The Impact of Thermal Treatment on Structural Changes of Teak and Iroko Wood Lignins

: Thermal modiﬁcation is an environmentally friendly method to improve dimensional stability, durability, and aesthetic properties of wood. Changes in lignin as one of the main wood components markedly inﬂuence wood product properties and recycling possibilities of thermowood at the end of its life cycle. Teak and iroko wood samples were thermally treated at the temperatures of 160 ◦ C, 180 ◦ C and 210 ◦ C following the Thermowood process. Dioxane lignin was isolated from treated and untreated wood and analysed by nitrobenzene oxidation (NBO), size exclusion chromatography (SEC) and Fourier transform infrared spectroscopy (FTIR). The yields of both acid-insoluble and dioxane lignins increased with an increasing treatment temperature. Dioxane lignins are GS-types containing more guaiacyl units compared to syringyl ones with S / G ratios of 0.91 and 0.84, respectively. In the process of thermal modiﬁcation, several degradation and condensation reactions were observed. The cleavage of methoxyl groups and side chains, oxidation reactions, cleavage of the β -O-4 ether linkage and cross-linking radicals arising at higher temperatures were all conﬁrmed. However, during the thermal treatment, teak lignin changed in a di ﬀ erent way than iroko lignin, e.g., the molecular weight of iroko lignin decreased at all applied temperatures while it increased at 180 ◦ C and 210 ◦ C in teak lignin, and the change in S / G ratio and the cleavage of alkyl-aryl bonds are di ﬀ erent in both wood species.


Introduction
Wood is a natural, renewable, mostly inexpensive and easily accessible material. Modification is applied to overcome weak points of the wood material that are mainly related to moisture sensitiveness, hardness, low dimensional stability, low resistance to wear, UV irradiation, to bio-deterioration against fungi, termites, and marine borers [1]. Wood thermal treatment is an economical and environment friendly pretreatment. It can be performed in range of 160 to 260 • C in different environments, e.g., air, vacuum, nitrogen and oil. Its main objective is to improve dimensional stability and durability, but sometimes thermal treatment is used to change the aesthetic properties of wood. The wood colour becomes darker depending on the treatment temperature, time, and environment [2][3][4].
Tropical woods generally have a great resistance to biological agents and mechanical wear, and a good dimensional stability and interesting wood figure. Many species of exotic wood are used for exterior constructions, decking, claddings, garden furniture or special plywood [5,6]. Teak is one of the most valuable woods, but its use is limited by scarcity and high cost. It is currently used in the construction of boats, flooring, decorative objects and veneer. Because of its colour and durability, iroko is suggested as a substitute for teak. Its durability makes it suitable for boat building, piles, other marine work, and railroad crossties. Other uses include joinery, furniture, and cabinetwork [7][8][9].

Lignin Isolation
The lignin isolation from the wood samples was performed using mild acidolysis and dioxane as a solvent by modified method [24]. Extracted wood flour with size ≤ 0.355 mm (10 g) was placed in a 500 mL three-necked flask fitted with a dropping funnel, a nitrogen bubbler and a reflux condenser. An amoj tof 100 mL of dioxane/water (9:1, v/v) with 1.82 g HCl (equivalent to 0.2 M) was slowly added. The mixture under nitrogen atmosphere was refluxed at 80 • C for 5 h. Then, the mixture was cooled to room temperature, filtered, neutralized by addition of ammonium acetate, and concentrated in vacuum at the temperature of 40 • C. Lignin was precipitated by slowly dropping of dioxane solution into the cold water.

Nitrobenzene Oxidation
Alkaline nitrobenzene oxidation (NBO) was carried out in a stainless-steel bomb by modified method ofĎurkovič et al. [25]. An amount of 50 mg of dioxane lignin was reacted with 5 mL of 2 M NaOH and 0.4 mL of nitrobenzene at the temperature of 180 • C for 2 h. The stainless-steel bomb was then cooled down immediately with cold water. The excess of nitrobenzene was extracted with CH 2 Cl 2 two times, and the organic phase was removed. The aqueous phase acidified by 2 M HCl to pH around 2.5 was extracted with CH 2 Cl 2 thrice. Then the organic phase was collected and dried with anhydrous Na 2 SO 4 , gently evaporated in nitrogen stream, and dissolved in methanol/water (1/1). NBO products were analysed by high performance liquid chromatography (HPLC) in an Agilent 1200 apparatus equipped with diode array detector at 240 nm, Kinetex column 2.6 µm C18 (100 × 4.6 mm), mobile phase water/methanol/propionic acid (88/4/8/0.1), flow rate 1 mL·min −1 , temperature 35 • C. The quantification was performed with the external calibration using standards obtained from Sigma-Aldrich (St. Louis, MO, USA).

Size Exclusion Chromatography (SEC)
The molecular weight distribution (MWD) of lignin was evaluated by a modified method [25]. The lignin samples were dissolved in N,N-dimethylformamide (DMF, 2 mg·mL −1 ), filtered through a 0.45 µm filter and placed in a 2 mL autosampler vial prior to SEC analysis. The separation was performed on an Agilent 1200 HPLC system equipped two PolarGel-M columns (7.5 mm × 300 mm) and a diode array detector, using dimethylformamide + 2% trifluoroacetic acid + 0.1% LiBr as the mobile phase (1.0 mL·min −1 ) at 35 • C. A calibration curve was constructed with polystyrene standards from 500 to 98,900 g·mol −1 . The data were collected using ChemStation for LC 3D systems (version Rev. B.04.03 [21]) software and calculated with the Clarity GPC module (version 7.4.01.088).

Fourier Transform Infrared Spectroscopy (FTIR)
FTIR spectra were recorded on a Nicolet iS10 spectrometer equipped with Smart iTR ATR accessory using diamond crystal. Four measurements for each sample were performed. Each spectrum was obtained by 32 scans in the range from 4000 cm −1 to 800 cm −1 with a resolution of 4 cm −1 . Measured spectra were baseline corrected and analysed in absorbance mode by OMNIC 8.0 software.

Statistical Analysis
The data were subjected to an analysis of variance (ANOVA), and the significant differences between the means were determined using Duncan's multiple range test with a p-value of 0.05.

Results and Discussion
The advantage of using dioxane lignin for structural studies are minimal structural changes of lignin during its preparation and good extraction yields [26,27].

Klason and Dioxane Lignin Yields
Lignin content in teak wood ranges from 30 to 43% [28][29][30][31]. Unmodified (reference) wood in this work has lignin content within this range (35.42%), as reported previously [32]. The apparent insoluble lignin content increased during thermal modification, when compared to the reference sample and the sample treated at the temperature of 210 • C by 14% and 27% for teak and iroko, respectively. Duncan's multiple range test shows that Klason lignin yields change significantly for both wood species (for iroko KL from the temperature of 180 • C). This is mainly due to hemicellulose degradation, which leads to condensation products as acid-insoluble residues [33]. The amount of lignin that can be extracted with dioxane depends to a large extent on the degree of its condensation [34]. The results of dioxane lignin yields (Table 1) show that the yield of lignin from wood increased 2.1-fold and 1.7-fold for teak and iroko, respectively, and these changes are significant. Similarly, the slight increase of dioxane lignin yields was observed at heat sterilization of pine wood in the temperatures up to 120 • C [25]. On the other hand, yields of dioxane lignin in spruce heated wood increased up to temperatures of approx. 210 • C, and with an increased temperature (up to 270 • C), it rapidly decreased [34].

Nitrobenzene Oxidation Products Analysis
Teak has a GS-lignin, with a monomeric composition determined by Py-GC/MS(FID) of H:G:S 1:34:24 (sapwood) and 1:29:23 (heartwood) and S/G ratios of 0.71 and 0.81, respectively [28,35]. Similar values-50.6% guaiacyl units, 46.0% syringyl units, 3.4% p-hydroxyphenyl units, with S/G ratio 0.91-were found by nitrobenzene oxidation (Table 2). On the other hand, Faix et al. (1987) [36] found up to 70% of guaiacyl structural units in teak lignin. The ratio of syringyl units to guaiacyl (S/G) decreases with an increasing temperature, the yield of NBO products significantly increases except at the highest temperature. The results indicate the predominance of degradation reactions at lower temperatures and condensation reactions at the temperature of 210 • C. Like teak lignin, iroko has a GS-lignin, containing more guaiacyl units compared to syringyl ones with an S/G ratio of 0.84 (Table 3). Some authors reported higher values of the S/G ratio for iroko lignin, 0.92 (by Py-GC-MS) and 1.005 (by FTIR) respectively [37,38]. Differences may be caused by other methods of the isolation of lignin as well as different methods of analysis. The yield of NBO products gradually increased up to 180 • C, then decreased, however, the changes are mostly insignificant. The increase in NBO products may suggest that iroko lignin treated at lower temperatures may be structurally modified to an easily degradable form. A similar trend was observed in the S/G ratio, indicating preferential contribution of G-type lignin to the condensation reaction as reported by Kim et al. (2014) [39]. Table 3. Nitrobenzene oxidation products of iroko dioxane lignins (mean ± SD, %).

Molecular Weight Determination
Lignin isolated from the untreated teak wood has a molecular weight of about 7100 g·mol −1 ; in the available literature, no adequate data were found for comparison. In the thermal treatment of teak wood, its molecular weight (MW) and polydispersity (PD) decreased at a temperature of 160 • C (Table 4). Elution profiles confirmed a cleavage of the intermediate molecular portions of the lignin macromolecule (Figure 1). At higher temperatures (180 • C and 210 • C), both degradation and condensation reactions occur simultaneously, resulting in increased polydispersity and expanded peaks in the chromatogram (Figure 1). Changes in MW and PD are significant, except the difference in PD between temperatures 180 • C and 210 • C. Condensation reactions are predominant at 210 • C, leading to a lignin molecular weight increase and to a decrease in the yield of NBO products. The reduction in M n and M w indicates that the lignin molecules were fragmented into smaller units. Oxidized lignin can undergo re-polymerization as can be observed in our GPC data and as reported by other authors [40,41]. Simultaneous degradation and condensation reactions were also confirmed by the heating of MWL lignin isolated from poplar wood [39]. The lignin thermal degradation leads to the arising of phenoxy radicals through a hydrogen atom abstraction process. They could undergo coupling reactions with the corresponding C5-centered mesomeric radical derivatives to generate new 4-O-5 and 5-5' linkages, respectively [42,43].
Appl. Sci. 2020, 10, x FOR PEER REVIEW 5 of 12 temperatures may be structurally modified to an easily degradable form. A similar trend was observed in the S/G ratio, indicating preferential contribution of G-type lignin to the condensation reaction as reported by Kim et al. (2014) [39].

Molecular Weight Determination
Lignin isolated from the untreated teak wood has a molecular weight of about 7100 g·mol −1 ; in the available literature, no adequate data were found for comparison. In the thermal treatment of teak wood, its molecular weight (MW) and polydispersity (PD) decreased at a temperature of 160 °C (Table 4). Elution profiles confirmed a cleavage of the intermediate molecular portions of the lignin macromolecule (Figure 1). At higher temperatures (180 °C and 210 °C), both degradation and condensation reactions occur simultaneously, resulting in increased polydispersity and expanded peaks in the chromatogram (Figure 1). Changes in MW and PD are significant, except the difference in PD between temperatures 180 °C and 210 °C. Condensation reactions are predominant at 210 °C, leading to a lignin molecular weight increase and to a decrease in the yield of NBO products. The reduction in Mn and Mw indicates that the lignin molecules were fragmented into smaller units. Oxidized lignin can undergo re-polymerization as can be observed in our GPC data and as reported by other authors [40,41]. Simultaneous degradation and condensation reactions were also confirmed by the heating of MWL lignin isolated from poplar wood [39]. The lignin thermal degradation leads to the arising of phenoxy radicals through a hydrogen atom abstraction process. They could undergo coupling reactions with the corresponding C5-centered mesomeric radical derivatives to generate new 4-O-5 and 5-5' linkages, respectively [42,43].     . The differences may be due to different growth locality, lignin isolation procedure, and the chromatographic method itself. During the thermal treatment, iroko lignin changes in a different way than teak lignin. At 160 • C, low molecular weight lignin fragments are degraded (Figure 2), resulting in a slight increase in molecular weight and a negligible decrease in polydispersity. Degradation reactions are predominant at 180 • C and the molecular weight decreases by almost 10% when compared to the original sample. Changes in MW are significant, some changes in PD are significant, while others are not. The temperature of 210 • C causes both condensation and degradation reactions (which prevail) and the molecular weight decreases (Figure 2, Table 5). A similar trend was observed at the thermal modification of oak wood [44]. Lignin isolated from iroko wood has a different molecular weight (about 38% lower) and polydispersity (about 44% lower) when compared to teak lignin. These values (Mw = 4500 g·mol −1 , Mn = 2100 g·mol −1 , PD = 2.1) are lower than those published by Zikeli et al. (2019) [37]-(Mw = 13,900 g·mol −1 , Mn = 2300 g·mol −1 , PD = 5.7). The differences may be due to different growth locality, lignin isolation procedure, and the chromatographic method itself. During the thermal treatment, iroko lignin changes in a different way than teak lignin. At 160 °C, low molecular weight lignin fragments are degraded (Figure 2), resulting in a slight increase in molecular weight and a negligible decrease in polydispersity. Degradation reactions are predominant at 180 °C and the molecular weight decreases by almost 10% when compared to the original sample. Changes in MW are significant, some changes in PD are significant, while others are not. The temperature of 210 °C causes both condensation and degradation reactions (which prevail) and the molecular weight decreases ( Figure  2, Table 5). A similar trend was observed at the thermal modification of oak wood [44].

FTIR Spectra Analysis
The most frequent changes in lignin structure are determined in the FTIR spectra around the bands: 1720, 1600, 1500, 1460, 1420, 1370, 1326, 1266, 1220 and 1120 cm −1 [45,46]. A wide band in the range of 3300-3500 cm −1 and bands of about 2940 cm −1 and 2850 cm −1 can also be used to detect changes in lignin. The changes in both wood species, teak and iroko, are similar, thus their FTIR spectra will be evaluated together.
Broad bandwidth low absorbance around 3400 cm −1 (OH stretching vibration in primary or secondary alcohols, as well as in phenolic groups in lignin) [4,47] predominantly decreases with an increasing temperature. The reduction of hydroxyl groups is accompanied by condensation of lignin and oxidation to C=O groups [48,49] and was observed in thermally treated teak wood [50]. Bands  Table 5. Results of macromolecular characteristics of lignins from iroko wood. Data represent mean ± SD.

FTIR Spectra Analysis
The most frequent changes in lignin structure are determined in the FTIR spectra around the bands: 1720, 1600, 1500, 1460, 1420, 1370, 1326, 1266, 1220 and 1120 cm −1 [45,46]. A wide band in the range of 3300-3500 cm −1 and bands of about 2940 cm −1 and 2850 cm −1 can also be used to detect changes in lignin. The changes in both wood species, teak and iroko, are similar, thus their FTIR spectra will be evaluated together.
Broad bandwidth low absorbance around 3400 cm −1 (OH stretching vibration in primary or secondary alcohols, as well as in phenolic groups in lignin) [4,47] predominantly decreases with an increasing temperature. The reduction of hydroxyl groups is accompanied by condensation of lignin and oxidation to C=O groups [48,49] and was observed in thermally treated teak wood [50]. Bands near of 2937 and 2840 cm −1 (arising from C−H stretching vibrations in methyl, methylene and methoxyl groups and tertiary CH groups) [51][52][53] show only insignificant changes (especially in teak wood). The band found in the carbonyl region between 1700-1730 cm −1 (C=O stretching in unconjugated functional groups of aldehydes, ketones, esters and carboxylic acids) [51,52,54]. Our results showed an increase in the absorbance at 1715 cm −1 (teak) and 1712 cm −1 (iroko) (Figures 3 and 4). The increase may be confirmation of the more pronounced cleavage of the β-alkyl-aryl ether bonds and the subsequent formation of new C=O groups in the lignin structure. This results in an increase in the amount of acetyl and carboxyl groups from the lignin [52,55]. The mentioned changes are due to splitting of aliphatic side chains in lignin accompanied by cleavage of β-O-4 bonds during its heat treatment [56,57]. Similarly, the cleavage of the β-O-4 linkages and the splitting of the aliphatic methoxyl chains from the aromatic ring in lignin were observed at steam-heat treated teak wood [58].
This trend suggests the relative thermal stability of the lignin aromatic structures [63].
At bands of 1460 cm −1 and 1422 cm −1 (asymmetric C−H deformations in lignin and carbohydrates) [51,64], a slight decrease with increase of used treatment temperature was noted. Thus, the decrease of absorbance in iroko was more pronounced than in teak (this may be due to a more notable cleavage of methoxyl groups in the iroko). According to Cheng et al. [65], the decrease absorbance may be due to a decrease in the methoxyl groups in lignin. The peak absorbance around 1327 cm −1 (C−O vibration in syringyl derivatives-condensed structures in lignin) [38,62] increased permanently, except for the iroko at the highest temperature, where a decreased absorbance was observed. This suggests the condensation reactions in lignin structure (more pronounced in teak than iroko wood) as observed by Li et al. (2015) [58] for teak wood. A similar trend was also found for the band absorbances around 1265 cm −1 and 1220 cm −1 (C−O stretching of guaiacyl units) [63,66]. Considering the absorbance values at 1265-1267 cm −1 , it can be assumed that the proportion of guaiacyl units in the teak wood lignin is higher than in iroko. In contrast, the NBO results show a higher S/G ratio in teak lignin compared to iroko lignin. This difference may be because NBO provides the results only for non-condensed lignin substructure, while FTIR and SEC analyses show the properties of the whole dioxane lignin. Larger amounts of guiacyl units in teak lignin according to the FTIR method indicate their greater condensation, resulting in lower yields of NBO products and higher S/G ratio relative to iroko lignin.
The band at 1120 cm −1 (C−H vibration in lignin) is typical of syringyl units [67,68]. This band markedly decreased with an increasing temperature in both wood species. Compared to the band near 1220 cm −1 , there is a notable decrease in absorbance, which may be due to higher thermal stability of guaiacyl structures against syringyl ones, as reported by Ragauskas et al. [69].

Conclusions
During the thermal treatment of teak and iroko wood, the apparent insoluble lignin content increased, when compared to the reference sample and the sample treated at the temperature of 210 °C by 14% and 27%, respectively. The yields of dioxane lignin from wood increased 2.1-fold for teak An increase in the band at 1665 cm −1 (C=O stretching in various conjugated groups in lignin) [59,60] was observed at teak wood at 210 • C (this band did not appear at any temperature value for the iroko wood). This trend may be an indicator of the cleavage of α-alkyl-aryl ether linkages [25]. The peaks near 1600 cm −1 and 1500 cm −1 (C=C aromatic skeletal vibrations) [61,62] show only small changes with increasing temperature (a more pronounced decline is noticeable in the iroko wood). This trend suggests the relative thermal stability of the lignin aromatic structures [63].
At bands of 1460 cm −1 and 1422 cm −1 (asymmetric C−H deformations in lignin and carbohydrates) [51,64], a slight decrease with increase of used treatment temperature was noted. Thus, the decrease of absorbance in iroko was more pronounced than in teak (this may be due to a more notable cleavage of methoxyl groups in the iroko). According to Cheng et al. [65], the decrease absorbance may be due to a decrease in the methoxyl groups in lignin. The peak absorbance around 1327 cm −1 (C−O vibration in syringyl derivatives-condensed structures in lignin) [38,62] increased permanently, except for the iroko at the highest temperature, where a decreased absorbance was observed. This suggests the condensation reactions in lignin structure (more pronounced in teak than iroko wood) as observed by Li et al. (2015) [58] for teak wood. A similar trend was also found for the band absorbances around 1265 cm −1 and 1220 cm −1 (C−O stretching of guaiacyl units) [63,66]. Considering the absorbance values at 1265-1267 cm −1 , it can be assumed that the proportion of guaiacyl units in the teak wood lignin is higher than in iroko. In contrast, the NBO results show a higher S/G ratio in teak lignin compared to iroko lignin. This difference may be because NBO provides the results only for non-condensed lignin substructure, while FTIR and SEC analyses show the properties of the whole dioxane lignin. Larger amounts of guiacyl units in teak lignin according to the FTIR method indicate their greater condensation, resulting in lower yields of NBO products and higher S/G ratio relative to iroko lignin.
The band at 1120 cm −1 (C−H vibration in lignin) is typical of syringyl units [67,68]. This band markedly decreased with an increasing temperature in both wood species. Compared to the band near 1220 cm −1 , there is a notable decrease in absorbance, which may be due to higher thermal stability of guaiacyl structures against syringyl ones, as reported by Ragauskas et al. [69].

Conclusions
During the thermal treatment of teak and iroko wood, the apparent insoluble lignin content increased, when compared to the reference sample and the sample treated at the temperature of 210 • C by 14% and 27%, respectively. The yields of dioxane lignin from wood increased 2.1-fold for teak and 1.7-fold for iroko. Both teak and iroko lignins are GS-types containing more guaiacyl units compared to syringyl ones with S/G ratios of 0.91 and 0.84, respectively. In the process of thermal modification, several degradation and condensation reactions were observed. The cleavage of methoxyl groups and side chains, oxidation reactions, cleavage of the β-O-4 ether linkage and cross-linking radicals arising at higher temperatures as reported by other authors at various conditions of lignin treatment were all confirmed. However, during the thermal treatment, teak lignin changes in a different way than iroko lignin, e.g., the molecular weight of iroko lignin decreased at all applied temperatures while it increased at 180 • C and 210 • C in teak lignin, and the change in S/G ratio and the cleavage of alkyl-aryl bonds are different in both wood species. This can lead to various alterations of the physical and mechanical properties in the process of thermal modification and in the recycling of treated wood.