With a BDE of 82.4 kcal mol⁻¹, the unsubstituted guaiacol (1) is an efficient DPPH^• scavenger (Table 1). Compounds 2, 3, 4, and 6 (creosol, ethylguaiacol, propylguaiacol, and eugenol, respectively) exhibited higher antioxidant activity in both DPPH^• and ABTS^•+ assays, as seen from the RDI value higher than 1.00 (Table 1). This result agrees with the corresponding BDE values, which are lower than 82.4 kcal mol⁻¹ for these three compounds (Table 1). This effect can be attributed to a positive inductive effect of the alkyl chain thus increasing the electrophilicity of this lignin structural unit. The length of saturated alkyl chains (methyl, ethyl or propyl) had negligible effect on BDEs and DPPH^• scavenging activity. Compounds 1–4 exhibited RDI values higher than 1, meaning that more than one radical can be scavenged per OH group. This also means that these guaiacol derivatives may form oxidative products (e.g., dimers) also behaving as efficient scavengers. The spin density distribution of the aryloxy radicals formed after HAT from the guaiacyl moieties highlighted multiple reactive sites for the further dimerization reactions in solution (Figure 3).

Compounds 5 and 6 are special cases because of possessing delocalized π electrons in the side-chain at the para position to phenolic OH group. Both compounds differ from the location of the double bond (Figure 1). In 5, the double bond is conjugated with the phenyl ring, allowing the radical delocalization after HAT over the phenyl ring and the side-chain. The longer the delocalization, the more stable the radical, and the lower the BDE. Compound 5 indeed exhibits a low BDE value of 77.7 kcal mol⁻¹ (Table 1). This makes 5 efficient at scavenging both DPPH^• and ABTS^•+ free radicals. Bortolomeazzi and coauthors evidenced that 5 may form oxidatively-induced biphenyl dimers, following fast kinetics [47]. These dimers appeared 90 times less active than 5 in terms of antioxidant efficacy, most probably because of the lowering of electrophilicity by acquiring the electron-accepting substituent in the 5th position of the guaiacyl unit (i.e., increasing BDE values). This is confirmed by the O–H BDE of 81.9 kcal mol⁻¹ as obtained for the most abundant dimer dehydrodiisoeugenol (15), as seen in Table 1. Some contradiction between a very low BDE and the RDI value lower than 1 of isoeugenol is therefore rationalized by the quick formation of less active isoeugenol dimers. The high rate constant of dimerization may also cause quick self-termination between two isoeugenol radicals, explaining the fact that its stoichiometric factor is lower than 1 against DPPH^• [20]. The double bond in 6 is not conjugated with the phenyl ring, therefore decreasing the π-conjugated path compared to 5. As a direct consequence, the BDE of the former is higher than that of the latter (81.0 and 77.7 kcal mol⁻¹, respectively). However, surprisingly, 6 is more active than 2–4 at scavenging DPPH^• (Table 1). Dimers can also be formed by two eugenolyl radicals [47]. The most abundant dimer, dehydrodieugenol (16), is approximately twice as active as the monomer (5) in terms of free radical scavenging. The BDE of 16 is 80.1 kcal mol⁻¹, confirming its better activity. As steric hindrance may prevent planarity of the phenyl rings and extension of the π-conjugated path in the dimer, the increase in activity is mainly due to the presence of two phenolic OH groups, each one being prone to HAT and to efficient scavenge more than one free radical per one phenolic group. Additionally, 16 reacts quicker with DPPH^• than 6 [47]. The 6 is therefore able to scavenge free radicals multiple times due to the formation of dimers, thus explaining the difference between its high BDE and the stoichiometric factors against DPPH^•.

Surprisingly, both 5 and 6 showed a similar and efficient antioxidant activity against ABTS^•+ (RDI values 1.45 and 1.44 for 5 and 6, respectively). This can be rationalized by the kinetically preferential SPLET mechanism in water. Both compounds exhibit low ETE values (105.7 and 107.3 kcal mol⁻¹, respectively), which correlate with their efficient antioxidation activity in water against ABTS^•+.

Vanillin, acetovanillone, and propiovanillone (7–9) are guaiacol derivatives substituted with a chain bearing a carbonyl group in the α-position. These three compounds almost entirely lost DPPH^• scavenging activity (Table 1) and exhibited high BDE values ranging from 85.0 to 85.3 kcal mol⁻¹. The carbonyl group, being highly electronegative, attracts the density of electrons, thus reducing the π-conjugation and the stability of the aroxyl radical (Figure 3).

The length of the side-chain (one to three carbon atoms for 7 to 9) had no influence on BDEs and antioxidant activities. Interestingly, these three compounds still exhibited ABTS^•+ scavenging activities (0.57, 0.54, and 0.45 for 7, 8, and 9, respectively). As expected from the high BDE values, these activities were rather low, much lower than that of compounds 1–6. The discrepancy between both assays may come from the thermodynamic balance of scavenging reactions. In the case of DPPH^•, DPPH-H exhibits a BDE of 80.4 kcal mol⁻¹ in polar solvent [42], hence HAT from compounds 7–9 to DPPH^• is thermodynamically unlikely, therefore forbidding almost all primary scavenging action. Moreover, compounds 7, 8, and 9 have pKa of 7.40, 7.81, and 7.98 (Table 1), reflecting acidity of the phenolic OH group. In buffered solution (pH = 7.35, see Experimental section), this corresponds to the presence of 44%, 11%, and 5% phenolates (deprotonated forms of the phenolic compounds), for the three compounds, respectively. Therefore, the SPLET mechanism may also proceed, as reflected by the ETE descriptors. Indeed, ABTS^•+ electron affinity is −131 kcal mol⁻¹ and the ETE values of 7, 8, and 9 are 121.4, 119.5, and 118.9 kcal mol⁻¹, respectively.

Homovanillic acid (10) and vanillylmandelic acid (11) are guaiacyl derivatives substituted with a chain containing a carboxylic acid in β-position. The pK_a of both compounds are 4.41 and 3.43, respectively, in water solution corresponding to the carboxylic acid group [48]. Thus, in all assays assessed in water, such as the ABTS^•+ assay, the monoanionic form must be considered for interpretation. In pure methanol, the carboxylic acid pK_a of both compounds is higher than 9 [49]. Therefore, concerning the DPPH^• assay performed in methanol, the protonated form is predominant. The antioxidant activity of 10 and 11 is similar to that of 1 (RDI of 1.18 and 0.98 for both compounds, respectively, as depicted in Table 1). Compound 10 appears slightly more active than 1 and 11, in agreement with slightly lower BDE values (BDE of 82.0, 82.4, and 82.5 kcal.mol⁻¹ for the three compounds, respectively); however, both from theoretical and experimental points of view these differences are almost insignificant. In 10 and 11, the antioxidant constructive effect of the alkyl chain appears counterbalanced by the antioxidant destructive effect of the carboxylic acid moiety. It is worth noting that the hydroxyl group in the α-position of 11 has a slight detrimental effect. Low phenolic BDEs for the monoanionic forms (78.7 and 79.0 kcal mol⁻¹ for 10 and 11, respectively) rationalized the rather high ABTS^•+ scavenging (RDI of 1.67 and 0.92, respectively). The slightly lower BDE of 10 may explain its somewhat better activity with respect to 11, however this difference is not very significant. A better explanation is given by ETE, which appears slightly lower for 10 than for 11. This confirms, at least partially, the contribution of SPLET mechanism, i.e., facilitated electron transfer from the carboxylated form to ABTS^•+. Vanilglycolic acid (12) bears both the properties of a carbonyl group in α-position and a carboxylic acid in the β-position. This compound has no scavenging activity against DPPH^•, which is fully supported by the high O–H BDE value (86.9 kcal mol⁻¹). In water, 12 has two low pK_a (pK_a1 = 1.60 and pK_a2 = 7.54 [43]), meaning that in the ABTS^•+ assay, this compound is mainly mono-deprotonated (carboxylate) and slightly bi-deprotonated (carboxylate and phenolate). The first ETE value (energy required to remove an electron from the carboxylate) is high thus making SPLET inefficient. The second ETE (energy required to remove an electron from the phenolate) is significantly lower, allowing SPLET towards ABTS^•+. Therefore, the remaining ABTS^•+ scavenging activity is most probably due to the small percentage of bi-deprotonated form of 12.

Ferulic acid (13) is a very well-known antioxidant polyphenol. Experimentally, it showed very good DPPH^• scavenging activity (RDI = 1.23), in agreement with previously reported values [50,51]. Its dihydrogenated analog dihydroferulic acid (14) exhibits the same DPPH^• scavenging activity. As for compounds with carboxylic acid moieties in β-position, these compounds exhibit low pK_a values in water (4.56 for 13) but their pK_a values in methanol are expected to be much higher [49]. Therefore, 13 and 14 are most likely present in their protonated form in methanol. Hence, the similar BDEs of the protonated forms (81.8 and 81.1 kcal mol⁻¹, respectively) rationalize similar DPPH^• scavenging activities for both compounds behaving by PCET. In water, both 13 and 14 are deprotonated from the carboxylic acid moieties (first pK_a value is 4.56 for 13 [48]). Therefore, the higher antioxidant activity of 13 against ABTS^•+can be explained by the BDE of its deprotonated form (ferulate, BDE of 77.6 kcal mol⁻¹) compared to 14 (dihydroferulate, BDE of 79.5 kcal mol⁻¹). Additionally, the larger π-conjugated path (delocalization of electron density) in ferulate with respect to dihydroferulate also accounts for the higher activity of ferulate. In their respective radical forms, the spin density of the radical is only located on the phenyl ring for dihydroferulatoxyl radical, whereas it also spans the conjugated side-chain for ferulatoxyl radical (Figure 3). A higher delocalization of the spin density induces higher stability of the radical, thus in agreement with lower BDE and higher antioxidant activity. It should be noted that the effect of the π-conjugation is different in 13 and 14, compared to 5 and 6, due to the presence of the carboxylic moiety.

Compound 17 (divanillin) is the C5–C5 dimer formed from two vanillin molecules (Figure 1). It possesses two active symmetrical OH groups. Compounds 18 and 19 are arylether guaiacol dimers found in lignins. In both cases, the two guaiacyl units are β-O-4’linked. Thus, both compounds have only one active phenolic OH group. These three dimers exhibit carbonyl moieties in α-position on their para side-chains, so they share the same properties as their corresponding monomers (compounds 7 and 8). Moreover, no additional π-conjugation arises from the dimerization process, even for divanillin in which steric hindrance prevents coplanarity of both phenyl rings. Therefore, these compounds show the absence or just weak activity against DPPH^• and ABTS^•+, respectively, as confirmed by the high O–H BDE values (84.8, 86.0, and 86.5 kcal mol⁻¹ for 17, 18, and 19, respectively). It is worth noting that even though 19 has a carboxylic acid moiety that can be deprotonated in solution, it is located on the guaiacyl moiety without the active OH group so with almost no influence on the BDE (Table 1).

Since BDE allows direct describing of the structure–antioxidant activity relationship, this parameter was chosen to quantify the contribution of each structural descriptor. For this purpose ΔBDE values were calculated among derivatives with or without of the corresponding chemical pattern (Table 2). The ΔBDE was calculated for each pair of compounds shown in Table 2 and the average ΔBDE value of the involved pairs of compounds was taken for each structural descriptor (Table 2).

The presence of α-carbonyl group affects the reactivity of phenolic OH group much stronger in comparison to other structural descriptors, as can be seen by the biggest increase of BDE (highest ΔBDE value) (Table 2). This rationalizes that compounds containing α-carbonyl groups are almost inactive at scavenging DPPH^• radicals (e.g., RDI of compound 7 was 0.02, see Table 1). Because of this on/off effect, the ΔBDE of α-carbonyl groups was taken as a reference to normalize the impact of the other descriptors (Table 2).

The quantitative impact of different structural descriptors (chemical substituents) on the reactivity of the guaiacyl OH group (Table 2) was used to rationalize the antioxidant activity of the oligomeric and polymeric lignin products and to find the most appropriate way for modification of technical lignins to obtain highly efficient phenolic antioxidants. Soda lignins used as a raw material in this study are heterogeneous both from point view of their chemical composition and the structure with a rather broad molecular weight distribution. Regarding the practical application of these antioxidants, numerous studies have highlighted the advantage of isolation of relatively uniform lignin fractions for these purposes, e.g., obtained by solvent fractionation. The most prospective approach is reported to obtain oligomeric fractions isolated from technical lignins by dichloromethane (CH₂Cl₂) [22,23,26,27,28,29,31,32,35]. Here, different soda lignins and their CH₂Cl₂ fractions were characterized in terms of the chemical composition, molecular weight distribution, content of functional groups, and other structural properties (Table 3).

Their antioxidant activity was examined and compared to that of the synthetic antioxidants TBHQ and Irganox 1010 (Figure 4), which are highly effective and widely used commercial phenolic antioxidants.

Despite numerous advantages of CH₂Cl₂-fractioning, including a significantly higher content of phenolic OH moieties, lower heterogeneity in terms of composition, and the molecular weight distribution in comparison to the parent lignins, a noticeable part of phenolic OH groups in CH₂Cl₂ fractions is inactive in radical scavenging reactions (i.e., RDI < 1, see Table 3). The antioxidant activity assay of the oligomeric lignin CH₂Cl₂ fractions shows that they are not competitive in comparison to the commercial phenolic antioxidants TBHQ and Irganox (Table 3). This behavior could be explained by the negative effect of the structural descriptors mentioned in Table 2 within the lignin structure, being the α-C=O the most suspicious. Hence, this proposition was verified by more detailed analysis of lignins and the corresponding CH₂Cl₂ fractions.

The analysis of lignins in strongly basic aprotic solvents allow to discriminate and quantify the phenolic hydroxyls in noncondensed and condensed structures [52,53]. As an example, the analysis of black alder lignin and its corresponding CH₂Cl₂ fraction is envisaged, but very similar features were find for other lignins in study. The ¹H-NMR spectra in highly basic solution (HMPT-d₁₈) show the difference between the black alder parent lignin and the corresponding CH₂Cl₂ fraction (Figure 5).

The phenylpropane units containing phenolic OH revealed the characteristic resonances in the spectra range of 8.6 to 11.0 ppm [52]. CH₂Cl₂ fraction showed an enhanced content of α-carbonyl (αCO) (signal at 10.5–10.7 ppm [52]) and aldehyde groups (CHO) (signals at 10.2–10.4 ppm [52]) in the units containing the phenolic groups (Figure 5). They also have an increased content of carboxyl groups (signal at 11.46 ppm [52]) in comparison to parent lignin. According to the data, obtained by density functional theory (DFT) calculations (Table 2), the presence of these oxygen-containing groups in the side-chain, especially carbonyl groups, is the major reason for the inactivation of lignin CH₂Cl₂ fractions in radical scavenging reactions. The results of ¹H-NMR analysis were confirmed by the Py-GC/MS, which shows the significantly higher yield of α-carbonyl group-containing compounds in the lignin derived products for CH₂Cl₂ fractions in comparison to the parent lignins (Table 3).

Due to the aforementioned structural features, the elimination of α-carbonyl is prone to increase drastically the antioxidant activity of lignin fractions. To tackle this hypothesis, the chemical reduction of α-carbonyl moieties was performed. The most commonly used reaction for the reduction of carbonyl groups of lignins is by using NaBH₄; however, in such case the carbonyl groups are reduced to aliphatic OH groups which also have slightly negative effect on the antioxidant activity (Table 2). Instead, the chemical transformation of carbonyl groups into CH₂ fragments would not only avoid the negative effect of the former but also bring the positive effect of the latter, see Table 2. Therefore, this modification of lignin products was performed according to the reaction sequence shown in Figure 6.

The elimination of the carbonyl groups was also controlled by FTIR spectroscopy (Figure 7). The variation of the intensity of the band assigned to the carbonyl groups vibration in the normalized spectra of flax, black alder, and ash-tree soda lignin CH₂Cl₂ fractions after their modification decreased from 76.2 to 19.5%, from 29.7 to 11.3%, and from 28.0 to 13.0%, respectively (Figure 7).

The elimination of the carbonyl groups was also confirmed by Py-GC/MS showing a significant decrease in the yield of the lignin-derived compounds containing the carbonyl group with a simultaneous increase in the yield of the compounds containing α-CH₂ fragments (Figure 8).

The decrease of the relative content of phenolic OH groups conjugated with carbonyl groups (with respect to the total OH phenolic) was also confirmed using differential UV/Vis spectroscopy (Table 4).

As it was predicted, the elimination of carbonyl groups drastically increases the reactivity of lignins phenolic OH groups making them competitive with respect to the commercial phenolic antioxidants TBHQ and Irganox 1010 (Table 4).

Commercially available synthetic antioxidants inhibit significantly the biodegradation of polymeric materials. Conversely, lignins can stimulate their decomposition under natural environment after the end of their service life. Therefore, replacing synthetic antioxidants with natural ones would promote the production of environmentally friendly materials. For example, it was shown that the addition of lignin (4.2% to 9.3% in weight) in the poly(ethylene adipate) urethane films significantly improved their degradability by fungi enzymes (e.g., by laccase isolated from the Aspergillus sp.) [13].

In this work, the influence of different antioxidants on thermooxidative destruction of polyurethane (PU) films at 310 °C in air was studied for the demonstration purposes. Accordingly, flax soda lignin modification products were compared to Irganox 1010, a highly effective antioxidant against thermooxidative degradation of polyurethanes [54]. The kinetic curves of the thermooxidative degradation of the polyurethane films show that the addition of the antioxidant decreases the weight loss rate, especially in the first stage of thermooxidation destruction (Figure 9), which proceeds by radical mechanism [55].

Without antioxidant, 50% of the weight loss of the material is reached after 32.5 min of incubation. With 2% of flax soda lignin CH₂Cl₂ fraction 50% of the weight loss of the material is reached after 74.5 min of incubation. Adding 2% of Irganox or modified flax soda lignin CH₂Cl₂ fraction increases this time to 100 or 111.5 min, respectively (Figure 9).

The antioxidant activity was also assessed as integrals of the kinetic curves (Figure 9). The integral antioxidant activity as portion from the activity of Irganox was 0.6 for flax soda lignin CH₂Cl₂ fraction and 1.3 for modified flax soda lignin CH₂Cl₂ fraction. This indicated that the modified flax soda lignin CH₂Cl₂ fraction performed 30% better inhibition effect than Irganox. This confirms a highly efficient antioxidant activity of this lignin derivative, and outlines the promising application of lignins in environmentally friendly materials.

Model compounds 1–11, 13, 14, and 17 (Figure 1) used in this study are commercially available:

1.: 2-methoxyphenol (guaiacol), ≥ 98%, Sigma Aldrich, CAS: 90-05-1

2.: 2-methoxy-4-methylphenol (4-methylguaiacol), ≥ 99%, Acros organics, CAS 93-51-6

3.: 2-methoxy-4-ethylphenol (p-ethylguaiacol), ≥ 98%, SAFC, CAS 2785-89-9

4.: 2-metoxy-4-propylphenol (4-propylguaiacol), ≥ 99%, Sigma Aldrich, CAS 2785-87-7

5.: 2-methoxy-4-propenylphenol (isoeugenol), 99%, Sigma Aldrich, CAS 97-54-1

6.: 4-allyl-2-methoxyphenol (eugenol), ≥ 98%, Sigma Aldrich CAS 97-53-0

7.: 4-hydroxy-3-methoxybenzaldehyde (vanillin), 99%, Sigma Aldrich, CAS 121-33-5

8.: 4’-hydroxy-3’-methoxyacetophenone (acetovanillone), 98%, Acros organics, CAS498-01-2

9.: 1-(4-hydroxy-3-methoxyphenyl)-1-propanone (propiovanillone), ≥ 95%, Bide Pharmatech, CAS 1835-14-9

10.: 4-hydroxy-3-methoxyphenylacetic acid (homovanilic acid), Acros organics, CAS 306-08-1

11.: (RS)-hydroxy(4-hydroxy-3-methoxy-phenyl)acetic acid (vanillylmandelic acid), 99%, Acros organics, CAS 55-10-7

13.:  (E)-3-(4-hydroxy-3-methoxy-phenyl)prop-2-enoic acid (trans-ferulic acid), 99%, Sigma Aldrich, CAS 537-98-4

14.: 3-(4-hydroxy-3-methoxyphenyl) propanoic acid (dihydroferulic acid), ≥ 96%, Sigma Aldrich, CAS 1135-23-5

17.: 6,6′-dihydroxy-5,5′-dimethoxy-[1,1′-biphenyl]-3,3′-dicarboxaldehyde (divanillin), Fluorochem, CAS 2092-49-1

Compound 12 (4-hydroxy-3-methoxyphenyl)(oxo) acetic acid (or vanilglycolic acid) was synthesized at the Latvian State Institute of Wood Chemistry by the oxidation of compound 8 by nitrobenzene in alkaline medium at 100 °C, using the method reported by Mottern [56] with minor modifications. Compound 8 (35 g), nitrobenzene (27 g), and NaOH (25 g) were mixed in 75 mL water at 100 °C for 24 h. Aniline and small amount of nitrobenzene were distilled with water vapor; the cooled solution was acidified, treated with active oxygen and filtered. The solution was cooled and compound 12 was crystallized. The amount of compound 12 was 34 g (85%) (T_melting = 133 °C) and its purity was 85%, as checked chromatographically (GC, gas chromatography).

Compounds 18 and 19 (Figure 1) were synthesized by the methods described previously [57,58,59]. Their purity was 88 and 90%, respectively, as checked by GC.

The soda sulfur-free lignins were produced on industrial scale from annual (nonwood) plants using the patented lignin precipitation process applied to black liquor from soda pulping [60]. Soda lignin from flax (Linum usitatissimum) was kindly supplied by the Granit SA (Lausanne, Switzerland).

The wood based soda lignins were obtained as a by-product from the black liquor of soda pulping (the IWC pilot equipment). The wood chips of black alder (Alnus glutinosa) and the ash-tree (Fraxinus excelsior) were pre-extracted with acetone to remove the extractives according to TAPPI standard T 280 pm-99. The residual content of the extractives was 0.15% on the oven-dried wood. The 280 g of chips were cooked in the digester with 2240 mL of 5% sodium hydroxide (Sigma Aldrich, Steinheim, Germany) which corresponds to an alkali charge of 42% based on the dry wood. Once the reactor reached the operating temperature of 160 °C, it was maintained for 3 h. The pulp was filtered, and the filtrate was acidified to pH 2 to precipitate the lignin part. Lignin was filtered off and washed with a hot water until neutral reaction. The obtained lignin was dried under vacuum at 40 °C. The yields of soda lignins from black alder and ash tree were ~20%.

Both industrial and laboratory lignins were used as raw materials for modification to obtain the highly effective lignin-derived antioxidants. The oligomeric fraction of lignin was obtained by fractionation with CH₂Cl₂ as described previously [61].

The O–H bond dissociation enthalpy (BDE) of each phenol moiety (ArO-H) was calculated based on the enthalpies balance as follows.

It is noteworthy that the lower the BDE, the easier the cleavage of the O–H bond, and the higher the antioxidant activity. Similarly, the electron transfer enthalpy (ETE) was calculated as follows.

DFT calculations were performed with the B3P86 functional in Gaussian program as it was shown to accurately describe polyphenol BDEs [39,40,62,63]. The 6-31+G(d,p) basis set was used as providing very similar results compared to the larger and more computationally demanding 6-311+G(2d,3pd) basis set. Ground-state geometries were confirmed by a vibrational frequency analysis that indicated the absence of imaginary frequency. Enthalpies were calculated at 298 K and 1 atm. As many lignin fragments possess carboxylic acid moieties that can be deprotonated in aqueous media, all protonation states were evaluated. The solvent effect was considered using the integral-equation-formalism polarizable continuum model (IEF-PCM). Continuum models consider the molecular system embedded in a shape-adapted cavity surrounded by a dielectric continuum characterized by its permittivity (ε = 32.61 for methanol). Calculations in methanol reproduced the conditions of the DPPH^• assay. All calculations were performed with the Gaussian09^® software from Gauss Inc, USA [64].

Chemicals used for the antioxidant evaluation (DPPH^•, ABTS^•+ and solvents) were of analytical grade (Sigma–Aldrich). All solutions were prepared freshly before the measurements.

DPPH^• (2,2-diphenyl-1-picrylhydrazyl) was dissolved in methanol at a final concentration of ~10–4 mol L⁻¹ and incubated in the dark at room temperature for 16 h. The exact concentration of DPPH^• in terms of absorbency at 515 nm was calculated from a calibration curve (A₅₁₅ = 8.3 × C_DPPH^• + 0.001).

A stock solution of ABTS (2 mM) was prepared in 50 mM phosphate buffered saline (PBS) made of 8.18 g NaCl, 0.27 g KH₂PO₄, 3.58 g NaHPO₄ x11 H₂O, and 0.15 g KCl in 1 L of distilled water. When necessary, the pH of the solutions was adjusted at 7.4 with 0.1 M NaOH. The ABTS^•+ solution was produced by mixing 50 mL of ABTS stock solution with 0.2 mL of K₂S₂O₈ (70 mM) aqueous solution. The mixture was kept in the dark at room temperature during hours 15 to 16 before use. To evaluate the antioxidant capacity, the ABTS^•+ solution was diluted with PBS to reach the absorbance at 734 nm of 0.800 ± 0.030 [65]. The exact concentration of ABTS^•+ was calculated from the calibration curve (A₇₃₄ = 23.85 × C_ABTS^•+ − 0.007).

A 0.03 mL DMSO solution of the lignin or relative phenolic compounds was added to 3 mL of the prepared the DPPH^• or the ABTS^•+ solution. Five different concentrations were used for each derivative. The decrease in DPPH^• or ABTS^•+ concentration was followed at a 515 or 734 nm respectively until the reaction’s completion using Perkin Elmer UV/VIS spectrometer Lambda 650 equipped with a thermally controlled cell at 22 °C. All experiments were done in triplicate. The DPPH^• or ABTS^•+ concentrations at steady state (i.e., primary reaction completed and no change in solution for most of compounds) were plotted as a function of the molar concentrations of the lignins or relative phenolic compounds. The antioxidant activity was expressed as a radical deactivation index (RDI) which corresponds to the number of deactivated free radicals per molecule of tested compounds in the case of lignin modeling compounds, or as the number of deactivated free radicals per phenolic OH group in the case of technical lignins and their fractions. The obtained results are shown as the average with a confidence interval (α = 0.05 level of significance). The RDI index is independent of the free radical concentration. The higher the RDI, the higher the free radical scavenging activity.

The procedure describing the reduction of aldehyde and ketone functions to methylene groups, using the tosylhydrazone derivatives under very mild conditions [68] was adapted for lignin. The yield of the modified lignin was ~75%. Recrystallization from the aqueous methanol provides analytically pure product with the yield ~64%.

Thermooxidative destruction of the model polyurethane (PU) elastomer (control and containing 2% of lignin sample films) was studied at 310 °C under isothermal regime by TGA in air atmosphere (flow rate 50 mL min⁻¹) using the Metler Toledo Star System TGA/ADTA 851e device. A 8–10 mg sample was used. The PU elastomer films were obtained as described previously [33].