Revisiting the condensation reaction of lignin in alkaline pulping with quantitativity part III: predominant formation of α-β-type over α-5-type condensation product in soda cooking treatments of apocynol and creosol

Abstract

Progress in the condensation of lignin is believed to interfere with delignification in alkaline pulping processes without any clear evidence, which motivated us to revisit it quantitatively. This study is the 3rd in the series which evaluates the condensation reactions of lignin in model systems of soda cooking processes using 4-(1-hydroxyethyl)-2-methoxyphenol (apocynol, Ap) and 2-methoxy-4-methylphenol (creosol, Cr). Ap was primarily converted to 2-methoxy-4-vinylphenol (vinylguaiacol, Vg) via the quinone methide intermediate to establish equilibrium before condensation reactions proceeded. Only the α-5-type condensation product between Ap and Cr (ApCr, 1-(4-hydroxy-3-methoxyphenyl)-1-(2-hydroxy-3-methoxy-5-methylphenyl)ethane) and the α-β-type condensation product between Ap and Vg or two molecules of Vg (ApVg, trans-1,3-bis(4-hydroxy-3-methoxyphenyl)but-1-ene) were identified without detecting any self-condensation products of Ap. The α-β-type condensation has not been well known and is an important finding of this study. The formation of ApVg was over 10 times faster than that of ApCr, which demonstrates that the α-β-type condensation is a major mode in soda cooking. However, because origins of α-β-type condensation substructures, such as C₆-C₂-type enol ethers, do not exist in native lignin, the results support our previous conclusion that the condensation reactions of lignin progress less frequently than previously believed.

Introduction

Condensation is a major reaction mode of phenolic substructures in lignin under alkaline pulping conditions. Gierer reviewed the condensation reactions of lignin and showed the formation of α-5-, α-1-, and 5-CH₂-5-type products under these conditions [1] on the basis of the results of Gierer and co-workers in their model experiments, as shown in Fig. 1 [2,3,4,5,6,7,8,9]. The quinone methide intermediate (QM) is primarily generated as an electrophile by releasing the hydroxide anion (HO⁻) and reacts with a nucleophilic aromatic nucleus during condensation. Other researchers also supported the progress of these condensation reactions [10, 11]. Moreover, the condensation products described above were confirmed to be stable under these conditions [12, 13]. In light of the previous knowledge and the fact that lignin polymerizes during condensation, it is unquestioningly believed that condensation of lignin interferes with delignification under alkaline pulping conditions.

Fig. 1

Condensation reactions of lignin proposed by Gierer et al. [1,2,3,4,5,6,7,8,9] The chemical species enclosed by the solid and dotted circles are nucleophiles and electrophiles, respectively. The square shows the carbon numbering of a lignin unit in lignin chemistry

However, it was reported that the β-O-4 bond cleavage of a trimeric model compound of condensation products of lignin, which consists of both β-O-4 and α-5 interunit bonds, is faster than that of the common β-O-4-type model compound in soda cooking treatment (Fig. 2). [14] Another trimeric model compound carrying both β-O-4 and α-1 interunit bonds also underwent β-O-4 bond cleavage despite the slower rate than that of the common β-O-4-type model compound in the same treatment (Fig. 2) [14]. Various studies showed that the β-O-4 bond cleavage in a soda cooking process of a woody sample progresses more efficiently when p-cresol and its analogs are condensed to lignin in an acidic pretreatment before the soda cooking process [15,16,17,18,19,20,21,22,23]. These observations suggest a possibility that condensation of lignin accelerates the β-O-4 bond cleavage occurring near the α-position at which condensation has previously progressed and hence does not always interfere with delignification. Furthermore, the literatures described in the previous paragraph were not based on the quantitative, but only qualitative analyses. Thus, we have revisited the condensation reactions of lignin under alkaline pulping conditions with quantitativity, in previously published reports [24, 25].

Fig. 2

Progress of the β-O-4 bond cleavage in model compounds of α-5- and α-1-type condensation substructures. [14]

In our previous studies, we employed the simplest system using C₆-C₁-type phenolic lignin model compounds, 4-hydroxymethyl-2-methoxyphenol (vanillyl alcohol, Va, Fig. 3) and 2-methoxy-4-methylphenol (creosol, Cr, Fig. 3), under soda cooking conditions [24, 25]. The condensation reactions in this system are as follows: The specific QM is primarily generated from Va (QM_Va, Fig. 3) and electrophilically reacts with an aromatic nucleus as a nucleophile to afford a condensation product. Other QMs can also be generated from condensation products with a specific aromatic nucleus carrying both phenolic and benzyl hydroxy groups. We successfully identified and quantified the α-5-type condensation product between Va and Cr (2-(4-hydroxy-3-methoxybenzyl)-6-methoxy-4-methylphenol, VaCr, Fig. 3), the α-5-type self-condensation product of Va (2-(4-hydroxy-3-methoxybenzyl)-4-hydroxymethyl-6-methoxyphenol, VaVa, Fig. 3), the α-1-type self-condensation product of Va (bis(4-hydroxy-3-methoxyphenyl)methane, VaVa’, Fig. 3) as dimers, and the α-5-/α-1-type self-condensation product of Va (2,4-bis(4-hydroxy-3-methoxybenzyl)-6-methoxyphenol, VaVaVa’, Fig. 3) as a trimer. The results were summarized as follows: (i) The formation rates of the dimers are in the order VaCr ≈ VaVa > VaVa’, comprehensively showing that Va prefers to progress to self-condensation reactions rather than condensation with Cr (owing to the existence of two routes in the former); (ii) VaVaVa’ is generated from VaVa rather than VaVa’ via the reaction of the QM derived from VaVa (QM_VaVa, Fig. 3) with Va; (iii) The dimers are not good nucleophiles in reactions with QM_Va, QM_VaVa, and others; (iv) Because of iii), the condensation reaction of lignin may progress less readily than expected during actual soda cooking processes.

Fig. 3

Summary of the condensation reactions in the soda cooking treatment of Va and Cr observed in our previous reports. [24, 25] The chemical species enclosed by the solid and dotted circles indicate nucleophiles and electrophiles, respectively

In the present study, a C₆-C₂-type phenolic lignin model compound, 4-(1-hydroxyethyl)-2-methoxyphenol (apocynol, Ap, Fig. 4), was treated alone or together with Cr under soda cooking conditions. The results were compared with those obtained in our previous studies, in which Va and Cr were used, to examine the effect of increasing the steric factor of the side chain on the condensation reaction [24, 25]. The disappearance of Ap and Cr was quantitatively examined, and the condensation products were identified and quantified. Although substructures with the same structures as Ap or Cr do not exist in native lignin, these compounds would undergo condensation reactions much more frequently than any common substructures in native lignin owing to their small steric factors. Thus, these compounds are good model compounds for examining the reactions and mechanisms of lignin condensation in detail.

Fig. 4

Summary of the condensation reactions in the soda cooking treatment of Ap and Cr observed in this report. The chemical species enclosed by the solid and dotted circles indicate nucleophiles and electrophiles, respectively

Materials and methods

Materials

All chemicals, except Ap, were purchased from Fujifilm Wako Pure Chemical Industries Co. (Osaka, Japan), Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan), or Sigma-Aldrich Japan K. K. (Tokyo, Japan), and used without further purification. Deionized water (H₂O) used was degassed in advance by sonication under reduced pressure. Ap was synthesized from 4-acetyl-2-methoxyphenol (acetoguaiacone, AG). It was reduced with sodium borohydride in sodium hydroxide (NaOH) solution. The isolated reduction product, consisting mostly of Ap, was recrystallized from a mixture of n-hexane (n-C₆H₁₄) and ethyl acetate (EtOAc) (3/1, v/v). The structure and high purity of Ap and Cr were confirmed by proton nuclear magnetic resonance spectroscopy (¹H-NMR, JNM-A500, 500 MHz, JEOL Ltd., Tokyo, Japan) using acetone-d₆ with a drop of deuterium oxide (D₂O) as the solvent.

Soda cooking for isolation and identification of condensation products

A solution (25 mL) containing NaOH (1.35 mol/L), Ap (0.10 mol/L), and Cr (0.25 mol/L) was prepared for simulating a soda cooking process. This solution was equally divided into five portions, each of which was transferred into a stainless-steel autoclave (10 mL volume, Taiatsu Techno® Co., Tokyo, Japan), and the air present in the headspace was replaced with nitrogen gas. The five autoclaves were immersed in an oil bath at 150 °C and left to soak with shaking for 180 min.

All autoclaves were removed, subsequently immersed in an ice/water bath, and content of each was neutralized with acetic acid (AcOH, 0.8 mL). The combined neutralized solution was extracted with dichloromethane three times and then with EtOAc once, followed by washing with H₂O and brine, and drying over anhydrous sodium sulfate (Na₂SO₄). The dried solution was concentrated using an evaporator under reduced pressure to obtain a brown syrup (1.30 g). The syrup was fractionated by flash chromatography (Isolera, Biotage Japan Ltd., Tokyo, Japan) using a mixture of n-C₆H₁₄/EtOAc (2/1, v/v) as the eluent. The removal of the eluent from each of the two fractions using an evaporator under reduced pressure afforded a light brown syrup. Each light brown syrup was further purified by preparative thin-layer chromatography (PTLC) using a mixture of n-C₆H₁₄/EtOAc (3/1, v/v) as the mobile phase. During the separation of either light brown syrup, an area of the PTLC surface, where condensation products could exist, was extracted with EtOAc and successively evaporated to obtain a colorless syrup (181 mg or 88 mg). Further isolation was performed because the amount of syrup (88 mg) obtained in this procedure was not sufficient for use.

Another reaction solution containing the same additives as the above solution (with slightly different concentrations), except for the absence of Cr, was also prepared and reacted in a manner similar to the above solution. The obtained reaction solutions were subjected to procedures similar to those described above to isolate the condensation products. Finally, two colorless syrups were obtained with amounts of 235 and 73 mg from 2.0 g of Ap.

Each of the four obtained syrups was dissolved in acetone-d₆ with an aliquot of D₂O, and then analyzed by ¹H-NMR, ¹³C-NMR, ¹³C-NMR DEPT135, 2D-NMR ¹H-¹H COSY, 2D-NMR ¹H-¹³C HSQC, 2D-NMR ¹H-¹³C HMBC (JNM-A500), and liquid chromatography/mass spectrometry (LC/MS, LC-2010C_HT/LCMS-2020, Shimadzu Co., Kyoto, Japan). Conditions for the LC-2010C_HT apparatus were as follows: a column for high-performance liquid chromatography (HPLC), Luna 5 μm C18(2) 100 Å (length: 150 mm, inner diameter: 4.6 mm, particle size: 5.0 μm, Phenomenex, Inc., Torrance, CA, USA), was used at an oven temperature of 40 °C with a solvent flow rate of 0.2 mL/min. The solvent and gradient were methanol (MeOH)/H₂O (v/v) from 15/85 to 45/55 for 5 min, 45/55 to 50/50 for 55 min, 50/50 to 75/25 for 0 min and maintained for 7.5 min, and 75/25 to 15/85 for 0 min and maintained for 7.5 min, giving a total time of 75 min. Electrospray ionization (ESI) was used for MS analysis.

Soda cooking for quantitative analysis

A reaction solution (5.0 mL) containing NaOH, Ap (10 mmol/L), and Cr was prepared using degassed H₂O and transferred into a stainless-steel autoclave. The initial concentrations of the components and systems employed in this study are listed in Table 1. The system employed in our previous studies was Cr0–Cr75 [24, 25]. The air in the headspace was replaced with nitrogen gas. Seven other autoclaves with the same content were also prepared. All eight autoclaves were immersed in an oil bath with shaking at 150 °C, and each was soaked for 0, 5, 10, 20, 30, 45, 60, or 120 min with shaking. All reactions were conducted thrice to confirm reproducibility.

Table 1 Reaction systems employed in this work

The same procedure was repeated for another reaction solution containing NaOH (1.01 mol/L) and 2-methoxy-4-vinylphenol (10 mmol/L, vinylguaiacol, Vg, Fig. 4), as listed in Table 1.

Quantification

The autoclave was removed from the oil bath at each specified reaction time and immediately immersed in an ice/water bath. AcOH (0.6 mL) was added for neutralization, followed by the addition of a MeOH solution (5.0 mL) containing the internal standard compound (IS), 3-ethoxy-4-hydroxybenaldehyde (ethyl vanillin). After shaking the autoclave thoroughly, a portion of the content was filtered through a membrane filter, and the filtrate was analyzed by HPLC (LC-2010C_HT) equipped with a UV–Vis detector for quantification (280 nm). The conditions for HPLC analysis were same as those described above.

Results and discussion

Identification of reaction products

Ap and Cr were together or Ap was solely treated under conditions similar to system Cr25’ or Cr0’ for quantitative analysis, respectively, to isolate and identify reaction products including condensation products, except that the scale was larger than that of Cr25’ or Cr0’, respectively. As described above, four colorless syrups were isolated with amounts of 181 mg in the co-treatment of Ap (420 mg) with Cr (864 mg), 88 mg in the co-treatment of Ap with Cr, 235 mg in the sole treatment of Ap (2.0 g), and 73 mg in the sole treatment of Ap. These four syrups were identified as 1-(4-hydroxy-3-methoxyphenyl)-1-(2-hydroxy-3-methoxy-5-methylphenyl)ethane (the α-5-type condensation product between Ap and Cr, ApCr, Fig. 4), trans-1,3-bis(4-hydroxy-3-methoxyphenyl)but-1-ene (the α-β-type condensation product between Ap and Vg or between two molecules of Vg, ApVg, Fig. 4), ApVg, and Vg, respectively, based on the observed NMR and MS spectra (Additional file 1: Figs. S1, S2, and S3, respectively, in the additional file). Both syrups (88 and 235 mg) corresponded to ApVg. The peaks in the ¹H-NMR, ¹³C-NMR, and MS spectra were assigned below. The letters “Ar_A-” and “Ar_B-” in the spectral data indicate the aromatic nuclei of ApCr or ApVg labeled as “A” and “B”, respectively, in Fig. 3. The MS spectra are described below. When the reaction solution obtained from each of Cr0’–Cr75’ at 120 min for quantitative analysis was analyzed by LC/MS (Fig. 5), the MS spectra of the peaks appearing at 21.8, 59.3, and 62.4 min were identical to those of isolated Vg, ApCr, and ApVg, respectively.

Fig. 5

HPLC chromatograms of the reaction solutions obtained at a reaction time of 120 min in systems Cr0’: a Cr10’: b, Cr25’: c, Cr50’: d, Cr75’: e, and Vg: f

¹H-NMR of Vg: δ 3.84 (s, 3H, Ar-OCH₃), 5.04 (dd, 1H, J = 1.2 & 10.9 Hz, Cβ-H (cis of Cα-H)), 5.61 (dd, 1H, J = 1.1 & 17.6 Hz, Cβ-H (trans of Cα-H)), 6.63 (dd, 1H, J = 10.9 & 17.8 Hz, Cα-H), 6.79 (d, 1H, J = 8.1 Hz, ArC5-H), 6.89 (dd, 1H, J = 2.0 & 8.3 Hz, ArC6-H), 7.07 (d, 1H, J = 2.0 Hz, ArC2-H). ¹³C-NMR of Vg: δ 56.1 (-OCH₃), 109.9 (ArC2), 111.0 (Cβ), 115.7 (ArC6), 120.6 (ArC5), 130.5 (ArC1), 137.7 (Cα), 147.6 (ArC4), 148.4 (ArC3). MS in positive mode of Vg (m/z (detected ion, relative intensity)): 355 ([2 M + MeOH + Na]⁺, 96), 301 ([2 M + H]⁺, 100). MS in negative mode of Vg (m/z (detected ion, relative intensity)): 149 ([M–H]⁻, 100).

¹H-NMR of ApCr: δ 1.50 (d, 3H, J = 7.2 Hz, Ar_ACβ-H₃), 2.16 (s, 3H, Ar_BCα-H₃), 3.76 (s, 3H, Ar_A-OCH₃), 3.78 (s, 3H, Ar_B-OCH₃), 4.47 (q, 1H, J = 7.2 Hz, Ar_ACα-H), 6.52 (d, 1H, J = 1.7 Hz, Ar_BC2-H), 6.58 (d, 1H, J = 1.7 Hz, Ar_BC6-H), 6.71 (m, 2H, Ar_AC5-H and Ar_AC6-H), 6.87 (s, 1H, Ar_AC2-H). ¹³C-NMR of ApCr: δ 21.1 (Ar_BCα), 21.2 (Ar_ACβ), 37.4 (Ar_ACα), 56.2 (Ar_A-OCH₃), 56.2 (Ar_B-OCH₃), 110.4 (Ar_BC2), 112.5 (Ar_AC2), 115.3 (Ar_AC5), 120.6 (Ar_AC6), 120.6 (Ar_AC6), 128.5 (Ar_BC1), 133.4 (Ar_BC5), 138.5 (Ar_AC1), 141.8 (Ar_BC4), 145.4 (Ar_AC4), 147.6 (Ar_BC3), 147.9 (Ar_AC3). MS in positive mode of ApCr (m/z (detected ion, relative intensity)): 599 ([2 M + Na]⁺, 100), 343 [M + MeOH + Na]⁺, 34), 329 ([M + H₂O + Na]⁺, 17), 311 ([M + Na]⁺, 99). MS in negative mode of ApCr (m/z (detected ion, relative intensity)): 287 ([M − H]⁻, 100).

¹H-NMR of ApVg: δ 1.37 (d, 3H, J = 7.2 Hz, Ar_ACβ-H₃), 3.51 (qu(qd), 1H, J = 6.8 Hz, Ar_ACα-H), 3.80 (s, 3H, Ar_A-OCH₃), 3.81 (s, 3H, Ar_B-OCH₃), 6.26 (dd, 1H, J = 6.6 & 15.8 Hz, Ar_BCβ-H), 6.31 (d, 1H, J = 16.1 Hz, Ar_BCα-H), 6.71 (dd, 1H, J = 2.0 & 8.2 Hz, Ar_AC6-H), 6.74 (d, 1H, J = 8.1 Hz, Ar_BC5-H), 6.76 (d, 1H, J = 8.0 Hz, Ar_AC5-H), 6.82 (dd, 1H, J = 2.0 & 8.2 Hz, Ar_BC6-H), 6.86 (d, 1H, J = 1.7 Hz, Ar_AC2-H), 7.01 (d, 1H, J = 2.0 Hz, Ar_BC2-H). ¹³C-NMR of ApVg: δ 22.0 (Ar_ACβ), 43.0 (Ar_ACα), 56.1 (Ar_A-OCH₃), 56.2 (Ar_B-OCH₃), 109.9 (Ar_BC2), 111.8 (Ar_AC2), 115.7 (Ar_AC5 or Ar_BC5), 115.7 (Ar_AC5 or Ar_BC5), 120.2 (Ar_AC6), 120.2 (Ar_BC6), 128.7 (Ar_BCβ), 130.6 (Ar_BC1), 133.7 (Ar_BCα), 138.3 (Ar_AC1), 145.6 (Ar_AC4), 146.8 (Ar_BC4), 148.2 (Ar_AC3), 148.4 (Ar_BC3). MS in positive mode of ApVg (m/z (detected ion, relative intensity)): 623 ([2 M + Na]⁺, 26), 355 ([M + MeOH + Na]⁺, 100), 341 ([M + H₂O + Na]⁺, 37), 323 ([M + Na]⁺, 80). MS in negative mode of ApVg (m/z (detected ion, relative intensity)): 299 ([M − H]⁻, 100).

Reaction products appearing at retention times of 11.4, 12.5, and 13.2 min in Fig. 5 were identified as 4-hydroxy-3-methoxybenzaldehyde (vanillin, V), AG, and 2-methoxyphenol (guaiacol, G) by confirming that their authentic compounds appeared at the same retention times in the HPLC analyses and that their MS spectra were the same as those of the peaks appearing at 11.4, 12.5, and 13.2 min, respectively. The spectral data are presented in the additional file. The formation of these compounds will be explained in detail later.

Reactions progressing in the systems employed in this report

Figure 4 shows the presumed reactions that occur in the systems employed in this study. In systems Cr0’–Cr75’, QM_Ap is primarily generated from Ap. QM_Ap is converted to Vg via β-proton abstraction. Besides, QM_Ap, as an electrophile, reacts at the Cα (benzyl) carbon with a nucleophile that is mainly Cr at the aromatic C5 carbon or Vg at the Cβ carbon of the side chain. Subsequent deprotonation affords ApCr or ApVg, respectively. In the latter reaction, the QM precursor (QM_ApVg) is primarily generated and can react with HO⁻ to afford the adduct, 1,3-bis(4-hydroxy-3-methoxyphenyl)butan-1-ol (ApVg’). The reactions progressing in system Vg are mostly the same as those in system Cr0’, except that the starting compound is Vg and QM_Ap is primarily generated from Vg.

Interconversion between Ap and Vg

Figures 6a–6f show the time course of the disappearance of the starting compound, Ap or Vg, and the formation of the quantified reaction products, Ap, Vg, ApCr, and ApVg, in systems Cr0’–Cr75’ and Vg, respectively. The values of all data points are listed in Additional file 1: Table S1 in the additional file. Figures 6g and 6h show the time course of the disappearance and formation of Ap and Vg, respectively, in all the systems. Figure 6i shows the disappearance of Cr in all the systems. Figures 6k and 6l show the formation of ApCr and ApVg, respectively, in all the systems.

Fig. 6

Time courses of the changes in the concentrations of Ap (filled circles, filled diamonds, and filled squares), Vg (open circles, open diamonds, and open squares), ApCr (open bow-tie-shaped marks, open triangles, and open inverted triangles), and ApVg (filled bow-tie-shaped marks, filled triangles, and filled inverted triangles) in systems Cr0’: a (solid black lines), Cr10’: b (solid red lines), Cr25’: c (dotted red lines), Cr50’: d (solid blue lines), Cr75’: e (dotted blue lines), and Vg: f (dotted black lines), and in the yields of Ap: g, Vg: h, Cr: i, ApCr: j, and ApVg: k in all the systems. Filled circles, open circles, open bow-tie-shaped marks, and filled bow-tie-shaped marks show the 1st trials. Filled diamonds, open diamonds, open triangles, and filled triangles show the 2nd trials. Filled squares, open squares, open inverted triangles, and filled inverted triangles show the 3rd trials

ApVg was formed with a yield of lower than 3 mol% before a reaction time of 20 min in systems Cr0’ and Vg. ApCr and ApVg were formed similarly with yields of lower than 3 mol% before the reaction time in systems Cr10’ and Cr25’, although the yields of ApCr were slightly higher than 3 mol% before the reaction time in the other two systems. The disappearance of Ap in system Cr0’ or Vg in system Vg was much faster than the formation of ApVg before the reaction time and was almost quantitatively accompanied by the formation of Vg or Ap, respectively. These phenomena comprehensively show that the chemical reaction that progressed in all the systems before a reaction time of 20 min could be limited to the interconversion between Ap and Vg via QM_Ap regardless of the starting compound (Ap or Vg). The ratio of the amounts of Vg to Ap remaining after a reaction time of 20 min in each system was almost constant between 1.3 and 1.4, although their absolute amounts gradually decreased owing to the progress of the condensation reactions. Therefore, the interconversion between Ap and Vg via QM_Ap represents the pre-equilibrium step of the condensation reactions; hence, it was much more rapid than the condensation reactions in the systems employed in this study. In actual soda cooking processes, this type of equilibrium corresponds to that between a C₆-C₂-type enol ether and C₆-C₂-type β-O-4 substructures. Although the establishment of this equilibrium is not general, a few previous reports supposed or referred to it [26, 27].

This pre-equilibrium is described by formula (1), where k_Ap and k_−Ap are the rate constants of the conversion of Ap to QM_Ap and the reverse conversion, respectively, and k_Vg and k_−Vg are the rate constants of the conversion of Vg to QM_Ap and the reverse conversion, respectively.

$${\mathbf{Ap}}\underset{{k_{{ - {\text{Ap}}}} }}{\overset{{k_{{{\text{Ap}}}} }}{\rightleftharpoons}} {\mathbf{QM}}_{{{\mathbf{Ap}}}} \underset{{k_{{ - {\text{Vg}}}} }}{\overset{{k_{{{\text{Vg}}}} }}{\leftrightharpoons}}{\mathbf{Vg}}$$

(1)

Ap and Vg do not decrease owing to this interconversion, although these compounds disappear owing to the progress of condensation reactions after establishing the pre-equilibrium. Equations (2) and (3) thus work thereafter when [Ap], [QM_Ap], and [Vg] show the concentrations of Ap, QM_Ap, and Vg, respectively.

$$d\left[ {{\mathbf{Ap}}} \right]/dt = - k_{{{\mathbf{Ap}}}} \left[ {{\mathbf{Ap}}] + k_{{ - {\mathbf{Ap}}}} } \right[{\mathbf{QM}}_{{{\mathbf{Ap}}}} ] = 0$$

(2)

$$d\left[ {{\mathbf{Vg}}} \right]/dt = - k_{{{\mathbf{Vg}}}} \left[ {{\mathbf{Vg}}] + k_{{ - {\mathbf{Vg}}}} } \right[{\mathbf{QM}}_{{{\mathbf{Ap}}}} ] = 0$$

(3)

By arranging these equations, the following Eq. (4) is obtained because the value of [Vg]/[Ap] is approximately 1.3.

$$\left[ {{\mathbf{Vg}}} \right]/[{\mathbf{Ap}}] = \left( {k_{{{\mathbf{Ap}}}} /k_{{{\mathbf{Vg}}}} } \right)(k_{{ - {\mathbf{Vg}}}} /k_{{ - {\mathbf{Ap}}}} ) \approx {1}.{3}$$

(4)

In our previous report, the formation rates of Ap and Vg from QM_Ap were examined by generating QM_Ap from the benzyl methyl ether of Ap (2-methoxy-4-(1-methoxyethyl)phenol) and detecting Ap and Vg as almost exclusive reaction products under conditions using the same concentration of NaOH as in this report (1.0 mol/L) at a temperature of 95 °C [28]. The ratio [Vg]/[Ap] during this reaction was presumed to be smaller than 0.1, because the yield of Ap was always higher than 90% based on the amount of the disappearing starting compound. Furthermore, the interconversion between Ap and Vg was slow at 95 °C, which was much lower than the temperature employed in this study (150 °C). Thus, the value of 0.1 is regarded to correspond to that of k_−Vg/k_−Ap. When 0.1 is used as the value of k_−Vg/k_−Ap in Eq. (4), despite the different temperatures employed between this and our previous reports, the value of k_Ap/k_Vg is calculated to be approximately 13. The values of k_Ap/k_−Ap and k_−Vg/k_−Vg cannot be calculated only from the above discussion. A schematic of the potential energy diagram for this interconversion is shown in Fig. 7. The formation of Ap is kinetically more favorable than that of Vg when these compounds are generated from QM_Ap, which is expressed by the activation energy of the route from QM_Ap to Ap (TS_Ap) lower than that from QM_Ap to Vg (TS_Vg). Vg is thermodynamically more stable than Ap in this equilibrium via QM_Ap, which is expressed by the free energy of Vg lower than that of Ap.

Fig. 7

Schematic potential energy diagram for the interconversion between Ap and Vg as the pre-equilibrium of condensation reactions in this study

Condensation reactions and predominant formation of α-β-type condensation product

In order to determine how frequent side reactions including further condensations were, it is necessary to determine what percentage of the whole reaction products, ApCr and ApVg, they represented. In system Cr0’, the starting compound, Ap, was converted to Vg, ApVg, and side products, including further condensation products. In systems Cr10’–Cr75’, it was converted to Vg, ApCr, ApVg, and side products, including them. In system Vg, Vg was converted to Ap, ApVg, and side products, including them. The total molar yield of Ap, Vg, ApCr, and twice the ApVg was 97% (± 0.3), 91% (± 3.2), 89% (± 1.4), 89% (± 1.2), 90% (± 2.2), or 102% (± 1.1) (the values in parentheses: standard deviations obtained by three duplications) based on the initial molar amount of the starting compound, Ap or Vg, at the end of the reaction (120 min) in system Cr0’, Cr10’, Cr25’, Cr50’, Cr75’, or Vg, respectively (see Additional file 1: Table S1). The formation of a molecule of ApVg consumes two molecules of Vg or each molecule of Ap and Vg, either of which was originally the starting compound but interconverted to each other during the reaction owing to the establishment of pre-equilibrium, resulting in usage of twice amount of ApVg in the calculation. Thus, it was confirmed that side reactions, including further condensation, progressed only slightly in all systems.

The characteristic findings in this report are as follows: (i) No self-condensation products of Ap were formed even in system Cr0’, as well as in the others, as described above; (ii) ApVg was formed exclusively in systems Cr0’ and Vg, and its formation was greater than that of ApCr in systems Cr10’ and Cr25’, but slightly less significant than that of ApCr even in system Cr50’; (iii) The formation of ApVg gradually decreased with increasing initial concentration of Cr despite the small decreases, whereas that of ApCr gradually increased; (iv) The formation of ApVg in system Vg exhibited almost the same behavior as that in system Cr0’.

The α-5-type and α-1-type self-condensation products of Ap, 1-[2-hydroxy-5-(1-hydroxyethyl)-3-methoxyphenyl]-1-(4-hydroxy-3-methoxyphenyl)ethane and 1,1-bis(4-hydroxy-3-methoxyphenyl)ethane, respectively (Fig. 4), were expected to form. However, as per finding i), these compounds were not isolated. Because further condensation of dimers was confirmed to be a minor reaction, as described above, the absence of these self-condensation products of Ap does not result from their further condensation, but from their small formation.

Moreover, finding (i) is quite different from the result obtained in our previous studies, in which soda cooking treatments of Va were conducted under similar conditions as those in this report [24, 25]. The Va self-condenses more frequently than condensation with coexisting Cr. The absence of self-condensation of Ap observed in this report results from either of the low reactivity of Ap as a nucleophile or that of QM_Ap as an electrophile, or both. The low reactivity of QM_Ap as an electrophile seems most decisive than the others, because of the following observations: a) The condensation reactions in system Cr75’ (Fig. 6e), in which the nucleophile was mostly Cr (75 mmol/L), progressed slower than those in system Cr75 employed in our previous reports (Table 1), also in which the nucleophile was mostly Cr (75 mmol/L) [24, 25]. This shows that nucleophilic addition of Cr to QM_Ap is slower when compared with that to QM_Va. b) The concentration of QM_Ap present in system Cr75’ is presumed to be higher than that of QM_Va in system Cr75. This presumption is based on both the Hammond postulate and our previous result which states that the formation of QM_Ap is much faster than that of QM_Va in alkaline treatment using the same concentration of NaOH as this report at a lower temperature of 95 °C [28]. The Hammond postulate says that the stability of a compound is higher than that of another compound under certain conditions when the formation of the former compound is faster than that of the latter compound from an original specific compound under these conditions. It is thus confirmed that the low reactivity of QM_Ap, as an electrophile, in the condensation reactions in observation a) is not apparent, but absolute. The low reactivity is attributed to the presence of methyl group at the β-position of QM_Ap. The presence lowers the reactivity via the steric and electronic factors and results in the above-described absence of the self-condensation products of Ap.

The finding (ii) shows that Vg reacts at its β-carbon with QM_Ap much more rapidly than Ap reacts at its aromatic C5-carbon with it. In system Cr50’, the formation rate of ApVg was similar to that of ApCr (Fig. 6d). Because the concentration of Vg was always lower than 5 mmol/L in system Cr50’ (Fig. 6d), in which the concentration of Cr was equal to or just slightly lower than 50 mmol/L (Fig. 6i), the rate of the reaction of Vg with QM_Ap can be at least more than 10 times of that of Cr with QM_Ap. Notably, 1-(4-hydroxy-3-methoxyphenyl)-1-(2-hydroxy-3-methoxy-5-vinyl)ethane, which is the α-5-type condensation product between QM_Ap and Vg, did not form at all. Thus, Vg reacts exclusively at its β-carbon in the side chain with QM_Ap. Although ApVg’, which is the HO⁻ adduct of QM_ApVg, was expected to form (Fig. 4) similarly to the formation of Ap from QM_Ap, it was not detected at all. Although this absence cannot be rationally explained only on the basis of the results obtained in this report, the thermodynamic stability of ApVg’ may be much lower than that of ApVg. The cis counterpart of ApVg was not formed. Although the cis counterpart must be thermodynamically less stable than ApVg, it is unclear whether the difference in their thermodynamic stabilities is sufficiently large to result in an absence.

The finding (iii) naturally shows that the reactions of Vg and Cr with QM_Ap are competitive with each other. However, the large increase in the concentration of Cr with the variation from system Cr0’ to Cr75’ did not remarkably affect the competitive reaction of Vg with QM_Ap. Thus, the reaction of Vg with QM_Ap is much faster than that of Cr with QM_Ap, as described above.

The finding (iv) supports the establishment of an equilibrium between Ap and Vg before the clear progress of the condensation reactions.

It was thus clarified that Ap undergoes the condensation reactions much slower than Va. This low reactivity of Ap is mainly attributed to the steric factor of the side chain larger than that of Va. However, because even the steric factor of the side chain of Ap is much smaller than that of native lignin and residual lignin in being cooked pulp, it is suggested that condensation reactions of lignin in soda cooking processes progress less frequently than those we have believed for far.

Because the major condensation product, ApVg, was the α-β-type, the formation of which has not been well known as a condensation mode, its formation is a new finding of this report. In this study, the α-β-type condensation progressed much faster than the most common condensation mode, the α-5-type. The α-β-type condensation can progress between two phenolic substructures, either of which has a conjugated carbon–carbon double bond at the side chain. Such phenolic substructures with unsaturated side chains may not exist with large amounts in residual or dissolved lignin in alkaline cooking processes when compared with the amount of aromatic nuclei, most of all which are reaction sites in the α-5-type condensation. Candidates for such phenolic substructures with unsaturated side chains are C₆-C₂-type enol ethers and stilbenes, which are produced from the QMs derived from β-O-4 and β-5 substructures, respectively, by releasing the γ-position as the formaldehyde (HCHO) molecule in alkaline cooking processes, as well as coniferyl alcohol, which can be produced as the counterpart of releasing phenoxides in the common β-O-4 bond cleavage in kraft cooking processes, and others. When it is also taken into consideration that the steric factor is larger in actual alkaline cooking processes than in this model experiment, the predominant progress of the α-β-type condensation found in this report supports our previous suggestion that the progress of the condensation reactions of lignin, at least whose modes are the common α-5- and α-1-types, is not necessarily an important factor in suppressing delignification in the later stages of alkaline cooking processes [25]. However, the condensation reaction of lignin certainly progresses when steric factor is not large. Thus, it may not result in polymerization of lignin, but in structural alteration of residual and dissolved lignin, which will affect their usability in subsequent biomass utilization.

Formation of minor monomeric products

It was presumed in our previous report that QM_Va oxidizes Va to V, whereas QM_Va itself is reduced to Cr in systems Cr0–Cr75 [24]. Hence, the formation of AG in this study may be explained by the progress of this type of reaction, namely, the oxidation of Ap by QM_Ap (Fig. 8), despite the lack of detection of the reduction product of QM_Ap (4-ethyl-2-methoxyphenol, ethylguaiacol). The yields of AG were 1.0% (± 0.0), 0.9% (± 0.0), 0.9% (± 0.1), 0.8% (± 0.0), 0.8% (± 0.1), and 0.9% (± 0.1) in systems Cr0’, Cr10’, Cr25’, Cr50’, Cr75’, and Vg, respectively, at the end of the reactions based on the initial molar amount of Ap or Vg (the values in parentheses are standard deviations obtained by the three duplications). These yields were clearly lower than those of V in our previous report [24], which suggests that QM_Ap has a reactivity lower than QM_Va in this type of reaction. Anyway, the formation of AG is a minor reaction.

Fig. 8

Possible formation reactions of AG, G, and V in this study

It was also presumed in our previous report that the precursor produced primarily in the formation of the α-1-type self-condensation product of Va can be rearranged to afford G and QM_Va by releasing the HCHO molecule in systems Cr0–Cr75 [24]. Thus, as shown in Fig. 8, the formation of G in this study can be explained by the progress of this type of reaction, namely, the rearrangement of the precursor of the α-1-type self-condensation product of Ap to afford G and QM_Ap by releasing the acetaldehyde (CH₃CHO) molecule, despite the lack of detection of the α-1-type self-condensation product of Ap. The precursor may exclusively undergo rearrangement to afford G, QM_Ap, and CH₃CHO without progressing to the α-1-type self-condensation product. The yields of G were 5.4% (± 0.1), 4.7% (± 0.1), 5.0% (± 0.1), 5.2% (± 0.3), 4.5% (± 0.7) and 5.2% (± 0.2) in systems Cr0’, Cr10’, Cr25’, Cr50’, Cr75’, and Vg, respectively, at the end of the reactions based on the initial molar amount of Ap or Vg, which were comparable to those detected in our previous report [24].

A possible reaction for the formation of V involves the oxidation of Vg or ApVg by oxygen, possibly contaminating the system. It was shown that the guaiacoxystyrene (4-hydroxy-3-methoxystyrene) substructure, which is Vg itself or a constituent of ApVg, is quite labile to oxygen oxidation under alkaline conditions, affording V in high yield [29, 30]. The yields of V were only 0.4% (± 0.0), 0.5% (± 0.0), 0.5% (± 0.1), 0.5% (± 0.1), 0.4% (± 0.0) and 0.7% (± 0.2) in systems Cr0’, Cr10’, Cr25’, Cr50’, Cr75’, and Vg, respectively, at the end of the reactions based on the initial molar amount of Ap or Vg, indicating that the formation of V is a minor reaction.

Conclusions

The condensation reaction of lignin was quantitatively examined in a model system in which a phenolic lignin model compound, Ap, was reacted with or without another phenolic lignin model compound, Cr, under soda cooking conditions with various initial concentrations of Cr.

Ap was primarily converted to Vg via QM_Ap to establish an equilibrium between these compounds before the progress of condensation reactions. The ratio, [Vg]/[Ap], showed a value of 1.3–1.4 in this equilibrium.

Only ApCr, the α-5-type condensation product between Ap and Cr, and ApVg, the α-β-type condensation product between Ap and Vg or between two molecules of Vg, were isolated and identified without any self-condensation products of Ap. This is in contrast to the same soda cooking treatment of Va with or without Cr conducted in our previous reports, where the self-condensation reactions of Va preferentially progressed [24, 25]. The formation of ApVg was over 10 times faster than that of ApCr. Thus, the α-β-type condensation was suggested to be the major mode in condensation reaction of lignin. This type of condensation has not yet been well known, and hence is a new finding in this report.

This new finding, as well as others, suggests that the progress of the condensation reaction of lignin is not necessarily an important factor in suppressing delignification in the later stages of alkaline cooking processes.