UV–vis spectroscopy as a rapid method for evaluation of total phenolic hydroxyl structures in lignin

2.1 Materials

Lignin samples were derived from different sources such as Norway spruce (a softwood) and eucalypt (Rose Gum Eucalyptus grandis, a hardwood), and were isolated from wood using various methods including kraft pulping, organosolv pulping, and oxygen delignification following kraft pulping. The kraft lignin samples (see Table 1) – kraft spruce (SL) and kraft eucalyptus (EL) – were obtained using LignoBoost technology (Tomani 2010) and were used without additional preparation.

Methods for preparing samples of acetone fractionated of EL, eucalyptus organosolv lignin and oxlignin are described in Section 2.2.1.

All common chemicals used in various analyses, such as NaOH, H₂SO₄, H₃BO₃, and NaH₂PO₄, were of analytical grade and purchased from Sigma-Aldrich.

2.2 Methods

2.3 Analytical techniques

2.3.1 UV–vis spectroscopy

Ultraviolet absorption spectra were determined with a UV–Vis spectrophotometer (UV-2550, Shimadzu, Tokyo, Japan). Matched 1-cm quartz absorption cells were used.

The difference spectrum was determined by measuring the absorbance of the alkaline solution (1 or 2) relative to that of the neutralized solution which was placed in the reference cell of the spectrophotometer as the “blank”. Readings were taken from 220 to 450 nm, spaced 0.5 nm apart.

Difference spectra were plotted in terms of absorbance (ΔA) or absorptivity (Δa, L^.g^−1.cm⁻¹):

(2) ∆ a = Δ A l · C

where ΔA – difference absorbance readings, C, g^.L⁻¹ – concentration of the lignin samples solutions; l, cm – path length.

The content of hydroxyl groups was calculated with four Δε-IDUS methods that were previously developed (Chen et al. 2022; Gärtner et al. 1999; Goldschmid 1954; Lin and Dence 1992; Zakis 1994). The equations used to calculate the phenolic content were derived from the differential molar absorptivitie (∆ε, L^.mol^−1.cm⁻¹):

(3) ∆ ε = Δ A l · C *

where ΔA – difference absorbance readings, C ^* , (mol of –OH) ^.L⁻¹ - concentration of the solutions of phenolic model compounds (Table 2); l, cm – path length.

Table 2:

Model compounds used in varieties of Δε-IDUS method.

Method	Model phenol	Type of Ph-OH according Figure 1	References
Δε-IDUS 1		I	Goldschmid (1954)
Δε-IDUS 2		I	Lin and Dence (1992)
		II
		III
		IV
Δε-IDUS 3		I	Zakis (1994)
		II
		III
		IV
Δε-IDUS 4		I	Chen et al. (2022)
		II
		III
		IV

The concentration of the solutions of phenolic model compounds were calculated according to:

C * = m M r · V

Where m, g – weighted amount of the model compound; M _r , g^.mol⁻¹ – molecular weight of model solution equivalent to one –OH group, V, L – the total volume of solution used for absorbance reading.

The Δε-IDUS 1 method was based on equation obtained in the work (Goldschmid 1954). Only one lignin alkali solution (with buffer pH 12) was used, and calculation was based on the value of maximal absorptivity (∆a_max) of the band centered near 300 nm.

, (4) c ( O H Total ) = ∆ a max ∆ ε max

where c (OH_Total), mol^.g⁻¹ – total number of Ph-OH; Δa, L^.g^−1.cm⁻¹ – the maximal value of difference absorptivity of lignin samples and ∆ε_max, L^.mol^−1.cm⁻¹ – the maximal value of difference molar absorptivity of the model compounds. Both number were calculated according equations (2) and (3) at λ_max of the band centered near 300 nm.

The Δε-IDUS 2 method was based on the proposal of Lin and Dence (Goldmann et al. 2017; Lin and Dence 1992). This method also uses only one ionization level of Ph-OH (with alkali solution 1 with buffer pH 12) and describe six possible model structures. The formula for calculation were:

c ( OH C 5 − substituted ) = 0.112 · ∆ a 320

c ( OH Uncond ) = 0.223 · ∆ a 300 + 0.078 · ∆ a 315 + 0.043 · ∆ a 350 + 0.032 · ∆ a 370 + 0.056 · ∆ a 380

c C ( O H Conjugated ) = 0.078 · ∆ a 315 + 0.043 · ∆ a 350 + 0.032 · ∆ a 370 + 0.056 · ∆ a 380

c ( OH Unconjugated ) = 0.223 · ∆ a 300 + 0.112 · ∆ a 320

(5) c ( OH Total ) = 0.223 · ∆ a 300 + 0.078 · ∆ a 315 + 0.043 · ∆ a 350 + 0.032 · ∆ a 370 + 0.056 · ∆ a 380 + 0.112 · ∆ a 320

Where c (OH_name), mol^.g⁻¹ – number of varied –OH structures per gram of substance; name – unconjugated (structure I + III), conjugated (structure IIa + IVa), weakly acidic C5-substituted (structure III), uncondensed (structure I + IIa) and total (structure I + IIa + III + IVa) (Figure 1).

Figure 1:

Types of phenolic structures determined by the UV method (Gärtner et al. 1999; Zakis 1994) I – unconjugated-uncondensed. IIa, IIb – conjugated-uncondensed. III – unconjugated-condensed. IVa, IVb – conjugated-condensed.

The Δε-IDUS 3, the third variation used, was based on the work (Gärtner et al. 1999; Zakis 1994), was utilize two ionization levels and two alkali solutions, 1 and 2, with buffer pH 12 and with 0.2 molL⁻¹ NaOH respectively. Calculation based on the formula:

c ( O H C 5 − substituted ) = 0.250 · ( ∆ a N a O H 300 − ∆ a p H 12 300 ) + 0.107 · ( ∆ a N a O H 360 − ∆ a p H 12 360 )

c ( OH Uncondensed ) = 0.250 · ∆ a p H 12 300 + 0.107 · ∆ a p H 12 360

c ( OH Conjugated ) = 0.048 · ∆ a NaOH 360

c ( OH Unconjugated ) = 0.250 · ∆ a NaOH 300 + 0.059 · ∆ a N a O H 360

(6) c ( OH Total ) = 0.250 · ∆ a N a O H 300 + 0.107 · ∆ a N a O H 360

where c (OH_name), mol^.g⁻¹ – number of varied–OH structures per gram of substance; name – unconjugated (structure I + III), conjugated (structure IIa + IVa), C5-substituted (structure III, IVa and IVb), uncondensed (structure I + IIa) and total (structure I + IIa + III + IVa + IVb) (Figure 1).

Δε-IDUS 4 was used the implementation made by (Chen et al. 2022). According to this work the total Ph-OH content was calculated as a sum:

(7) c ( O H Total ) = c I + c I I + c I I I + c I V

Where C_I – C_IV, mmol g⁻¹ are contributions of different types of phenolic structures (structures I–IV according to Figure 1).

These numbers were calculated by measuring the difference absorption (ΔA) at 300, 320, 350, and 370 nm of super alkaline (0.2 N NaOH) solution of model compounds listed in Table 2 and solving the linear system of equations:

∆ A 300 = 3619 c I + 586 c I I I − 6728 c I V − 4480 c I I

∆ A 320 = 3039 c I I I − 11428 c I V + 391 c I I

∆ A 350 = 334 c I I I + 5436 c I V + 24010 c I I

∆ A 370 = 22590 c I V + 10658 c I I

For each lignin sample, six parallel measurements were made, varying the initial weighted amount of sample in the stock solution and the aliquot of the stock solution taken to prepare the diluted solution. The result of the analysis was presented as a Confidence Limit (CL):

, (8) C L = c ( O H Total ) ‾ ± ∆ c ( O H Total )

c ( O H Total ) ‾ = ∑ c i ( O H Total ) n

∆ c ( O H Total ) = t P , n · s n

(9) s = ∑ ( c i ( O H Total ) − c ( O H Total ) ‾ ) n − 1

where c _i (OH_Total) and c ( O H Total ) ‾ , mmol^.g⁻¹ – running and mean value of total Ph-OH content, ∆c (OH_Total), mmol^.g⁻¹ – confidence interval or margin of error, t_P,n = 2.57 for confidence degree (P = 0.95) and number of measurements (n = 6).

3.1 Analysis of Δε-IDUS approaches of UV–vis method

The most popular approach of UV–vis spectroscopy method of determination of Ph-OH groups in lignins was used in this work, namely ionization difference spectroscopy (Δε-IDUS, Δε-method according to Goldmann et al. (2017). The Δε-IDUS method is based on the fact that Ph-OH of lignin can absorb electromagnetic radiation in the UV range. On the other hand, they dissociate in an alkaline environment, which causes a shift in the absorption spectrum toward longer wavelengths (Jablonsky et al. 2015). It is well known that the non-phenolic, aliphatic (Lundquist and Parkås 2011) lignin units produce identical UV–vis spectra in the neutral and alkaline solutions (Aulin-Erdtman 1952; Goldschmid 1954; Lin and Dence 1992). Therefore, the difference between the spectra of lignin solutions with the same concentration but different pH, the difference absorption spectrum (Δε-spectrum), represents only the absorbance from the phenolic units that are ionized under the alkaline condition. So, it became the basis for creating a group of methods for UV determination of the Ph-OH content in lignins (Cybulska et al. 2012; Goldmann et al. 2017; Jablonsky et al. 2015; Stark et al. 2016).

Two approaches to obtaining the Δε-spectrum are known in the literature. According to the first (Lin and Dence 1992), absorption spectra of alkaline and neutralized lignin solutions were measured separately. Then the Δε-spectrum was calculated by subtracting the UV–visible spectrum of the sample dissolved in the neutral solution from the comparison in the alkaline solution. The second possibility (Goldschmid 1954) consists of measuring the absorbance of the alkaline solution relative to the neutralized solution, which was placed in the reference cell of the spectrophotometer as a “blank”. In the presented work, the second method was used, as it is less time-consuming.

Although the Δε-IDUS method is notable for its simplicity, rapidity, and cost-effectiveness, positioning it as the most promising candidate for quick phenolic –OH group analysis (Goldmann et al. 2017) without any preliminary treatment, the challenge lies in selecting appropriate solvents, pH level, and wavelengths for characterizing different types of phenolic–OH groups (Chen et al. 2022).

The best solvent should completely dissolve the sample and interfere with the adsorption of lignin in the UV range of wavelengths. In addition, as was shown in (Ruwoldt et al. 2022), the nature of the solvent significantly affects the peak location and amplitude of the Δε-spectra due to the change of Hansen solubility parameters. According to Chen et al. (2022) 2-methoxyethanol/water/acetic acid and 2-methoxyethanol/water was the ideal solvent system for UV–vis determination of Ph-OH in technical lignins. Ruwoldt et al. (2022) suggested that DMSO, ethylene glycol, and a mixture of propylene carbonate or ethanol and water are less hazardous alternatives to traditional lignin solvents in UV–visible spectrometry. In addition, Ruwoldt et al. (2022) showed that the shape of spectra of lignin samples were similar when alkali and neutralized water or 2-methoxyethanol/water solutions were used. Water has a cutoff below 200 nm and is the best solvent for UV measurements. Water is the safest solvent also. Therefore, all studies in the presented work were carried out in aqueous solutions.

According to recommendations (Chen et al. 2022; Goldschmid 1954; Gärtner et al. 1999; Lin and Dence 1992; Ruwoldt et al. 2022; Zakis 1994), solutions with pH 6.0, pH 12, and 0.2 N NaOH were used. Generally accepted Ph-OH groups were non-dissociated at pH 6, mainly dissociated at pH 12, and fully dissociated at 0.2 NaOH.

Initially, the quantification of Ph-OH was based on the difference in absorption at 300 and 380 nm between phenolic units in neutral and alkaline solutions. The content of ionizing phenol hydroxyl groups can be quantitatively evaluated by comparing the values of substances studied at certain wavelengths to the values of the respective model compounds. In the original work by (Zakis et al. 1975; Zakis 1994) all phenolic structures were divided into four groups (I, IIa, III, and IVa types), as shown in Figure 1. So, only α-carbonyl groups were mentioned as a source of conjugation, and the maximum absorbance was measured at 360 nm. In technical lignin samples, the presence of other conjugated structures such as stilbenes (Yamasaki et al. 1981) and enol ethers (Gellerstedt and Lindfors 1987), may be significant, and the maximum absorbance in the 350–370 nm region can be assumed to reflect the content of all types of conjugated structures. So, measurement of the analytical signal and calculation of the total content of phenolic groups was carried out as described in Section 2.3.1 according to equation (4)–(7).

Difference spectra of sample solutions in units of absorption (ΔA) and absorptivity (Δa, equation (2)) are presented in Figure 2.

Figure 2:

Difference spectra of sample solutions. (a) In units of absorptivity. (∆a = a_pH12-a_pH6) (b) in units of absorptivity. (∆a = a_0.2MNaOH-a_pH6) (c) in units of absorption. (ΔA = A_{0.2 M NaOH} – A_{pH 6}).

There are three groups of peaks in spectra in Figure 2. First group with maxima at wavelengths between 230 and 260 nm. This wavelength range is common for technical lignins and originates from the presence of aromatic rings and non-conjugated phenolic –OH groups in the lignin structure. Second group: The peaks in this group have maxima at wavelengths around 300 nm. The maxima in this range indicate the presence of unconjugated Ph-OH in the lignin samples. Third group: The peaks in this group have maxima at wavelengths between 350 and 380 nm. The presence of maxima in this range indicates the presence of conjugated phenolic –OH groups in the lignin samples. So, the absorbance at this wavelength range can be valuable in identifying and characterizing the chemical structures present in the samples being analyzed. Specifically, the presence of peaks in the second and third ranges (300–360 nm and 350–380 nm, respectively) indicates the occurrence of unconjugated and conjugated phenolic –OH groups in the lignin samples. This spectroscopic data can be useful for understanding the composition and properties of the lignin samples in various applications. The values of the total Ph-OH content calculated from Δε-IDUS 1 – Δε-IDUS 4 methods were demonstrated in Table 3 and Figure 3a.

Figure 3:

Results (a) and relative uncertainty toward the ³¹P NMR (b) of total Ph-OH content determined by various Δε-IDUS methods.

It can be seen that the calculated amount of Ph-OH differed depending on the method. In particular, the value of the total phenolic content increases by 1–4 mmolg⁻¹ for all samples when moving from method Δε-IDUS 1 to method Δε-IDUS 4. The minimum values of C(OH_total) were observed in the case of calculations using method Δε-IDUS 1. This can be explained by the fact that Goldschmid (1954) used as the model compounds (Table 2), eugenol and conidendrin, which have a type I structure (Figure 1) according to (Gärtner et al. 1999; Zakis 1994). So, using this approach we will only be able to estimate the content of unconjugated uncondenced Ph-OH groups. Methods Δε-IDUS 2 and Δε-IDUS 3 use model compounds that represent the full spectrum of phenolic OH groups of lignins. It is this fact that ensures an increase in the specified number of OH groups. The increase in the result when going from Δε-IDUS 2 to Δε-IDUS 3 can be explained by a more complete dissociation of C5-substituted OH. The thing is that Lin and Dence (1992) use only one ionization level (pH 12) and describe six possible model structures, of which two types cannot be predicted accurately due to ionization resistance caused by steric hindrance. C5-substituted (or so-called weakly acidic) structures have been assumed by (Goldmann et al. 2017) to be condensed, although this is not always the case. The variation Δε-IDUS 3 utilizes two ionization levels, the first solution being alkaline at pH 12 and the second super-alkaline at 0.20 molL⁻¹ NaOH. Their approaches differ only conceptually with respect to the model structures. Zakis (1994) describes four structure types of phenolic units based on conjugation (α-carbonyl) and C5-substitution (which resist ionization at pH 12). Gärtner et al. (1999) utilize the same model structures of Zakis et al. (1975) (i.e. do not recalibrate the method) and reinterpret the conjugated structures to include the structures with an α-β double bond. Gärtner (Gärtner et al. 1999) also added a term to the calculation equations which had been missing from the method by Zakis (1994). It is possible, however, that other structures also affect the degree of ionization and the difference in absorption in the two alkaline solutions is therefore preferably interpreted as reflecting the proportion of weakly acidic phenols in general. The proportion of “weakly acidic” phenols can thus be estimated by using the two alkaline solutions. The increase in the calculation result when using Δε-IDUS 4 can be explained by two factors. First, it was the choice of wavelengths for measuring the light absorption of the lignin solution, which corresponded to the characteristic maximum of each model compound: 300 nm for the model I, 320 for II, 370 for III, and 350 for IV). Second, Chen et al. (2022) proposed a mathematical model that accounted for the additional contribution of each type of phenolic group to the total light absorption of the solution.

A second trend that can be seen when analyzing the data in Table 3 is the improvement in precision of determination when going from Δε-IDUS 1 to Δε-IDUS 4. The relative standard deviation (% rsd) was used to express precision (Milne et al. 1992):

where s was calculated according equation (9).

It can be seen that the % rsd in these series decreases from 15–95.3 % for Δε-IDUS 1 to 4.4–5.0 % for Δε-IDUS 4 method. The explanation for the reduction of the indeterminate error lies in the peculiarities of the measurement and mathematical processing of the analytical signal. In particular, all calculations in method Δε-IDUS 1 were based on the maximal absorptivity of the solution at only one wavelength. In addition, its location was not clearly defined. It was only defined as the maximum of a band with a center near 300 nm. Instead, a multi-point wavelength measurement was adopted in Δε-IDUS 4.

In our opinion, the Δε-IDUS 4 model most fully describes the spectrochemical properties of the phenolic groups of lignin. As can be seen from Table 4 method Δε-IDUS 3 was the most time-effective method. Therefore, Δε-IDUS 3 and Δε-IDUS 4 seem to be the most promising for rapid assessment of the total Ph-OH content in lignin in industrial laboratories.

3.2 The reference method of Ph-OH quantification

To control the accuracy of the results of the Ph-OH determination in lignins by the UV–vis method, classical methods of FTIR (Ruwoldt et al. 2022), non-aqueous titration (El Mansouri and Salvado 2007), aminolysis (Gärtner et al. 1999) and ³¹P NMR (Chen et al. 2022) were used. In our opinion, the last of the listed methods is the best option for use as a referee. The ³¹P NMR is a powerful tool for the structural elucidation of Ph-OH. This accurate, quantitative, and highly reproducible analytical procedure, which quantifies and differentiates different types of hydroxy, phenolic, and carboxyl groups in lignins, has now become a common analytical tool (Wen et al. 2013). Quantitative ³¹P NMR makes it possible to reliably identify unsubstituted, o-monosubstituted, and o-dissubstituted phenols (Granata and Argyropoulos 1995), aliphatic –OH, and part of the carboxylic acid in lignins with a wide application potential. At the same time, ³¹P NMR is not sensitive to impurities that are not active after phosphorylation. It is the most mentioned method (Serrano et al. 2018) for determining phenols in lignins. However, the preparation of the lignin samples is time-consuming, extremely expensive, and needs professional maintenance. Thus, the ³¹P NMR method cannot quantify the Ph-OH content in lignin quickly, economically, and extensively, but it is widely used as a reference when comparing the results obtained by other methods.

In this study, the results obtained by modified Δε-IDUS methods were compared with ones from ³¹P NMR.

The ³¹P NMR spectra of all samples are presented in Figure 4, the results of calculations based on these spectra can be seen in Table 5. As can be seen in Figure 3a, the results of the NMR determination of total phenolic content agree with published data (Gordobil et al. 2016; Tagami et al. 2019). The relative uncertainty of data obtained (rsd, %, calculated according equation (10)) does not exceed 9.5 % for all samples, except for OSE, which was not completely dissolved during the sample preparation process. This fact showed that the NMR method is reliable and the result can be used as a reference for quantitative determination of Ph-OH by UV–vis methods.

Figure 4:

³¹P-NMR spectra of lignin samples. Condensed OH include syringyl and 5-substituted Ph-OH.

3.3 Comparison of ³¹P NMR and Δε-IDUS methods

It can be observed that the types of phenolic–OH groups, as determined by ³¹P NMR (Figure 5) and UV–vis methods (Figure 1), are different. As a result, we can compare only the total quantity of phenolic –OH groups. The main criteria for comparing was the quality of the obtained results (the coincidence of the amount of Ph-OH determined by variety Δε-IDUS and ³¹P NMR methods). The relative error (RE) parameter was used in the presented work to express the accuracy of the results obtained. RE was calculated as:

Where c ( O H ∆ ε − IDUS ) ‾ and c (OH_31PNMR), mmol^.g⁻¹ – values of Ph-OH total content determined by modified UV–Vis and ³¹P MNR method respectively.

Figure 5:

Types of phenolic structures determined by the ³¹P-NMR method.

In the presented work, a relative error of ± 10 % was considered an acceptable deviation of the tested method. A deviation of ≤10 % was considered to be sufficient to recommend the method for further use.

One can see in Figure 3b that the Δε-IDUS 1 and Δε-IDUS 2 methods gave significantly lower (compared to the ³¹P NMR result) values of the Ph-OH content. Such a tendency in case Δε-IDUS 1 can be explained by a natural limitation of the applied model. In particular, Goldschmid (1954) calculated the content of Ph-OH using equations based on only non-conjugated phenolic groups. The main reason for underestimated results in the case of the Δε-IDUS 2 method is the incomplete dissociation of phenolic hydroxides when using an alkaline buffer solution with a pH of 12 (Gärtner et al. 1999; Goldmann et al. 2017). Also, some authors assumed (Zakis 1994) that phenolic groups are partially ionizied in the neutral solvent (when lignin is dissolved in a buffer with pH 6). However, subsequent studies and our observations did not confirm this fact.

The extractive approach used for isolation dramatically affects the structure of the resulting lignin (Adler 1957). Argyropoulos et al. (2021) show two lignin structures that differ in the isolation approach used (see Figure 6). Some considerations regarding the impact of the extraction method could be highlighted. First, kraft lignin is an alkylated, highly fragmented, and condensed lignin, while organosolv lignin has a structure similar to ground wood lignin (isolated using the Björkman approach) (Björkman 1956, 1957a, 1957b). Accordingly, the nature of phenolic hydroxyl groups is also different.

Figure 6:

Representative structures for technical lignins (Argyropoulos et al. 2021), kraft (a) and organosolv (b).

UV–vis-spectroscopy determines only some phenolic structures, so the phenolic groups might be underestimated. The equations used in the calculations are based on molar extinction values of model compounds representing typical phenolic structures present in lignin (Table 2). It is not therefore recommended to use the UV–vis method on highly modified lignins, oxlignin particularly, or lignins of very exotic origin since these may contain structures that have not been included in the model.

Results presented in Table 5 and Figure 3b show a good correlation between the modified Δε-IDUS 3 with two alkali solutions and the ³¹P NMR method for EL, ELA, SL, and OS samples. When using the Δε-IDUS 3 method, a significant deviation of the results is observed only for samples of organosolv lignins (OSS and OSE). The most probable reason is that the reference compounds chosen for this model do not correspond to the peculiarities of the structure of organosoluble lignins. The deterioration of the analysis result in the case of organosolve eucalyptus lignin (OSE) can be additionally influenced by its incomplete solubility in NaOH solution with pH 12.

The Δε-IDUS 4 method gives the best results (see Table 5) for the determination of total Ph-OH in spruce lignin samples (OS, OSS, and SL). As it is known (Ragauskas et al. 2014), the lignin of softwood, especially spruce, consists mainly of G units. In contrast, eucalyptus hardwood contains a significant proportion of S units. Large errors in the results of the analysis of eucalyptus lignins were expected. These are obviously explained by the natural limitation of the method, which was developed specifically for the quantification of guaiacyl-type Ph-OH.

Considering the satisfactory accuracy, the absence of complex preliminary sample preparation and environmental friendliness modified Δε-IDUS 3 and Δε-IDUS 4 could be recommended for rapid evaluation of total phenolic hydroxyl structures in common types of technical lignins and in lignins originated from softwood, respectively.