Characterisation of Kraft Pulps by Size-exclusion ...8571/FULLTEXT01.pdfV Characterisation of dissolved kraft lignin by capillary zone electrophoresis Elisabeth Sjöholm, Nils-Olof

Elisabeth Sjöholm ”Characterisation of Kraft Pulps by Size-Exclusion Chromatography and Kraft LigninSamples by Capillary Zone Electrophoresis”, Doctoral Thesis, 1999

Characterisation of

Kraft Pulps by Size-exclusion Chromatography

and

Kraft Lignin Samples by Capillary Zone Electrophoresis

Elisabeth Sjöholm

Doctoral Thesis

Royal Institute of TechnologyDepartment of Pulp and Paper Chemistry and Technology

Division of Wood Chemistry

Stockholm 1999


Characterisation of

Kraft Pulps by Size-exclusion Chromatography

and

Kraft Lignin Samples by Capillary Zone Electrophoresis

Elisabeth Sjöholm

1999

Doctoral Thesis

Supervisor, Prof. Anders Colmsjö, University of Stockholm

Royal Institute of TechnologyDepartment of Pulp and Paper Chemistry

and Technology

Swedish Pulp and Paper Research Institute


Akademisk avhandling som med tillstånd av Kungliga Tekniska Högskolan i Stockholmframlägges till offentlig granskning för avläggande av teknologie doktorsexamen fredagenden 1 oktober 1999 kl 10.15 i STFI-salen, STFI, Drottning Kristinas väg 61, Stockholm.Avhandlingen försvaras på engelska.

Elisabeth SjöholmStockholm 1999-09-01KTH Högskoletryckeriet


To Matz,

Jonas

and Anna


Coming together is a beginning

Keeping together is a progress

Working together is a success!

Stellan Hjertén


Abstract

In the present thesis two analytical methods for characterisation of underivatised pulp andlignin samples obtained from kraft pulping of hardwood and softwood are evaluated. The firstmethod is the use of lithium chloride/N,N-dimethylacetamide (LiCl/DMAc) for dissolutionand size-exclusion chromatography (SEC) of pulps. The applicability of LiCl/DMAc-SEC isdemonstrated for birch wood kraft pulps with different relations between zero-span tensilestrength and viscosity. The second method concerns the applicability of capillary zoneelectrophoresis (CZE) for characterising black liquor and isolated lignin samples with respectto mobility, i.e charge density. A method for determining the average mobility (µav) of themobility distributions to ease comparison between samples is also presented.The solubility in LiCl/DMAc and the elution behaviour in SEC differ between hardwood andsoftwood kraft pulps. Hardwood kraft pulps are completely dissolved in LiCl/DMAc, whereashigh amounts of lignin and presence of glucomannan restrict the solubility of softwood kraftpulps. The undissolved fraction of softwood kraft pulps consists of larger amounts ofmannose and lignin but has a diminished xylose content compared with the initial pulp. Xylanand cellulose of hardwood kraft pulps are fairly well separated by LiCl/DMAc-SEC. Incontrast, the molecular weight distributions (MWD) of softwood kraft pulps are morecomplex. It was found that the hemicelluloses of softwood pulps elute over the entiremolecular weight range, indicating various degrees of association with cellulose. The neutralmonosaccharide composition implies that associations between galactoglucomannan andcellulose increase with decreasing amount of galactose. The elution behaviour of softwoodkraft pulp xylan suggests that this xylan consist of substructures with varying propensity forassociating with cellulose and/or mannan. In absence of associations between cellulose andhemicellulose, cellulosic solutions of LiCl/DMAc consist of cellulose aggregates, which isseen as a shoulder on the high molecular weight end of the MWD of cellulose. According tothe profiles of the MWD and light scattering measurements, it is possible to break theseaggregates by mechanical treatment of the solutions, without causing severe cleavage of theglycosidic bonds. The relation between MWD and zero-span tensile strength was studied onhardwood kraft pulps degraded by gamma irradiation, oxygen/alkali or alkali. For alltreatments, the MWD of cellulose is shifted to a lower molecular weight range as degradationproceeds. In the chemically treated pulps, a shoulder on the low molecular-weight end of thecellulose distribution gradually develops, which is not seen for the gamma treated pulps. Theobserved decrease in shape factor/fibre strength of the chemically treated-pulps is proposed tobe due to a combination of heterogeneous degradation and removal of hemicellulose whereasthe decrease in Mw of cellulose is of minor importance.The mobility distributions obtained by CZE differ between black liquor, isolated dissolvedlignin and isolated residual lignin. The µav measured at pH 12 reveals that the residual ligninisolated from pine wood kraft pulp has a significantly lower charge density than the ligninremoved from the pulp throughout the cook. At the end of the kraft cook of birch, the µav ofthe isolated residual lignin is about the same as that of the isolated dissolved lignin, whichsuggests that the solubility is sufficient for the pulp lignin to be dissolved. Comparisonsbetween the µav at pH 12 and pH 10 indicate that the isolated dissolved lignin samplesobtained in the middle of the cook have a lower acidity than the other samples. The observeddifference in µav between black liquor and isolated dissolved lignin may be due toassociations between lignin fragments and carbohydrate polymers in the black liquor.

Keywords: Kraft pulps, Black liquor, Cellulose, Hemicellulose, Lignin, LiCl/DMAc,Dimethylacetamide, Size-exclusion chromatography, Capillary zone electrophoresis


List of publications

This thesis is based on the following papers, referred to by their Roman numerals in the text.

I Characterisation of the cellulosic residues from lithium chloride/N,N-dimethylacetamide dissolution of softwood kraft pulpElisabeth Sjöholm, Kristina Gustafsson, Bert Pettersson and Anders ColmsjöCarbohydr. Polym., 32(1), 1997, 57-63

II Influence of the carbohydrate composition on the molecular weight distribution ofkraft pulpsElisabeth Sjöholm, Kristina Gustafsson, Fredrik Berthold and Anders ColmsjöAccepted for publication in Carbohydr. Polym.

III Aggregation of cellulose in lithium chloride/N,N-dimethylacetamideElisabeth Sjöholm, Kristina Gustafsson, Bengt Eriksson, Wyn Brown and AndersColmsjöAccepted for publication in Carbohydr. Polym.

IV Fibre strength in relation to molecular weight distribution of hardwood kraft pulp.Part 1. Degradation by gamma irradiation, oxygen/alkali or alkaliElisabeth Sjöholm, Kristina Gustafsson, Erik Norman, Torbjörn Reitberger and AndersColmsjöSubmitted to Holzforschung

V Characterisation of dissolved kraft lignin by capillary zone electrophoresisElisabeth Sjöholm, Nils-Olof Nilvebrant and Anders ColmsjöJ. Wood Chem. Technol., 13(4), 1993, 529-544

VI A method for evaluating lignin mobility distributions obtained by capillary zoneelectrophoresisElisabeth Sjöholm, Erik Norman, Störker Moe and Anders ColmsjöSubmitted to J. Wood Chem. Technol.

VII Charge density of lignin samples from kraft cooking of birchElisabeth Sjöholm, Erik Norman and Anders ColmsjöSubmitted to J. Wood Chem. Technol.


CONTENTS

1 INTRODUCTION................................................................................................................... 1

2 BACKGROUND..................................................................................................................... 3

2.1 Wood fibre structure............................................................................................................. 3

2.2 Structure of wood polymers ................................................................................................. 42.2.1 Cellulose ........................................................................................................................... 42.2.2 Hemicellulose................................................................................................................... 52.2.3 Lignin................................................................................................................................ 6

2.3 Chemical reactions of wood polymers during kraft pulping................................................ 82.3.1 Delignification.................................................................................................................. 82.3.2 The main reactions of the carbohydrate polymers ...................................................... 9

2.4 Size-exclusion chromatography (SEC)............................................................................... 102.4.1 Principles........................................................................................................................ 112.4.2 Solvent-mobile phase selection..................................................................................... 122.4.3 Lithium chloride/N,N-dimethylacetamide (LiCl/DMAc) .......................................... 13

2.5 Capillary electrophoresis (CE) .......................................................................................... 132.5.1 Instrumentation............................................................................................................. 142.5.2 Principles of CZE .......................................................................................................... 15

3 METHODS............................................................................................................................ 17

3.1 Pulp characterisations........................................................................................................ 17

3.2 Determination of fibre strength and fibre deformation ..................................................... 18

3.3 Dissolution of pulp samples in LiCl/DMAc........................................................................ 18

3.4 Static light scattering ......................................................................................................... 18

3.5 Black liquor and isolation of lignin samples...................................................................... 19

3.6 Chemical characterisation of isolated lignin samples ....................................................... 20

3.7 SEC of pulps....................................................................................................................... 20

3.8 SEC of lignin samples......................................................................................................... 21

3.9 CZE .................................................................................................................................... 21


4 DISCUSSION ....................................................................................................................... 23

4.1 Characterisation of kraft pulps dissolved in LiCl/DMAc (papers I-IV) ............................ 234.1.1 Differences between softwood and hardwood kraft pulps (papers I and II) ........... 234.1.2 Aggregation of cellulosic samples dissolved in LiCl/DMAc (paper III) ................... 294.1.3 Molecular weight distribution of birch wood kraft pulp samples with different

relations between zero-span tensile strength and pulp viscosity (paper IV) ........... 33

4.2 Characterisation of kraft lignin samples by CZE (papers V-VII) ..................................... 394.2.1 Electrophoretic characterisations of lignin-derived compounds .............................. 394.2.2 Charged groups and molecular weight........................................................................ 404.2.3 Charge density distributions ........................................................................................ 424.2.4 Average charge density ................................................................................................. 464.2.5 Influence of pH on mobility.......................................................................................... 49

5 CONCLUSIONS................................................................................................................... 53

6 ACKNOWLEDGEMENTS .................................................................................................. 55

7 REFERENCES...................................................................................................................... 57

APPENDIX I: List of abbreviations ....................................................................................... 69APPENDIX II: Pulp samples used in papers I-IV................................................................. 70APPENDIX III:Data for pulps corresponding to the cooks from which lignin samples in

papers V and VI were obtained .................................................................... 71APPENDIX IV:Data for pulps corresponding to the cooks from which lignin samples in

paper VII were obtained................................................................................ 72


1

1 INTRODUCTION

Kraft pulping is the most common process for production of pulps for paper production. In

Sweden about 6600 kton of kraft pulps are produced annually. The purpose of the process is

to liberate the individual fibres and remove lignin with sulphide and hydroxide ions as the

active delignifying species. The first patents on the use of sulphides in wood pulping were

obtained by Eaton in 1870 and 1871. The kraft process was developed by Dahl in 1884 as a

modification of the soda process. In this patent, sodium carbonate was replaced in the

recovery cycle by the cheaper sodium sulphate, which is reduced to sodium sulphide during

combustion of black liquor. Efforts to improve the kraft pulping process are continuously

carried out so as to produce pulp with pre-determined properties in high yields. Much

research has been focused on e.g. the incomplete delignification, the degradation and loss of

carbohydrates and what implication these factors have on the final pulp properties. In order to

study the process, it is important to characterise pulp and the organic components of black

liquor at various stages of the cook. In addition, the final pulp needs to be characterised to i)

control the process and achieve a uniform quality ii) evaluate the influence of process changes

iii) obtain information describing the papermaking potential. In this context, it is necessary to

improve known methods for chemical characterisation and apply new analytical techniques.

Chemical structure and molecular weight are important parameters in relation to the physical

properties of polymeric materials. Bulk properties, such as viscosity of pulp solutions and

kappa number of pulps, are frequently used to describe the extent of carbohydrate degradation

and delignification, respectively. To obtain a more detailed information about the pulp the

constituent polymers need to be separated. The main obstacle for chromatographic separation

of the pulp is its limited solubility in common solvents and that the solvent used has to be

compatible with the chromatographic matrix. Proper chromatographic separation of the lignin

fragments in black liquor is also difficult to achieve, due to its complexity, high alkalinity and

tendency to adsorb irreversibly on common column matrices.

The aim of this study was to i) evaluate the use of lithium chloride/N,N-dimethylacetamide

(LiCl/DMAc) to dissolve and characterise underivatised kraft pulps by size-exclusion

chromatography (SEC) and ii) to examine the applicability of capillary zone electrophoresis

to characterise underivatised lignin from the kraft process.

2



3

2 BACKGROUND

2.1 Wood fibre structureThe main morphological constituent in wood is the supporting tissue (Sjöström, 1993a). In

softwoods, tracheids are responsible for support and water transport and ray parenchyma cells

store nutrition. The hardwoods have more differentiated cells. Libriform cells and fibre

tracheids are the supporting tissue, vessels conduct water and the storage tissue consists of

parenchyma cells. Nutrition and water are exchanged through pores in the living cell wall.

The amount and composition of the main molecular components, cellulose, hemicellulose and

lignin, vary between cell types. These constituents are also distributed differently in the cell

wall layers, Figure 1, around the central cavity called the lumen. Hemicellulose and lignin

form a matrix around the microfibrils of cellulose, which are arranged in an irregular way in

the primary wall (P), but have distinct directions in each secondary wall layer. The main part

of the lignin is located in the thickest secondary wall layer (S2) and the remaining part in the

compound middle lamella (ML+P).

Figure 1. Schematic view showing the main layers of a woody cell surrounding lumen (L);middle lamella (ML), primary wall (P), outer (S1), middle (S2) and inner (S3) secondary wall.Adopted from Kleen (1993).

As reported by Meier (1961), the composition of carbohydrate polymers differs between the

layers in pine and birch fibres. In pine (Pinus sylvestris) the S1 and the outer parts of S2 are

richest in cellulose whereas the relative amount of glucomannan increases toward the lumen.

The xylan is more irregularly distributed but the concentration is highest in the S3 layer. In

birch (Betula verrucosa), the cellulose content is highest in the inner part of S2 and in S3

whereas the outer parts of the cell wall (S1 and outer parts of S2) have a higher relative

4


amount of glucuronoxylan. However, due to experimental difficulties, the distribution of

polysaccharides through the cell wall cannot be considered settled (Meier, 1985).

2.2 Structure of wood polymers

The relative composition of polymers in the fibre depends on wood species, as illustrated in

Table 1.

Table 1. Relative composition of wood polymers (% of dry wood weight) (Sjöström, 1993b).

Species Cellulose Glucomannana Glucuronoxylanb Lignin Total

Pinus sylvestris (pine) 40.0 16.0 8.9 27.7 91.9

Betula verrucosa (birch) 41.0 2.3 27.5 22.0 92.8a Including galactose and acetyl in pine. b Including arabinose and acetyl in birch.

All of the polymers consist of a mixture of molecules of varying chain length and are thus

polydisperse. The common features of the carbohydrate polymers are the glycosidic bonds

between the monomers and the presence of a non-reducing and a reducing end group.

2.2.1 Cellulose

Cellulose is a linear polymer composed of β-D-(1→4)-glucopyranose units, Figure 2

(Sjöström, 1993c). The degree of polymerisation (DP) of wood cellulose is around 10,000.

The hydroxyl groups of cellulose form intramolecular and intermolecular hydrogen bonds.

The cellulose chains are aggregated in bundles to elementary fibrils. They are assembled to

microfibrils, which are the smallest morphological structure of the fibre. The existence of the

elementary fibril is however debated. There is a general agreement that cellulose contains

both ordered and less ordered regions although the exact arrangement in the microfibrils is

still under debate. The ordered cellulose may exist in more than one crystalline form i.e.

polymorphs. The dominant polymorph in higher plants such as cotton and wood is cellulose Iβ

(Atalla and VanderHart, 1984; VanderHart and Atalla, 1984) which is transformed to

cellulose II during e.g. swelling in strong alkali (mercerisation). Cellulose II can also be

synthesised by a mutant of Acetobacter xylinum (Kuga et al., 1993). The cellulose chains are

parallel in cellulose I and antiparallel in cellulose II. The cellulose II is more densely packed

and more strongly interbonded than the cellulose I (Krässig, 1996).


5

OOH

OH

OH

CH2OH

OOH

OH

CH2OH

O

CH2OH

OH

OH

O

O

OH

Non-reducing end Reducing endn-2

Figure 2. Principal structure of cellulose.

2.2.2 Hemicellulose

Because of their complex structure, most hemicelluloses are thought to be unordered in wood.

Isolated softwood hemicelluloses have a DP of around 100 and hardwood hemicellulose a DP

of around 200 (Sjöström, 1993c). Besides this difference in DP, the monosaccharide

composition in the hemicelluloses differs. In softwoods, the major hemicellulose is

galactoglucomannan, Figure 3.

OOH

OH

OH

CH2OH

OOH

OH

CH2OH

OOR

CH2OH

OR

OOR

CH2

ORO

OR

CH2OH

OROOO

O

O

O

R

4-β-D-Glcp-1 4-β-D-Manp-1 4-β-D-Manp-1 4-β-D-Manp-1

α-D-Galp-1

6 2,3

Figure 3. Principal structure of galactoglucomannan in softwood, showing the proportions ofthe galactose-rich fraction. R= CH3CO or H.

The partially acetylated β-D-glucomannan is further divided in two sub-types depending on

amount of α-D-galactose substituents. For the low-galactose type, often referred to as

glucomannan, the ratio between galactose:glucose:mannose is around 0.1:1:4 and for the high

galactose type the ratio is around 1:1:3. The β-D-(1→4)-xylopyranose in softwood is on

average substituted with two 4-O-methyl-α-D-glucuronic acid groups at C-2 and one α-L-

arabinose group at C-3 per ten xylose units. In hardwoods, O-acetyl-4-O-methylglucurono-β-

D-xylan is the principal hemicellulose component, Figure 4. To the partially acetylated xylan

6


backbone one 4-O-methyl-α-D-glucuronic group is (1→2) linked per ten xylose units. The

reducing end of birch wood xylan also contains α-D-galactouronic and α-L-rhamnose units.

The structure can be written: -β-D-Xylp-(1→4)-β-D-Xylp-(1→3)-α-L-Rhap-(1→2)- α-D-

GalpA-(1→4)-D-Xyl (Johansson and Samuelson, 1977a). The small amount of β-D-(1→4)

glucomannan present in hardwoods has a ratio between glucose:mannose of 1:1-1:2.

OOH

O

OOR

ORO

OH

OH

OOH

OH

COOH

H3CO

OO

OOH

OH

OO

O

R

7

4-β-D-Xylp-1 4-β-D-Xylp-14-β-D-Xylp-1 4-β-D-Xylp-12

4-O-Me-α-D-GlclpA-1

72,3

Figure 4. Principal structure of hardwood xylan. R= CH3CO or H.

2.2.3 Lignin

Lignin has a much more complex structure than the carbohydrate polymers. It has a branched

structure and differs between softwood and hardwood as well as morphological origin

(Sjöström, 1993d). The precise structure is not completely known, but some fundamental

features concerning monomer composition and bonds are known. Lignin is composed of three

basic types of monomers referred to as p-hydroxyphenylpropane (p-hydroxyphenyl), 3-

methoxy-4-hydroxyphenylpropane (guaiacyl) and 3,4-dimethoxy-4-hydroxyphenylpropane

(syringyl), Figure 5. The propane chain of the monomers, is substituted by hydroxyl and

carbonyl groups in various amounts.

Softwood lignin is mainly composed of guaiacyl units but contains small amounts of p-

hydroxyphenyl units as well. Hardwood lignin is a mixture of guaiacyl and syringyl units in

various proportions. In birch wood, the overall proportion of syringyl:guaiacyl units is

roughly equal (Adler, 1977). The ratio of syringyl:guaiacyl lignin is higher in the secondary

wall (S) than in the middle lamella (ML) region of fibre and than in vessels (Fergus and

Goring, 1970; Hardell et al., 1980a; Saka and Goring, 1988). In addition, the lignin in the

ML of birch has been reported to be highly condensed (Meshitsuka and Nakano, 1985).


7

Spruce lignin has a higher content of phenol groups in the S than in the ML region (Yang and

Goring, 1980; Hardell et al., 1980b; Whiting and Goring, 1982).

CH2OH

HCOH

HCOH

OH

OCH3H3CO

CH2OH

HCOH

HCOH

OH

CH2OH

HCOH

HCOH

OH

OCH3

p-hydroxyphenyl guaiacyl syringyl

2

34

5

61

α

β

γ

Figure 5. Fundamental type-structures of the monomers in lignin. Designation of the carbonatoms are illustrated beside the p-hydroxyphenyl monomer.

The phenol groups in lignins are either free or etherified. Softwood contains larger

concentrations of free phenol groups than hardwood (Adler, 1977). In contrast to hardwood,

softwood also contains carbonyl groups (Adler, 1977; Chen, 1992). Carboxyl groups have

been detected in milled wood lignin (Ekman and Lindberg, 1960), but if present in native

lignin the amount is small.

C

C

C

O

CO

C

C

C

O

CO

C

C

C

O

C

O

C

O

C

C

C

O

C

C

C

O

C

C

C

O

C

C

C

O

O

C

C

C

O

C

C

C

O

ββ-O-4 αα-O-4 ββ-5

2-8% (7%)50% (60%) 9-12% (6%)

ββ-1 ββ-ββ 5-54-O-5

7% (7%) 2% (3%) 4% (7%) 10% (5%)

Figure 6. Common bonds in softwood and hardwood lignin. The figures refer to the relativefrequency in Picea abies and in Betula verrucosa (within parenthesis) (Sjöström, 1993d).

8


Another feature that contributes to the structural heterogeneity of lignin is the different types

and relative frequency of bonds. Some of the bonding patterns are shown in Figure 6 together

with their relative abundance in softwood and hardwood. The β-O-4-arylether bond is the

dominant type of bond. The branched structure of lignin is due to the presence of more than

two inter-connecting linkages in some of the monomers.

2.3 Chemical reactions of wood polymers during kraft pulping

The main purpose of chemical pulping is to remove the lignin thus releasing the fibres.

Hydroxide and sulphide ions are the active chemicals in the white liquor, which is added to

the wood chips. The kraft cook is carried out at 160-170°C. About 90% of the initial amount

of lignin is removed from the wood but some of the carbohydrate polymers are degraded

and/or removed as well.

2.3.1 Delignification

The kraft cook can be divided into three phases; initial, bulk and residual delignification

phase, according to different delignification rates (Wilder and Daleski, 1965; Kleinert,

1966; Lémon and Teder, 1973). It has been shown that the secondary wall (S) of fibres

delignifies faster than the middle lamella (ML) region (Whiting and Goring, 1981; Fergus

and Goring, 1969) and vessels (Fergus and Goring, 1969). The rate of delignification is

attributed to the chemical structure of the lignin in the various regions (see section 2.2.3)

rather than the accessibility.

From studies on model compounds with guaiacyl units, possible chemical reactions have been

correlated to the different delignification-phases (Gierer, 1980). The rapid initial

delignification phase is diffusion controlled (Olm and Tistad, 1979) and includes cleavage of

α- and β-aryl ether bonds in phenolic lignin units, an example of sulphidolytic cleavage is

shown in Figure 7. During the bulk delignification phase when the main part of the lignin is

removed, non-phenolic α- and β-aryl ether bonds are cleaved as well, Figure 8 (Ljunggren,

1980). The lower delignification rate in the residual phase is ascribed to cleavage of carbon-

carbon bonds.

Condensation reactions may also take place between lignin fragments and between lignin and

carbohydrate moieties during the cook. These reactions retard lignin dissolution and become

more important at the end of the cook, due to the lower concentration of hydroxide ions which

otherwise suppress formation of condensed structures (Gierer, 1980).


9

H2COH

HC

HC

O

R2O

OCH3

O

OCH3

OR1

H3CO

H2COH

HC

HC

O

R2O

OCH3

H3CO

H2COH

HC

CH

O

O R2

OCH3

H3CO

H2COH

CH

CH

O

H3CO

H2COH

CH

HC

O

H3CO

SS

H

+ HS

-

R2-

- OR1

S0-

Figure 7. Cleavage of phenolic arylether bonds in lignin during kraft pulping. R1= H oranother monomer, R2= lignin molecule. The first step shows alkaline cleavage of α-arylether.Sulphidolytic β-arylether cleavage is shown on the bottom line (Gierer, 1980).

H2COH

HCOH

CHO

H3CO

H2CO

HC

CH

O

O R

OCH3

H3CO

O

O

H2COH

CH

CHO

H3CO

O

O R

OCH3

-OH+

Figure 8. Alkaline cleavage of non-phenolic β-aryl-ether bonds in lignin during kraft pulping.R = rest of lignin molecule (Gierer, 1980).

2.3.2 The main reactions of the carbohydrate polymers

The amount and composition of the carbohydrate polymers are changed continuously during

pulping. The extent of these transformations affects the yield and the properties of the final

pulp. The main carbohydrate reactions include deacetylation, end-wise degradation (peeling)

10


and alkaline hydrolysis (Sjöström, 1993e). The hemicelluloses are deacetylated and the

amount of uronic acid groups is reduced during the initial delignification phase. The yield loss

during the initial delignification phase is ascribed to the peeling reaction and direct

dissolution of hemicelluloses. The peeling reaction starts at about 100 °C (Gierer, 1980) and

proceeds at the reducing end until the competing stopping reaction has formed a stable acid

end group, examples shown in Figure 9 (Johansson and Samuelson, 1975).

COOOH

RO

OH

R'

OH

RO COO

CH3

a) b)

R' = CH 2OH o r H

R = carbohydrate chain

Figure 9. The main acid end groups in polysaccharides formed by the stopping reaction,metasaccharinic acid (a) and 2-C-methylglyceric acid (b).

Glucomannan is sensitive to peeling, whereas xylan is primarily dissolved into the black

liquor without being extensively degraded. Some of the dissolved xylan is later readsorbed

onto the fibre surface. The decreased solubility is due to deacetylation and decrease in uronic

acid groups in the xylan and the decrease in alkalinity of the cooking solution (Yllner and

Enström, 1956; Hansson and Hartler, 1969).

The rate of the alkaline hydrolysis becomes pronounced at temperatures above 150 °C, i.e.

during the bulk delignification phase. This degradation is most severe to cellulose because of

its high DP. Alkaline hydrolysis also creates new reducing end-groups which are subjected to

peeling reactions, which further decreases the yield.

OOH

R1O

O

CH2OH

CHOHO

OH

OH

CH2OH

OR2 OOH

R1OO

CH2OH

R1OOHOR2-

Figure 10. Alkaline hydrolysis of glycosidic linkages in carbohydrate polymers. R1 and R2 isthe rest of the carbohydrate molecule.

2.4 Size-exclusion chromatography (SEC)

The molecular weight (M) is related to a polymer's physical properties, with optimum M

depending on chemical structure and application (Stevens, 1990). SEC is the most common

technique for characterisation of polymers with respect to molecular weight distribution


11

(MWD). Originally soft gels with aqueous-based eluants were used for the analysis of water-

soluble polymers (Porath and Flodin, 1959) and the technique was called gel permeation

chromatography (GPC) or gel filtration chromatography (GFC). The introduction of synthetic

resins (Moore, 1964) made it possible to use organic eluants. Today, GPC is used to denote

the use of a soft or semirigid organic stationary phase, whereas SEC denotes the use of either

an organic gel or a rigid inorganic gel matrix (Malawer, 1995). In spite of this, the term GPC

is used interchangeably with SEC.

2.4.1 Principles

In size-exclusion chromatography (SEC) polymers are separated according to hydrodynamic

size using columns packed with porous particles, commonly polystyrene-divinylbenzene.

Sample molecules with a large size will elute earlier as compared to smaller molecules, which

are able to diffuse into the pores to a greater extent. All sample molecules above a critical size

are excluded from the pores and will co-elute in the same volume, Vo. Below a critical

molecular size, depending on the total permeation volume, Vp, it is not possible to separate

molecules of different size. The selective permeation volume, Vs, which is available for

separation is defined as

Vs = Vp - Vo [1]

The molecular weight of the sample decreases logarithmically with increasing elution volume.

When mass sensitive detectors such as differential refractive index (RI) and ultraviolet

absorbance (UV) detectors are used, the relationship between elution volume and molecular

weight is commonly established by using monodisperse standards of known molecular

weight. Average molecular weights are commonly reported for polymers with broad MWD.

The number-average molecular weight (Mn) and weight-average molecular weight (Mw) are

defined by

Mn = ∑ niMi/∑ ni [2]

Mw = ∑ niMi2/∑ niMi [3]

where ni is the number of molecules and Mi the molecular weight at the ith retention volume.

The SEC Mn and Mw correspond to the values obtained by osmometry and light scattering

photometry methods, respectively.

Pullulan is frequently used as a standard to obtain the MWD of cellulosic samples. Because of

its linear structure pullulan has about the same relation between molecular weight and

12


hydrodynamic volume, as cellulose (Striegel and Timpa, 1995). Polystyrene standards are

commonly used to evaluate lignin samples, in spite of the obvious difference in structure

between the two. Thus, the reported molecular weights obtained from SEC in this thesis are

relative to the molecular weight of the used standards having the same hydrodynamic volume,

i.e. elution volume as the sample.

2.4.2 Solvent-mobile phase selection

Under non-ideal SEC conditions, interactions can occur which will contribute to the observed

retention time (Barth, 1987). To obtain a separation solely based on differences in molecular

size, the conditions should be chosen so that associations between sample molecules and

sample-stationary phase interactions are eliminated. The same solvent is commonly used to

dissolve the sample and as the mobile phase. A suitable solvent should i) completely dissolve

the sample without degradation ii) prevent the sample polymers from adsorbing onto the

stationary phase iii) be compatible with the stationary phase and iv) be stable. Common

organic solvents meet the two latter requirements but, like cellulose, wood pulp is insoluble in

common solvents unless derivatised. The major concerns in using derivatives in SEC are

incomplete derivatisation or uneven distribution of the substituents along the chain.

Tricarbanilated cellulose samples are easily dissolved in tetrahydrofuran (THF) (Evans and

Wallis, 1987) and tricarbanilate is the most common derivative used for SEC-

characterisations of cellulose. There are, however, several disadvantages with the procedure.

Unbleached pulp samples have to be delignified prior to derivatisation (Schroeder and

Haigh, 1979), which excludes information about the lignin in pulp samples. The two solvents

used for the carbanilation reaction are dimethylsulfoxide (DMSO) or pyridine. It has been

reported that DMSO degrades high molecular weight samples (Evans et al., 1989) and that

incomplete derivatisation in pyridine leads to aggregation of cellulose (LaPierre and

Bouchard, 1997). Low molecular weight cellulose may be lost when methanol precipitation

is applied (Wood et al., 1986). Even if the precipitation has been optimised for cellulose

(Evans et al., 1989) a certain loss of hemicellulose cannot be excluded.

Due to limits of the DP range over which cellulose is soluble, degradation of cellulose or

adverse effects on the column packing material, only a few solvent systems have been used

for SEC characterisations of underivatised cellulose and pulp samples. One is cadoxen [tris

(ethylenediamine) cadmium dihydroxide] (Eriksson et al., 1966; Bao et al., 1980; Schwald

and Bobleter, 1988). Cadoxen is a clear colourless solvent and forms stable solutions of

cellulose (Brown, 1967). Some disadvantages in using cadoxen are that hardwood pulps have


13

a limited solubility (Sjöström and Enström, 1966), it includes a toxic compound (Cd) and is

time-consuming to prepare.

2.4.3 Lithium chloride/N,N-dimethylacetamide (LiCl/DMAc)

One promising solvent for characterisation of cellulose and wood pulp by SEC is lithium

chloride/N,N-dimethylacetamide (LiCl/DMAc). Separation has been achieved by using

crosslinked polystyrene-divinylbenzene particles as stationary phase. The preparation of

sample solutions suitable for SEC is relatively simple but requires activation either in water

followed by solvent exchange (Ekmanis, 1986) or by refluxing DMAc (Timpa, 1991).

Originally, the solvent was used to dissolve chitin (Austin, 1977) and for the preparation of

cellulose derivatives (McCormick and Lichatowich, 1979; McCormick, 1981; Kvernheim

and Lystad, 1989). This solvent is considered to be non-degrading (Turbak et al., 1981;

Turbak, 1983) although a slight decrease in the viscosity of cellulose solutions after 30 days

has been reported (McCormick et al., 1985). LiCl/DMAc is a non-derivatising solvent for

cellulose (Nehls et al., 1995). The solvent is generally described as a complex between

lithium and DMAc with a free chloride anion, [Li•xDMAc]+Cl-, with x≈4 (Dawsey and

McCormick, 1990). The dissolution of cellulose is thought to proceed by exchange of one

molecule of DMAc by one hydroxyl group of cellulose (Morgenstern et al., 1992). Due to its

basicity, the chloride weakens the hydrogen bonds of cellulose and thereby participates in the

dissolution. A number of structural models of the solvent and for the solvent-cellulose

complex have been proposed and reviewed recently (Morgenstern and Kammer, 1996;

Striegel, 1997). Examples of structural models are shown in Figure 11.

Cell

DMAc

O

H

Cl-

Li+

DMAcDMAc

CH3

H3C

CH3

C

O

N

Li+

a) b)

Figure 11. Model of LiCl/DMAc complex (a) according to Turbak (1984) and LiCl/DMAc-cellulose complex (b) according to Morgenstern et al. (1992).

2.5 Capillary electrophoresis (CE)

In capillary electrophoresis separation is achieved by applying a high voltage across a

capillary filled with an electrolyte or a gel. It was originally developed from slab gel

14


electrophoresis for biomedical applications. Fundamental works showed the benefit of

performing electrophoresis in a free solution using a capillary (Hjertén, 1967; Virtanen,

1974; Mikkers et al., 1979). The advantage of using a capillary is its large surface area-to-

volume ratio, which ensures heat to be efficiently dissipated and thus makes it possible to use

high voltages. Modern CE relies on the work by Jorgenson and Lukacs (1981a; 1981b) who

demonstrated high separation efficiencies by decreasing the inner diameter of the silica

capillaries to less than 100 µm. Commercial instruments were introduced about ten years ago

and the technique has grown rapidly ever since. Today, CE includes several sub-techniques,

which makes it possible to separate a variety of analytes. Capillary zone electrophoresis

(CZE) is the most common of these sub-techniques. Separation is achieved in a free solution,

i.e. in absence of a matrix, by applying a voltage across a capillary filled with an electrolyte

with or, as in this thesis, without additives.

2.5.1 Instrumentation

The basic CE-system consists of an injector, electrolyte vessels connected by a quartz

capillary, a high voltage supply and a detector as illustrated in Figure 12.

HV

Detector

Samplevial

CapillaryInjection

µ µ0

Figure 12. Schematic of a CE-system, illustrating the electrolyte vessels at the anodic/injection site (+) and the cathodic/ detector site (-), the capillary position during injection(dotted line) and during separation (solid line), on-line detection and high voltage supply(HV). The arrows indicate direction and relative magnitude of the electrophoretic mobility (µ)and electroosmotic flow (µ0), respectively, that influence a negatively charged solute.

There are several methods of injection and detection but in the following only those used in

this thesis will be commented. The sample is introduced by hydrodynamic injection, by

applying vacuum at the cathodic end of the capillary. Voltages up to 30 kV are possible on

commercial CE-instruments, but in practise excess heat generated by the current limits the


15

applied voltage. The most used detector is the on-column UV/visible (VIS) absorbance

detector. Since the path length is defined by the inner diameter of the capillary the

concentration sensitivity of absorbance detection is limited, although the mass sensitivity is

high.

2.5.2 Principles of CZE

The migration velocity (v) of a charged solute is given by

v = µappU/ lL [4]

where µapp is the apparent mobility, U is the applied voltage, l is the capillary length to the

detector and L is the total length of the capillary (Oda and Landers, 1997). Sample

components that differ in charge-to-size ratio have different electrophoretic mobilities (µ),

which can be approximated by

µ = q/ 6πηr [5]

where q is the charge of the solute molecule. The denominator describes the friction of a

molecule with a spherical geometry obeying Stokes law. η is the viscosity of the electrolyte

and r the hydrodynamic radius of the solute. In fused silica capillaries, v of a charged solute is

also influenced by the electroendoosmotic flow (EOF). This flow is caused by the movement

of hydrated cations adsorbed onto the negatively charged surface of the silica at pH>2. The

EOF leads to a plug-like flow of liquid which is characteristic for CZE. The rate of the EOF-

mobility (µ0) depends on the zeta potential (ζ) across the diffuse mobile layer and the

immobile Stern layer which is closest to the silica surface, η and the dielectric constant (ε) of

the electrolyte as given by

µ0 = ζε/4πη [6]

Thus, the composition of the electrolyte, i.e. nature and concentration of ions, determines the

magnitude of the EOF. For a given type of electrolyte with constant ionic strength, the EOF

increases as the number of dissociated silanol groups increases, i.e. with increase in pH

(Lukacs and Jorgenson, 1985; Vindevogel and Sandra, 1991). EOF increases with

decreasing ionic strength due to an increase in the thickness of the double layer and thereby ζ

Fujiwara and Honda, 1986). Although the mobility of a negatively charged molecule is

directed towards the anode it is swept to the cathode by the EOF during electrophoresis. In

presence of EOF, the migration time (tr) and µ of a solute is given by

tr = lL/U(µ+µ0) [7]

16


µ = lL/U (1/tr-1/t0) [8]

where t0 is the migration time of a neutral molecule. The number of theoretical plates (N) can

be increased by increasing the voltage

N = µU/2D [9]

where D is the diffusion coefficient of the solute. Since polymers have a lower D can be

separated with higher efficiency than low molecular weight compounds. The applied voltage

is, however, limited due to Joule heating which increases the peak width (σ2tot) and thereby

decreases resolution. The terms contributing to σ2tot are given by

σ2tot = σ2

int + σ2diff + σ2

wall + σ2T [10]

where the subscripts designate the variance due to length of the sample plug (int), the variance

due to longitudinal diffusion (diff), the variance due to interactions between analyte and

capillary wall (wall) and the variance due to temperature (T) (Oda and Landers, 1997). The

σ2tot can be minimised by keeping the ratio between length of the sample plug linj and L below

3% (Huang et al., 1989) and using an appropriate electrolyte of high concentration and low

ionic strength (Hjertén and Liao, 1995; Reijenga et al., 1996).


17

3 METHODS

3.1 Pulp characterisations

The viscosity of pulps dissolved in cupridiethylenediamine (CED) was determined according

to SCAN-CM 15:88 (1988) and the brightness of the bleached pulps was measured according

to SCAN-C11:75 (1975). The kappa number, was determined according to SCAN-C 1:77

(1977). Although the kappa number is commonly used to describe the degree of

delignification, it is rather a measure of all structures oxidised by potassium permanganate,

i.e. does not exclusively reflect the amount of lignin. Unsaturated structures, such as

hexeneuronic acid in hemicellulose also contribute to the kappa number (Li, 1999). This is in

particular true for birch wood kraft pulps. The characteristics of the pulps used in papers I-IV

are shown in appendix II.

The total recovery of the dissolved lignin samples and the efficiency of the acid dioxane

extraction were calculated from the obtained amount of lignin compared to the amount of

acid-insoluble lignin (Klason lignin) of wood and pulps, respectively (papers VI-VII).

Klason lignin was gravimetrically determined after swelling the pulp in 72% sulphuric acid at

30 °C for one hour, followed by dilution to 3% and hydrolysis at 125 ºC, 1.4 bar for one hour.

Acid-soluble lignin was determined by measuring the absorbance of the hydrolysate at 205

nm, using a correction factor for the degradation products of the carbohydrates (Schöning

and Johansson, 1965).

The neutral carbohydrate composition of the pulps was determined after extraction with

dichloromethane to remove extractives followed by acid hydrolysis, as described above,

essentially according to Theander and Westerlund (1986). The hydrolysate was reduced,

the corresponding alditols were acetylated and analysed by gas chromatography using 2-

deoxy-galactose as an internal standard. Separation was performed on a fused silica column

DB-1 (0.17µm x 0.32mm x 25m) (J&W Scientific). The injector temperature was 260 °C and

the temperature of the detector 290 °C. Helium was used as a carrier gas, flow 1.85 mL/min,

split ratio 1:20. The temperature program was 160 °C for 2 minutes, followed by a

temperature increase of 2 °C/minute up to a final temperature of 250 °C. To obtain the neutral

composition of pulp samples fractionated by SEC (paper II), the DMAc was removed by

vacuum evaporation at room temperature. A small portion of deionised water was added to

dissolve the salt followed by addition of concentrated sulphuric acid to a concentration of 3%.

Acid hydrolysis was performed at 125 ºC, 1.4 bar for one hour. The monosaccharide

18


composition of the hydrolysate was determined as described above. The overall recovery of

the carbohydrate analysis for the fractions collected from the unbleached hardwood kraft pulp

(HP) sample was 62% and for the fractions collected from the bleached softwood kraft pulp

(BSP) sample 84%.

3.2 Determination of fibre strength and fibre deformation

Zero-span tensile strength was measured according to Cowan (Boucai, 1971) using a Pulmac

Zero-span Tester. Rewetted handsheets were used to minimise interfiber bonding. Fibre

deformations of the pulp samples were measured as shape factor using the STFI FiberMaster

(Karlsson and Fransson, 1994). The shape factor is the ratio between projected length and

real fibre length. A shape factor of 100% corresponds to a completely straight fibre, the value

decreases as the fibre curl increases. Four pulp samples were also PFI-mill beaten by 3000

revolutions prior to measurement of the zero-span tensile strength and fibre curl.

3.3 Dissolution of pulp samples in LiCl/DMAc

Pulps were thoroughly washed with deionised water and were activated in deionised water at

+4 °C for one hour. Excess water was removed and the pulps were solvent-exchanged three

times with methanol and then three times with DMAc with an intermediate equilibration of

the pulp for 30 minutes. In the standard procedure, 8% LiCl/DMAc was added to give a

concentration of 0.8% with respect to the pulp. The sample solution was allowed to stand for

five days at +4 °C. The pulps were then diluted with DMAc to a final concentration of 0.5%

with respect to LiCl and 0.05% with respect to the pulp. In order to remove any undissolved

material and dust, the final solutions were centrifuged at 87,000g for 30 minutes. In order to

obtain molecular-dispersed solutions, the final solutions were treated in a Retsch vibratory

mill type MM-2 for 30 minutes (paper IV). In this paper centrifugation was omitted.

3.4 Static light scattering

Light scattering measurements were made on LiCl/DMAc solutions of cotton linter (paper

III). The centrifuged samples were filtered (0.7 µm, Whatman GF/F). Static light scattering

measurements were made using a Hamamatsu photon counting device with a 35 mW He-Ne-

laser (633 nm). Toluene was used as reference (Rtol=13.59x10-4/m at 633 nm). dn/dc was

measured in a differential refractometer with Rayleigh optics at 25 °C; dn/dc=0.091 mL/g.

The weight-average molecular weight (Mw) was obtained by measuring the angular


19

dependence of the scattered light, Rθ, of different sample concentrations (c). As the angle (θ)

and c approach zero the following relationship holds

Kc/ Rθ = 1/Mw [11]

K=4π2n02(dn/dc)2/NAλ4, where n0 is the refractive index of the solvent, dn/dc is the refractive

index increment measured for the present system at the wavelength used, NA is Avogadro's

constant and λ the wavelength. R90 is the Rayleigh ratio at angle 90° determined using

[I-I0/Itol]Rtol(ns/ntol)2. Here ns is the solvent refractive index and ntol that of toluene.

Rtol=4.0x105/cm. I is the measured solution intensity, I0 that of the solvent and Itol that of

toluene.

3.5 Black liquor and isolation of lignin samples

Flow-through kraft cooks of pine (papers V, VI) and birch (paper VII), were performed in

order to obtain lignin samples for CZE characterisations. Dry chips were impregnated with

deionised water for one hour to facilitate access of the cooking liquor. White liquor, with a

composition of hydrogen sulphide and sodium hydroxide to simulate conventional kraft

cooking conditions, was continuously added. The initial temperature was 65 ºC (birch wood)

and 70 ºC (pine wood), raised by 1 ºC/min to 165 ºC (birch wood) and 170 ºC (pine wood). If

nothing else is stated below, the cooking time was 210 minutes. Pulp data for pine wood

cooks are shown in appendix III and for birch wood cooks in appendix IV.

Fractions of black liquor were continuously collected during two cooks of each wood source

i) for direct characterisation and ii) for isolation of dissolved lignin. Dissolved lignin was

isolated by acid precipitation (pH 4.5) of the black liquor fractions essentially according to

Gellerstedt and Lindfors (1984). The acidified samples were placed in a freezer overnight to

ease coagulation of lignin. The following day, the precipitates were separated by

centrifugation at 10,700g for 15 minutes, washed twice by chilled deionised water and then

freeze-dried. The acid precipitates were extracted by n-pentane to remove elemental sulphur

and extractives. After freeze-drying, the precipitates were extracted with dioxane :deionised

water (9:1) to separate lignin from co-precipitated carbohydrates.

Residual lignin was isolated from pulps from three different cooks of each type of wood

interrupted after 90, 150 and 210 minutes (birch wood) and after 90, 150 and 220 minutes

(pine wood). Since the 90 minute-cooks were stopped before the wood chips were completely

defibrated, these pulps were milled to 4 mm granulates prior to further treatment. All pulp

samples were extracted by acetone to remove extractives. The residual lignin was obtained by

20


acid dioxane extraction, i.e. 0.1 M hydrogen chloride in dioxane-water (9:1) for two hours at

88 °C following the method of Gellerstedt et al. (1994).

Fresh black liquor samples were diluted by an equal volume of deionised water and

characterised by CZE within two hours. The isolated samples were dissolved in the pH 12-

electrolyte over night (under nitrogen) and diluted by an equal volume of deionised water to a

final concentration of 4 mg/L. All lignin samples were filtered (0.45 µm, Chrompack,

Acrodisc LC 13) prior to the CZE experiments.

3.6 Chemical characterisation of isolated lignin samples

Phenol group determinations of the isolated lignin samples were carried out by aminolysis

according to Månsson (1983). This method also provided the acetylated samples for

characterisation by size-exclusion chromatography (SEC). Carboxyl groups were determined

by quantitative 1H-NMR. Spectra of lignin (20-30mg) in DMSO-d6 (500 µl) were recorded on

a Bruker 500 MHz spectrometer using a 45° pulse, 7507 Hz sweep width, 3 s delay, 2 s

acquisition time and a time domain of 29k. 512 scans were recorded, and the spectrum was

processed without any zero-filling or window function. Trimethylsilyl propionic acid

2,2,3,3-d4 (TSP) was used as internal shift and quantitative standard. The spectra were

analyzed as described in literature (Li and Lundquist, 1994; Froass et al., 1996). The RSD

of the carboxyl group determination was 10% and the minimum detectable concentration

(MDC) was 0.1 mmol/g (signal-to-noise-ratio of 3).

3.7 SEC of pulps

Filtered samples (0.5 µm, Millex-SR, Millipore) were characterised on a Size Exclusion

Chromatography (SEC) system consisting of an automatic sampler, AS-4000A (Hitachi), with

an L-6200A pump (Hitachi) and a refractive index detector, RI-71 (Shodex). The injection

volume was 100 µL and the separations were performed at 80 °C with 0.5% LiCl/DMAc (1

mL/min) on four 20 µm Mixed-A columns preceded by a guard column (Polymer

Laboratories). Pullulan standards, 1.66M, 853k, 380k, 186k, 100k, 48k, 23.7k, 12.2k, 5.8k

and 738 Dalton (Polymer Laboratories), were used to calibrate the system. The linear

correlation coefficient (r2) between the Mp of the standards and elution time was 0.997. Data

acquisition and calculations were carried out using PL Caliber 7.04 software (Polymer

Laboratories).


21

3.8 SEC of lignin samples

40 µg of acetylated lignin dissolved in tetrahydrofuran (THF) was characterised at room

temperature using THF as mobile phase. The SEC system consisted of a Rheodyne 7125

injector, a set of three columns connected in series, 104Å, 500Å, 100Å (ultrastyragel, Waters),

and a Waters 510 pump (1 mL/min). The solutes were detected by a Waters 410 refractive

index detector. Polystyrene (PS) standards 350k, 115k, 34.5k, 10.2k, 3.1k, 1.05k and 580

Dalton (Polymer Laboratories) were used to calibrate the columns. Data acquisition and

calculations were carried out using Baseline (Waters).

3.9 CZE

CZE was performed on an Applied Biosystem (Model 270A-HT) instrument with a 72 cm x

50 µm I.D. fused silica capillary. The sample vials were filled with 50 µl of each sample

solution and 5 µl of n-C14 was added to prevent oxidation. To minimise evaporation, the

sample compartment was kept at a constant temperature of +7 °C. The samples were

introduced into the capillary by vacuum injection for 1.5 s preceded by 0.5 s injection of

pyridine (0.1% v/v) as neutral marker (NM). A voltage of 20 kV was applied and the

temperature of the separation compartment was kept at 30 °C. After each run the capillary

was conditioned by applying 30 kV for 2 minutes, flushing with 0.1 M sodium hydroxide for

2 minutes and with electrolyte for 3 minutes. Detection was carried out at 254 nm, 50 cm

from the injection site. Electropherograms were registered and calculations were performed

using ELDSPRO Labdata system (Chromatography Data Systems AB, Kungshög, Sweden).

22



23

4 DISCUSSION

In the first part of this section selected findings on dissolution, MWD profiles and

applicability of SEC-characterisation of kraft pulps are discussed (papers I-IV). Data on the

pulps used are shown in appendix II. The second part of this section concern the applicability

of capillary zone electrophoresis for lignin characterisation with respect to charge density

(papers V-VII). Data on the pulps corresponding to the cooks from which lignin samples

were obtained are shown in appendix III (papers V and VI) and appendix IV (paper VII).

4.1 Characterisation of kraft pulps dissolved in LiCl/DMAc (papers I-IV)

4.1.1 Differences between softwood and hardwood kraft pulps (papers I and II)

A number of underivatised cellulosic samples have been characterised by SEC using

LiCl/DMAc as solvent and mobile phase. Examples are cotton fibres (Timpa, 1991), purified

cellulose samples (Kennedy et al., 1990), wood pulps from the sulphite process (Ekmanis,

1986; Kennedy et al., 1990) and wood pulps from the kraft process (Kennedy et al., 1990;

Westermark and Gustafsson, 1994; Hortling et al., 1995; Silva and Laver, 1997). For

kraft pulps there is however some important differences between softwood and hardwood

pulps with respect to solubility as well as the shapes of MWD. In contrast to hardwood kraft

pulps, softwood kraft pulps are incompletely dissolved in LiCl/DMAc (Karlsson and

Westermark, 1994). The undissolved fraction is gel-like and renders a pellet on

centrifugation. In addition, we have observed that the dissolved fractions of softwood kraft

pulps increase the operating pressure during chromatography, which indicates that some of

the sample adsorbs onto the stationary phase. This means that only a part of the pulp

dissolved in LiCl/DMAc is available for SEC-characterisation. Attempts to increase the

availability for the solvent by swelling the pulp in e.g. ammonia were not successful, neither

did activation with hot DMAc, according to the one-pot procedure dissolution reported by

Timpa (1991), improve dissolution.

To obtain information about the incomplete dissolution and in order to improve the ability to

chromatograph kraft pulps, the amount of undissolved fraction from different types of never-

dried softwood kraft pulps was determined (paper I). As shown in Figure 13 lignin exerts a

negative impact on the solubility of the pulp when the kappa number is high. The limited

solubility cannot solely be explained by the presence of lignin since even the fully bleached

24


pulp could not be completely dissolved in LiCl/DMAc. At very high kappa numbers the

limited dissolution is probably due to the stiffness of the fibres, i.e. low swellability, rather

than the chemical structure of lignin. This is also supported by the observation that isolated

lignin samples are easily dissolved in LiCl/DMAc (Sjöholm et al., 1999a, 1999b).

0

10

20

30

40

50

60

70

80

90

% R

esid

ue

Prep. + DMAcK-number 18, n=6

DMAcK-number 18, n=4

MeOH + DMAcK-number 18, n=5

Pectinase +DMAcK-number 18, n=6

Mannanase + DMAcK-number 18, n=5

DMAcbleached, n=4

DMAcK-number 79, n=5

Figure 13. Influence of sample preparation and kappa number (k-number) on the amount ofLiCl/DMAc undissolved fraction of softwood kraft pulps. The error bars indicate the standarddeviation, n=number of samples and prep.=preparative samples handling.

Hemicellulose isolated from softwood did however differ in solubility (paper I):

arabinoglucouronoxylan was almost completely dissolved after the addition of water whereas

glucomannan, (Glc:Man/1:5.9) formed a gel on addition of LiCl/DMAc. In accordance with

this, the solubility was significantly improved when unbleached softwood pulp was treated by

mannanase prior to dissolution, Figure 13. In general, enzymes are too large to penetrate the

fibre wall (Viikari et al., 1990). It has been reported that deacetylation of glucomannan

(Hansson, 1970) and of galactoglucomannan (Thornton et al., 1994) leads to adsorption onto

fibres. The hemicelluloses in kraft pulps contain no acetyl groups, since these are easily

hydrolysed in the highly alkaline cooking conditions. It is thus reasonable to assume that the

improved solubility after treatment with mannanase is due to degradation of

galactoglucomannan adsorbed on the surface of the fibre.

The lignin content and neutral carbohydrate composition in the undissolved fraction differ

from those of the initial pulp, i.e. solvent exchanged pulp, Table 2. The glucose, mannose and

lignin content was higher in the undissolved fraction, whereas the lower content of arabinose


25

and xylose indicates that arabinoglucuronoxylan is more easily dissolved in LiCl/DMAc.

There is no obvious connection between the limited solubility and the cellulose content since

cotton linter is completely dissolved in LiCl/DMAc. The coincident higher concentration of

lignin and mannan may be due to association and or bonds between the two, since there are no

obstacles to dissolving isolated lignin samples. The ratio galactose:mannose was lower in the

undissolved fraction than in the initial pulp. This may reflect that galactoglucomannan is more

easily dissolved in LiCl/DMAc than glucomannan.

Table 2. The composition of neutral carbohydrates and lignin in initial pulp and residue,respectively. Initial pulp = unbleached softwood kraft pulp, residue = undissolved fraction ofthe initial pulp obtained from a preparative sample dissolution in LiCl/DMAc. Acid sol. =acid soluble. The monosaccharide data are the % of analysed carbohydrates and the ligninfigures are the % of sample.

Sample Glc

%

Gal

%

Man

%

Ara

%

Xyl

%

Klason

lignin (%)

Acid sol.

lignin (%)

Recovery1

%

Initial pulp 86.8 0.2 5.0 0.4 7.6 2.1 0.5 96.3

Residue 90.9 0.1 7.1 0.3 1.6 3.3 0.5 93.41 of carbohydrates and lignin.

In water systems, gelling and adsorption is favoured by low amounts of galactose substituents

(Overbeeke et al., 1987; Thornton et al., 1994). Glucomannans having a high ratio of

glucose:mannose self-associate and render the hemicellulose more "cellulose-like" which

facilitates adsorption onto cellulosic fibres (McCleary, 1990). It is however difficult to draw

any conclusion from the lower ratio of glucose:mannose observed in the residue compared to

the initial pulp, since any traces of cellulose will disguise the true ratio between the

hemicellulose components. There is no single explanation for the limited solubility of

softwood kraft pulp in LiCl/DMAc. At very high kappa numbers it seems reasonable that the

stiffness of the fibres, i.e. low swellability, restricts dissolution. The formation of gel,

probably due to glucomannan, seems to be the main reason for the incomplete dissolution of

pulps with kappa numbers at and below 18. To increase the solubility of softwood kraft pulps

it appears that the (galacto)glucomannan either has to be either removed or specific additives

have to be used to facilitate dissolution in LiCl/DMAc. For the time being, unbleached

softwood pulps characterised by LiCl/DMAc-SEC should be evaluated with care since the

composition of the dissolved fraction may not be representative of the pulp.

26


In general, the MWD profile of the dissolved fraction of softwood kraft pulps (SP) differs

from hardwood kraft pulps (HP) (paper II). The profiles obtained by RI detection of an

unbleached SP and an unbleached HP are shown in Figure 14. The MWD of the SP

corresponds to about 62% of the pulp, whereas that of the HP corresponds to a completely

dissolved sample. The MWD of the HP renders two fairly well separated distributions

whereas the MWD of the SP is more complex.

3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8

logM (relative pullulan)

dw/d

logM

Softwood pulp

Hardwood pulp

RI-detection of unbleached kraft pulps

Figure 14. MWD profiles obtained by RI detection of the dissolved fraction of a softwoodkraft pulp (SP) and a hardwood kraft pulp (HP).

3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8


dw/d

logM

Hardwood pulp

Softwood pulp

UV-detection of unbleached kraft pulps

Figure 15. MWD profiles obtained by UV detection (295 nm) of the dissolved fraction of asoftwood kraft pulp (SP) and a hardwood kraft pulps (HP).


27

This characteristic difference remains even for bleached pulp samples where about 92% of the

corresponding SP is dissolved (paper II), suggesting that the difference is caused by

hemicelluloses rather than the lignin. This was further supported by the MWD profile

obtained by UV detection of softwood and hardwood kraft pulps which, although differently

accentuated, both render two distributions, Figure 15.

Examination of the neutral monosaccharide compositions in different elution volumes

indicates that, for hardwood kraft pulps, the high molecular distribution originates from

cellulose while the low molecular weight distribution originates from xylan, Figure 16. Since

bleaching does not change the characteristic differences in the MWDs between hardwood and

softwood kraft pulps, a bleached softwood kraft pulp (BSP) was used in order to increase the

amount of sample available for carbohydrate determination. In contrast to the hardwood

carbohydrate polymers, hemicelluloses of softwood kraft pulps elute over the entire molecular

weight (M) range. The relative concentration of mannose decreases during elution, i.e. with

decreasing M, whereas galactose, xylose and arabinose increase during elution, i.e. with

decreasing M, Figure 17.

15 17 19 21 23 25 27 29 31 33 35

Elution time (min)

Diff

eren

tial r

efra

ctio

n

Glc 100 19Xyl 77Gal 4

Carbohydrate composition (%) of HP

Figure 16. Chromatogram showing the relative carbohydrate composition (%) in differentelution volumes of an unbleached hardwood kraft pulp (HP).

28


15 17 19 21 23 25 27 29 31 33 35

Elution time (min)

Diff

eren

tial r

efra

ctio

n

Glc 86 83 68Xyl 4.0 9.7 24 Ara 0.3 0.8 1.6Man 8.1 5.3 4.6Gal 1.4 1.7 2.3

Carbohydrate composition (%) of BSP

Figure 17. Chromatogram showing the relative carbohydrate composition (%) in differentelution volumes of a bleached softwood kraft pulp (BSP).

The increasing ratio of galactose:mannose during elution implies differences between the two

main types of glucomannans in softwood having high and low galactose content, respectively.

These differences may either originate from the arrangement of the galactoglucomannan and

cellulose in the fibre wall or be due to different degrees of association in LiCl/DMAc.

Glucomannan has been proposed to be associated with the cellulose in the fibre wall (Salmén

and Olsson, 1998). In solution, glucomannans have also been reported to have a high affinity

for cellulose (McCleary, 1990; Iwata et al., 1998; Whitney et al., 1998). Furthermore, it has

been suggested that cellulose acts as a template for unsubstituted mannan segments which

may allow glucomannan to partly adopt a “cellulosic conformation” (Whitney et al., 1998).

The early elution of low galactose-content glucomannan indicated in Figure 17 may thus

reflect association with cellulose either already present in the fibre wall or formed in

LiCl/DMAc. The amount (Dea et al., 1977) and distribution of galactosyl substituents (Dea et

al., 1986) influence associations involving galactoglucomannans. Large amounts of galactose

substituents and low frequency of unsubstituted portions in the galactoglucomannan would be

expected to significantly hinder these associations. The increasing ratio of galactose:mannose

with decreasing molecular weight may thus indicate that galactose groups in the

galactoglucomannan aggravate association with cellulose.

The reason for the difference in elution behaviour between the xylan in hardwood and

softwood kraft pulps, as indicated by the neutral carbohydrate composition shown in Figures


29

16 and 17, is not clear. The presence of arabinose does not seem to prevent the xylan in BSP

to elute in the high molecular weight range, since the ratio arabinose:xylose is about the same

in all of the studied fractions. One explanation may be that the amount and/or distribution of

uronic acid substituents differ between the two pulps. During alkaline pulping most of the 4-

O-Me-α-D-glucuronic acid (MeGlcA) is converted to 4-deoxyhex-4-enuronic acid (HexA),

which partly is further degraded (Johansson and Samuelson, 1977b; Teleman et al., 1995;

Buchert et al., 1995). The remaining amount of HexA is higher in birch wood kraft pulps

than in unbleached pine wood kraft pulps (Gellerstedt and Li, 1996). Bleaching under acidic

conditions, as for the BSP sample in this study, decreases the amount of HexA (Buchert et

al., 1995; Bergnor-Gidnert et al., 1998) although MeGlcA and arabinose substituents are

fairly stable (Buchert et al., 1995). The amount of uronic acid groups in the BSP shown in

Figure 18 is therefore probably lower than in the unbleached HP shown in Figure 17,

although the amounts of uronic acids in the studied pulps were not determined. There is no

complete knowledge about how the uronic acid substituents in kraft pulps are distributed.

MeGlcA in birch wood (Havlicek and Samuelson, 1972; Rosell and Svensson, 1975;

Lindquist and Dahlman, 1998) and softwood (Shimizu et al., 1978; Comtat and Joseleau,

1981) is irregularly distributed along the xylan backbone and can thus not provide an

explanation of the observed differences in the LiCl/DMAc-SEC system.

It is possible that the xylan in the BSP consists of substructures differing in content of uronic

groups. A gradual decrease in uronic groups would then account for the association and

thereby the indicated co-elution with cellulose and/or glucomannan in BSP. A higher degree

of uronic acids would render the BSP-xylan more like the HP-xylan and consequently it

would elute in the low molecular weight range. However, this explanation needs to be

confirmed in future studies since the amounts of uronic substituents have not been determined

for the studied pulps.

4.1.2 Aggregation of cellulosic samples dissolved in LiCl/DMAc (paper III)

In contrast to the case for softwood kraft pulps, the MWD of hardwood kraft pulp (HP) that

corresponds to cellulose extends up to the high molecular weight region, Figure 14. The

weight-average molecular weight (Mw) relative pullulan of the cellulose distribution of HP

and bleached HP (BHP) is high. A cotton linter sample, which consists solely of cellulose,

with a similar elution volume as the cellulose portion of the hardwood kraft pulps was used to

study this behaviour, Figure 18.

30


3 4 5 6 7 8


dw/d

logM

HP

CLMw (LS)

4.4x106

BHP

Figure 18. Molecular weight distributions relative pullulan of cotton linter (CL), unbleached(HP) and bleached (BHP) hardwood kraft pulps. Sample solutions prepared by the standardprocedure. Mw (LS) = Mw determined by light scattering.

The true Mw of the cotton linter sample measured by laser light scattering (LS) was in close

agreement with the Mw determined relative pullulan using SEC but about thrice that of the Mw

calculated from the weight-average degree of polymerisation (DPw) given by the supplier,

Table 3. This indicates that the high values obtained by SEC are not due to different

relationships between molecular weight and elution volume for pullulan and cellulose.

Table 3. Mw relative pullulan of the cellulose portion of hardwood kraft pulps and of cottonlinter. For comparison Mw calculated from the DPw and as measured by light scattering (LS)of the cotton linter is also shown. The samples were dissolved by the standard procedure.

sample Mw

(relative pullulan)Mw

(calculated1)Mw

(LS)unbleached hardwood kraft pulp(HP)

3.0 x 106 - -bleached HP (BHP) 1.9 x 106 - -cotton linter (CL) 4.6 x 106 1.3 x 106 4.4 x 106

1 from the DPw 7950 specified by the supplier.

Provided that the value obtained from the supplier is correct, these observations indicate the

presence of aggregates of cellulose dissolved in LiCl/DMAc. This indication is further

supported by the presence of a more or less accentuated shoulder on the high molecular

weight region of the cellulose distributions. Others have also reported this behaviour of wood


31

pulps dissolved in LiCl/DMAc (Silva and Laver, 1997). By applying a curve-fitting

procedure, three Gaussian distributions were resolved from the MWD of the HP sample,

Figure 19. These correspond to hemicellulose (c1), cellulose (c2) and aggregates (c3).

3 4 5 6 7 8


dw/d

logM

MWD c1 + c2 + c3

c1

c2

c3

MWD

Figure 19. Molecular weight distribution (MWD) (normal line) and Gaussian curvesdeconvoluted from the MWD of unbleached hardwood kraft pulp (HP) dissolved by thestandard procedure, using an initial LiCl concentration of 8%. c1, c2, c3 = componentsdeconvoluted from the MWD (normal line), c1 + c2 + c3 = the sum of the deconvoluted curves(bold broken line).

The indicated aggregates may be due to unbroken intermolecular bonds present in the initial

cellulose sample and/or that such bonds have been formed during dissolution in LiCl/DMAc.

Factors that influence dissolution in LiCl/DMAc are i) activation to open up the fibre

structure ii) Mw and concentration of the cellulose sample iii) salt concentration iv) time of

dissolution and v) temperature (Turbak et al., 1981; McCormick et al., 1985; Dawsey and

McCormick, 1990). Activation in hot DMAc gave a similar MWD profile as our standard

procedure and attempts to obtain molecular-dispersed solutions of the HP sample by varying

the dissolving conditions failed. The area of c3 decreased with decreasing concentration of

LiCl in the order 10%>8%>6%, suggesting a decrease in the amount of aggregates (paper

III), but due to incomplete dissolution further reduction of LiCl concentration was not

possible. The position of the distribution on the logM scale is about the same and the ratio of

the peak value (Mp) between aggregates/cellulose is about seven irrespective of salt

concentration used. Stable aggregates of cellulose dissolved in LiCl/DMAc have been

32


reported to consist of about seven fully extended cellulose molecules with side-by-side

organisation (Terbojevich et al., 1985).

Since it was not possible to avoid the presence of aggregates, sample solutions prepared by

the standard procedure were mechanically treated in a vibratory mill to disrupt the forces that

keep the aggregates together. From the MWD profile it seems that this treatment

preferentially affects the high molecular weight fraction, although a certain chain cleavage

cannot be excluded. The minimum treatment time required to achieve deaggregated sample

solutions with a minimum of chain scission was 30 minutes, according to polydispersity

(Mw/Mn) (paper III) and Mw, Figure 20. After 30 minutes treatment time of the CL sample

solution, the Mw according to LS determination agrees fairly well with the Mw calculated

from the DPw obtained from the supplier, Table 3.

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

0 15 30 45 60 75 90 105 120

Treatment time (minutes)

Mw x

106

(rel

ativ

e pu

llula

n)

CLHP (hm)BHP (hm)

4.4x106 (LS)

1.25x106 (LS)

Figure 20. Influence of the time of mechanical treatment on the Mw (relative pullulan). CL =cotton linter, HP(hm) = the high molecular weight distribution of unbleached kraft wood pulpand BHP(hm) = the high molecular weight distribution of bleached kraft wood pulp. (LS) =Mw determined by light scattering.

The high molecular weight shoulder that can be discerned in the MWDs prepared by the

standard procedure is removed by the treatment, Figure 21. The high molecular weight

distribution of cellulose is narrower and its peak value (Mp) higher for both samples than for

their untreated counterparts. In contrast, the low molecular weight distributions are hardly

affected at all. The Mw (relative to pullulan) of a standard pullulan (853k), dissolved in 0.5%

LiCl/DMAc decreases by about 10% when mechanically treated for 30 minutes. This result


33

indicates that breakage of covalent bonds in the cellulose during treatment may only represent

a small fraction of the decrease in the Mw observed for the cellulose of the HP and BHP

samples. Although a certain degradation of the cellulose cannot be ruled out, the suggested

mechanical treatment of the solution seems to be a way to deaggregate LiCl/DMAc solutions

of hardwood kraft pulps.

3 4 5 6 7 8


dw/d

logM

HP-uc

BHP-30

HP-30

BHP-uc

Figure 21. MWDs relative pullulan of unbleached (HP) and bleached (BHP) hardwood kraftpulps. uc = solutions prepared by the standard procedure including ultracentrifugation, 30 =solutions prepared by the standard procedure with 30 minutes mechanical treatment (omittingultracentrifugation).

4.1.3 Molecular weight distribution of birch wood kraft pulp samples with different

relations between zero-span tensile strength and pulp viscosity (paper IV)

The effect of degradation on the carbohydrate polymers of a laboratory-made birch kraft pulp

was examined by subjecting the pulp to gamma-irradiation (γ), oxygen/alkali (O2/OH) and

alkali (OH), respectively. The impact of degradation on single fibre strength can be estimated

by zero-span tensile strength of rewetted pulp samples (Boucai, 1971; Gurnagul and Page,

1989). The treatments give pulps of different relationship between fibre strength and

viscosity, Figure 22. At a given viscosity the strength of the pulps decreases in the order γ >

O2/OH > OH and the difference becomes more accentuated as degradation proceeds. It is

notable that the strength of the γ-degraded pulps is hardly affected at all, although the

viscosity has decreased considerably during the treatment.

34


80

100

120

140

160

180

200

600 700 800 900 1000 1100 1200

Viscosity (mLg-1)

Zer

o-sp

an te

nsile

stre

ngth

(Nm

g-1

)

OH

O2/OH

γuntreated

Figure 22. Zero-span tensile strength versus viscosity for untreated birchwood kraft pulp andafter treatments by gamma irradiation (γ), oxygen/alkali (O2/OH) and alkali (OH),respectively, for various lengths of time.

Homogeneous degradation is predicted to decrease strength less than heterogeneous

degradation, which is assumed to occur at weak points irregularly localised along the fibre

(Gurnagul et al., 1992). In analogy with these findings, the γ-treatment presented in our

study illustrates homogeneous degradation, i.e. a small reduction in strength, whereas O2/OH

and OH treatments, respectively, represent an increasing degree of heterogeneity leading to a

greater loss of strength. Since a deformed fibre has a decreased ability to transfer load, zero-

span tensile strength reflects both fibre strength and amount of fibre deformation (Mohlin

and Alfredsson, 1990; Mohlin et al., 1996). There is a clear relationship between zero-span

tensile strength and shape factor for the three types of degradations, Figure 23. The shape

factor is not changed by γ-treatment. The degradation obtained by O2/OH decreases the shape

factor slightly whereas the impact of the OH treatment is severe. The alkali-treated pulp fibres

with lowest shape factor were straightened by PFI-milling but the zero-span tensile strength

did not improve indicating that the fibres were irreversibly damaged by the alkali treatment.


35

80

100

120

140

160

180

200

86 88 90 92 94 96

Shape factor (%)

Zer

o-sp

an te

nsile

stre

ngth

(Nm

g-1

)

OH

O2/OHγ Initial

pulp

Figure 23. Zero-span tensile strength versus shape factor for untreated birchwood kraft pulp(initial pulp) and after treatments by gamma irradiation (γ), oxygen/alkali (O2/OH) and alkali(OH), respectively, for various lengths of time. Unfilled symbols refer to beaten pulp sampleswhich correspond to pulp samples marked grey.

The MWDs of the pulps obtained by LiCl/DMAc-SEC are shown in Figures 24-26.

3 4 5 6 7

log M (relative pullulan)

dw/d

log

M

untreated and γ-irradiated pulps

Figure 24. MWDs of untreated birchwood kraft pulp (viscosity 1160 mL/g) and aftertreatment by gamma irradiation for various lengths of time to different viscosities; 1080, 960,880 and 690 mL/g, respectively. The arrows indicate decreased viscosity of the samples.

36


3 4 5 6 7


dw/d

log

Muntreated and O2/OH-degraded pulps

Figure 25. MWDs of untreated birchwood kraft pulp (viscosity 1160 mL/g) and aftertreatment by oxygen/alkali (O2/OH) for various lengths of time to different viscosities; 1020,920, 830 and 780 mL/g, respectively. The arrows indicate decreased viscosity of the samples.

3 4 5 6 7


dw/d

log

M

untreated and OH-degraded pulps

Figure 26. MWDs of untreated birchwood kraft pulp (viscosity 1160 mL/g) and aftertreatment by alkali (OH) for various lengths of time to different viscosities, 1030, 910, 780and 710 mL/g, respectively. The arrows indicate decreased viscosity of the samples.

The high molecular weight distribution reflects cellulose and the low molecular weight

distribution the hemicellulose and lignin. Because of the low amount of pulp lignin in the


37

untreated pulp (1.7% Klason lignin) the low-molecular distribution is referred to as

hemicellulose in the following. For all treatments, the molecular weight distribution of

cellulose is shifted to a lower molecular weight range. The intermediate molecular-weight

fraction that gradually develops in the MWD of the chemically treated pulps, Figures 25 and

26, is not observed for the γ-treated pulps, Figure 24. This may indicate that there are certain

positions in the cellulose chain that are cleaved during the chemical treatments; if so, these

positions are probably located in the vicinity of less ordered regions, i.e. nodes or weak points

of the fibre.

The different MWD profiles of the cellulose for radiochemical and chemical degradation

reflect differences in random and non-random chain scission. γ-irradiation mainly

depolymerizes through direct ionisation (Lind et al., 1997) although the action of hydroxyl

radicals formed in the surrounding water may also contribute (Ek et al., 1989). In contrast to

the γ-treatment, the extent of degradation by the chemical treatments will be governed by the

accessibility of cellulose, i.e. morphology of the fibre as well as the extent and position of

fibre damage. The initial reacting species during O2/OH treatment is oxygen itself and the

radicals formed, such as superoxide and hydroxyl radicals (Gierer, 1990). These reactive

species cause depolymerization of the O2/OH-treated pulps by introduction of alkali-labile

groups in the carbohydrate polymers. The observed depolymerization of cellulose caused by

the OH-treatment is due to alkaline hydrolysis.

As a consequence of the degradation of the cellulose, the chromatographic resolution between

the high and low molecular weight distributions gradually deteriorates. This affects the

evaluation of the chromatograms, especially with respect to hemicellulose. Degradation of

hemicellulose, which should be evidenced by a decrease in Mp and, in the case of removal of

hemicellulose, also decreased relative area of the low-molecular distribution, is thus partly

disguised by the chromatographic interference of (degraded) cellulose.

The peak molecular weight (Mp) of the hemicellulose distribution is essentially the same for

all of the pulp samples. The probability of demonstrating degradation caused by γ-irradiation

is smaller for hemicellulose than for cellulose, since the relative amount of hemicellulose in

the pulp is smaller and it also has a considerably lower molecular weight than cellulose. The

unaltered Mp of the low molecular weight portion of the chemically treated pulps may also

indicate that there is a critical molecular size below which birch kraft pulp hemicellulose is

dissolved in alkali. Other explanations of why the Mp seems to be unaltered may be that low

molecular weight fragments are lost during the solvent exchange that precedes the

38


chromatographic separation and/or that the decrease in chain length of the hemicellulose is

not severe enough to be revealed by the chromatographic system.

The relative area of the low molecular weight distribution of the γ-treated pulps remains about

the same as for the untreated pulp, illustrating that the hemicellulose has not been removed,

Figure 27. It appears that if the extraction of degraded hemicellulose is efficient enough, the

contribution from degraded cellulose to the low-molecular distribution will be overruled,

which may explain the decrease in relative area, observed for OH-pulps as degradation

proceeds. Fibre strength has been proposed to decrease with decreasing amount of amorphous

regions (Page, 1983) and increasing fibril angle in the S2 layer (Page et al., 1972). For kraft

fibres of low fibril angle, zero-span tensile strength increases proportionally up to an α-

cellulose content of about 80% (Page et al., 1985). The α-cellulose is considered to reflect the

amount of undegraded cellulose determined by its insolubility in alkali during specified

conditions (Wilson et al., 1952; Rånby, 1952; Tappi Standard, 1993). Based on the MWD

evaluations, the observed decrease in shape factor/fibre strength is probably due to

heterogeneous degradation and removal of hemicellulose, whereas the observed decrease in

Mw of the cellulose is of minor importance.

80

100

120

140

160

180

200

18 19 20 21 22 23 24 25 26 27 28 29 30

Area of low molecular weight distribution (% of MWD)

Zer

o-sp

an te

nsile

stre

ngth

(Nm

g-1

) untreated pulp γ

O2/OH

OH

Figure 27. Zero-span tensile strength versus amount of hemicellulose according to MWD.Untreated birchwood kraft pulp (X), after treatments by gamma irradiation (γ), oxygen/alkali(O2/OH) and alkali (OH), respectively, for various lengths of time.


39

4.2 Characterisation of kraft lignin samples by CZE (papers V-VII)

The rate and extent of delignification may be influenced by the accessibility and reactivity of

the lignin (Gierer, 1980; Whiting and Goring, 1981) as well as the hydrophilicity (Gierer,

1980; Ljunggren, 1980) and size (Ahlgren et al., 1971; Kerr and Goring, 1975) of the

lignin fragments. In this context, the content of phenol groups and the MWD of lignin are

commonly reported. To obtain information about the distribution of ionised groups between

lignin fragments of various sizes, the charge density of lignin samples was estimated by

capillary zone electrophoresis (CZE).

4.2.1 Electrophoretic characterisations of lignin-derived compounds

When wood is degraded in nature, lignin is partly depolymerized, oxidised and becomes a

part of humic substances (HS) in soil (Schevenko and Bailey, 1996). Electropherograms

obtained by CZE have been used for profile characterisation of HS (Rigol et al., 1994; Fetsch

and Havel, 1998; Schmitt-Kopplin et al., 1998a; Schmitt-Kopplin et al., 1998b). Due to

the carboxylic acid groups in HS, the characterisations have mainly been performed at low

pHs (3-5). Different molecular mass fractions of humic acids have also been characterised

with respect to their charge density distribution and the fractions quantified by CZE (Rigol et

al., 1998). Lignins have been studied by traditional electrophoretic methods. Lignin

sulphonates has been studied by paper and free boundary electrophoresis (Goring et al.,

1958) and fractions of milled wood lignin by glass fibre sheet electrophoresis (Lindgren,

1958). So far, CZE has only been applied to low molecular weight lignin-derived sulphonates

(Dahlman and Månsson, 1996) and degradation products from lignin (Masselter et al.,

1995; Maman et al., 1996; Hiermann and Radl, 1998) using phosphate or borate buffers.

The applicability of CZE for characterising kraft lignin samples was shown by using Gly-

NaOH as electrolyte (paper V). The major benefit of the method is that the high pH permits

underivatised samples to be studied and electrostatic adsorption onto the capillary wall is

avoided. The aromatic structure of lignin makes it possible to detect low amounts of sample.

Bandbroadening is minimized by injecting small volumes of sample (linj/L was below 0.3%)

and using a low conductivity electrolyte. The conductivity (κ) of the used electrolyte, 100

mM Gly-NaOH, is about 3.0 x 10-6 Ω-1cm-1 at pH 10 and about 5.4 x 10-6 Ω-1cm-1 at pH 12

which makes it possible to apply a high voltage without excess Joule heating. The samples

were isolated by acid precipitation of black liquor fractions collected from a flow-through

kraft cook of pine wood. The obtained electropherograms show a mobility distribution which

reflects the charge density distribution of lignin. The distribution of the lignin is broad and is

40


shifted towards longer migration time, i.e. the mobility increases, as the cook proceeds. The

increase in charge density is probably due to the increased concentration of phenol groups

since the MWDs of the studied fractions shifts towards higher M range.

Birch wood and pine wood differ in chemical composition and behave differently during

pulping with respect to when defibration occurs, rate of delignification and amount of residual

lignin. Although a number of studies have been performed, the reason for these differences is

not completely clear. In order to study these differences, the change in charge density of black

liquor, dissolved lignin (isolated from black liquor) and residual lignin (isolated from pulp)

during flow-through kraft cooking of pine wood and birch wood, respectively, was further

studied (papers VI-VII).

4.2.2 Charged groups and molecular weight

Black liquor was collected in 30-minutes intervals. The figures in the sample designations of

the pine wood black liquor (Sb) and the dissolved lignin (Sd) (Table 4) and of the

corresponding birch wood Hb and Hd (Table 5), respectively, refer to the middle of the

collection time range. The figures in the sample designations of the residual lignin refer to the

cooking time of pine wood (Sr) and birch wood (Hr), respectively. The transition point

between initial and bulk delignification is assumed to occur during the 60-90 minutes interval

and the samples obtained after 180 minutes cooking time correspond to the residual

delignification phase. The amounts of phenol and carboxyl groups, as well as fragment size,

will influence the mobility of kraft lignin samples. The concentration of charged groups and

the number-average (Mn) of the isolated lignin samples derived from pine wood and birch

wood are shown in Table 4 and Table 5, respectively.

Table 4. Concentration of phenol and carboxyl groups and number-average molecular weight(Mn) of isolated lignin samples from kraft cooking of pine wood.

Isolated lignin samples Recovery%

Phenol groupsmmol/g

Carboxyl groupsmmol/g

Mna

(relative PS)Dissolved lignin Sd 45 2.4 n.d.b 790

Sd 75 2.6 0.46 980Sd 105 2.7 0.47 1300Sd 135 2.7 0.37 1500Sd 165 2.8 0.71 1600Sd 195 59c 3.0 <0.10 1700

Residual lignin Sr 90 13d 2.2 <0.10 3400Sr 150 50d 2.7 <0.10 4600Sr 220 53d 3.2 0.18 3500

a acetylated samples. b not determined. c of the lignin removed from the pulp. d of the pulp lignin.


41

Table 5. Concentration of phenol and carboxyl groups and number-average molecular weight(Mn) of isolated lignin samples from kraft cooking of birch wood.Isolated lignin samples Recovery

%Phenol groups

mmol/gCarboxyl groups

Mmol/gMn

a

(relative PS)Dissolved lignin Hd 45 1.7 0.39 760

Hd 75 2.2 0.41 820Hd 105 2.3 0.39 830Hd 135 2.6 0.18 910Hd 165 2.8 0.16 970Hd 195 70c 2.5 n.d.b 1200

Residual lignin Hr 90 65d 0.86 <0.10 59000Hr 150 61d 2.3 <0.10 12000Hr 210 >100d 2.5 n.d.b 5300

a acetylated samples. b not determined. c of the lignin removed from the pulp. d of the pulp lignin.

The recovery of the dissolved lignin reflects the acid-precipitable fraction of the black liquor,

i.e. polymeric fragments. The increase in concentration of phenol groups in the Sd/Hd

samples during the cooks reflects cleavage of aryl-ether bonds in lignin. These reactions seem

to become less frequent at the end of the birch wood cook. The observed difference in

recovery of residual lignin samples between the two types of wood pulps is probably due to

the presence of syringyl units in the birch wood pulp (Lundquist, 1973). This is also

indicated by a higher methoxyl content in the Hr-samples (Sjöholm et al., 1999a) than in the

corresponding Sr-samples (Sjöholm et al., 1999b). The concentration of phenol groups of the

Sr/Hr samples reflects cleavage of aryl-ether bonds during the cook but there is also a

contribution due to the applied isolation method. The residual lignin samples were isolated

from the pulps by acid dioxane extraction. This isolation method has been applied to

softwood kraft pulps with various lignin contents (Gellerstedt and Lindfors, 1984; Jiang

and Argyropoulos, 1999) and to birch kraft pulp from a completed cook (Gellerstedt et al.,

1994).

Studies on milled wood lignin have shown that mainly aryl ether bonds are cleaved during

acidolysis (Lundquist and Lundgren, 1972). During the weaker acidic conditions applied to

the pulps in the present study, the α-aryl ether bonds are probably completely cleaved

whereas the cleavage of β-aryl ether bonds may contribute less to fragmentation. The extent

of aryl ether cleavage in pulp lignin subjected to the acid extraction conditions used in this

study has so far not been reported and no significant change in the structure of residual lignin

has been demonstrated (Gellerstedt et al., 1994; Jiang and Argyropoulos, 1999).

Because of the non-oxidative conditions during kraft cooking, the amount of carboxyl groups

in respective lignin sample is considerably lower than the amount of phenol groups. Carboxyl

42


groups have been detected in dissolved kraft lignin and are considered to be aliphatic in nature

(Ekman and Lindberg, 1960; Enkvist, 1961; Marton and Adler, 1963; Mörck et al.,

1986). A certain contribution from carboxyl groups in carbohydrates is also possible, since

guaiacyl compounds substituted with an aliphatic chain with α-hydroxy carboxyl groups have

also been identified in black liquor (Löwendahl et al., 1978; Gierer and Wännström, 1984).

It is possible that similar structures exist in polymeric lignin fragments as well.

The influence of size on mobility depends on the molecular weight (M) and conformation of

the lignin, i.e. the friction it exerts against movement. In the present study, the number-

average molecular weight (Mn) of acetylated samples was determined relative polystyrene by

SEC using THF as mobile phase. The Mn data are thus only relative. All of the residual lignin

samples have a higher Mn than the dissolved lignin samples. According to the change in Mn,

the size of the dissolved lignin fragments increase during the cooks whereas the change in Mn

of birch wood residual lignin indicates degradation of the pulp lignin. The non-uniform

change of Mn with respect to cooking time of the pine wood residual lignin may be due to the

low efficiency of the acid dioxane extraction of the pulp from the beginning of the cook. At

this stage of delignification the pulp is not completely defibrated. In contrast to the Hr 90

sample, it is probable that the Sr 90 sample represents easily accessible lignin fragments of

fairly low molecular size. The high Mn value of the Hr 90-sample may partly be due to

condensation reactions between reactive lignin structures during the acid isolation procedure.

However, acid dioxane extraction as used in this study, has not previously been reported for

isolation of residual lignin from highly lignified hardwood pulps. No evidence for

condensation has been found for residual lignin isolated from birch kraft pulps obtained at the

end of the cook (Gellerstedt et al., 1994).

4.2.3 Charge density distributions

In CZE the separation is based on differences in mobilities and consequently the peak width

increases as the migration time increases. To compensate for this, the integrated peak area is

commonly divided by the migration time when low molecular weight compounds are to be

quantified (Kitagishi, 1997). The inconstant velocity of the lignin samples was corrected in a

similar way by recalculating the detector signal according to

zi = Bi + hi × tk/ti [12]

where Bi, hi and ti are the baseline signal, the peak height and the migration time, respectively,

of each datapoint i of the distribution. tk is a constant equal to the migration time of the peak


43

maximum of the integrated distribution and is used to obtain a peak maximum of about the

same intensity as the original distribution (paper VI). Time-corrected electropherograms are

shown in the following.

The lignin samples were characterised at pH 12 to achieve comprehensive dissociation of the

phenol groups. Because of their negative charge, lignins have longer migration times than the

added neutral marker (NM), and the migration time increases with increasing charge density.

The profiles of the mobility distribution differ between the studied types of samples. Some

electropherograms of residual lignin (Sr/Hr) dissolved lignin (Sd/Hd) and black liquor (Sb),

are shown in Figures 28-31. Due to the lower detector response of the birch wood derived

samples, the intensity in the electropherograms is set to half of the intensity for pine wood

derived samples. All of the residual lignin samples (Sr/Hr) have a fairly smooth mobility

distribution, Figures 28 and 29. The shapes of the Sr-samples are Gaussian. The distribution

of the birch residual lignin from the beginning of the cook contains low mobility components.

The relative amount of these components decreases and the shape become close to Gaussian

as the cook proceeds.

4 6 8 10 12 14 16 18 20 22

Sr 90

NM

Migration time (min)

1 mAU

24 26

Sr 150 Sr 220

Figure 28. Mobility distributions of residual lignin isolated from pine wood kraft pulps. Thesamples correspond to the beginning (Sr 90), the middle (Sr 150) and the end (Sr 220) of thecook. Capillary, 0.72 m x 50 µm I.D.; applied voltage 20 kV; electrolyte, 0.1 M glycine-NaOH, pH 12; temperature 30 °C; detection 254 nm.

44


4 6 8 10 12 14 16 18 20 22 24 26


1 mAU

Hr 210

Hr 150

Hr 90

NM

Figure 29. Mobility distributions of residual lignin isolated from birch wood kraft pulps. Thesamples correspond to the beginning (Hr 90), the middle (Hr 150) and the end (Hr 210) of thecook. Conditions as in Figure 28.

The dissolved lignin samples (Sd/Hd) have some well-discernible peaks superimposed on

their distributions. The shapes of the mobility distributions deviate from Gaussian for the Hd-

samples because of contribution from low mobility components and the same is true for Sd 75

which corresponds to the beginning of the cook.

6 8 10 12 14 20 2218164


Sd 75

NM

1 mAU

2624

Sd 135

Sd 195

Figure 30. Mobility distributions of dissolved lignin isolated from black liquor collected inthe beginning (Sd 75), in the middle (Sd 135) and at the end (Sd 195) of a kraft cook of pinewood. Conditions as in Figure 28.


45

4 6 8 10 12 14 16 18 20 22

1 mAU


NM

24 26

Hd 195

Hd 135

Hd 75

Figure 31. Mobility distributions of dissolved lignin isolated from black liquor collected inthe beginning (Hd 75), in the middle (Hd 135) and at the end (Hd 195) of a kraft cook of birchwood. Conditions as in Figure 28.

Since the sample concentrations are equal in the isolated samples, a difference in detector

response reflects differences in the lignin structure. The birch wood-derived samples have a

lower detector response compared to their pine wood counterparts. However, the low detector

response of the Hr 90-sample is also due to the much higher Mn, implying a higher viscosity

(η) of the sample solution compared to the other residual lignin samples. The injected volume

decreases with increasing η and thus a smaller amount is introduced into the capillary.

The black liquor has a broad mobility distribution (papers VI and VII). The pine wood-

derived black liquor samples are shown in Figure 32. The shape of these mobility

distributions is less distinct compared to the isolated lignin samples. Superimposed peaks are

present in the entire mobility distribution, which indicates that black liquor contains low

molecular weight components having phenol and possibly carboxyl groups. In contrast to the

46


isolated samples, the differences in detector response of the black liquor mainly reflect

different concentrations of lignin since these samples were diluted in the same way prior to

characterisation.

4 10 12 14 16 18 20 226 8


NM

Sb 75

Sb 135

Sb 195

1 mAU

24 26

Figure 32. Mobility distributions of black liquor collected in the beginning (Sb 75), in themiddle (Sb 135) and at the end (Sb 195) of a kraft cook of pine wood. Conditions as in Figure28.

4.2.4 Average charge density

Average electrophoretic mobilities have been calculated from the migration times of HS

(Schmitt-Kopplin et al., 1998b) by approximating the distribution of the mobility to be

Gaussian. Due to the differences in the profiles this procedure is more difficult to apply to the

studied lignin samples. To ease comparison of the mobility between the different types of

lignin samples, the average mobility (µav) of the distribution was determined after establishing

the centre of gravity of time-corrected mobility distributions (tg) (paper VI).

tg = ∫ft

t

i dttft0

)( / ∫ft

t

dttf0

)( [13]

where t0 is the initial and tf is the final migration time of the distribution, ti the time value of

each data point i and f(t) the corrected detector signal as a function of time. µav was

determined by replacing tr in the common equation [8] with tg

µav = lL/U (1/tg-1/tNM) [14]


47

Lignin only contains anions and thus µav will be negative, but is reported as |µav | in the

following.

The relation between µav and concentration of phenol groups agrees for the isolated pine

wood-derived samples and for the residual lignin isolated from the birch kraft pulp. Apart

from the Hr-samples, the µav cannot be predicted from the ratio between concentration of

charged groups and Mn. To compare the influence of M on mobility, similar conformations of

the samples must be assumed, which may not be true for the samples in this study,

considering the ir different origins and isolation methods. In addition the conformation of

acetylated lignin in THF probably differs from that of underivatised lignin in alkali. The

magnitude of the friction, i.e. the denominator in [5], depends on the geometry of the

molecule. For spheres, the friction is proportional to the cubic root of the molecular weight

(M), for random coiled polymers to the square root of the M (Giddings, 1991).

Cooking time (min)

40 60 80 100 120 140 160 180 200 220

Ave

rage

mob

ility

x 1

0-8 (

m2 /V

s)

2.6

2.8

3.0

3.2

3.4

3.6

3.8

black liquorresidual lignin

dissolved lignin

Figure 33. Average mobility (µav) at pH 12 of black liquor and isolated lignin samplescorresponding to different cooking times of pine wood kraft cook. µav was calculatedaccording to equation 14. The 95% confidence limits of the mean were calculated on everysample from triplicates.

48


Cooking time (min)

40 60 80 100 120 140 160 180 200 220

Ave

rage

mob

ility

x 1

0-8 (

m2 /V

s)

2.6

2.8

3.0

3.2

3.4

3.6

3.8

dissolved lignin

black liquor

residual lignin

Figure 34. Average mobility (µav) at pH 12 of black liquor and isolated lignin samplescorresponding to different cooking times of birch wood kraft cook. µav was calculatedaccording to equation 14. The 95% confidence limits of the mean were calculated on everysample from triplicates.

The dissolved lignin samples have the highest mobility at each time throughout the pine wood

cook. In the beginning of the cook the µav decreases in the order dissolved lignin>black

liquor>residual lignin for both wood sources. At the end of the pine wood cook, the average

charge densities of the Sb and Sr samples are similar, whereas the relation between the lignin

samples from birch wood kraft cooks are Hd≈Hr>Hb. The charge density of the black liquor

samples has about the same trend as the dissolved lignin samples although it is lower for the

former. According to equation 5, an increase in size decreases the mobility of a particular

solute. The lower mobility of black liquor compared to isolated dissolved lignin samples may

be due to associations between lignin and carbohydrate fragments.

According to these results, the residual lignin isolated from pine wood kraft pulp has a

considerably lower charge density than the lignin removed from the pulp. The introduction of

either phenol or carboxylic groups into pulp lignin would probably ease the removal of


49

residual lignin during the cook. Limited solubility of lignin fragments is one suggested

explanation to the declining delignification rate at the end of kraft cooking of birch

(Gellerstedt et al., 1988; Gellerstedt et al., 1994). Since the charge density is about the same

for Hr and the lignin removed from that pulp, the declining rate of delignification does not

seem to be explained by a limited solubility of the residual birch lignin. However, this

conclusion needs to be verified by using some other isolation method for the residual lignin.

4.2.5 Influence of pH on mobility

Differences in acidity of lignin fragments may influence their solubility and thus dissolution

from the fibre. The phenol groups in lignin do not have a common pKa-value (Lindberg,

1959; Shtreis and Nikitin, 1967; Woerner and McCarthy, 1987). Lignin-related phenol

groups differ in acidity over a wide range (pKa 6.2–11.3) depending on substitution (Ragnar

et al., 1999). For example, carbonyl groups formed in the lignin during pulping (Hortling et

al., 1997) may increase the acidity of the lignin, whereas a carbon substituent in position 5

(see Figure 5) would decrease the acidity of the phenol groups (Zakis, 1994). An extensive

dissociation of the phenol groups can be expected at pH 12 whereas at pH 10 the degree of

dissociation should be considerably lower. To study the effect of dissociation, the µav at pH 12

was compared to the µav at pH 10, Figure 35 and Figure 36. Since the ionic strength is higher

at pH 12 than at pH 10 in the electrolyte used, the shift in µav only reflects relative differences

in pKa. If the phenol groups in the studied samples have the same pKa-value the shape of the

µav graph at pH 12 should essentially be retained. This is true for the residual samples and the

black liquor samples from the pine wood kraft cook.

The main difference between the µav -graphs is caused by a larger decrease in mobility at pH

10 for the dissolved lignin samples (Sd/Hd) that correspond to the middle of the cook. This

may indicate a higher pKa-value for the phenol groups in the lignin fragments dissolved

during the bulk delignification phase, compared to the other samples. To determine the pKa of

lignin accurately, it is however important to keep the ionic strength constant and to measure

the mobility over a wider pH range.

50


Cooking time (min)

40 60 80 100 120 140 160 180 200 220

Ave

rage

mob

ility

x 1

0-8 (

m2 /V

s)

2.4

2.6

2.8

3.0

3.2

3.4

3.6

dissolved lignin

black liquorresidual lignin

Figure 35. Average mobility (µav) at pH 10 of black liquor and isolated lignin samplescorresponding to different cooking times of pine wood. µav calculated according to equation14. The 95% confidence limits of the mean were calculated on every sample from triplicates.

The shape of the µav -graph for the black liquor samples from birch cooking is also changed

when lowering the pH of the electrolyte. At pH 10, the relation between the average

mobilities is Hb>Hd≈Hr, i.e. different to that at pH 12. This means that the black liquor

samples contain lignin fragments of much higher charge density than the dissolved lignin

samples at pH 10. The modest decrease in mobility of the Hb samples indicate the presence of

acid groups, which are completely dissociated even at pH 10. Such groups may be carboxylic

acids in dissolved carbohydrates associated to dissolved lignin fragments. During the isolation

of lignin from black liquor, such highly polar components are removed.


51

Cooking time (min)

40 60 80 100 120 140 160 180 200 220 240

Ave

rage

mob

ility

x 1

0-8 (

m2 /V

s)

2.4

2.6

2.8

3.0

3.2

3.4

3.6

black liquor

dissolved lignin

residual lignin

Figure 36. Average mobility (µav) at pH 10 of black liquor and isolated lignin samplescorresponding to different cooking times of birch wood. µav calculated according to equation14. The 95% confidence limits of the mean were calculated on every sample from triplicates.

52



53

5 CONCLUSIONS

The composition of hemicellulose in kraft pulps influences the solubility in LiCl/DMAc and

the chromatographic behaviour in LiCl/DMAc-SEC. Hardwood kraft pulps are completely

dissolved in LiCl/DMAc, whereas softwood kraft pulps are only partially dissolved in this

solvent system. Hence, softwood kraft pulps characterised by LiCl/DMAc-SEC should be

evaluated with care since the composition of the dissolved fraction may not be representative

for the pulp. The hemicellulose (mainly xylan) and cellulose of hardwood kraft pulps are

fairly well separated. In contrast, the hemicelluloses of softwood pulps elute over the entire

molecular weight range, indicating various degrees of associations with cellulose. It is thus

more difficult to interpret differences in the MWDs of softwood kraft pulps than in those of

hardwood kraft pulps. In absence of associations between cellulose and hemicellulose,

cellulosic solutions of LiCl/DMAc consist of cellulose aggregates. According to the MWD

profiles and light scattering measurements, it is possible to break these aggregates by

mechanical treatment of the solutions, without severe cleavage of the glycosidic bonds. The

complete solubility and better separation of the carbohydrate polymers makes hardwood kraft

pulps suitable to use when the effect of different types of degradation are to be studied by

LiCl/DMAc-SEC. The MWD-profiles and strength of degraded hardwood kraft pulps depend

on type of degradation. The MWD of pulps degraded with oxygen/alkali or alkali indicate that

there are certain positions in the cellulose chains that are cleaved during treatment. The

hemicellulose fraction is more difficult to evaluate due to the chromatographic interference

with cellulose, as the degradation becomes severe. The observed decrease in shape

factor/fibre strength of pulps degraded with oxygen/alkali or alkali is proposed to be due to a

combination of heterogeneous degradation and removal of hemicellulose whereas the

decrease in Mw of cellulose is of minor importance.

CZE is well suited for obtaining information about the charge density distribution of lignin.

By using glycine-sodium hydroxide as electrolyte at pH≥ 10 electrostatic adsorption onto the

capillary wall can be avoided and underivatised samples can be studied using high voltages.

The mobility distribution, i.e charge density, differs between black liquor, isolated dissolved

lignin and isolated residual lignin. The proposed method for determining the average mobility

(µav) of the distribution makes it easier to compare the charge density of different types of

lignin samples. The µav measured at pH 12 reveals that the residual lignin isolated from pine

wood kraft pulp has a significantly lower charge density than the lignin removed from the

54


pulp during the cook. This implies that introduction of charged groups into the pulp lignin

derived from pine should ease its extraction from the pulp. At the end of a kraft cook of birch,

the µav of the isolated residual lignin is about the same as that of the isolated dissolved lignin,

which suggests that the solubility is sufficient for pulp lignin to be dissolved. This needs

however to be confirmed using other isolation methods for residual lignin, since the applied

method probably leads to an overestimation of the charge density of the pulp lignin.

Comparisons between the µav at pH 12 and pH 10 indicate that the isolated dissolved lignin

samples obtained in the middle of the cook have a lower acidity than the other samples. The

difference in µav of black liquor compared to isolated dissolved lignin may be due to

associations between lignin fragments and carbohydrate polymers in the black liquor. CZE

may thus be used in studies concerning associations between carbohydrates and lignin.

This thesis demonstrates the importance of understanding the interplay between complex

samples and applied analytical method to interpret the results properly. The findings

presented provide a basis for using LiCl/DMAc-SEC and CZE for studies concerning, e.g.,

the

• interactions between the pulp polymers

• delignification/extraction of lignin from the pulp

• relation between pulp properties and composition of the constituting polymers

in order to improve the kraft pulping process.


55

6 ACKNOWLEDGEMENTS

I would like to thank the following without whose help this thesis would not be possible:

My supervisor Anders Colmsjö for his continuous support, sharp mind and humour,

Göran Gellerstedt for always keeping his door open for discussions,

My "chief" Mikael Lindström for support, advice and encouragement,

Kristina Gustafsson for being a sincere friend, for professional skill and for invaluable

support throughout this work, in spite of the winding and harsh road,

Erik "LZ" Norman for his friendly and calm attitude, pulping knowledge, criticism and

occasional co-operation which has been a reliable source of labour, laughs, "que" and !

Fredrik Berthold for his bright and open mind, valuable criticism and for sharing my

interest about the chemistry of pulp fibres - I really appreciate our discussions!

Bert Pettersson for taking part in paper I and for being such a nice guy,

Wyn Brown, University of Uppsala, for light scattering measurements and valuable

discussions about cellulose solutions,

Bengt Eriksson for his educational and mathematical skills, especially his

"deconvoluting" contribution to paper III,

Nils-Olof Nilvebrandt, for valuable criticism and friendly consultations

Störker Moe, University of Trondheim, for 1H-NMR measurements,

Olof Dahlman for valuable criticism of this thesis,

Christina "Lucky Luke" Hörnell for her careful linguistic revision and pleasant e-mails,

Former and present colleagues for sharing their professional skill with me, special

thanks are due to Britt Ahlstedt, Stig Andersson, Nermina Begovic, Rickard Berggren,

Dag Ehrengård, Ylva Frisk, Eva Hellberg, Asha Ismail, Kristina Isberg, Helena

Lennholm, Eva-Lisa Lindfors, Roland Mörck, Annika Rydlund and Therese Wurgler,

Filipe de Sousa and Lillemor Johansson with whom I have had the pleasure of

discussing various aspects of life during our precious lunchtime,

To all my friends in the Small cups and in the BH for enjoyable happenings,

My sisters and "bother" for tough discussions, care and for being such lovely pains,

My parents for always supporting me in whatever I am doing - I love you!

Matz, Jonas and Anna for their patience with junk-food and my mental absence, and for

bringing joy to my life - I definitely love you!

Finally, I would like to express my gratitude to STFI for making it possible for this thesis

come true.

56



57

7 REFERENCES

Adler, E. (1977). Lignin chemistry-Past, Present and Future. Wood Sci. Technol., 11, 169-218.

Ahlgren, P.A., Yean, W.Q. and Goring, D.A.I. (1971). Chlorite delignification of sprucewood. Comparison of the molecular weight of the lignin dissolved with the size of pores inthe cell wall. Tappi, 54(5), 737-740.

Atalla, R.H. and VanderHart, D.L. (1984). Native cellulose: A composition of two distinctcrystalline forms. Science, 223, 283-285.

Austin, P.R. (1977). Chitinlösung und deren Verwendung. Deutsches Patentamt,Offenlegungsschrift 2,707,164.

Bao, Y.T., Bose, A., Ladish, M.R. and Tsao, G.T. (1980). New approach to aqueous gelpermeation chromatography of nonderivatized cellulose. J. Appl. Polym. Sci., 25, 263-275.

Barth, H.G. (1987). Nonsize Exclusion Effects in High-Perfomance Size ExclusionChromatography. In Chemistry of High-Temperature Superconductors, eds., D.L. Nelson,M.S. Whittingham and T.F. George. ACS Symp. Series, 352, American Chemical Society,Washington, D.C., pp. 29-46.

Bergnor-Gidnert, E., Tomani, P.E. and Dahlman, O. (1998). Influence on pulp quality ofconditions during the removal of hexeneuronic acids. Nord. Pulp Pap. Res. J.,13(4), 310-316.

Boucai, E. (1971). Zero-span tensile test and fibre strength. Pulp Paper Mag. Can., 72(10),73-80.

Brown, W. (1967). The cellulose solvent cadoxen. Svensk Papperstidn., 70(15), 458-461.

Buchert, J., Teleman, A., Harjunpää, V., Tenkanen, M., Viikari, L. and Vourinen, T. (1995).Effect of cooking and bleaching on the structure of xylan in conventional pine kraft pulp.Tappi J., 78(11), 125-130.

Chen, C.-L. (1992). Determination of carbonyl groups. In Methods in lignin chemistry, eds.S.Y. Lin and C.W. Dence. Springer-Verlag, Berlin Heidelberg, Germany, pp. 446-457.

Comtat, J. and Joseleau, J.-P. (1981). Mode of action of a xylanase and its significance for thestructural investigation of the branched L-arabino-D-glucurono-D-xylan from redwood.Carbohydr. Res., 95, 101-112.

Dahl, C.F. (1884). Process of manufacturing cellulose from wood. US Patent 296,935.

Dahlman, O. and Månsson, K. (1996). Analysis of low molecular weight lignin-derivedsulphonates by capillary electrophoresis. J. Wood Chem. Technol., 16(1), 47-60.

Dawsey, T.R. and McCormick, C.L. (1990). The lithium chloride/dimethylacetamide solventfor cellulose: A literature review. Rev. Macromol. Chem. Phys., C30(384), 405-440.

58


Dea, I.C.M., Morris, E.R., Rees, D.A., Welsh, E.J., Barnes, H.A. and Price, J. (1977).Associations of like and unlike polysaccharides: Mechanism and specificity ingalactomannans, interacting bacterial polysaccharides and related systems. Carbohydr. Res.,57, 249-272.

Dea, I.C.M., Clark, A.H. and McCleary, B.V. (1986). Effect of galactose-substitution-patternson the interaction properties of galactomannans. Carbohydr. Res., 147, 275-294.

Eaton, A.K. (1870). Improvement in the manufacture of paper pulp. US Patent 106,143.

Eaton, A.K. (1871). Improvement in the manufacture of paper pulp. US Patent 119,224.

Ek, M., Gierer, J., Jansbro, K. and Reitberger, T. (1989). Study on the selectivity of bleachingwith oxygen-containing species. Holzforschung, 43(6), 391-396.

Ekman, K.H. and Lindberg, J.J. (1960). On the origin of the infra-red bands in the 1720 cm-1

region in lignins. Pap. Puu, 42(1), 21-22.

Ekmanis, J.L. (1986). Gel permeation chromatographic analysis of cellulose. Polymer Notes,Waters Chromatography Division, Millipore Corp., Milford, Massachusetts, US, 1(3).

Enkvist, T. (1961). Work carried out in Helsingfors concerning useful products of lignin. Pap.Puu, 43(11), 657-662.

Eriksson, K.-E., Johanson, F. and Pettersson, B. (1966). Molecular weight fraction ofcellulose by gel filtration chromatography. Svensk Papperstidn., 70(19), 610-611.

Evans, R. and Wallis, A.F.A. (1987). Comparison of cellulose molecular weights determinedby high performance size exclusion chromatography and viscometry. In Proc. Fourth Int.Symp. Wood and Pulping Chemistry, Paris, April 27-30, Vol. 1, pp. 201-205.

Evans, R., Wearne, R. and Wallis, A.F.A. (1989). Molecular weight distribution of celluloseas its tricarbanilate by high performance size exclusion chromatography. J. Appl. Polym. Sci.,37, 3291-3303.

Fergus, B.J. and Goring, D.A.I. (1969). The topochemistry of delignification in kraft andneutral sulphite pulping of birch wood. Pulp Paper Mag. Can., 70, T314.

Fergus, B.J. and Goring, D.A.I. (1970). The distribution of lignin in birch wood as determinedby ultraviolet microscopy. Holzforschung, 24(4), 118-124.

Fetsch, D. and Havel, J. (1998). Capillary zone electrophoresis for the separation andcharacterization of humic acids. J. Chromatogr. A, 802, 189-202.

Froass, P., Ragauskas, A.J. and Jiang, J.-er. (1996). Chemical structure of residual lignin fromkraft pulp. J. Wood Chem. Technol., 16(4), 347-365.

Fujiwara, S. and Honda, S. (1986). Determination of cinnamic acid and its analogues byelectrophoresis in a fused silica capillary tube. Anal. Chem., 58, 1811-1814.


59

Gellerstedt, G. and Lindfors, E.-L. (1984). Structural changes in lignin during kraft pulping.Holzforschung, 38(3), 151-158.

Gellerstedt, G., Gustafsson, K. and Northey, R.A. (1988). Structural changes in lignin duringkraft cooking. Part 8. Birch lignins. Nordic Pulp Paper Res. J., 2, 87-94.

Gellerstedt, G., Pranda, J. and Lindfors, E.-L. (1994). Structural and molecular properties ofresidual birch kraft lignins. J. Wood Chem. Technol., 14(4), 467-482.

Gellerstedt, G. and Li, J. (1996). An HPLC method for the quantitative determination ofhexeneuronic groups in chemical pulps. Carbohydr. Res., 294, 41-51.

Giddings, J.C. (1991). Friction coefficients and molecular parameters. In Unified SeparationScience, John Wiley & Sons, Inc., New York, pp. 77-79.

Gierer, J. (1980). Chemical aspects of kraft pulping. Wood Sci. Technol., 14, 241-266.

Gierer, J. and Wännström, S. (1984). Formation of alkali-stable C-C bonds between ligninand carbohydrate fragments during kraft pulping. Holzforschung, 38(4), 181-184.

Gierer, J. (1990). Basic principles of bleaching. Part 1: Cationic and radical processes.Holzforschung, 44(5), 387-394.

Goring, D.A.I., Webb, T. and Sehon, A.H. (1958). Physicochemical studies of ligninsulphonates. Part 4. Paper and free electrophoresis. Svensk Papperstidn., 61(23), 1010-1015.

Gurnagul, D. and Page, D.H. (1989). The difference between dry and rewetted zero-spantensile strength of paper. Tappi J., 72(12), 164-167.

Gurnagul, D., Page, D.H. and Paice, M.G. (1992). The effect of cellulose degradation on thestrength of wood pulp fibres. Nord. Pulp Pap. Res. J., 7(3), 152-154.

Hansson, J.-Å. (1970). Sorption of hemicelluloses on cellulose fibres. Part 2. Sorption ofglucomannan. Holzforschung, 24(3), 77-83.

Hansson, J.-Å. and Hartler, N. (1969). Sorption of hemicelluloses on cellulose fibres. Part I.Sorption of xylans. Svensk Papperstidn., 72(17), 521-530.

Hardell, H.-L., Leary, G.J., Stoll, M. and Westermark, U. (1980a). Variations in ligninstructure in defined morphological parts of birch. Svensk Papperstidn., 83(3), 71-74.

Hardell, H.-L., Leary, G.J., Stoll, M. and Westermark, U. (1980b). Variations in ligninstructure in defined morphological parts of spruce. Svensk Papperstidn., 83(2), 71-74.

Havlicek, J. and Samuelson, O. (1972). Chromatography of oligosaccharides from xylan byvarious techniques. Carbohydr. Res., 22, 307-316.

Hiermann, A. and Radl, B. (1998). Analysis of aromatic plant acids by capillaryelectrophoresis. J. Chromatogr. A, 803, 311-314.

60


Hjertén, S. (1967). Free zone electrophoresis. Chromatogr. Rev., 9, 122-219.

Hjertén, S. and Liao, J.L. (1995). Electrophoresis in low conductivity buffers. US Patent5,464,517.

Hortling, B., Tamminen, T., Tenkanen, M., Teleman, A. and Pekkala, O. (1995).Investigations of residual lignin and its interaction with polysaccharides in chemical pulps. InIn Proc. Eigth Int. Symp. Wood and Pulping Chemistry, Helsinki, Finland, 6-9 June, Vol.1,pp. 231-238.

Hortling, B. Tamminen, T. and Kenttä, E. (1997). Determination of carboxyl and non-conjugated carbonyl groups in dissolved and residual lignins by IR spectroscopy.Holzforschung, 51(5), 405-410.

Huang, X., Coleman, W.F. and Zare, R.N. (1989). Analysis of factors causing peakbroadening in capillary zone electrophoresis. J. Chromatogr., 480, 95-110.

Iwata, T., Indrarti, L. and Azuma, J.-I. (1998). Affinity of hemicellulose for celluloseproduced by Acetobacter xylinum. Cellulose, 5, 215-228.

Jiang, Z.-H. and Argyropoulos, D.S. (1999). Isolation and characterization of residual ligninsin kraft pulps. J. Pulp Paper Sci., 25(1), 25-29.

Johansson, M.H. and Samuelson, O. (1975). End-wise degradation of hydrocellulose duringhot alkali treatment. J. Appl. Poly. Sci., 19, 3007-3013.

Johansson, M.H. and Samuelson, O. (1977a). Reducing end groups in birch xylan and theiralkaline degradation. Wood Sci. Technol., 11, 251-263.

Johansson, M.H. and Samuelson, O. (1977b). Epimerization and degradation of 2-O-(4-O-methyl-α-D-glucopyranosyluronic acid)-D-xylitol in alkaline medium. Carbohydr. Res., 54,295-299.

Jorgenson, J.W. and Luckacs, K.D. (1981a). Zone electrophoresis in open-tubular glasscapillaries. Anal. Chem., 53, 1298-1302.

Jorgenson, J.W. and Luckacs, K.D. (1981b). High-resolution separations based onelectrophoresis and electroosmosis. J. Chromatogr., 218, 209-216.

Karlsson, H. and Fransson, P.-I. (1994). STFI FiberMaster ger pappersmakarna nya muskler.Svensk Papperstidn., 97(10), 26-28.

Karlsson, O. and Westermark, U. (1994). Condensation reactions between wood polymersduring kraft pulping. In Proc. Tappi Pulping Conf., San Diego, Vol. 1, pp. 1-4.

Kennedy, J.F., Rivera, Z.S., White, C.A., Lloyd, L.L. and Warner, F.P. (1990). Molecularweight characterization of underivatized cellulose by GPC using lithium chloride-dimethylacetamide solvent system. Cellulose Chem. Technol., 24, 319-325.


61

Kerr, A.J. and Goring, D.A.I. (1975). The role of hemicellulose in the delignification ofwood. Can. J. Chem., 53, 952-959.

Kitagishi, K. (1997). Quantitative analysis, precision and reproducibility. In Handbook ofcapillary electrophoresis applications, eds. H. Shintani and J. Polonský, Chapman and Hall,London, pp. 25-55.

Kleen, M. (1993). In Characterization of wood and pulp using analytical pyrolysis andmultivariate data analysis. PhD thesis, Royal Institute of Technology, Stockholm, Sweden, p.7.

Kleinert, T.D. (1966). Mechanism of alkaline delignification. I. The overall reaction pattern.Tappi J., 49(2), 53-57.

Krässig, H.A. (1996). The fibre structure. In Cellulose. Structure, accessibility and reactivity,Gordon and Breach Science publishers, Singapore, pp. 6-42.

Kuga, S., Takagi, S. and Brown, Jr. R.M. (1993). Native folded-chain cellulose II. Polymer,34(15), 3293-3297.

Kvernheim, A.L. and Lystad, E. (1989). Size-exclusion chromatography and methylationanalysis of cellulose in N,N-dimethylacetamide/LiCl. Acta Chem Scand., 43, 209-211.

LaPierre, L. and Bouchard, J. (1997). Molecular weight determination of softwood kraftcellulose. In Proc. Ninth Int. Symp. Wood and Pulping Chemistry, Montreal, Canada, June 9-12, Vol. 1, pp. L6-1-L6-6.

Lémon, S. and Teder, A. (1973). Kinetics of the delignification in kraft pulping. I. Bulkdelignification of pine. Svensk Papperstidn., 72(11), 407-414.

Li, J. (1999). Practical utilization in chemical pulping. In Towards an accurate determinationof lignin in chemical pulps, PhD thesis, Royal Institute of Technology, Stockholm, Sweden,1999, pp. 52-60.

Li, S. and Lundquist, K. (1994). A new method for the analysis of phenolic groups in ligninsby 1H NMR spectrometry. Nordic Pulp Paper Res. J., 9(3), 191-195.

Lind, J., Merenyi, G. and Nilvebrant, N.-O. (1997). Hydroxyl radical induced viscosity loss incellulose fibres. J. Wood Chem. Technol., 17(1&2), 111-117.

Lindberg, J.J. (1959). Über einige Versuchen zur Charakterisierung von Wasserstoffbrückender Lignine. Finska Kemists. Medd., 68(1), 5-17.

62


Lindgren, B.O. (1958). The lignin-carbohydrate linkage. Acta Chem. Scand., 12, 447-452.

Lindquist, A. and Dahlman, O. (1998). Characterization of pulp hemicelluloses usingMALDI-TOF-MS. In Proc. Fifth European Workshop on Lignocellulosics and Pulp, 30 Aug-2 September, Aveiro, Portugal, pp. 483-486.

Ljunggren, S. (1980). The significance of aryl ether cleavage in kraft delignification ofsoftwood. Svensk Papperstidn., 83(13), 363-369.

Lukacs, K.D. and Jorgenson, J.W. (1985). Capillary zone electrophoresis: effect of physicalparameters on separation efficiency and quantitation. J. High Res. Chromatogr. Chromatogr.Communications., 8, 407-411.

Lundquist, K. (1973). Acid degradation of lignin. Part VIII. Low molecular weight phenolsfrom acidolysis of birch lignin. Acta Chem. Scand., 27, 2597-2606.

Lundquist, K. and Lundgren, R. (1972). Acid degradation of lignin. Part VII. The cleavage ofether bonds. Acta Chem. Scand., 26, 2005-2023.

Löwendahl, L., Petersson, G. and Samuelson, O. (1978). Phenolic compounds in kraft blackliquor. Svensk Papperstidn., 81(12), 392-396.

Maman, O., Marseille, F., Guillet, B., Disnar, J.-R. and Morin, P. (1996). Separation ofphenolic aldehydes, ketones and acids from lignin degradation by capillary zoneelectrophoresis. J. Chromatogr. A, 755, 89-97.

Marton, J. and Adler, E. (1963). Reactions of lignin with methanolic hydrochloric acid. Adiscussion of some structural questions. Tappi, 46(2), 92-98.

Masselter, S., Zemann, A. and Bobleter, O. (1995). Analysis of lignin degradation products bycapillary electrophoresis. Chromatographia, 40(1/2), 51-57.

McCleary, B.V. (1990). Comparison of endolytic hydrolases that depolymerize 1,4-β-D-mannan, 1,5-α-L-arabinan and 1,4-β-D-galactan. In Enzyme in biomass conversion, eds. G.F.Leatham and M.E. Himmel. ACS Symp. Ser. 460, American Chemical Society, Washington,D.C., pp. 437-449.

Malawer, E.G. (1995). Introduction to size exclusion chromatography. In Handbook of sizeexclusion chromatography, ed. C.-s. Wu. Marcel Dekker, Inc., New York, pp.1-24.

McCormick, C.L. and Lichatowich, D.K. (1979). Homogeneous solution reactions ofcellulose, chitin and other polysaccharides to produce controlled-activity pesticide systems. J.Polym. Sci., Polym. Lett. Ed., 17, 479-484.

McCormick, C.L. (1981). Novel cellulose solutions, U.S. Patent 4,278,790

McCormick, C.L., Callais, P.A. and Hutchinson, B.H. Jr. (1985). Solution studies of cellulosein lithium chloride and N,N-dimethylacetamide. Macromol., 18, 2394-2401.


63

Meier, H. (1961). The distribution of polysaccharides in wood fibres. J. Polym. Sci., 51, 11-18.

Meier, H. (1985). Localization of polysaccharides in wood cell walls. In Biosynthesis andbiodegradation of wood polysaccharides, ed. T. Higucci. Academic Press, London, pp. 43-50.

Meshitsuka, G. and Nakano, J. (1985). Structural characteristics of compound middle lamellalignin. J. Wood Chem. Technol., 5(3), 391-404.

Mikkers, F.E.P., Everearts, F.M. and Verheggen, Th.P.E.M. (1979). Concentrationdistributions in free zone electrophoresis. J. Chromatogr., 169, 1-10.

Mohlin, U.-B. and Alfredsson, C. (1990). Fibre deformation and its implications in pulpcharacterization. Nordic Pulp Paper Res. J., 5(4), 172-179.

Mohlin, U.-B., Dahlbom, J. and Hornatowska, J. (1996). Fibre deformation and sheetstrength. Tappi J., 79(6), 105-111.

Moore, J.C. (1964). Gel permeation chromatography I. A new method for molecular weightdistribution of high polymers. J. Polym. Sci., Part A, 2, 835-843.

Morgenstern, B., Kammer, H.-W., Berger, W. and Skrabal, P. (1992). 7Li-NMR study oncellulose/LiCl/N,N-dimethylacetamide solutions. Acta Polym., 43, 356-357.

Morgenstern, B. and Kammer, H.-W. (1996). Solvation in cellulose-LiCl-DMAc solutions.Trends Polym. Sci., 4(3), 87-92.

Månsson, P. (1983). Quantitative determination of phenolic and total hydroxyl groups inlignins. Holzforschung, 37, 143-146.

Mörck, R., Yoshida, H. A., Kringstad, K. and Hatakeyama, H. (1986). Fractionation of kraftlignin by successive extraction with organic solvents. I. Functional groups, 13C-NMR-spectraand molecular weight distribution. Holzforschung, 40, 51-60.

Nehls, I., Wagenknecht, W. and Philipp, B. (1995). 13C-NMR spectroscopic studies ofcellulose in various solvent systems. Cellulose Chem. Technol., 29, 243-251.

Oda, R.P. and Landers, J.P. (1997). Introduction to capillary electrophoresis. In Handbook ofcapillary ellectrophoresis, ed. J.P. Landers. CRC Press LLC, Florida, pp. 1-47.

Olm, L. And Tistad, G. (1979). Kinetics of the initial stage of kraft pulping. SvenskPapperstidn., 82(15), 458-464.

Overbeeke, N., Fellinger, A.J. and Termorshuizen, G.H.M. (1987). Production of an α-galactosidase enzyme for modification of guar gum. In Proc. Fourth Eur. Congr. Biotechnol.,eds. O.M. Neijssel, R.R. van der Meer and K.Ch.A.M. Luyben, Elsevier Publishers B.V.,Amsterdam, Netherlands, Vol. 1, p. 509.

Page, D.H., El-Hosseiny, F., Winkler, K. and Bain, R. (1972). The mechanical properties ofsingle wood-pulp fibres. Part I: A new approach. Pulp Paper Mag. Can., 73(8), 72-77.

64


Page, D.H. (1983). The origin of the difference between sulphite and kraft pulps. J. PulpPaper Sci., 9(3), TR15-20.

Page, D.H., Seth, R.S. and El- Hosseiny, F. (1985). Strength and chemical composition ofwood pulp fibres. In Papermaking Raw Materials, ed. V. Punton. Mechanical EngineeringPublications Limited, London, pp. 77-91.

Porath, J. and Flodin, P. (1959). Gel filtration: A method for desalting and group separation.Nature, 183, 1657-1659.

Ragnar, M., Lindgren, C.T. and Nilvebrant, N.-O. (1999). On the dissociation constants ofphenolic groups in lignin structure. In Proc. Tenth Int. Symp. Wood and Pulping Chemistry,Yokohama, Japan, 7-10 June, Vol. 2, pp. 154-157.

Reijenga, J.C., Verheggen, T.P.E.M., Martens, J.H.P.A. and Everaerts, F.M. (1996). Buffercapacity, ionic strength and heat dissipation in capillary electrophoresis. J. Chromatogr. A,744, 147-153.

Rigol, A., López-Sánchez, J.F. and Rauret, G. (1994). Capillary zone electrophoresis ofhumic acids. J. Chromatogr. A, 664, 301-305.

Rigol, A., Vidal, M. and Rauret, G. (1998). Ultrafiltration-capillary zone electrophoresis forthe determination of humic acid fractions. J. Chromatogr. A, 807, 275-284.

Robert, D.R., Bardet, M., Gellerstedt, G. and Lindfors, E.-L. (1984). Structural changes inlignin during kraft cooking. Part 3. On the structure of dissolved lignin. J. Wood Chem.Technol., 4(3), 239-263.

Rosell, K.-G. and Svensson, S. (1975). Studies of the distribution of the 4-O-methyl-D-glucuronic acid residue in birch xylan. Carbohydr. Res., 42, 297-304.

Rånby, B. (1952). The physical characteristics of alpha-, beta- and gamma-cellulose. SvenskPapperstidn., 55(4), 115-124.

Saka, S. and Goring, D.A.I. (1988). The distribution of lignin in white birch wood asdetermined by bromination with TEM-EDXA. Holzforschung, 42(3), 149-153.

Salmén, L. and Olsson, A.-M. (1998). Interaction between hemicelluloses, lignin andcellulose: Structure-property relationships. J. Pulp Paper Sci., 24(3), 99-102.

SCAN-C11:75 (1975). Brightness. In SCAN Test method, Secretariat, Scandinavian Pulp,Paper and Board Testing Committee, Box 5604, SE-114 86 Sockholm, Sweden.

SCAN-C 1:77 (1977). Kappa number. In SCAN Test method, Secretariat, Scandinavian Pulp,Paper and Board Testing Committee, Box 5604, SE-114 86 Sockholm, Sweden.

SCAN-CM 15:88 (1988). Viscosity in cupriethylenediamine solution, In SCAN Test method,Secretariat, Scandinavian Pulp, Paper and Board Testing Committee, Box 5604, SE-114 86Sockholm, Sweden.


65

Schmitt-Kopplin, P., Hertkorn, N., Schulten, H.-R. and Kettrup, A. (1998a). Structuralchanges in a dissolved soil humic acid during photochemical degradation processes under O2

and N2 atmosphere. Environ. Sci. Technol., 32, 2531-2541.

Schmitt-Kopplin, Ph., Garrison, A.W., Perdue, E.M., Freitag, D. and Kettrup, A. (1998b).Capillary electrophoresisin the analysis of humic substances. Facts and artifacts. J.Chromatogr. A, 807, 101-109.

Schevenko, S.M. and Bailey, G.W. (1996). Life after death: lignin-humic relationshipsreexamined. Crit. Rev. Environ. Sci. Technol., 26(2), 95-153.

Schroeder, L.R. and Haigh, F.C. (1979). Cellulose and wood pulp polysaccharides. Gelpermeation chromatographic analysis. Tappi, 62(10), 103-105.

Schwald, W. and Bobleter, O. (1988). Characterization of nonderivatized cellulose by gelpermeation chromatography. J. Appl. Polym. Sci., 35, 1937-1944.

Schöning, A.G. and Johansson, G. (1965). Absorptiometric determination of acid-solublelignin in semichemical bisulfite pulps and in some woods and plants. Svensk Papperstidn.,68(18), 607-613.

Shimizu, K., Hashi, M. and Sakurai, K. (1978). Isolation from a softwood xylan ofoligosaccharides containing two 4-O-methyl-D-glucuronic acid residues. Carbohydr. Res., 62,117-126.

Shtreis, G.B. and Nikitin, V.M. (1967). Zh. Prikl. Khim., 40(8), 1814-1819. (In Russian).

Silva, A.A. and Laver, M.L. (1997). Molecular weight characterization of wood pulpcellulose: Dissolution and size exclusion chromatographic analysis. Tappi J., 80(6), 173-180.

Sjöholm, E., Gustafsson, K. and Colmsjö, A. (1999a). J. Liq. Chromatogr. Relat. Technol.,22(11), 1663-1685.

Sjöholm, E., Gustafsson, K. and Colmsjö, A. (1999b). J. Liq. Chromatogr. Relat. Technol.,22(18). In press.

Sjöström, E. and Enström, B. (1966). Spectrophotometric determination of the residual ligninin pulp after dissolution in cadoxen. Svensk Papperstidn., 69(15), 469-476.

Sjöström, E. (1993a). The structure of wood. In Wood Chemistry. Fundamentals andApplications, Academic Press, San Diego, California, USA, pp. 1-20.

Sjöström, E. (1993b). Appendix. In Wood Chemistry. Fundamentals and Applications,Academic Press, San Diego, California, USA, p. 249.

Sjöström, E. (1993c). Wood polysaccharides. In Wood Chemistry. Fundamentals andApplications, Academic Press, San Diego, California, USA, pp. 51-70.

66


Sjöström, E. (1993d). Lignin. In Wood Chemistry. Fundamentals and Applications, AcademicPress, San Diego, California, USA, pp. 71-91.

Sjöström, E. (1993e). Wood pulping. In Wood Chemistry. Fundamentals and Applications,Academic Press, San Diego, California, USA, pp. 114-165.

Stevens, M.P. (1990). Molecular weight and polymer solutions. In Polymer chemistry. Anintroduction. 2nd edition, Oxford University Press, New York, USA, pp. 40-69.

Striegel, A.M. and Timpa, J. (1995). Molecular characterization of polysaccharides dissolvedin Me2NAc-LiCl by gel-permeation chromatography. Carbohydr. Res., 267, 271-290.

Striegel, A.M. (1997). Theory and applications of DMAC/LiCl in the analysis ofpolysaccharides. Carbohydr. Polym. 34, 267-274.

Tappi Standard, T 203 om-93, (1993). Alpha-, beta- and gamma-cellulose in pulp. In Tappitest methods, 1998-1999, Tappi press, Technology Park, Atlanta, USA.

Teleman, A. Harjunpää, V., Tenkanen, M., Buchert, J., Hausalo, T. Drakenberg, T. andVuorinen T. (1995). Characterisation of 4-deoxy-β-L-threo-hex-4-enopyranosyluronic acidattached to xylan in pine kraft pulp and pulping liquor by 1H and 13C NMR spectroscopy.Carbohydr. Res., 272, 55-71.

Terbojevich, M., Cosani, A., Conio, G., Ciferri, A. and Bianchi, E. (1985). Mesophaseformation and chain rigidity in cellulose and derivatives. 3. Aggregation of cellulose in N,N-dimethylacetamide-lithium chloride. Macromol., 18, 640-646.

Theander, O. and Westerlund, E.A. (1986). Studies on dietary fiber. 3. Improved proceduresfor analysis of dietary fiber. J. Agric. Food Chem., 34, 330-336.

Thornton, J., Tenkanen, M., Ekman, R., Holmbom, B. and Viikari, L. (1994). Possibility ofincreasing mechanical pulp yield by enzymatic treatment. Holzforschung, 48, 436-440.

Timpa, J.D. (1991). Application of universal calibration in gel permeation chromatographyfor molecular weight determinations of plant cell wall polymers: Cotton fiber. J. Agric. FoodChem., 39, 270-275.

Turbak, A.F., El-Kafrawy, A., Snyder, F.W. and Auerbach, A.B. (1981). Solvent system forcellulose. U.S. Patent 4,302,252.

Turbak, A.F., (1983). Newer cellulose solvent systems. In Wood and Agricultural Residues,ed. E.J. Soltes. Academic Press, New York, pp. 87-99.

Turbak, A.F. (1984). Recent developements in cellulose solvent systems. Tappi J., 67(1), 94-96.

VanderHart, D.L. and Atalla, R.H. (1984). Studies of microstructure in native celluloses usingsolid state 13C NMR. Macromolecules, 17, 1465-1472.


67

Viikari, L. Kantelinen, A., Rättö, M. and Sundquist, J. (1990). Enzymes in pulp and paperprocessing. In Enzyme in biomass conversion., eds. G. Leatham and M. Himmel. ACS Symp.Ser. 460, American Chemical Society, Washington, D.C., pp. 13- 21.

Vindevogel, J. and Sandra, P. (1991). Simultaneous pH and ionic strength effects and bufferselection in capillary electrophoretic techniques. J. Chromatogr., 541, 438-488.

Virtanen, R. (1974). Zone electrophoresis in a narrow-bore tube employing potentiometricdetection. A theoretical and experimental study. Acta Polytechnica Scand., 123, 1-67.

Westermark, U. and Gustafsson, K. (1994). Molecular size distribution of wood polymers inbirch kraft pulps. Holzforschung, 48, 146-150.

Whiting, P. and Goring, D.A.I. (1981). The topochemistry of delignification shown bypulping middle lamella and secondary wall tissue from black spruce. J. Wood Chem.Technol., 1(2), 111-122.

Whiting, P. and Goring, D.A.I. (1982). Chemical characterization of tissue fractions from themiddle lamella and secondary wall of black spruce tracheids. Wood Sci. Technol., 16, 261-267.

Whitney, S.E.C., Brigham, J.E., Darke, A.H., Reid, J.S. and Gidley, M.J., (1998). Structuralaspects of the interaction of mannan-based polysaccharides with bacterial cellulose.Carbohydr. Res., 307, 299-309.

Wilder, H.D. and Dalseki, E.J. (1965). Delignification rate studies. Part II of a series on kraftpulping kinetics. Tappi J., 48(5), 293-297.

Wilson, K., Ringström, E. and Hedlund, I. (1952). The alkali solubility of pulp. SvenskPapperstidn., 55(2), 31-37.

Woerner, D.L. and McCarthy, J.L. (1987). Ultrafiltration and light scattering evidence forassociation of kraft lignins in aqueous solutions. In Proc. Fourth Int. Symp. Wood andPulping Chemistry, Paris, 27-30 April, Vol. 1, pp.71-76.

Wood, B.F., Conner, A.H. and Hill, C.G. (1986). The effect of precipitation on the molecularweight distribution of cellulose tricarbanilate. J. Appl. Polym. Sci., 32, 3703-3712.

Yang, J.-M. and Goring, D.A.I. (1980). The phenolic hydroxyl content of lignin in sprucewood. Can. J. Chem., 58, 2411-2414.

Yllner, S. and Enström, B. (1956). Studies of the adsorption of xylan on cellulose fibresduring the sulphate cook. Svensk Papperstidn., 59(6), 229-232.

Zakis, G.F. (1994). Hydroxyl groups. In Functional analysis of lignins and their derivatives,Tappi Press, Atlanta, Georgia, p. 13.

68



69

APPENDIX I: List of abbreviations

BHP bleached hardwood kraft pulpBSP bleached softwood kraft pulpCE capillary electrophoresisCED cupriethylenediamineCL cotton linterCZE capillary zone electrophoresisD diffusion coefficientDMAc N,N-dimethylacetamideDP degree of polymerisationDPw weight-average DPEOF electroosmotic flowGFC gel filtration chromatographyGly glycineGPC gel permeation chromatographyHexA 4-deoxyhex-4-enuronic acidHP unbleached hardwood kraft pulpHS humic substancesK- Kappa numbernumberL total length of the capillaryl length from injector site to the

detectorM molecular weightMn number-average molecular

weightMp peak molecular weightMw weight-average molecular

weightMDC minimum detectable

concentrationMeGlcA 4-O-Me-α-D-glucuronic acidML middle lamella (see Figure 1)MWD molecular weight distributionN number of theoretical platesNM neutral markerPS polystyreneq chargeRI refractive indexr hydrodynamic radiusr2 correlation coefficientRSD relative standard deviationS secondary wall of the fibreS2 the thickest layer of the

secondary wallSEC size-exclusion chromatography

SP unbleached softwood kraft pulpTHF tetrahydrofurantNM migration time of pyridine to

determine t0tr migration timet0 migration time of EOFU voltageUV ultravioletVp total permeation volumeVs selective permeation volumeV0 void volumeVIS visiblev migration velocityε dielectric constantζ zeta potentialη viscosityµ electrophoretic mobilityµapp apparent mobilityµ0 mobility of electroosmotic flowµav average electrophoretic mobility

70


APPENDIX II: Pulp samples used in papers I-IV

A summary of the characteristics of the pulps used in papers I-IV is shown in Table 1.

Industrially produced hardwood kraft pulp (HWKP) and softwood kraft pulp (SWKP) were

used in papers I-III. A laboratory-made birch wood kraft pulp (labHWKP) was used in

paper IV. This pulp was treated with gamma irradiation, oxygen/alkali or alkali for various

times. For a description of the treatment conditions, see paper IV.

Table 1. A summary of the characteristics of the pulps used in papers I-IV. OD = oxygen

treated and bleached with chlorine dioxide, OZ = oxygen treated and bleached with ozone.

Pulp Used in

paper

Viscosity (dm3/kg) Kappa number Brightness

(%)

unbleached SWKP I - 79 -

unbleached SWKP (SP) I, II, III 950 18 -

OD SWKP (BSP) I, II 610 - 86

unbleached HWKP (HP) II, III 1130 15 -

OZ HWKP (BHP) II, III 820 - 88

labHWKP IV 1160 15 -


71

APPENDIX III: Data for pulps corresponding to the cooks from which lignin samples in

papers V and VI were obtained

Black liquor and isolated lignin samples were obtained from flow-through kraft cooks of pine

wood (Pinus sylvestris). A summary of the characteristics of wood and pulps is shown in

Table 1 and the samples corresponding to the pulps are shown in Table 2.

Table 1. Pulp yield (% of wood) of the cooks, viscosity and lignin content of wood and pulps

used to obtain lignin samples. The figures refer to the cooking times.

Kappa number afterWood/pulp1 Pulp yield

(%)

Viscosity

(mL/g)

Klason

Lignin (%) cooking extraction2

PulpI 2103 43.8 - - 16 -

WoodII - - 26.8 - -

PulpII 210 44.8 1200 2.6 22 -

WoodIII - - 25.9 - -

PulpIII 90 69.7 - 20.8 83 76

PulpIII 150 49.1 1370 4.9 39 18

PulpIII 220 45.1 1160 1.9 16 6.9

PulpIII 210 44.4 1140 - 22 -1 Designation according to Table 2. 1 Acid dioxane extraction. 3 According to Robert et al., 1984.

Table 2. Abbreviations for pulp and lignin samples.

Pulp Corresponding to the cook from which Paper

PulpI 210 dissolved lignin was isolated from black liquor fractions V

PulpII 210 dissolved lignin (Sd) isolated from black liquor fractions VI

PulpIII 90 residual lignin (Sr 90) was isolated from pulp VI



PulpIII 210 black liquor (Sb) fractions were obtained for direct

characterisations

VI

72


APPENDIX IV: Data for pulps corresponding to the cooks from which lignin samples in

paper VII were obtained

Black liquor and isolated lignin samples were obtained from flow-through kraft cooks of

birch wood. A summary of the characteristics of wood and pulps is shown in Table 1 and the

samples corresponding to the pulps are shown in Table 2.

Table 1. Pulp yield (% of wood) of the cooks, viscosity and lignin content of wood and pulps

used to obtain lignin samples. The figures refer to the cooking times.

Kappa number afterWood/pulp1 Pulp yield

(% of wood)

Viscosity

(mL/g)

Klason lignin

(%) cooking extraction2

WoodI - - 19.7 - -PulpI 210 44.5 1290 <0.1 9.2 -

WoodII - - 18.3 - -

PulpII 90 64.6 - 13.9 84 27

PulpII 150 48.8 1380 0.5 13 2.9

PulpII 210 46.7 1210 0.1 9.2 2.7

PulpII 210 47.5 1270 - 9.3 -1 Designation according to Table 2. 2 Acid dioxane extraction.

Table 2. Abbreviations for birch pulp and lignin samples.

Pulp Corresponding to the cook from which

PulpI 210 dissolved lignin (Hd) was isolated from black liquor

PulpII 90 residual lignin (Hr 90) was isolated from pulp



PulpII 210 black liquor (Hb) fractions were obtained for direct characterisations

Documents

Characterisation of Kraft Pulps by Size-exclusion ...8571/FULLTEXT01.pdfV Characterisation of dissolved kraft lignin by capillary zone electrophoresis Elisabeth Sjöholm, Nils-Olof