = 0.03°) and a

scan of maximum 1200 s. The measurements were performed in transmission mode at room temperature using CuKα1 radiation

(λ = 1.54056 Å) selected by means of a Ge(111) monochromator. From these XRD patterns the relative crystalline/amorphous

character of the polymers was acquired.

» Thermal analyses

Thermogravimetric analysis (TGA) was performed while heating under N2, O2 or ambient atmosphere using a TA Instruments TGA

Q500. About 10 mg of the dried sample was heated at 10oC.min-1 to 750oC and kept isothermal for 15 min at a flow rate of 20

mL.min-1. TGA allows determination of the degradation temperature (Td) and simultaneously indicates the presence/absence of

residual solvents. Melting points (Tm) of the crystals were determined using glass capillaries in a Stuart Scientific SMP3 melting

point apparatus and confirmed by differential scanning calorimetry (DSC) on a TA Instruments DSC Q200 by cycling between 40

and 180oC at heating/cooling rates of 10oC.min-1 under N2 atmosphere. DSC experiments further inform about initial crystallinity,

9

crystallisation exotherms, glass-transition temperatures (Tg) and melt temperatures (Tm) of polymers and were conducted by

cycling between 25 and 320oC at heating/cooling rates of 10 or 20oC.min-1 under N2 atmosphere (50 mL.min-1).

In vitro oestrogenic potency screening

The experiments were conducted as reported previously by Witters et al. (2010)14 with slight modifications.

» MELN cells

MELN cells (provided by INSERM, Montpellier, FR; Balaguer et al. (1999)15) are oestrogen-sensitive human breast cancer cells

(MCF-7) stably transfected with the oestrogen-responsive gene (ERE-βGlo-Luc-SVNeo) carried by integrated plasmids. In addition

to the antibiotic resistance selection gene (SVNeo), these plasmids also contain oestrogen-responsive elements to which the

oestrogen receptor (hERα)-ligand complex can bind, hence inducing the transcription of the luciferase reporter gene. MELN cells

were cultured in DMEM:F12 medium with GlutaMaxTM I supplemented with 1 % penicillin/streptomycin (all Gibco, ThermoFisher,

Ghent, BE), 1 mg.mL-1 G418 sulphate (Invivogen, Toulouse, FR) and 7.5 % fetal bovine serum superior (Biochrome, Gentaur,

Kampenhout, BE). The cell line was maintained in an incubator at 37oC, a relative humidity of 95 % and a CO2 concentration of 5

%.

» Exposure of cells

A standard set-up has been developed to expose MELN cells and measure ER-transactivation for xeno-oestrogenic compounds. In

order to decrease the background signal, cells were adapted to charcoal/dextran treated fetal calf serum (Gibco, ThermoFisher,

Ghent, BE). Cells were seeded at a density of 8 x 105 cells per well, in oestrogen-free black 96-well plates with transparent bottoms

(Costar). Cells were maintained in 100 µL test medium for 24 h. Serial dilutions of the test compounds were made in oestrogen-

free dimethyl sulfoxide (DMSO). Dilutions of the test compound were added to the test medium and 100 µL of each concentration

was added to three replica wells. The final solvent concentration was always 0.1 vol%. Cells were treated with the test compounds

for 19-20 h. Each bissyringol compound was studied in a range finding experiment (10-4 – 6.1.10-9 M; see Table S6), and subsequent

repeat experiments in an appropriate working range to determine EC50 (see Data Analyses). In each experiment, for each

concentration three replica wells were tested. Test compounds were assessed in comparison to a positive assay control, the

natural hormone 17β-oestradiol (E2), and the known positive industrial compound, BPA. Details in Table S7.

» Luciferase assay

At the end of the incubation period, the remaining medium is removed for analysis of cell damage using the CytoTox-OneTM

Homogenous Membrane Integrity Assay (Promega) as previously described by Berckmans et al. (2007).16 Next, cells were lysed by

adding 30 µL reporter lysis buffer (Promega, Leiden, NL) in each well. After shaking plates for 25 min, plates were frozen (-80oC)

for minimum 1 h and maximum 1 week. After thawing the plates, luminescence was measured using a luminometer (Luminoskan)

after injection of 50 µL luciferase reagent (Promega, Leiden, NL) in each well. Results are expressed as relative light units (RLU).

» Data analysis

The results are presented as induction of ER activation expressed as percentage of luciferase induction by the vehicle control (set

at 100 %). Averaged results from three independent experiments were graphed ± standard deviation (SD) using Graphpad Prism

software (version 7.03, 2017) with mean EC50 values determined by fitting a four-parameter sigmoidal dose-response curve (cf.

Hill equation). To obtain mean EC50 values for partial dose-response curves, the fit was constrained at the top value. Few

compounds exhibited cytotoxicity at the highest concentrations (0.5.10-4 – 10-3 M; see Table S6); if present, these results were

excluded in the fit. The EC50 values, the half-maximal effective concentration for ER activation, allow to rank the compounds for

their potency (i.e. higher EC50, less potent). To calculate the relative oestrogenic potency (REP), the EC50 for reference compound

oestradiol was divided by the EC50 for each bisphenol, and this number was expressed as percent. Statistically significant

differences for logEC50 and Emax values were determined by one-way ANOVA (α < .05) followed by pairwise comparisons with

Tukey’s HSD post-hoc analysis.

10

Synthesis and characterisation of polyesters

Polyesters were prepared on a (milli)gram scale via interfacial polymerisation with terephthaloyl chloride, according to methods

reported previously with slight modifications.17,18 In a round-bottom flask (10 mL) pure m,m’-bissyringol (0.8 mmol) was

deprotonated in an aqueous solution of NaOH (3 mmol; 120.7 mg in 3 mL). This mixture was continuously stirred in an ice-cooled

water bath kept at 10oC. After complete dissolution (ca. 30 min.), a catalytic amount of benzyltriethylamonium chloride (41 µmol;

9.3 mg) was added. A separate solution of terephthaloyl chloride (0.8 mmol; 0.1641 g) in DCM (4-5 mL) was added after 30 min

and the reaction mixture was vigorously stirred at 1400 rpm for 90 min at 10oC. The reaction mixture was poured into hot water

(150 mL, at 60oC); the precipitated polymer was collected on a Büchner funnel and washed with water (3 x 15 mL). The polymer

was then dissolved in chloroform and reprecipitated into ice-cooled methanol (150 mL). The polymer was again Büchner filtered,

washed with cold methanol (3 x 15 mL), and dried on the filter for 30 min. The solid was further dried in vacuo (25oC, ~2 mbar)

overnight yielding a white powder (see Fig. S10). The polymer yield is calculated from the theoretical molar mass of the repeating

units, being 478.5 and 534.6 g.mol-1 for TPC/BSF-4M and TPC/BSF-4P, respectively, under the assumption that the total amount

of end-groups is negleglible compared to the total polymer weight.

GPC/SEC analyses on the polyesters were performed on a Shimadzu system (LC-10ADvp pump unit, CTO-10Avp column oven and

SCL-10Avp gradient controller) with Agilent PLgel 5 µm MIXED-D (300 x 7.5 mm) column, equipped with Shimadzu SDP-10Avp UV-

VIS detectors (at 254 and 330 nm) and RID-10 detector, using tetrahydrofuran (THF) or hexafluoroisopropanol (HFIP) as a mobile

phase (1 mL/min) at 30oC. Polyesters were dissolved in THF or HFIP (5 mg/mL) and left to stand for 24 h at room temperature

prior to being measured. Before the actual GPC measurement, the solution was consecutively filtered over two Millex® FH filters

(PTFE, 0.45 μm and 0.2 μm). Solutions were injected with a Fortuna® Optima glass syringe (1 mL). For THF, the system was

calibrated with narrow polydispersity polystyrene standards, and molecular weights (Mn and Mw) are reported as polystyrene

equivalents. For HFIP, the system was calibrated with narrow polydispersity poly(methyl methacrylate) standards, and converted

to polystyrene equivalents.19

» poly[methylene bis(4-n-propylsyringol) terephthalate] – TPC/BSF-4P

Yield: 0.4274 g (99 %). M.p. decomposes before melting; 1H NMR (300 MHz, CDCl3, 25oC, TMS): δH = 0.96 (t, 6H; -CH2CH3), 1.51

(sex, 4H; -CH2CH2CH3), 2.68 (t, 4H; -ArCH2CH2-), 3.41 (s, 6H; -2-OCH3), 3.77 (s, 6H; -6-OCH3), 3.97 (s, 2H; -ArCH2Ar-), 6.59 (s, 2H; -

m-ArH) and 8.32 ppm (s, 4H; -ArH); 13C NMR (100 MHz, CDCl3, 25oC, TMS): δC = 163.5 (CCO), 151.5 (C2), 149.9 (C6), 140.2 (C4), 133.7

(CTPC), 131.3 (C1), 130.4 (CTPC), 126.2 (C3), 108.5 (C5), 60.8 (Ce*), 56.1 (Ce), 36.1 (Cb), 24.2 (Cc), 24.1 (Ca) and 14.3 ppm (Cd); FTIR (KBr):

ṽmax = 1747 cm-1 (C=O stretch); Molecular weight (GPC in THF, 30oC, 254 nm, PS): Mn = 24368 and Mw = 43013 g.mol-1, PDI = 1.8

» poly[methylene bis(4-methylsyringol) terephthalate] – TPC/BSF-4M

Yield: 0.3482 g (91 %). M.p. decomposes before melting; 1H NMR (300 MHz, CDCl3, 25oC, TMS): δH = 2.29 (s, 6H; -ArCH3), 3.49 (s,

6H; -2-OCH3), 3.77 (s, 6H; -6-OCH3), 3.98 (s, 2H; -ArCH2Ar-), 6.58 (s, 2H; -m-ArH) and 8.34 ppm (s, 4H; -ArH); 13C NMR (100 MHz,

CDCl3, 25oC, TMS): δC = 163.6 (CCO), 151.4 (C2), 149.8 (C6), 135.8 (C4), 133.6 (CTPC), 131.2 (C1), 130.4 (CTPC), 125.9 (C3), 109.6 (C5),

61.0 (Ce*), 56.1 (Ce), 24.6 (Ca) and 20.6 ppm (Cb); FTIR (KBr): ṽmax = 1745 cm-1 (C=O stretch); Molecular weight (GPC in THF, 30oC,

254 nm, PS): Mn = 6339 and Mw = 20402 g.mol-1, PDI = 3.2

11

II. Tables

Table S1. Literature yields of 4-n-propylsyringol (PS) obtained from direct catalytic hydrogenolysis on (transgenic) hardwood

substrates with hydrogen as reductant unless indicated otherwise.

Substrate Lignin YieldMonomers

Selectivity YieldPS Catalytic system Ref.

wt% wt%lignin SPS % wt%lignin wt%wood

hardwood

Birch (Betula platyphylla) 19a 45.5 76 34.8 6.6h Rh/C, H3PO4, dioxane, water 20

Birch 19.8b 46.5 39 18.2 3.6h Ni-W2C, ethylene glycol 21

Birch 19.8b 44.1 45 19.9 3.9h Pt/C, water 21

Birch 19c 49.0 73 36.0 6.8h Ni/C, ethylene glycol, Ar 22

Birch 19c 60.8 61 37.2 7.1h Ni/C, methanol, Ar 22

Poplar (Populus) 19d 54e 55 29.7 5.6h Pd/C, ZnCl2, methanol 23

Birch (Betula papyrifera) 16d 52e 69 35.9 5.7h Pd/C, ZnCl2, methanol 23

Birch (Betula pendula) 20c 62.5f 62 38.5f 7.5 H2-act. Pd/C, ethanol, water, Ar 24

Poplar (Populus x canadensis) 21.2a 43.9g 46 20.4 3.7 Ru/C, methanol 3

Birch (Betula pendula) 19.5a 51.4g 72 37.0 7.1 Ru/C, methanol 3,4

transgenic hardwood Poplar (Populus spp., 717-F5H)j 20d 36e,i 83 29.9 6.0h Pd/C, ZnCl2, methanol 23

Poplar (Populus spp., F5H)k 14.2a 77 64 49.0 7.0h Ru/C, methanol 25 Lignin content as determined by a Klason lignin method, b Van Soest method, c TAPPI standard method (T 222 om-02) or d acetyl bromide-soluble lignin (ABSL) analysis. e Yield is calculated using the initial mass of lignin and the mass of the products, factoring in the loss of two atoms of oxygen for each mole of product produced. f Isolated yields, assuming that propenyl aryls can be hydrogenated to propyl aryls with 100 % selectivity. g Expressed as carbon yield. h Assumes complete delignification. i Yield of methylparaben not in calculation. j Contains 73% S-units, as determined by DFRC analysis k Contains 98.3% S-units, as determined by NMR, DFRC and thioacidolysis.

Table S2. Design parameters for a certain split of PS and PG in the second (vacuum) column as estimated with the FUG-method.

Design parameters 99.9 mol% split PS 99 mol% split PG

95 mol% split PS 90 mol% split PG

minimal number of equil. stages (Nmin) 31 14 actual number of stages (Nact) 57 26 minimal reflux ratio (Rmin) 7.77 6.89 actual reflux ratio (Ract) 10.10 8.96

Table S3. Process parameters for SEP001 and SEP002 after optimisation.

Process parameters SEP001 SEP002

distillate-to-feed ratio (D:F) 0.942 0.366 feed stage 5 15 condenser pressure (in bar) 1.01 0.10 reflux ratio (R) 3.00 10.10 number of stages, (N) 10 57

12

Table S4. Components present in the monomer-enriched oil dissolved in n-hexane with indication of (i) the availability of physical property parameters in the Aspen Plus® database, (ii) the boiling points and (iii) the concentrations in n-hexane (75 mL).

Component CAS

number available in Aspen

Properties® Boiling point at 760 mmHg

Molecular weight

Composition

of lignin oil Concentration

in n-hexane Ref.

yes/no oC g.mol-1 wt% g.L-1 -

n-hexane 110-54-3 yes 68.73 86.18 - solvent 26

4-ethylguaiacol 2785-89-9 no 236.5 152.19 0.72 0.63 27

isoeugenol 97-54-1 yes 254.8 164.22 1.29 1.13 26

syringol 91-10-1 yes 261.0 154.17 1.35 1.17 26

4-n-propylguaiacol (PG) 2785-87-7 yes 268.2 166.22 21.33 18.59 26

4-ethylsyringol 14059-92-8 no 273.2a 182.22 2.80 2.44 28,29

4-methylsyringol 6638-05-7 no 278.4a 168.19 0.72 0.63 30,31

4-n-propylsyringol (PS) 6766-82-1 no 300.8 196.24 66.22 57.69 32,33

4-prop-1-enylsyringol 6635-22-9 no 324.5a 194.23 0.98 0.85 34 a As approximated by a pressure-temperature nomograph (based on the Clausius-Clapeyron equation) under the assumption that the heat of vaporisation is constant and independent of pressure. Boiling points as stated in literature for 4-ethyl-, 4-methyl- and 4-prop-1-enylsyringol are 106.0oC (at 2 mmHg), 145.5oC (at 14 mmHg) and 180.0oC (at 11 mmHg), respectively.

Table S5-1. Distillation resistance for the monomer-enriched oil (6.53 g) dissolved in n-hexane (75 mL) obtained from RCF on native birch (hard)wood.

Solvent Concentration Tb ∆T Ωfeed Ωproduct

wt% oC oC oC-1 oC-1

n-hexane 88.32 68.7 167.8 0.53 5.15

4-ethylguaiacol 0.08 236.5 18.3 0.00 0.04

isoeugenol 0.15 254.8 6.2 0.02 0.24

syringol 0.16 261.0 7.2 0.02 0.22

4-n-propylguaiacol (PG) 2.49 268.2 5.0 0.50 4.87

4-ethylsyringol 0.33 273.2 5.2 0.07 0.65

4-methylsyringol 0.08 278.2 22.4 0.00 0.03

4-n-propylsyringol (PS) 7.73 300.8 23.7 0.33 3.19

4-prop-1-enylsyringol 0.11 324.5 - - -

dimers 0.35 - - - -

undefined 0.18 - - - -

sum 100 1.5 14.4

4-n-propyls 10.22

Table S5-2. Distillation resistance for a theoretical monomer-enriched oil (6.53 g) dissolved in n-hexane (75 mL) obtained from RCF on high-syringyl poplar (hard)wood.a

Solvent Concentration Tb ∆T Ωfeed Ωproduct

wt% oC oC oC-1 oC-1

n-hexane 88.32 - - 0.53 -

4-ethylguaiacol - 236.5 18.3 - -

isoeugenol - 254.8 6.2 - -

syringol - 261.0 7.2 - -

4-n-propylguaiacol (PG) 0.72 268.2 5.0 0.14 1.41

4-ethylsyringol 0.53 273.2 5.2 0.11 1.04

4-methylsyringol - 278.2 22.4 - -

4-n-propylsyringol (PS) 10.44 300.8 23.7 0.44 4.31

4-prop-1-enylsyringol - 324.5 - - -

dimers - - - - -

undefined - - - - -

Sum 100 1.2 11.9

4-n-propyls 11.16

a Calculated from the monomer composition as reported by Luterbacher et al.25 (2016) and rescaled to the concentrations stated in Table S5-1.

13

Table S5-3. Distillation resistance for an unextracted lignin oil obtained from RCF on high-syringyl poplar (hard)wood.a

Solvent Concentrationa Tb ∆T Ωfeed Ωproduct

wt% oC oC oC-1 oC-1

n-hexane - - - - -

4-ethylguaiacol - 236.5 18.3 - -

isoeugenol - 254.8 6.2 - -

syringol - 261.0 7.2 - -

4-n-propylguaiacol (PG) 6.15 268.2 5.0 1.23 1.29

4-ethylsyringol 4.56 273.2 5.2 0.91 0.96

4-methylsyringol - 278.2 22.4 - -

4-n-propylsyringol (PS) 89.29 300.8 23.7 3.77 3.95

4-prop-1-enylsyringol - 324.5 - - -

dimers - - - - -

undefined - - - - -

sum 100 5.9 6.2

4-n-propyls 95.44

a Calculated from the monomer composition as reported by Luterbacher et al.25 (2016)

Table S6. Additional experimental details of the reporter cell line MELN-hERα for E2, BPA, BSF-4M and BSF-4P.

Compound Range testeda / M

Cytotoxicityb / M min. max.

17β-oestradiol (E2) 4.57.10-13 1.00.10-09 -

BPA 1.00.10-10 1.00.10-03 1.00.10-03

BSF-4Mc 6.10.10-09 1.00.10-03 1.00.10-03

BSF-4Pc 1.53.10-09 2.50.10-05 1.00.10-04 a Maximum range in preliminary test, range is refined and more narrow in repeat tests. b Lowest concentration with cytotoxicity by lactate dehydrogenase (LDH) assay and/or visual microscopy. c Top value constrained.

14

III. Figures

Figure S1. Configuration of industrial scale distillation unit for the separation of PS (and PG) with denoted stream nomenclature.

Figure S2. Purity (top) and recovery efficiency (bottom) of PS (and PG) as function fo the number of equilibrium stages

15

Figure S3. (A) Sample of 4-n-propylsyringol (≥98 %) as obtained from RCF of raw birch wood chips. The liquid lignin oil, containing

phenolic monomers (50 C%), dimers (15-20 C%) and oligomers (30-35 C%) was subjected to a threefold reflux extraction with n-

hexane to extract the monomers. After removal of n-hexane by vacuum distillation, this monomer-enriched oil was fractionated

by column chromatography on silica-gel with 25 vol% acetone-heptane as sustainable mobile phase. * Signal in Figure 1 originates

from asymmetric condensation with residual 4-ethylsyringol. (B) Crystals from BSF-4P obtained by gradual evaporation from a

diethyl ether-heptane solution.

Figure S4. Crude reaction mixtures of 4-methyl- (A) and 4-n-propylsyringol (B) before reaction, after reaction and after decantation

and dissolution acetone or ether. Density observations agree with the differences between the calculated density for the aqueous

acidic phase (1.038 g.mL-1) and the available densities for 4-methyl- (1.105 g.mL-1) and 4-n-propylsyringol (1.059 g.mL-1). The

solubility of the organic product phase increases with chain length.

1 cm

A B

16

Figure S5. Comparison of crystallisation (blue) and precipitation (red) as purification technique as illustrated by GPC/SEC (left)

benchmarked against the raw reaction product mixture (black). Attempts to precipitate the raw product mixture obtained for 4-

n-propyl derivatives yielded a suspension and are therefore not shown. Melting point assessment by DSC (right). Notice the

crystallisation tendency for BSF-4M.

Figure S6. Mass spectra of the m,m’-dimeric products formed from MS (left) and PS (right) as detected by an ion trap mass

spectrometer. M+• is molecular ion peak, representing the molecular weight of the dimers. The most abundant peak identified

represents the molecular weight of the most stable ion fragment from each dimer.

17

Figure S7-1. 1H NMR spectrum of 3,3’-methylenebis(2,6-dimethoxy-4-methylphenol) in CDCl3 at 300 MHz.

Figure S7-2. 13C (bottom) and 13C DEPT-135o (top) NMR spectra of 3,3’-methylenebis(2,6-dimethoxy-4-methylphenol) in CDCl3 at

400 MHz.

18

Figure S7-3. 2D 1H,13C HMBC NMR spectrum of 3,3’-methylenebis(2,6-dimethoxy-4-methylphenol) in CDCl3 at 600 MHz.

ppm

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 ppm

20

30

40

50

60

70

80

90

100

110

120

130

140

150

ppm

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 ppm

20

30

40

50

60

70

80

90

100

110

120

130

140

150

3.9

3

6.43 5.30

3.8

4

3.5

7

2.14

b

a

e

f

e*

20.0 24.4

56.1

108.6

125.9 128.0

136.4

145.2 146.0

b a

e

3

6

4

5

1

2

60.1 e*

5

19

Figure S8-1. 1H NMR spectrum of 3,3’-methylenebis(2,6-dimethoxy-4-n-propylphenol) in CDCl3 at 300 MHz.

Figure S8-2. 13C (bottom) and 13C DEPT-135o (top) NMR spectra of 3,3’-methylenebis(2,6-dimethoxy-4-n-propylphenol) in CDCl3

at 400 MHz.

20

Figure S8-3. 2D 1H,13C-HMBC NMR spectrum for 3,3’-methylenebis(2,6-dimethoxy-4-n-propylphenol) in CDCl3 at 600 MHz.

ppm

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 ppm

20

30

40

50

60

70

80

90

100

110

120

130

140

150

ppm

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 ppm

20

30

40

50

60

70

80

90

100

110

120

130

140

150

ppm

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 ppm

20

30

40

50

60

70

80

90

100

110

120

130

140

150

3.9

4

6.45 5.30

3.8

5

3.5

1

0.86

b a

e

f

e*

c

d

1.34 2.54 14.2

23.7

56.1

107.8

126.2

132.8 136.5

145.4 145.9

b

a

e

3

6

4

5

1

2

60.0 e*

24.5

35.5

c

d

5

21

Figure S9. FT-IR spectra of bis(4-alkylsyringol)s derived from MS (top) and PS (bottom) measured via the KBr pellet procedure for

solid samples.

Figure S10. Top view in glass-filters containing the dried bisphenol- and bissyringol-derived polyesters TPC/BPA (left), TPC/BSF-

4M (middle) and TPC/BSF-4P (right) after precipitation and filtration. Pictures are taken under similar lighting conditions.

22

Figure S11. 1H NMR spectra of poly[methylene bis(4-alkylsyringols) terephthalate]s derived from 4MS (top) and PS (bottom) in

deuterated chloroform (CDCl3, δH = 7.24 ppm) at 300 MHz. The presences of terephthalic esters can be confirmed by the

disappearance of the singlet 1H resonance at 5.30 ppm (2H; -OH) and the appearance of a singlet 1H resonance at 8.33 ppm (4H;

-ArH).

Figure S12. 13C NMR spectra of poly[methylene bis(4-alkylsyringols) terephthalate]s derived from 4MS (top) and PS (bottom) in

deuterated chloroform (CDCl3, δc = 77.2 ppm) at 400 MHz. The presences of terephthalic esters can be confirmed by the

appearance of three extra 13C resonances at 130.4 (aromatic), 133.6 (aromatic) and 163.5 ppm (carbonylic).

23

Figure S13. Structural confirmation of polyester (PE) formation by FT-IR spectroscopy measured via the KBr pellet procedure for

solid samples. Spectra of corresponding bis(4-alkylsyringol)s in light-grey. Notice the loss of the phenolic -OH stretch (3100 – 3600

cm-1) with subsequent introduction of a C=O stretch (± 1746 cm-1). From top to bottom: TPC/BSF-4M and TPC/BSF-4P.

Figure S14. Loss of crystallinity as observed in PXRD patterns of poly[methylene bis(4-alkylsyringol) terephthalate]s powders

derived from MS (top) and PS (bottom).

24

IV. References 1. S. Burkhardt, L. Kumar, R. Chandra and J. Saddler, Biotechnol. Biofuels, 2013, 6, 90. 2. C. W. Dence, in Methods in Lignin Chemistry, Springer, 1992, pp. 33-61. 3. S. Van den Bosch, W. Schutyser, R. Vanholme, T. Driessen, S.-F. Koelewijn, T. Renders, B. De Meester, W. J. J. Huijgen, W.

Dehaen, C. M. Courtin, B. Lagrain, W. Boerjan and B. F. Sels, Energy Environ. Sci., 2015, 8, 1748-1763. 4. S. Van den Bosch, W. Schutyser, S.-F. Koelewijn, T. Renders, C. M. Courtin and B. F. Sels, Chem. Commun., 2015, 51, 13158. 5. D. G. Drueckhammer, S. Q. Gao, X. Liang and J. Liao, ACS Sustain. Chem. Eng., 2013, 1, 87-90. 6. J. D. Seader, E. J. Henley and D. K. Roper, Separation Process Principles: Chemical and Biochemical Operations, John Wiley &

Sons, Hoboken, 3rd edn., 2011. 7. R. H. Perry and D. W. Green, Perry's Chemical Engineers' Handbook, McGraw-Hill, New York, 8th edn., 2008. 8. J.-P. Lange, ChemSusChem, 2017, 10, 245-252. 9. E. C. Carlson, Chem. Eng. Prog., 1996, 92, 35-46. 10. S. I. Sandler, Chemical, biochemical, and engineering thermodynamics, John Wiley & Sons, 2017. 11. R. Turton, R. C. Bailie, W. B. Whiting and J. A. Shaeiwitz, Analysis, synthesis and design of chemical processes, Pearson

Education, Inc., New Jersey, 3rd edn., 2008. 12. K. Chuang and K. Nandakumar, Encyclopedia of Separation Science, 2000, 1135-1140. 13. S.-F. Koelewijn, S. Van den Bosch, T. Renders, W. Schutyser, B. Lagrain, M. Smet, J. Thomas, W. Dehaen, P. Van Puyvelde, H.

Witters and B. F. Sels, Green Chem., 2017, 19, 2561-2570. 14. H. Witters, A. Freyberger, K. Smits, C. Vangenechten, W. Lofink, M. Weimer, S. Bremer, P. H. J. Ahr and P. Berckmans,

Reprod. Toxicol., 2010, 30, 60-72. 15. P. Balaguer, F. François, F. Comunale, H. Fenet, A.-M. Boussioux, M. Pons, J.-C. Nicolas and C. Casellas, Sci. Total Environ.,

1999, 233, 47-56. 16. P. Berckmans, H. Leppens, C. Vangenechten and H. Witters, Toxicol. In Vitro, 2007, 21, 1262-1267. 17. Z. Ping, W. Linbo and L. Bo-Geng, Polym. Degradation Stab., 2009, 94, 1261-1266. 18. S. S. Kuhire, S. S. Nagane and P. P. Wadgaonkar, WO Pat., 2015/140818 A1, 2015. 19. S. Mori, Anal. Chem., 1989, 61, 1321-1325. 20. N. Yan, C. Zhao, P. J. Dyson, C. Wang, L.-T. Liu and Y. Kou, ChemSusChem, 2008, 1, 626-629. 21. C. Li, M. Zheng, A. Wang and T. Zhang, Energy Environ. Sci., 2012, 5, 6383-6390. 22. Q. Song, F. Wang, J. Cai, Y. Wang, J. Zhang, W. Yu and J. Xu, Energy Environ. Sci., 2013, 6, 994-1007. 23. T. Parsell, S. Yohe, J. Degenstein, T. Jarrell, I. Klein, E. Gencer, B. Hewetson, M. Hurt, J. I. Kim, H. Choudhari, B. Saha, R.

Meilan, N. Mosier, F. Ribeiro, W. N. Delgass, C. Chapple, H. I. Kenttamaa, R. Agrawal and M. M. Abu-Omar, Green Chem., 2015, 17, 1492-1499.

24. M. V. Galkin and J. S. Samec, ChemSusChem, 2014, 7, 2154. 25. L. Shuai, M. T. Amiri, Y. M. Questell-Santiago, F. Héroguel, Y. Li, H. Kim, R. Meilan, C. Chapple, J. Ralph and J. S. Luterbacher,

Science, 2016, 354, 329-333. 26. Aspen Technology, Aspen Properties® - Physical property database, (accessed 2017-05-17, 2017). 27. SRC, Inc., PHYSPROP Database - CAS Query: 2785-89-9 http://www.srcinc.com/what-we-do/environmental/scientific-

databases.html, (accessed 2017-05-17, 2017). 28. H. Schultes, Ber. Dtsch. Chem. Ges. B, 1936, 69B, 1870-1873. 29. Parchem - fine & specialty chemicals, 4-ethylsyringol, http://www.parchem.com/chemical-supplier-distributor/4-ethyl-

syringol-009351.aspx, (accessed 2017-05-17, 2017). 30. K. Freudenberg and H. H. Hubner, Chem. Ber., 1952, 85, 1181-1191. 31. Alfa Aesar, Thermo Fisher Scientific Inc., B21801 2,6-dimethoxy-4-methylphenol, 97%,

https://www.alfa.com/en/catalog/B21801/, (accessed 2017-05-17, 2017). 32. ChemSrc. - Shanghai Ji Zhao Information Technology Co., Ltd, compound: 2,6-dimethoxy-4-propylphenol,

http://www.chemsrc.com/en/cas/6766-82-1_687803.html, (accessed 2017-05-17, 2017). 33. The Good Scents Company, 4-propylsyringol, 4-propyl-2,6-dimethoxyphenol,

http://www.thegoodscentscompany.com/data/rw1037721.html, (accessed 2017-05-17, 2017). 34. H. Richtzenhain, Chem. Ber., 1948, 81, 260-265.