Thermochemical modification of beech wood with ammonium hydroxide

The beech wood under examination (Fagus sylvatica L.) originated from the Bieszczady Mountains in Poland, where it underwent a seasoning period of two years. Before commencing the experiments, the wood was dried at 105 ℃ for 12 h in a laboratory dryer to a moisture content of 10%. Test samples measuring 20 × 20 × 30 mm were extracted from the prepared boards. For consistency in the test outcomes, boards with a straight grain devoid of defects such as knots, cracks, and other visible blemishes were deliberately chosen. The orientation of the sample structure was aligned with the anatomical planes of the log, transverse, radial and tangential, as illustrated in Fig. 1.

Five series of samples of 20 pieces each were prepared.

With a thermochemical modification of wood, it is necessary to prove that a chemical bond has been formed with the wood cell wall polymers. A straightforward test involves determining the weight and volume change of the wood before and after modification (Hill 2006). Table 1 shows the average results of the WPG (the index indicating the percentage change in weight) and BC (the index of percentage change in volume) measurements.

The most extensive weight loss (Table 1) was found for the modification of beech wood with a 10% ammonia water solution for 24 h, and the smallest for the modification with a 5% NH₄OH solution for 12 h. On the contrary, the percentage bulking coefficient (BC) showed the opposite trend: its value decreases with increasing concentration of NH₄OH concentration and increasing modification time.

Under the influence of an aqueous solution of ammonia, there is a reduction in the dimensions of samples. Literature data confirm that ammonia significantly affects the reduction of these dimensions, especially in deciduous species. Oniśko and Matejak (1971) found that under the influence of a 25% aqueous ammonia solution, the shrinkage of the beech wood was 11.4% in the tangential direction and 5.2% in the radial direction, while for the oak wood, it was 14.4% and 9.8%, respectively. On the contrary, the pine samples exhibited minimal dimensional changes, with only slight variations in length and width (1.8% and 1.7%, respectively). The different effect of ammonia on deciduous and coniferous species is probably related, on the one hand, to the different availability of cell membranes, on the other hand, to the different solubility of the wood substance in it, resulting, in turn, from the different chemical composition and submicroscopic structure of individual wood species.

There were no significant changes in the density of modified beech wood (Table 1), even though the results obtained for beech wood by Bariska (1975) indicate a change in the density of the wood substance under the influence of ammonia of different durations.

With increasing ammonia concentration, the color of the treated samples gradually deepened and assumed different shades. Figure 2 shows the appearance of the ammonia modification wood under different treatment conditions.

Table 2 shows the measured values of the CIELAB color system.

The values presented in Table 2 were compared statistically. A two-way ANOVA analysis shows a statistically significant difference in average L* values for both contraction variants (F(2) = 0.232, p < 0.001) and for modification times (F(1) = 1.785, p < 0.001), though the interaction between these terms was also significant (F(1 × 2) = 0.196, p < 0.001). The p-value corresponding to the F-statistic of the two-way ANOVA is lower than 0.05, which strongly suggests that one or more variants of the modification results with significantly changed color coefficients. There are 5 modification variants, including the control sample, so we applied Tukey’s HSD test to each of the 10 pairs to pinpoint which exhibits statistically significant difference. The critical value of the Tukey-Kramer HSD Q statistic based on the k = 5 variants and ν = 70 degrees of freedom for the error term, for significance levels α = 0.01 and 0.05 (p-values) in the Studentized Range distribution. We obtain these critical values for Q, for α of 0.01 and 0.05, as \({Q}_{\text{c}\text{r}\text{i}\text{t}\text{i}\text{c}\text{a}\text{l}}^{\alpha =0.01, k=5, v=70}=4.7862\) and \({Q}_{\text{c}\text{r}\text{i}\text{t}\text{i}\text{c}\text{a}\text{l}}^{\alpha =0.05, k=5, v=70}=3.9600\). Table 3 summarizes the results of the comparisons. Comparisons between each pair of variants are shown for the parameters L*, a*, b*, c, and h. The results indicate that for the comparison between variant A (control) and other variants (B, C, D, and E), there is statistical significance (** p < 0.01) in terms of parameter L*. However, the comparisons are generally insignificant for the other parameters (a*, b*, c, and h). The comparison between variant B and other variants (C, D, and E) shows statistical significance (* p < 0.05) for parameters b* and c. For the comparison between variant C and other variants (D and E), there is statistical significance (* p < 0.05 and ** p < 0.01) for parameters L*, b*, c, and h. Comparisons between variants D and E indicate statistical significance (** p < 0.01) for parameters b*, c, and h. Overall, Table 3 provides an overview of the statistical significance between different variants for the specified parameters, indicating which comparisons are statistically significant and which are not.

Table 4 shows changes in lightness (L*), chroma from green to red (a*), and chroma from blue to yellow (b*) on the final color change (ΔE*) compared to the control unmodified sample.

The total color change ΔE* of the tested samples ranged from 19.33 to 21.09 units. According to the data presented by Barański et al. (2017) evaluation criteria (ΔE* > 12), it can be concluded that the tested samples obtained a different color as a result of modification. The recorded changes in ΔE* of the tested materials were caused primarily by changes in the L* brightness coordinate. The decreased values of the L* parameter concerning unmodified control samples indicate that the samples’ brightness decreased due to modification with ammonia. Trials darkened. Changes in other color coordinates are not statistically significant.

Similar results in terms of changing the color of wood under the influence of ammonia were observed by Stachowiak-Wencek et al. (2020). Beechwood was characterized by a lower degree of discoloration than black locust and oak wood. The total change in color (ΔE*) in Robinia wood ranged from 39.00 to 41.94, while oak wood ranged from 26.85 to 33.27, depending on the particular modification variant used.

The color of the wood is mainly derived from the chromophore groups in lignin; in addition, the pigment of the extract, tannin, resin, and other substances also significantly affect the color of the wood. According to existing research, the leading chemical cause of wood discoloration is the oxidative condensation reaction of tannins, pigments, alkaloids, sugars, phenols, and other organic substances in wood (Weigl et al. 2012). Under the action of ammonia, the wood’s resins, tannins, polyphenols, and anthocyanins undergo chemical reactions to deepen the color of the wood (Zhang et al. 2021). Considering the chemical composition of beech wood (Fengel and Wegener 2003), it can be assumed that lignin and extractives are mainly responsible for the color change.

The FTIR technique was used to investigate structural changes in the beech wood resulting from ammonia modification, and Figs. 3 and 4 display the results of these measurements. The bands from 2000 to 800 cm^− 1 collectively depict the composition of wood components. Notably, the spectra of all samples (as shown in Fig. 3) exhibited minimal disparities. Subtle variations were discernible between the spectra derived from the reference beech wood samples and those treated with NH4OH.

As illustrated in Fig. 3, the absorption peak at 3440 cm^− 1 corresponds to the vibration of hydroxyl groups. Across all modifications and the control sample, distinct bands appear at 2860 cm^− 1 and 2930 cm^− 1, aligning with lignin’s asymmetric C-H stretching vibrations originating from the benzene. A faint band in the 2100–2250 cm^− 1 range following ammonia modification suggests the potential existence of triple or cumulative double bonds.

The absorption peak intensity at 1740 cm^− 1 associated with the non-conjugated carbonyl group, diminishes indicating a relative reduction in hemicellulose content. Furthermore, the absorption peak at 1235 cm^− 1, associated with syringyl ring breathing and C-O stretching in lignin and xylan, decreases as well, in line with the findings of Pandey and Pitman (2003). Consequently, reducing the number of hydroxyl and carboxyl groups on the hemicellulose’s main and side chains suggests a reaction between NH₄OH and hemicelluloses, decreasing hemicellulose content.

Several studies, including Owen and Pawlak (1989), Pawlak and Pawlak (1997), Miklečic et al. (2012), and Stachowiak-Wencek et al. (2020) reported an additional band at 1650 cm^− 1 in wood after ammonia modification. This band is attributed to the reaction of ammonia with carboxylic ester groups, forming amides. Such a band is visible in the results of our research for all ammonia modified samples. Figure 4 shows the magnified wavelengths from Fig. 3 in the 1750 − 1550 cm^− 1 range with visible absorption peaks at 1740 cm^− 1and 1650 cm^− 1.

Additionally, the peak at 1040 cm^− 1 corresponds to the C − O stretching vibration of secondary lignin alcohols and aliphatic ethers. The main chromophoric and auxochromic groups in wood, including carbonyl, carboxyl, hydroxyl groups, and double carbon-carbon bonds present in lignin and extracts, may react with ammonia gas, as proposed by Zhao et al. (2020). Analysis of the absorption peak absorbance indicates an apparent reduction in peak intensity as the concentration and duration of ammonia treatment increase. Comparable findings were reported by Zhang et al. (2021).

Table 5 summarizes the ability of the tested material variants to withstand compressive loads applied parallel to the grain (along the longitudinal direction).

As depicted in Table 5, the modifications applied to the samples appear to correlate with a reduction in the material’s load-bearing capacity. Notably, the variations in the tested thermochemical modification variants are highlighted in Fig. 5, showcasing differences in the longitudinal modulus of elasticity (MOE), measured in parallel to grain compression.

The p-value (0.005) corresponding to the F-statistic (4.2948) of one-way ANOVA is lower than 0.05, suggesting that one or more variants of thermochemical modification significantly change the modulus of elasticity (MOE) value. Post hoc comparisons using the Tukey HSD test indicated that the mean MOE after modification with 10% of NH₄OH by 12 hours (E_mean = 13,360 MPa, SD = 477) and the mean MOE after modification with 10% of NH₄OH by 24 hours (E_mean = 13,390 MPa, SD = 523) were significantly different from the mean MOE of unmodified beechwood (E_mean = 10,898 MPa, SD = 398). However, MOE after modification with 5% NH₄OH for 12 hours (E_mean = 11,663 MPa, SD = 536) and after modification with 5% NH₄OH for 24 hours (E_mean = 12,450 MPa, SD = 648) did not significantly differ from the mean MOE of the unmodified beechwood.

The results summarized in Table 5 indicate that the treatment with ammonia hydroxide increases the wood’s MOE and compressive strength. MOE increased from 10,898 MPa (unmodified control sample) to 11,663 − 13,390 MPa, while the compressive strength decreased from 77.10 MPa (unmodified control sample) to the range of 81.56-107.19 MPa, depending on the treatment variant. Treatment of beechwood with ammonia hydroxide increases the strength of the wood samples in compressive strength tests. Ammoniacal plasticization of beech wood is observed. According to Weigl and Müller (2009), higher efficiency of the plasticization mechanism can be achieved when hydroxyl groups in the wood structure are adsorbed to water molecules from water vapor. The thermoplastic behavior of lignin plays the primary role in this process. The water causes swelling of the substrates and increases the volume of the call-wall matrix, which, in combination with heat, increases the volume of voids around the relaxing segments of the cell wall matrix molecules. This leads to increased molecular mobility of the matrix molecules and the macroscopic property of plasticity. Ammonia molecules could be connected directly to lignin, cellulose, and hemicelluloses in the cell wall and disintegrate H-bonds and van der Waals forces between these structures; therefore, cohesion between cellulose, lignin, and hemicelluloses decreases. Ammonia treatment may also cause changes in crystallinity, with a partial disintegration of the hemicelluloses. Ammonia penetrates not only into amorphous areas but also, according to the literature, partially into crystalline areas. Due to the leaching of solutes and the deep penetration of ammonia and water molecules into the structure of the cell membrane, an additional capillary system appears in it, and the inner surface of the wood is increased. Microfibrils gain greater freedom of movement relative to each other, which externally manifests itself in the form of plasticization of the wood tissue (Bariska 1969; Besold and Fengel 1983; Yatsu et al. 1986).