Activation reactions of the softwood kraft lignin

(1)

Dissertations in Forestry and Natural Sciences

HANNA PAANANEN

ACTIVATION REACTIONS OF THE SOFTWOOD

KRAFT LIGNIN

PUBLICATIONS OF

THE UNIVERSITY OF EASTERN FINLAND

(2)

(3)

Hanna Paananen

ACTIVATION REACTIONS OF THE SOFTWOOD KRAFT LIGNIN

Publications of the University of Eastern Finland Dissertations in Forestry and Natural Sciences

No 405

University of Eastern Finland Joensuu

2020

Academic dissertation

To be presented by permission of the Faculty of Science and Forestry for public online examination on December, 10, 2020, at 12 o’clock noon

(4)

Grano Oy Jyväskylä, 2017 Editor: Nina Hakulinen

Distribution: University of Eastern Finland / Sales of publications www.uef.fi/kirjasto

ISBN: 978-952-61-3656-1 (nid.) ISBN: 978-952-61-3657-8 (PDF)

ISSNL: 1798-5668 ISSN: 1798-5668 ISSN: 1798-5676 (PDF)

(5)

Author’s address: Hanna Paananen

University of Eastern Finland Department of Chemistry P.O. Box 111

80101 JOENSUU, FINLAND email: hanna.paananen@uef.fi

Supervisors: Professor Emerita Tuula Pakkanen, Ph.D.

80101 JOENSUU, FINLAND email: tuula.pakkanen@uef.fi Professor Mika Suvanto, Ph.D.

80101 JOENSUU, FINLAND email: mika.suvanto@uef.fi

Reviewers: Principal Scientist Tarja Tamminen, Ph.D.

VTT Technical Research Centre of Finland Tietotie 2

02150 ESPOO, FINLAND email: tarja.tamminen@vtt.fi

Professor Emeritus Raimo Alén, D.Sc.

University of Jyväskylä Department of Chemistry P.O. Box 35

40014 JYVÄSKYLÄ, FINLAND email: raimo.j.alen@jyu.fi

Opponent: Professor Ilkka Kilpeläinen, Ph.D.

University of Helsinki Department of Chemistry P.O. Box 55

00014 HELSINKI, FINLAND email: ilkka.kilpelainen@helsinki.fi

(6)

(7)

Paananen, Hanna

Activation reactions of the softwood kraft lignin Joensuu: University of Eastern Finland, 2020 Publications of the University of Eastern Finland Dissertations in Forestry and Natural Sciences 2020; 405 ISBN: 978-952-61-3656-1 (print)

ISSNL: 1798-5668 ISSN: 1798-5668

ISBN: 978-952-61-3657-8 (PDF) ISSN: 1798-5676 (PDF)

ABSTRACT

Lignin is the second most abundant biopolymer worldwide and has raised interest in several fields due to its high aromatic content. Kraft lignin is mostly a byproduct of the pulping industry and is used as an energy source. Due to the aromatic content, lignin can be applied as a substitute for phenol in phenolic resins and in fine chemical production. Lignin has a large molar mass, which complicates its utilization; therefore, the lignin material is often depolymerized, up to the monomeric units, or activated by increasing the number of reactive sites. In this study, softwood kraft lignin hydroxymethylation was studied with a straightforward differential scanning calorimetry method, which can depict the reactivity of lignin. Para- formaldehyde was used as the formaldehyde source because of its simple composition and easy usability.

Alkaline oxidative depolymerization was used for the activation of softwood kraft lignin. In this study, the changes in structure and reactivity were examined. Three different bases (NaOH, KOH, and NH4OH) were applied, and the chemical compositions of the reaction products were characterized with Fourier-transform ion cyclotron resonance (FT-ICR) mass spectrometry and nuclear magnetic resonance spectroscopy (³¹P, ¹³C, and ¹H-¹³C HSQC NMR) methods. Alkaline oxidation was found to decrease the molar mass. The formation of small carboxylic acids was observed, which indicates partial degradation of the lignin structure. Depolymerized lignin showed a notable reactivity in hydroxymethylation.

The simultaneous hydroxymethylation of softwood kraft lignin and phenol was studied in the temperature range of 80-120 °C. The methylolation was followed using the ³¹P, ¹³C, and ¹H-¹³C HSQC NMR methods. Lignin had a lower reactivity than phenol due to the lower number of reactive sites and steric hindrance present in the large macromolecular structure. Methylene bridges were mainly formed between the phenolic structures.

(8)

Universal Decimal Classification: 547.992; 543.5; 66.095.253

Library of Congress Subject Headings: Lignin; Reactivity (Chemistry); Structure-activity relationships (Biochemistry); Polyoxymethylene; Calorimetry; Methylation; Oxidation;

Mass spectrometry; Nuclear magnetic resonance spectroscopy.

(9)

ACKNOWLEDGMENTS

This study was carried out in the Department of Chemistry, University of East- ern Finland from 2016 to 2020. Financial support from the Faculty of Science and Forestry (SCITECO) of the University of Eastern Finland, the European Regional Development Fund (EVIM project), and The North Karelia Regional Fund of the Finnish Cultural Foundation are gratefully acknowledged.

I am most grateful to my supervisor Prof. Emerita Tuula Pakkanen for the op- portunity to work in the field of materials chemistry. Your supervision and motiva- tion have helped me through the studies. I sincerely thank Docent Leila Alvila for the support and guidance provided during my studies. Your door has always been open, and you always made time for engaging in helpful discussions. I also would like to thank Mika Suvanto for the collaboration.

My appreciation goes to the staff members of the Department of Chemistry, especially Päivi Inkinen, Dr. Sari Suvanto, and Martti Lappalainen. It was nice to come to work every day! Special thanks go to Dr. Ville Nissinen; you are the best of the best workmate. Paavo Auvinen, Pauliina Nevalainen, and Niko Kinnunen – thank you for the pleasant conversations.

Many thanks to my close friends and family for your support and encourage- ment.

Joensuu, 15.10.2020 Hanna Paananen

(10)

LIST OF ABBREVIATIONS

APCI Atmospheric-pressure chemical ionization APPI Atmospheric-pressure photoionization DMSO Dimethylsulfoxide

DSC Differential scanning calorimetry ESI Electrospray ionization

F Formaldehyde

F/P Formaldehyde/phenol ratio FT-ICR Fourier-transform ion cyclotron

G Guaiacyl unit

GPC Gel permeation chromatography

H p-Hydroxyphenyl unit

HMBC Heteronuclear multiple bond correlation HSQC Heteronuclear single quantum coherence

IR Infrared

IS Internal standard

L Lignin

LPF Lignin-phenol-formaldehyde Mn Number-average molecular mass Mw Weight-average molecular mass NMR Nuclear magnetic resonance

P Phenol

PF Phenol-formaldehyde

PFA Paraformaldehyde

S Syringyl unit

Tg Glass transition temperature TGA Thermogravimetric analysis TOCSY Total correlation spectroscopy

UV Ultraviolet

(11)

(12)

LIST OF ORIGINAL PUBLICATIONS

This thesis is based on data presented in the following articles, referred to by the Roman Numerals I-III.

I Paananen H, Pakkanen T. T (2020) Kraft lignin reaction with paraformaldehyde. Holzforschung, 74(7): 663-672.

II Paananen H, Eronen E, Mäkinen M, Jänis J, Suvanto M, Pakkanen T. T (2020) Base-catalyzed oxidative depolymerization of softwood kraft lignin. Industrial Crops and Products, 152: 112473.

III Paananen H, Alvila L, Pakkanen T. T Hydroxymethylation of softwood kraft lignin and phenol with paraformaldehyde, submitted for publication.

(13)

AUTRHOR’S CONTRIBUTION

I) The author planned the experimental work and wrote the manuscript under the guidance of the supervisor. The author carried out the experimental work.

II) The author developed the experimental work and wrote the manuscripts under the guidance of the supervisor and coauthors. The mass spectrometry measurements of the lignin samples were conducted in cooperation with the Professor Janne Jänis group. The author carried out the IR, NMR, and DSC experiments.

III) The author planned the experimental work and wrote the manuscripts under the guidance of the supervisors. The author performed the experimental work.

(14)

1 INTRODUCTION

1.1 LIGNIN OCCURRENCE AND ISOLATION METHODS

Lignin was discovered in the beginning of 1800. A.P. de Candolle, a Swiss bota- nist, isolated lignin in 1813 and described the fibrous material as nonsoluble in water or alcohol, but soluble in alkaline solutions.(Sjöström, 1993) Lignin is the second most abundant biopolymer worldwide. Lignin appears in woody and herbaceous plants. The wood species are typically divided into two categories, i.e., softwoods and hardwoods. Wood materials consist of cellulose (40-60%), hemicelluloses (10- 40%), lignin (15-30), and extractives. The relative amounts and compositions of these components vary depending upon the plant species. In softwoods, the lignin content is higher than in hardwoods and herbaceous crops. In softwoods, the lignin content is 21-29%; in hardwoods, it is 18-25%; and herbaceous crops have lignin contents of 15-24%.(Cao et al., 2018; Ponnusamy et al., 2019; Schutyser et al., 2018b)

The existence of lignin was found, when wood components were separated into cellulose and lignin. In plants, lignin acts as an adhesive to form a rigid structure between cellulose and hemicelluloses. Lignin also provides resistance to water en- tering the cell wall and protects against microorganisms. The occurrence of lignin in plants is presented in Figure 1. From the time of its discovery, lignin has been studied, and the main target has been to determine the representative structure and, later, to find utilization targets.

Figure 1. Lignin in plants.(Zakzeski et al., 2010)

The lignin structure depends on, for example, the plant and separation method.(Cao et al., 2018; Ek et al., 2016) Lignin can be isolated from plants using different methods. The isolation method influences the molar mass and structure of the lignin. The isolation methods can be divided into physical, chemical, and bio-

(18)

logical methods. The chemical isolation methods are divided into the following four categories: alkaline, acidic, reductive, and other methods. Chemical isolation methods are powerful and do not need harsh conditions; therefore, these methods are most commonly used in industry.(Cao et al., 2018) The alkaline isolation methods are most often used. For example, kraft and soda lignins are isolated using alkaline conditions. The kraft method is the most common method used, producing up to 85% of the total lignin production. The lignin is precipitated from an alkaline solution (black liquor), which contains inter alia, sodium hydroxide and sodium sulfide.

The method is related to the pulping industry. The harsh alkaline methods cause the lignin depolymerization and repolymerization processes to occur during isolation, which affect the lignin structure. The weak interunit ether linkages are cleaved but the C-C linkages, such as 5-5, survive and are also formed during the kraft process. The covalently bound sulfur, caused by the process, is often detected in lignin.

In addition, carbohydrates and fatty acids are common impurities of kraft lignin.(Chakar and Ragauskas, 2004; Galkin and Samec, 2016) Soda pulping is used for nonwoody plants, and the method is similar to kraft pulping. The process conditions of soda pulping are gentle and therefore convenient for nonwoody biomass types with a low lignin contents and open structures.(Schutyser et al., 2018b;

Zakzeski et al., 2010)

Lignosulfonates are products from sulfite pulping, which occur typically in neutral conditions. Because of the addition of sulfonate groups to the lignin structure, the lignosulfonates have better water solubility at different pH values compared to kraft lignin. The kraft process has partially replaced the sulfite process.(Liu et al., 2019; Schutyser et al., 2018b; Zakzeski et al., 2010)

The popular acidic isolation method is organosolv pulping. The most acidic methods are performed in aqueous solutions, but the addition of organic solvent increases the delignification. The organosolv process is carried out in the presence of an organic solvent, mineral acids and/or water. The produced organosolv lignin has a good solubility in organic solvents. The organosolv lignin includes oligomeric lignin fragments due to depolymerization and condensation reactions in acidic media. Reductive catalytic fractionation produces depolymerized lignin oil. The delignification depends on the solvent used, the reaction time, and the temperature applied. The solvents used are alcohols containing water. In addition, an acidic cocatalyst can be used to improve the lignin extraction.(Lange et al., 2013)

The other lignin isolation methods include mechanical pretreatment and ionic liquid treatment. Milled wood lignin is isolated from wood using ball milling. The extraction is performed at room temperature, and lignin is extracted with an organic solvent. The technique provides lignin that is close to native lignin. The milling time is high, and the delignification degree is low. Ionic liquids are used for two

(19)

different ways: dissolution and pulping. Lignocellulose material is either fully dissolved or only lignin and hemicellulose are extracted with ionic liquid. In the dissolution approach, cellulose is first precipitated with an anti-solvent and then lignin. In the pulping approach, cellulose is left as a solid pulp. The anion used in the solution has an effect on the β-O-4 cleavage.(Liu et al., 2019; Pin et al., 2020;

Schutyser et al., 2018b; Xia et al., 2020)

1.2 LIGNIN STRUCTURE

The structure of native lignin consists of three different monolignols (Fig. 2). The different monolignols have different numbers of methoxy groups in the aromatic structure. The main monolignol type in softwood lignin is coniferyl alcohol; it can also contain a small amount of p-coumaryl alcohol, but usually, no sinapyl alcohol units are present. Hardwood lignin contains both coniferyl alcohol and sinapyl alcohol units. There can be equal amounts of these monolignols or higher levels of sinapyl alcohol. Hardwood lignin can also contain small amounts of p-coumaryl alcohol. Grass lignins contain all monolignol types. The p-coumaryl content is higher than in wood lignins, but coniferyl alcohol is the most common monomer present. The monolignol composition is presented in Table 1.(Duval and Lawoko, 2014) The lignin structure can also contain substantial amounts of phenolic compounds, such as hydroxycinnamates, p-hydroxybenzoate and tricin. The com- pounds originate from incomplete polymerization.(Ek et al., 2016)

Figure 2. Phenylpropanol units and the corresponding monolignols in the lignin structure.

Table 1. Monolignol composition in plants (%).(Ek et al., 2016)

Plant p-Coumaryl alcohol Coniferyl alcohol Sinapyl alcohol

Softwood <5 <95 None or trace

Hardwood 0-8 25-50 46-75

Grass 5-33 33-80 20-54

(20)

The monolignols are bound together via radical reactions. The radical reactions are initiated by laccase and peroxidase enzymes, which oxidize the phenolic OH- group and form a phenoxy radical. The monolignols are bound in different orders, and therefore, different interunit linkages can be formed due to the conjugated π- system of the monolignols. The β-O-4 ether bond is the most abundant bond type, but other types of ether and carbon-carbon bonds are also formed. The most common bond types and structural units are presented in Figure 3. The interunit bonds depend on the monolignol distribution because the occupied ortho position cannot take part in the formation of carbon-carbon bonds. Consequently, syringyl (S) unit lignin can have a lower number of carbon-carbon bonds compared to guaiacyl (G) unit lignin. The ether bonds are weaker compared to the carbon-carbon bonds and are more easily depolymerized.(Ek et al., 2016; Laurichesse and Avérous, 2014)

The monolignols form a huge macromolecular structure, and lignin does not have a primary structure due to the random polymerization. The reported average molar mass values are between 2500-10000 g/mol. Of course, the plant type and isolation method can cause structural changes.(Lange et al., 2013; Zakzeski et al., 2010)

Figure 3. A structural segment of softwood lignin and bond types between phenylpropane units.

(21)

1.3 LIGNIN CHARACTERIZATION

Lignin has a large network structure with a high aromatic content. The aromatic structures have several types of functional groups, and the aromatic structures are bound together via different linkages. Lignin itself consists of carbon, hydrogen, and oxygen, but impurities, such as inorganic salts and sulfur, are usually present in cases of technical lignins. The difficulties in the structural characterization are due to the low solubility, large molar mass, and heterogeneity. The valorization of lignin requires a comprehensive characterization of the lignin.

1.3.1 NMR methods

NMR methods, such as ¹H, ¹³C, ³¹P, and two-dimensional ¹H-¹³C HSQC, are widely used for lignin characterization. The ¹³C and ³¹P NMR methods are most frequently used. ¹H NMR is a sensitive method, but due to the complex structure of lignin, the proton signals are not easily identified. The ¹³C NMR method is more valuable, because the signals of different structures can be separated, and overall,

13C NMR is one of the most reliable characterization methods for lignin.(Balakshin and Capanema, 2015) Aromatic and aliphatic carbons can be identified and inte- grated. The three main regions in the ¹³C NMR spectrum of lignin are the aromatic carbons between 105-160 ppm, methoxy signals at approximately 58 ppm, and aliphatic carbons in the range of 50-20 ppm. Due to the complex structure of lignin, a large amount of the sample is needed (200-500 mg in 0.5 mL solvent). The most commonly used solvent in ¹³C NMR measurements is deuterated dimethyl sul- fokside. The lignin must be fully solubilized to achieve good quality spectra. Acety- lation of lignin yields better solubility. For lignin acetylation, an acetic anhydride and pyridine solution mixture is often used. To achieve quantitative spectra, the relaxation delay must be long enough. A relaxation reagent, for example, chromium(III) acetylacetonate, can be used to accelerate relaxation for taking NMR measurements.(Balakshin and Capanema, 2015; Capanema et al., 2004; Shi et al., 2019)

The ³¹P NMR method provides information on lignin’s aliphatic and phenolic hydroxyl groups. ³¹P NMR can provide quantitative information, and the data are usually compared with those obtained from ¹³C NMR.(Li et al., 2018b) Lignin is phosphitylated with the reagent 2-chloro-4,4,5,5-tetramethyl-1,3,2- dioxaphospholane.(Granata and Argyropoulos, 1995) The phosphitylation reaction is presented in Figure 4. Granata and Argyropoulos (Granata and Argyropoulos, 1995) have presented a sample preparation, where the lignin sample is dissolved in a pyridine/chloroform solvent mixture, to which the phosphitylation reagent, pyridine solution of chromium(III) acetylacetonate and an internal standard are added.

The solvent used was deuterated chloroform. The most commonly used internal

(22)

standards are cyclohexanol and cholesterol.(Argyropoulos, 2010; Balakshin and Capanema, 2015; Li et al., 2018b)

Figure 4. Lignin phosphitylation reaction.

The¹H and ¹³C NMR methods can be combined in a two-dimensional NMR method. The 2D NMR method is an effective tool to identify and quantify the aromatic structures and interunit linkages of lignin, such as S, G, and H monolignols and ether-linkages. The 2D NMR methods used for lignin analysis are, for example, heteronuclear single-quantum coherence (HSQC), heteronuclear multiple bond coherence (HMBC), and heteronuclear single quantum coherence – total correlation spectroscopies (HSQC-TOCSY). HSQC is the most commonly used 2D NMR method in the field of lignin research, since the method provides structural information, such as interunit linkages and polysaccharides. The sample preparation is similar to the normal ¹³C NMR sample preparation, but more experimental parameters have to be considered (e.g., coupling and pulse offsets).(Constant et al., 2016; Crestini et al., 2017; Lu et al., 2017) With the ¹H-¹³C HSQC NMR method, chemical structures appearing at low contents could be detected due to the higher sensitivity of the HSQC method compared to the normal ¹³C method.(Schutyser et al., 2018b)

1.3.2 Mass spectrometric analysis

Mass spectrometric analysis is fast and highly sensitive. Mass spectrometry has been used for the identification of lignin components and the study of structural changes after modification. Mass spectrometry methods are usually classified on the basis of ionization methods. The three main ionization methods for lignin analyses are atmospheric pressure photoionization (APPI), atmospheric-pressure chemical ionization (APCI), and electrospray ionization (ESI). The ESI method provides information on polar, oxygen-containing compounds. The APCI method resembles the ESI method but also ionizes small non-polar compounds. The APPI method ionizes neutral and basic compounds, such as neutral carbohydrates and phenols, and APPI is also suitable for conjugated structures.(Banoub et al., 2007) The ESI, APCI, and APPI methods provide complementary information, and the combina- tion of these methods yields comprehensive information about the lignin structure.(Kosyakov et al., 2016; Miettinen et al., 2017, 2015; Qi et al., 2020)

Mass spectrometry presents the mass to charge ratio of ions. The sample is con- verted to ions in the gas phase, and the ions are placed in high magnetic and elec- tric fields, which move the ions to the detector. The result is a plot of the ion abun-

(23)

dance versus the mass to charge ratio.(Stroobant et al., 2007) The solid samples, such as lignin, have to be dissolved. The solubility of lignin is poor, and the dimethyl sulfoxide solvent, which is widely used for lignin, is not suitable due to the high boiling point. Solvent mixtures, for example, water/methanol, have been used for taking lignin measurements.(Qi et al., 2020)

The high number of detected species is an advantage of mass spectrometric measurements. On the other hand, the data processing involved can be challenging.

The van Krevelen diagram is widely used for data presentation. The plot represents the hydrogen-carbon ratio vs. oxygen-carbon ratio as a color mapped diagram.(Qi et al., 2020) From the mass spectrum of a heterogeneous/polymeric mixture (e.g.

lignin), weight-average (Mw) and number-average (Mn) molecular masses can be computed.

1.3.3 Thermal analysis

Thermal analysis methods, such as thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC), are used to study the thermal stability of different lignins. The TGA measurement describes the lignin degradation during heating, in which the sample is heated and the mass loss is measured. The lignin starts to degrade at a low temperature of approximately 150 °C, where the cleavage of hydroxyl groups and ether bonds occurs. As the temperature is increased, alkyl sidechains, and carbon-carbon bonds start to degrade. The glass transition temperature, T^g, can be detected with DSC. The T^g is related to the molar mass. The higher the molar mass, the higher the T^g.(Laurichesse and Avérous, 2014; Sahoo et al., 2011)

The lignin ash content can be detected, for example, by heating the dried lignin sample to 575 °C for 3 h. The structure of lignin degrades, and the ash, containing inorganic compounds of calcium, potassium, and sodium, is weighed. The lignin isolation process has an effect on the inorganic material composition.(Miettinen et al., 2017; Sameni et al., 2013)

1.3.4 IR analysis

Infrared spectroscopy (IR) is a quick technique to qualitatively study the lignin structure and chemical composition. The method is straightforward and nonde- structive; additionally, the sample preparation is fast, and a solid sample can be used.(Boeriu et al., 2004; Watkins et al., 2015)

(24)

1.3.5 Gel permeation chromatography

Gel permeation chromatography (GPC) is used to determine the molar mass distribution of lignin. The method separates molecules on the basis of their molecular size, and the separated molecules are detected with a UV detector. Prior to the GPC analysis, the lignin is acetylated, usually with an acetic anhydride-pyridine mixture, to increase the solubility of the lignin. Tetrahydrofuran is the most commonly used solvent in the mobile phase.(Cao et al., 2012)

1.3.6 Klason lignin

Klason lignin describes the acid insoluble lignin content in biomass samples. The method has been described as direct and reliable for lignin analysis and determination. The material is treated with concentrated (72%) sulfuric acid for 4 hours, after which the acid solution is diluted to 3% and refluxed for 2 hours. The insoluble portion is dried and weighed.(Aldaeus et al., 2011; Chen, 2015; TAPPI, 2006) The acid-soluble portion is determined with UV spectrophotometry using the wave- length of 205 nm. The acid-soluble lignin mainly contains small lignin degradation products and lignin-carbohydrate compounds. The sum of Klason lignin (insoluble portion) and acid-soluble lignin is equal to the total lignin content.(TAPPI, 2006;

Yasuda et al., 2001)

1.4 LIGNIN REACTIVITY

The chemical reactivity of lignin depends on the structure; therefore, the lignin source and isolation method are important factors. The ratio of monolignols affects the reactivity. The large number of free reactive sites in the phenolic ring enhances the reactivity. The molar mass is also an impressive factor, and therefore, the isolation method has an effect on the reactivity of the lignin material. The straightforward determination of the phenolic hydroxyl group content using NMR methods can provide information about lignin reactivity.(Jiang et al., 2018; Shimizu et al., 2012) The modification of the lignin structure by functionalization increases the reactivity, and the determination of new active sites provides the reactivity.(Jiang et al., 2018; Laurichesse and Avérous, 2014; Podschun et al., 2015) The Mannich reaction has been used to measure lignin reactivity. The Mannich reaction is a condensation reaction, where lignin reacts with an amine and formaldehyde (Fig. 5). The Mannich reaction can be performed under neutral, acidic or alkaline conditions.

After the reaction, the nitrogen content is measured, and the nitrogen content describes the free ortho sites in the lignin structure.(Du et al., 2014; Wang et al., 2016) The hydroxymethylation of lignin by formaldehyde is a possible side reaction in the Mannich reaction.

(25)

Figure 5. Mannich reaction.

Hydroxymethylation of lignin also takes place at the free reactive sites, the ortho position of the aromatic structure or at the side chain. The reaction is presented in Figure 6. Hydroxymethylation is also used for lignin activation.(Malutan et al., 2008; Taverna et al., 2019)

Figure 6. Lignin hydroxymethylation at an aromatic ortho position and a side chain.

The modification of lignin by, for example, phenolation, etherification or esteri- fication enhances the lignin reactivity by increasing the number of reactive sites.

Phenolation is performed in acidic media, where phenol reacts with the lignin aromatic ring and side chains. The free reactive sites of phenol can react with formaldehyde. Phenolation has been found to be an effective method for the preparation of phenolic resins.(Hu et al., 2011; Jiang et al., 2018; Laurichesse and Avérous, 2014;

Podschun et al., 2015)

1.5 LIGNIN DEPOLYMERIZATION

The utilization of lignin with a high molar mass is difficult. There are several modification pathways depending on the final target. The most commonly used method in depolymerization reactions is the cleavage of ether bonds. Due to the heterogeneous lignin structure, it is difficult to find a selective method for lignin depolymerization. On the other hand, the lignin structure affects depolymerization, especially the β-O-4 content, which is crucial. The monomer yield correlates with the β-O-4 content. Therefore, hardwoods give higher monomer yields compared to softwoods. In depolymerization processes, repolymerization and the formation of

(26)

carbon-carbon bonds also occur. Impurities, such as sulfur, alkali metals, and resid- ual carbohydrates, deactivate some catalysts and reactions. The isolation process affects the lignin structure and, therefore, can influence the depolymerization effi- ciency. Mild isolation methods, for example, organosolv processes, produce more active lignin than the traditional processes, such as the kraft process, produce.

1.5.1 Oxidative depolymerization

Oxidative depolymerization involves lignin depolymerization in the presence of an oxidizing agent. The oxidants generally used are oxygen gas and hydrogen per- oxide. The target of depolymerization can be side chain or phenolic ring cleavage.

Metal complexes of vanadium, copper, and cobalt have been used as heterogeneous catalysts, but the depolymerization results have not been effective. Additionally, the use of other catalyst types has been reported. Oxidative depolymerization can occur in acidic or neutral media.(Hasegawa et al., 2011; Liu et al., 2019; Ma et al., 2018; Rahimi et al., 2014; Toledano et al., 2012)

Alkaline oxidative depolymerization has been found to produce selectively aromatic compounds, for example, aldehydes and acids. The number of lignin hydroxyl groups decreases and the number of acid groups increases.(Demesa et al., 2015; Kalliola et al., 2011; Ma et al., 2018; Rodrigues Pinto et al., 2011; Villar et al., 2001) The product distribution depends on the lignin source, but the well-known reaction product is vanillin. Vanillin is produced mainly from lignosulfonates. The production of vanillin started in 1936, and vanillin is used in industry and research.

Alkaline oxidation is most often carried out in the presence of oxygen gas in a NaOH solution (0.5-4 M). Other bases are also used, for example, KOH. A high pH value is needed to dissolve lignin and ionize free phenolic groups. A high pH is also needed for the deprotonation of reaction intermediates. The reaction temperature is usually 120-190 °C, and the oxygen pressure is 2-14 bar. All of the reaction parameters influence the formation of products. Increasing the temperature and pressure accelerates the formation of vanillin, and the maximum vanillin yield can be achieved faster.(Fache et al., 2016; Schutyser et al., 2018b; Wang et al., 2018;

Zakzeski et al., 2010) During alkaline oxidation, the solution pH and oxygen pressure decrease, and methanol formation occurs. The oxygen consumption was calculated as 0.23-0.27 g O2 for 1 g of softwood kraft lignin.(Kalliola et al., 2011) The oxidative depolymerization provides good yields without catalysts, but some catalyst can be added to increase the reaction selectivity, but it can also accelerate the reaction.

(27)

1.5.2 Reductive depolymerization

Reductive depolymerization has been divided into the mild, harsh, bifunctional, and liquid phase reforming types. Reductive depolymerization requires a redox catalyst and a reducing agent. Hydrogen is the most commonly used reducing agent, and it can appear in a reaction as a gas or as a hydrogen donating species.

Usually, the reducing agent appears as a solvent. If hydrogen gas is used, the process is called hydroprocessing, and if hydrogen exists in the solvent, the reaction is called liquid-phase reforming. In reductive depolymerization, lignin is depolymerized and deoxygenated.(Schutyser et al., 2018b; Tymchyshyn et al., 2020)

Mild hydroprocessing is performed under mild conditions and at low temperatures (≤ 300 °C) in the liquid phase with a metal catalyst. The reaction has good selectivity, and the typical reaction products are different p-substituted methoxy- phenols. The reaction conditions of harsh hydroprocessing are a high temperature (≥ 320 °C) and high hydrogen pressure (≥ 35 bar). The reaction is carried out without a solvent and with a solid catalyst. The main reaction products are phenols with and without methyl groups or alkyl chains. Although the reaction conditions are harsh, the yield and selectivity are not high.(Schutyser et al., 2018b) Mild conditions in depolymerization are also related to the ionic liquids. The depolymerization using ionic liquids can be oxidative, alkaline, acidic, or hydroprocessing depolymerization process.(Xia et al., 2020)

Transition metal catalysts have been used in lignin depolymerization processes.

Noble metals such as rhodium, ruthenium, and palladium have been used for lignin hydrotreatment, and the products are aromatic compounds. NiMo and CoMo catalysts have been used to achieve better yields of alkylphenolics and aromatic compounds. Sulfur-containing catalysts are effective, but the products contain a large amount of sulfur.(Hita et al., 2018)

Bifunctional hydroprocessing contains both acid and metal sites. The hydrolysis and dehydration reactions are catalyzed at acid sites, while hydrogenolysis and hydrogenation are catalyzed at metal sites. The method provides a high selectivity and reaction yield. The reaction products are cycloalkanes and C6-C18 compounds.

Liquid phase reforming is performed under an inert atmosphere and with a hydrogen-donating solvent or agent. The solvents used are tetralin, isopropanol, and formic acid. The reaction temperature is in the range of 150-400 °C.(Schutyser et al., 2018b)

(28)

1.5.3 Thermal depolymerization

In thermal depolymerization, also called pyrolysis, lignin is heated to high temperatures (400-800 °C) in the absence of oxygen. The temperature, heating rate and additives affect the product distribution. In any case, the reaction product distribution is wide, with oils, char, and gases being formed. The main products are carbon oxide gases (carbon monoxide and carbon dioxide), small hydrocarbons, volatile liquids (methanol, acetone), monolignols, and substituted phenols. The pyrolysis selectivity is low. Bio-oils are produced with fast pyrolysis. In the production process, the temperature is rapidly increased, and after 5-30 minutes, the temperature is decreased and the formed vapors are condensed. Approximately 80% of the lignin material turns into bio crude oil(Laurichesse and Avérous, 2014). The condensation stage is the main step in achieving the products because repolymerization can also occur.(Laurichesse and Avérous, 2014; Pandey and Kim, 2011; Schutyser et al., 2018b)

1.5.4 Catalyzed thermal depolymerization

Catalysis in lignin depolymerization is used for selective bond cleavage and high conversion, as well as to prevent further reactions. The catalysts decrease the activation energy of the reaction and enable the use of milder reaction conditions.

Different acids and bases are usually included in the group of catalysts. Zeolites and silica-alumina catalysts are widely used. H-ZSM-5 zeolite produces aromatic hydrocarbons, while silica-alumina provides aliphatic hydrocarbons.(Pandey and Kim, 2011) Zeolites (e.g. H-ZSM-5 and HY) can be used under different reaction conditions, for example, for pyrolysis to increase the yield of aromatic compounds.(Liao et al., 2020)

(29)

2 UTILIZATION OF LIGNIN

Lignin, having a highly aromatic structure, is of great interest in several fields.

The structure of lignin can be depolymerized into monomeric units, and therefore, it is suitable for fine chemical production, for example, vanillin has been widely produced from lignosulfonates instead of through synthetic routes. The increase in lignin reactivity by depolymerization or modification enables its use for, for example, polymer production.(Laurichesse and Avérous, 2014; Zakzeski et al., 2010) Lignin can replace petroleum-based phenol in phenolic resins, but lignin can also be polymerized with other comonomers, for example, epoxy resins, which are formaldehyde-free adhesives.(Li et al., 2018a) The production of biofuels has also re- ceived attention, the availability of lignin is good, and the costs are low.(Gordobil et al., 2016; Schneider et al., 2019)

Phenol is a toxic chemical, and petroleum-based phenol production is not sus- tainable or environmentally friendly.(Podschun et al., 2015; Yang et al., 2015a) There is a desire to replace phenol with other aromatic compounds, for example, lignin monolignols. Lignin monolignols have structures similar to phenol; therefore, lignin can be utilized for phenol-formaldehyde resin preparation. The syringyl unit does not have any vacant reactive sites, but p-hydroxyphenyl and guaiacyl monolignols have free reactive sites at the ortho positions, while monolignols have their para positions occupied. Due to the steric hindrance and lower number of reactive sites, the reactivity of lignin is lower than that of phenol.(Danielson and Simonson, 1998; Moubarik et al., 2013; Park et al., 2008; Taverna et al., 2019;

Turunen et al., 2003) The formaldehyde reaction with lignin can also occur at the side chain next to the carbonyl group or C-C double bond. The reaction is called the Tollens reaction.(Marton et al., 1966; Peng et al., 1993, 1992; Zhao et al., 1994) The lignin reactivity can be improved through modification, for example, depolymerization or phenolation.(Du et al., 2014; Jiang et al., 2018; Ma et al., 2018)

Phenol-formaldehyde (PF) resins were invented in the early 1900s and are cur- rently widely used as adhesives, for example, in the plywood industry. PF resins have several desirable properties, such as resistance to water, weathering, and high temperatures. Phenol reacts with formaldehyde in the presence of a catalyst, forming a polymeric structure. The first reaction step is methylolation, where methylol groups are formed at the ortho or para positions of the activated phenol. At the methylolation stage, mono-, di-, and trimethylol phenols are formed. The methylolation reaction takes place in the temperature range of 80-95 °C.(Gardziella et al., 2000) The para position of phenol is more reactive than the ortho position.(Grenier- Loustalot et al., 1994) The formaldehyde sources used in resin preparation are formaldehyde solution, paraformaldehyde and hexamethylenetetramine.(Gardziella et

(30)

al., 2000) The second step is condensation, where methylol groups react to form methylene or ether bridges. The methylolation and condensation reactions are presented in Figure 7. At the condensation stage, the dimeric structures build up, and the reaction proceeds, forming larger units. The curing stage is the third step in the polymerization process, where resin is heated and the final cross-linked structure is produced.(Gardziella et al., 2000; Grenier-Loustalot et al., 1996)

Figure 7. Methylolation and condensation reactions of phenol.

The base-catalyzed resins are called resols, and the most commonly used base catalyst in the reaction is NaOH. The resins prepared in acidic media are called novolacs. Resols have formaldehyde/phenol (F/P) ratios greater than 1, and novolacs have molar ratios less than 1. Overall, the reaction is pH dependent, and the pH value in the reaction has an effect on the reaction intermediates and products.(Gardziella et al., 2000; Luukko et al., 2001)

(31)

3 AIMS OF THE STUDY

The study is related to the replacement of phenol with lignin in phenol- formaldehyde resins. Softwood kraft lignin derived from spruce and pine is used in all the experiments.

1. The target is to obtain a straightforward reactivity assessment method for lignin materials. The utilization of paraformaldehyde (PFA) in lignin hydroxymethylation is studied. A commercial formaldehyde solution contains small amounts of methanol, which decelerates the polymerization process in resin preparation. Para- formaldehyde is used due to its simple composition compared to the commercial formaldehyde solution. PFA is a solid material that is easy to handle and store. The paraformaldehyde composition and reactions with lignin are studied with use of the NMR methods. The DSC method is applied to determine the thermodynamic parameters of the hydroxymethylation of lignin with paraformaldehyde.

2. Activation of the lignin material with oxidative alkaline depolymerization.

The aim is to depolymerize lignin into oligomeric phenolic compounds. The product identification and changes in structure are studied. The average molar masses of the pristine kraft lignin and depolymerized lignin samples are measured with mass spectrometry. The ¹³C and ¹H-¹³C HSQC NMR methods are used to study the composition of the depolymerized lignin samples. The reactivities of the depolymerized samples with paraformaldehyde are compared using the DSC method.

3. The simultaneous hydroxymethylation of kraft lignin and phenol is examined.

The formation of methylene linkages between lignin and phenol in the presence of paraformaldehyde is studied with the ³¹P, ¹³C, and ¹H-¹³C HSQC NMR methods.

The NMR methods provide structural information on the reaction progress. The effects of the composition of the lignin-phenol mixture and the reaction temperature are also examined.

(32)

4 EXPERIMENT

4.1 MATERIALS

Softwood kraft lignin has been derived from spruce and pine. The elemental analysis was carried out using an Elementar vario MICRO cube elemental analyzer.

The elemental composition of lignin is 63.4% carbon, 6.0% hydrogen, 2.4% sulfur, 0.0% nitrogen, and the calculated oxygen content is 28.2%. Lignin contains 98.3 wt%

acid-insoluble lignin (Klason lignin), 0.84 wt% extractives, and 2.2 wt% ash. The ash content was determined by heating the sample to 575 °C for 3h. Lignin was freeze- dried and stored in a desiccator.

4.2 PREPARATION OF PARAFORMALDEHYDE SOLUTION

Water was heated to 60 °C, and solid paraformaldehyde was added. H2O:PFA was used in a mass ratio of 5:1. After 10 minutes of stirring, the pH was adjusted to 8 using a 1.0 M NaOH solution, and the solution became clear after 15 minutes. The solution was cooled down and filtered. The solution was kept in a refrigerator and protected from light. The concentration of the prepared PFA solution was 5.4 M (13 wt%), which was determined using ¹³C NMR.

4.3 HYDROXYMETHYLATION REACTIONS OF LIGNIN AND PHE- NOL

The lignin, lignin-phenol and phenol reactions with paraformaldehyde in alkaline conditions were studied. For the reaction, weighed amounts of the paraformaldehyde, NaOH (1.0 M) and lignin or the lignin-phenol mixture or phenol were packed into a Teflon-lined, stainless steel autoclave with a volume of 60 ml. The molar ratio of the reagents used in the lignin-paraformaldehyde reaction was lignin: paraformaldehyde: NaOH = 2.6:2.6:1 (the molar mass of 200 g/mol used for the C⁹ unit of lignin(Wang et al., 2016)). The molar ratio for the lignin-phenol- paraformaldehyde reactions was lignin:phenol:PFA:NaOH = 1.3: 1.3: 2.6: 1 and 1.95:

0.65: 2.6: 1. For the samples of the lignin-phenol mixture, lignin and phenol were used in molar ratios of 1:1 and 3:1. The reactions of the phenol-paraformaldehyde reaction were performed with phenol:PFA:NaOH = 2.6:2.6:1. The total heating time was 45 minutes, and after heating, the autoclave was cooled in an ice bath. The reaction mixture was removed from the autoclave.

(33)

4.4 REACTIONS IN DSC

DSC was used to study the reactivity of the lignin samples. For the measurement, solid paraformaldehyde was added to 1.0 M NaOH solution after mixing, and a weighed amount of lignin was added. The total volume of the sample was 400 μL. Lignin and paraformaldehyde were used in molar ratios of 1:1 and 1:2, when the lignin:NaOH molar ratio was varied from 10:1 to 2.6:1 to examine the effect of NaOH on the reaction. The final reactivity studies were carried out with molar ratios of lignin:PFA:NaOH = 2.6:2.6:1. For the DSC measurement, approximately 25 mg of sample was placed in 30 μL stainless steel high-pressure capsules.

The samples were measured from 30 to 300 °C with a heating rate of 10 °C/minute.

The reaction enthalpies of the samples were measured with a Mettler Toledo DSC 823^e instrument. The reproducibility was verified with three parallel samples, and the average value is presented.

4.5 LIGNIN ALKALINE OXIDATIVE DEPOLYMERIZATION

In our study, the oxidant used was oxygen gas^, and the temperature was 100 °C.

Alkaline conditions were applied; therefore, the base acts as a catalyst and a solvent. The bases used in the study were NaOH, KOH and NH4OH. NaOH and KOH are strong bases, and NH4OH is a weak base. The base concentration used was 3.0 M in all experiments, and the total volume of the solution was 12 mL. For the reaction, 1.0 g of freeze-dried lignin was used. The reaction times were 2 and 6 hours.

The oxidation reactions were carried out in an autoclave. The starting materials were packed in an autoclave, and magnetic stirring was used to ensure proper mixing during the reaction. The autoclave was pressurized with oxygen gas before heating, and during the reaction, the pressure was monitored with a pressure gauge. The reaction mixture was then divided into two parts. One part was only freeze-dried, and the second part was neutralized with HCl and then freeze-dried.

4.6 PREPARATION OF NMR SAMPLES

The ¹³C and ¹H-¹³C HSQC NMR measurements were performed on Bruker Ul- trashield^TM AMX-400 and Jeol 500 MHz spectrometers. An inverse gated-pulse pro- gram was used to obtain the quantitative ¹³C NMR spectra. The two-dimensional spectra were measured from the same samples used in the one-dimensional ¹³C NMR measurements. The ¹H-¹³C HSQC NMR spectra were collected, with 1024 data points for ¹H and 256 for ¹³C NMR.

For the ³¹P NMR analysis, a freeze-dried lignin sample was dissolved in a pyridine-DMF solvent mixture, after which lignin was phosphitylated with 2-chloro- 4,4,5,5-tetramethyl-1,3,2-dioxaphospholane and a pyridine solution containing pyridine, cyclohexanol, and Cr(acac)3. CDCl3 was used as the deuterated solvent. The

(34)

sample was prepared just before the measurement. Bruker Ultrashield^TM AMX-400 and Jeol 500 MHz spectrometers were used in the measurement of the ³¹P NMR spectra.

4.7 PREPARATION OF MASS SPECTROMETRY SAMPLES

The lignin samples were dissolved in a 25% ammonia solution to prepare 5 mg/mL stock solutions. For the measurements, the sample stock solutions were diluted to 100 μg/mL with a toluene-methanol (1:10, v/v) solvent mixture. All mass spectrometric analyses were performed using a 12-T Bruker Solarix XR FT-ICR mass spectrometer with a positive-ion mode APPI ion source.

(35)

5 REACTION OF SOFTWOOD KRAFT LIGNIN WITH PARAFORMALDEHYDE

^Ⅰ

5.1 LIGNIN STRUCTURE

For the entire study, softwood kraft lignin was used. The structure of initial softwood kraft lignin was examined using NMR methods and mass spectrometry.

The extensive knowledge of the lignin structure was utilized in the interpretation of the reactivity and modification results. The softwood lignin used contains mainly guaiacyl (G) units and minor amounts of syringyl (S) and p-hydroxyphenyl units (H). The structural information about the lignin was obtained using ¹³C NMR (Fig.

8).

Figure 8. Quantitative ¹³C NMR spectrum of softwood kraft lignin. The spectrum is measured with DMSO-d6 as the solvent.

A closer study of the lignin structure was carried out with ¹H-¹³C HSQC NMR (Fig. 9). HSQC NMR has been found to be a highly sensitive method and to provide more information on the lignin structure. The alkyl region of the HSQC spectrum contains the signals of the C-H groups, methoxy groups, and extractives.(Chen et al., 2016; Schutyser et al., 2018b)

(36)

Figure 9. ¹H-¹³C HSQC NMR spectrum of softwood kraft lignin. The spectrum is measured with DMSO-d6 as the solvent (¹H δ = 2.5 ppm, ¹³C δ = 39.5 ppm).

The ³¹P NMR method has become an important and powerful technique for studying lignin. The free OH-groups of lignin were phosphitylated with 2-chloro- 4,4,5,5-tetramethyl-1,3,2-dioxaphospholane, and therefore, the method provided information on the alcohol and acid groups of the lignin (Fig. 10). According to the

31P NMR, the syringyl/guaiacyl molar ratio was 0.24; therefore, the main monolignol was guaiacyl, which has one free reactive site in the phenolic ring.

Figure 10. ³¹P NMR spectra of softwood kraft lignin. Cyclohexanol is used as an internal standard (δ = 145.0 ppm), and chloroform was used as the solvent.

(37)

The CHNOS elemental composition determined for softwood kraft lignin was C=63.4%, H=6.0%, N=0%, S=2.4%, and O=28.2%, where the amount of oxygen was calculated by taking the difference. The sulfur content originated from the kraft process. The purification of lignin removes extractives (fatty and resin acids) and carbohydrates. The determined klason lignin content was 98.3 wt% (TAPPI, 2006)^, and the extractive content was 0.84%.(TAPPI, 1999) A mass spectrometric measurement with Fourier-transform ion cyclotron resonance (FT-ICR) mass spectrometry coupled with an atmospheric-pressure photoionization (APPI) technique was used to determine the chemical composition and the average molar masses of the softwood kraft lignin (Fig. 11). The weight average molar mass determined for the lignin was 374.64 g/mol, and the number average molar mass was 329.41 g/mol with a polydispersity index of 1.14. This result suggests that APPI is not capable of ionizing the heaviest lignin molecules, thus underestimating the obtained Mw and Mn values.

Figure 11. APPI (+) FT-ICR mass spectra of pristine softwood kraft lignin.

5.2 REACTIVITY STUDY

Previously, the reactivity of lignin materials has been assessed using the Man- nich reaction.(Wang et al., 2016) The Mannich reaction is a condensation reaction of lignin with dimethylamine and formaldehyde. After the reaction, the nitrogen content of lignin is determined to obtain the lignin reactivity. A new differential scanning calorimetry -based assessment method was used to examine the lignin reactivity. An alkaline reaction of kraft lignin with formaldehyde was used as the test reaction. The NaOH solution acted as a catalyst and reaction medium for the lignin.

Paraformaldehyde was used as a source of formaldehyde. The aqueous paraformaldehyde solution mainly contains methylene glycol and its dimeric form. Para- formaldehyde is easy to handle, and the composition of the paraformaldehyde solution is simple compared to the commercial formaldehyde solution (Fig. 12). The commercial formaldehyde solution contains more polymeric forms, methanol, and hemiformals.

(38)

Figure 12. ¹³C NMR spectra of the formaldehyde solution (37%), aqueous paraformaldehyde solution (13 wt%), and aged paraformaldehyde solution. The spectra are measured in D2O.

The concentration of the commercial formaldehyde solution is 37 wt%, and that of the aqueous paraformaldehyde solution is 13 wt% (5.4 M). The advantage of the paraformaldehyde solution is the narrow product distribution and the absence of methanol. During one month of storage, no changes in the paraformaldehyde solution composition are observed.

The lignin and formaldehyde reaction was measured with differential scanning calorimetry. The DSC curve presents the heat flux as a function of temperature or time. The integration of the exothermic DSC signal describes the reaction enthalpy.

The reaction temperature and enthalpy values obtained from the DSC measurement are important process parameters required in plywood production. The exotherm is caused by the formation of methylol groups to the phenolic structure of lignin and possible methylene bridges. The influence of the NaOH catalyst could be studied with DSC. The NaOH solution acted as a solvent, but strongly affected the solubility of the lignin. The effect of the NaOH concentration was studied to increase the lignin:NaOH molar ratio from 10:1 to 2.6:1. The higher NaOH concentration shifted the reaction exotherm to lower temperatures (Fig. 13).

(39)

Figure 13. DSC curves of softwood kraft lignin-paraformaldehyde reactions. Influence of the NaOH concentration on the lignin-paraformaldehyde reaction at L:PFA molar ratios of 1:1 (above) and 1:2 (below).

The lignin-formaldehyde reaction was examined to obtain a better understand- ing of the lignin reactivity and reaction sites. The lignin hydroxymethylation reaction was studied more closely with the ³¹P, ¹³C, and HSQC NMR methods. The reaction was carried out in an autoclave, and lignin, PFA, and NaOH were used, with a molar ratio of 2.6:2.6:1. The studied reaction temperatures were 120, 140, and 150

°C. The formaldehyde’s reaction to lignin occurs mainly at the free reactive site, i.e., the C⁵ position of the guaiacyl unit. Figure 14 presents the ³¹P spectra of pristine softwood kraft lignin, HCl-treated lignin, and a sample from the lignin-PFA reaction (150 °C). The spectrum of the reaction product includes signals of unreacted formaldehyde and small signals caused by neutralization with HCl treatment (151.0-149.5 ppm). The paraformaldehyde reaction occured mainly at the free reactive sites of the phenolic ring of guaiacyl. The signal of the guaiacyl unit (140.5- 138.5 ppm) in the ³¹P NMR spectrum almost disappeared, and in the area of the C^5- substituted structures (144.5-140.5 ppm), a new signal appeared. The signal at 143 ppm is connected to a 5-substituted phenolic ring that includes phenolic ortho sub- stituted monolignols.

(40)

Figure 14. A ³¹P NMR spectrum of pristine softwood kraft lignin (black), HCl-treated lignin (blue) and the lignin-paraformaldehyde reaction product (red). The lignin paraformaldehyde reaction is heated to 150 °C. Cyclohexanol is used as an internal standard (IS, δ = 145.0 ppm).

The lignin-paraformaldehyde reaction samples were also studied using the ¹³C and ¹H-¹³C HSQC NMR methods. In Figure 15, the softwood kraft lignin and lignin- paraformaldehyde reaction products are compared. The signal at 60 ppm belonging to the β-O-4 structure disappeared, and a new signal at 58 ppm appeared. The signal at 58 ppm represents a methylol group bound to a guaiacyl structure. The spectra of the reaction products included signals of unreacted PFA and methylene gly- cols at 82 and 84 ppm. The C⁵ signal appeared at 115 ppm, and after the reaction, the signal almost disappeared. The formation of formic acid (HCOOH) was caused by the Cannizzaro reaction, which is a side reaction of paraformaldehyde in an alkaline medium. The formic acid signal appeared at 168 ppm in the ¹³C NMR spectra. The signal was broad, which could indicate an interaction with lignin or formic acid could appear as a sodium salt.

(41)

Figure 15. ¹³C NMR spectrum of softwood kraft lignin (black) and the reaction products of the lignin-PFA reactions obtained at 120 °C (blue), 140 °C (brown), and 150 °C (red). The samples are measured in DMSO (δ=39.50 ppm).

The assignment of the ¹³C NMR signals was confirmed with the help of ¹H-¹³C HSQC measurements. Figure 16 presents the HSQC spectra of pristine kraft lignin and a sample of the lignin-PFA 150 °C reaction product. The signal in ¹³C NMR at (4.5, 58.2) ppm appeared in the HSQC spectrum and another signal appeared at (4.5, 63.2) ppm, which were connected to methylol groups bound to the aromatic structure of guaiacyl in different lignin structures (A, B and C in Fig. 16).(Singh and Prathap, 1997; Zhao et al., 1994) The lignin signals Cγ, Aα, and Cα disappeared after the reaction with paraformaldehyde. The disappearance indicated the occurrence of the hydroxymethylation reaction.

The two signals of unreacted PFA, including methylene glycol and its dimeric form, were observed at δ=80-85 ppm. The signal at (4.7, 88.1) ppm belonged to the dimeric methylene glycol connected to the C5 position of the lignin aromatic structure. The two weak signals at (3.7, 70.8) ppm and (3.6, 71.8) ppm were assigned to methylol groups bound to the side chain of lignin.(Peng et al., 1992) The methoxy signal of lignin in the lignin-PFA reaction did not change. All of the characterization methods indicated the formaldehyde’s reaction to the lignin C5 position.

(42)

Figure 16. ¹H-¹³C HSQC spectra of softwood kraft lignin (black) and the product of the lignin- PFA 150 °C reaction (red). The samples are measured in DMSO (¹³C δ=39.5 ppm, ¹H δ=2.5 ppm).

(43)

6 ALKALINE OXIDATIVE DEPOLYMERIZA- TION OF SOFTWOOD KRAFT LIGNIN

^Ⅱ

Lignin has a large and complicated molecular structure, and therefore, the utilization of lignin can be difficult without modification. The base-catalyzed oxidation was examined to achieve oligomeric aromatic products and to increase the reactivity of the softwood lignin material. Several methods have been presented to produce small aromatic compounds, but in our study, the process parameters were adjusted to yield oligomeric products with the use of a simple method.

The alkaline depolymerization reactions of lignin were carried out with three different bases, two strong bases (NaOH, KOH) and a weak base (NH4OH). Reac- tion times of 2 and 6 hours were used in the study. One indicator of lignin demeth- ylation during oxidation was the formation of methanol because of the methoxy group cleavage.(Garrigues et al., 1997; Kalliola et al., 2011) The reaction products evaporated easily and were present in the gas phase of the autoclave. The gas phase IR spectrum measured from the gas phase of a NaOH-catalyzed 2 h depolymerization reaction is presented in Figure 17.

Figure 17. Gas phase IR spectrum of a NaOH 2 h oxidation reaction carried out in the autoclave. The gaseous sample is taken from the reaction after cooling.

To understand the depolymerization mechanism and extent, characterization of the reaction products is needed. The APPI FT-ICR mass spectrometry measurements of the pristine lignin and oxidized lignin samples provided information on the average masses, mass distribution, and identity some of the products. Figure 18 presents the mass spectrum of softwood kraft lignin and those of the NaOH and KOH 2 h oxidized samples. After oxidation, the components with high m/z values, i.e., over 500, disappear. Normally, the molar mass range from m/z 100 to 250 represents monomeric lignin units, while the m/z range of 250-400 represents dimeric

(44)

units, and those > m/z 500 represent higher-order lignin molecules.(Banoub et al., 2007; Echavarri-Bravo et al., 2019) The determined weight average molar mass of pristine lignin is 374.64 g/mol, and the number average molar mass value is 329.41 g/mol. After oxidation, the molar mass values of the samples decreased. Overall, the main signals of the 6 h reaction samples are below m/z 350. CHNOS elemental analysis was used to show the changes in the elemental composition of the samples after the depolymerization reaction. The oxygen content almost doubled, when NaOH and KOH were used as the catalysts, and the O/C ratio increased from 0.33 to 1.04. Pristine lignin did not contain nitrogen, but after the NH⁴OH-catalyzed oxidation, nitrogen was detected. The main oxygen classes in pristine lignin are O⁴ and O5, which originate from diphenolic structures. After oxidation, the most abundant oxygen classes were O² and O³, indicating an increase in the number of monomeric units. The absence of O⁷ and higher order oxygen classes after the oxidation reaction was attributed to the depolymerization of lignin.(Miettinen et al., 2015)

Figure 18 also presents the mass spectrometry results in the format of van Krevelen diagrams. The signals of softwood kraft lignin were in the range of H/C = 0.8-1.4 and O/C = 0.1-0.5. The diagram also showed that the pristine lignin contains some fatty acids. After the oxidation reaction, the lignin signals spread, indicating depolymerization of the lignin structure.

Figure 18. Mass spectra and van Krevelen diagrams of pristine softwood kraft lignin and two oxidized samples (samples from the NaOH and KOH 2h reactions).

(45)

The ³¹P NMR method was employed for the structural characterization of the neutralized and freeze-dried oxidation products. Pristine softwood kraft lignin contained mostly guaiacyl units and traces of syringyl and p-hydroxyphenyl units (Fig. 19). After the oxidation reaction, the H-units seemed to be unchanged, as do the G-units. The oxidation reactions affected the S-units, probably because of the aromatic ring oxidative cleavage. Notable changes occured in the area of the aliphatic hydroxyl groups (145-149 ppm). The signals of the aliphatic hydroxyl groups almost disappeared after 6 h of oxidation. This indicated the cleavage of aliphatic hydroxyl groups from the lignin structure during the oxidative depolymerization reaction. The formed small acids were not detected in the ³¹P NMR spectra due to the neutralization of the samples before the analysis.(de Menezes et al., 2017;

Demesa et al., 2015; Hasegawa et al., 2011; Ma et al., 2014; Rovio et al., 2011; Yang et al., 2020)

The quantitative ¹³C NMR spectra of the oxidized samples (Fig. 19) were measured directly from the reaction mixture to preserve small volatile compounds. The

13C NMR spectra of oxidized samples showed several narrow signals in the range of 165-190 ppm. The narrow signals were assigned to free acids that have no interaction with lignin. The signals in the range of 166-185 ppm belonged to esters. Acetic, formic, and oxalic acids were identified on the basis of pure acid samples. The acids originated from the syringyl and guaiacyl oxidative degradation.

Figure 19. ³¹P NMR (A) and ¹³C NMR (B) spectra of the softwood kraft lignin and the oxidized lignin samples. In the ³¹P NMR samples, CDCl3 is used as the solvent, and cyclohexanol is used as an internal standard (IS, δ = 145.0 ppm). In the ¹³C NMR samples, DMSO is used as the solvent (δ = 39.5 ppm).