diffusive masstransport of ions in wood

THESIS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

Diffusive Mass Transport of Ions in Wood

REDDYSURESH KOLAVALI

Department of Chemistry and Chemical Engineering

CHALMERS UNIVERSITY OF TECHNOLOGY

Gothenburg, Sweden 2016


REDDYSURESH KOLAVALI

ISBN 978-91-7597-403-3

© REDDYSURESH KOLAVALI, 2016.

Doktorsavhandlingar vid Chalmers tekniska högskolan

Ny serie Nr 4084

ISSN 0346-718X

Department of Chemistry and Chemical Engineering

Chalmers University of Technology

SE-412 96 Gothenburg

Sweden

Telephone + 46 (0)31-772 1000

Cover:

(a) Wood cell: The radial surface of a softwood tracheid, showing pits on the side walls and

the lumen within the hollow cell

(b) Schematic diagram of the ion exchange and chemical complexation likely to occur during

the sorption of metal ions onto wood material. (Mn+ M: cation, n: valance)

(c) Schematic diagram of the transport of chemicals/ions in softwoods in the longitudinal

direction.

Printed by Chalmers Reproservice

Gothenburg, Sweden 2016

iii


Reddysuresh Kolavali Department of Chemistry and Chemical Engineering

CHALMERS UNIVERSITY OF TECHNOLOGY

ABSTRACT

Understanding the conditions that control the mass transfer of various chemicals/ions in wood is

of great importance when it is subjected to various processes. Homogeneous impregnation with

chemicals/ions, for example, increases the uniformity of the treatments to which the wood is

subjected, reduces reaction times and may increase the yield of the final products. Wood material

is porous and anisotropic and the constituents of the cell wall (solid) are composed of a variety of

functional groups, so the properties of the cell wall may influence not only the mass transport of

chemicals through the pores in the wood but also how these chemicals interact with the cell wall’s

components. The mass transports involved when wood material is impregnated with chemicals are

advective (penetration) and diffusive (diffusion). The latter is a complex mass transfer with several

processes that may involve the diffusion of chemicals through the cell pores (lumen and pit pores)

and through the cell walls at certain conditions, as well as sorption at solid surfaces. Although

several concepts and theories currently exist, some aspects of this type of mass transfer remain

unclear.

The aim of this work is to develop a methodology to investigate the diffusive mass transport of

ions in wood, combining experimental and modelling work. Experiments were performed on

Norway spruce in the form of wood flour, pieces of wood and isolated components of wood. An

experimental methodology was developed to measure the concentration profiles of cations through

the porous structure of a piece of wood that considered the effect of sorption of the ions onto the

matrix of the solid wood. In order to differentiate the amount of ions distributed between the solid

wood and the external solution located in the wood pores (partition coefficient), additional

experimental methodology was developed, using wood flour samples, to investigate the sorption

of ions onto wood. The diffusion accessible porosity of the wood material was estimated from the

intrusion pore volume measured by mercury (Hg) porosimetry and the solid volume of the cell wall

was determined using a helium pycnometer. The impregnation of Norway spruce wood with

lithium chloride (LiCl) was investigated in this study. Spectroscopic analysis (XPS, FTIR and

AAS) was employed to investigate the potential interactions that occur on the surfaces of the wood

material upon LiCl treatment. Using the concentration profiles, the partition coefficients and

porosities measured, the effective diffusion coefficients along with tortuosity factors were

estimated, using a transport model which was developed in COMSOL® multiphysics modelling

software.

The findings in this thesis showed that, for the experimental conditions chosen, the methods

developed gave reasonable results. However, defects in the pieces of wood (micro-cracks)

remained and were detectable. It was observed that the sorption of Li+ ions in Norway spruce wood

flour was a spontaneous process that probably involved several types of interaction/bonding

between the various functional groups in the cell wall of the wood and the Li+/Cl- ions; interactions

with functional groups containing oxygen, for example, are identified as Li-O interactions.

Keywords: diffusion, impregnation, mass transport, sorption, wood

v

List of publications This thesis is based upon the studies presented in the following papers, referred to by Roman

numerals in the text.

I. Determination of the diffusion of monovalent cations into wood under isothermal

conditions based on LiCl impregnation of Norway spruce

Reddysuresh Kolavali and Hans Theliander

Holzforschung 2013; 67(5):559-565

II. Experimental determination of the diffusion of monovalent cation into wood: Effects

of micro-cracks, wood structure, impregnation time and temperature on local

concentration profiles


J-FOR, Journal of Science & Technology for Forest Products and Processes 2014; 4(2):

29-35

III. The sorption of monovalent cations onto Norway spruce: model studies using wood

flour and LiCl solution

Reddysuresh Kolavali, Merima Hasani and Hans Theliander

In manuscript

IV. The sorption of monovalent cations onto Norway spruce wood flour: molecular

interactions behind the LiCl impregnation

Reddysuresh Kolavali and Merima Hasani

Submitted for publication in Holzforschung

V. Solute sorption and diffusion in wood based on the LiCl impregnation of Norway

spruce


In manuscript

vi

Results relating to this work have also been presented at the following conferences:

1. Experimental determination of the diffusion of monovalent cation into wood under

isothermal conditions Kolavali Reddysuresh and Hans Theliander

(Poster presentation)

In conference proceedings. EWLP 2012, 12th European Workshop on Lignocellulosics

and Pulp, Espoo, Finland, August 27-30, 2012, pp 476-479

2. Experimental determination of the diffusion of monovalent cation into wood: Effects

of temperature and impregnation time on concentration profiles

Kolavali Reddysuresh and Hans Theliander

(Oral presentation)

In conference proceedings. ISWFPC 2013, 17th International Symposium on Wood, Fiber

and Pulping Chemistry, Vancouver (BC), Canada, June 12-14, 2013, Process Chemistry

Track, Chemistry of the fiber wall and its components

vii

Contribution report The author of the thesis is the main author of the papers below:

Paper I-II– The experiments were planned, and the results evaluated, together with Professor Hans

Theliander. The author performed all of the experiments and analyses involved.

Paper III– The experiments were planned, and the results evaluated, together with Professor Hans

Theliander and Dr. Merima Hasani. The author performed all of the experiments and analyses

involved, with the exception of the analysis of the metal ion concentrations in wood flour, before

and after acid treatment. The modelling work was also performed by the author.

Paper IV– The experiments were planned, and the results evaluated, together with Dr. Merima

Hasani. The author performed all of the experiments and analyses involved, with the exception of

the XPS experiments.

Paper V– The experiments were planned, and the results evaluated, together with Professor Hans

Theliander and Dr. Merima Hasani. The author performed all of the experiments and analyses

involved. The modelling work was also performed by the author.

ix

ContentsContentsContentsContents

1. Introduction and Outline …………………………………………………………………………………………… 1

1.1. Introduction ……………………………………………………………………………………………………………….. 1

1.2. Outline ……………………………………………………………………………………………………………………….. 2

2. Background ………………………………………………………………………………………………………………….... 3

2.1. The structure and composition of wood ……………………………………………………………………… 3

2.2. Impregnating wood with chemicals …………………………………………………………………………….. 8

2.3. Measurement of the diffusion of chemicals in wood …………………………………………………. 11

3. Research Objectives …………………………………………………………………………………………………… 15

4. Materials and Methods …………………………………………………………………………………………….. 17

4.1. Materials and chemicals ……………………………………………………………………………………………. 17

4.2. Measurement of density and porosity ………………………………………………………………………. 19

4.3. Studies of equilibrium sorption …………………………………………………………………………………. 20

4.4. Measurement of the concentration profiles ……………………………………………………………… 20

4.5. Characterization of the material ……………………………………………………………………………….. 22

4.6. Theoretical model formulation :

Estimating the transport properties of ions in wood …………………………………………………. 22

5. Results and Discussion ………………………………………………………………………………………………. 25

5.1. Development of an experimental procedure:

Concentration profiles of ions in pieces of wood ………………………………………………………. 25

5.2. Investigation of possible interactions

between Norway spruce wood and an aqueous LiCl solution ……………………………………. 27

5.3. Density and porosity measurement of Norway spruce wood ……………………………………. 33

5.4. Sorption equilibrium of ions onto solid portions of wood flour …………………………………. 36

5.5. Concentration profiles of ions in the porous structure of pieces of wood …………………. 39

5.6. Effective diffusion coefficients and tortuosity factors of ions in wood ………………………. 40

6. Contributions made to this field of research ……………………………………………………….. 43

7. Conclusions …………………………………………………………………………………………………………………… 45

8. Acknowledgements ……………………………………………………………………………………………………. 47

9. References ……………………………………………………………………………………………………………………. 49

Appendices …………………………………………………………………………………………………………………… 55

1

1. INTRODUCTION AND O1. INTRODUCTION AND O1. INTRODUCTION AND O1. INTRODUCTION AND OUTLINEUTLINEUTLINEUTLINE

1.1 Introduction1.1 Introduction1.1 Introduction1.1 Introduction Petroleum and other fossil reserves are considered as being non-renewable carbon resources because of their long geological recycling times (~ 200 million years) compared with biomass (~ <1 to 80 years) feedstocks, which are renewable and carbon-neutral resources (Liu et al., 2006). Wood lignocelluloses are biomass that are seasonal-independent; they are also the most abundant biomass on Earth, accounting for an annual production of ~170 billion metric tons in the biosphere. In Sweden, wood biomass resources have been the basis of significant industrial activities for several centuries. The dominant use of wood raw material is for pulp, paper and sawn timber products, although its use for energy purposes has increased during recent decades. However, sharper competition from countries with fast-growing forests means that competitiveness needs to be improved, and new and highly-refined products be found that complement existing pulp-based products. In this context, the “biorefinery” is the suitable concept for converting biomass into fuels, power and chemicals (Fernando et al., 2006). A biorefinery is, in many ways, analogous to a petroleum refinery, in that a single feedstock is fractionated into a multitude of commodity products, depending on societal necessity. The pulping industry is the most developed chemical technology for processing wood. It is, to some extent, already a biorefinery in which energy, cellulosic fibres and minor amounts of turpentine and tall oil are produced from wood. Interest in the biorefinery concept in this area has increased significantly in recent years mainly due to: (a) the trend of increasing costs for wood and decreasing prices for pulp and paper, (b) greater competition from low cost producers of pulp and paper in South America and Asia and (c) higher energy prices and new policy instruments that affect the production of materials such as non-petroleum-based materials and chemicals. It is not only in the production of paper pulp and in the fractionation of different wood components in wood-based biorefineries but also in operations in which solid wood is treated with different kinds of chemicals (e.g. preservatives, fire retardants and dimension-stabilizing chemicals) that chemicals must be transported into wood prior to a reaction. The wood has to be impregnated with the reactants, so an understanding of the conditions that control the mass transfer of various chemicals/ions in wood is of great importance in wood processing technology: homogeneous impregnation, for example, increases the uniformity of a treatment, reduces reaction times and may increase the yield of the final products. Wood material can be impregnated with chemicals by means of either advective (penetration) or diffusive (diffusion) mass transport (Akhtaruzzaman and Virkola, 1979). Penetration is defined as the flow of liquor into the gas/vapour-filled voids of the pieces of wood under a pressure gradient, and diffusion as the transport of chemicals/ions through the liquid within the pieces of wood under the influence of concentration gradients. The former occurs very fast and at reasonable pressure gradients, whilst the latter is a much slower phenomenon. Diffusion is therefore the controlling mechanism in most (but not all) wood conversion processes. Little is known about diffusive mass transport, however, partly because of the difficulties in determining the relevant diffusivities for a given system. The phenomenon that occurs during the diffusion of chemicals in wood therefore needs to be understood better. The transport of chemicals/ions in a heterogeneous, porous, hygroscopic, and anisotropic material such as wood is very complex; the diffusive transport of chemicals in it is restricted

2

due to the reduced cross-sectional area available for transport and tortuous pathways. In addition, the cell walls of wood consist of functional groups associated with hemicelluloses, lignin and cellulose, and solutes (chemicals) may therefore be subjected to sorption interactions with these components that reduce their rate of transport further. A wood cell (fibre) may be described as a pipe that is closed at both ends. The cells overlap each other and, under normal conditions, the mass transport between the fibres is via “pit pores”. The cell wall may be described as being a more or less swollen gel. When the fibres and the pit pores become filled with a liquid, the main diffusive mass transport of the solutes occurs through that liquid. The solutes may be sorbed onto the cell wall too, so there may also be mass transport at the cell wall (surface diffusion) and/or in the gel-like structure of the cell wall. The overall mass transport rate may thus be influenced by several different mass transport phenomena as well as by sorption phenomena (Figure 1.1).

Figure 1.1 Schematic diagram of the transport of chemicals in softwood in the longitudinal direction.

1.2 Outline 1.2 Outline 1.2 Outline 1.2 Outline This work is based on five papers, all of which can be found at the end of the thesis. The following chapter provides background information on the structure and composition of wood (Section 2.1) and its impregnation with chemicals (Section 2.2). It also provides some literature pertaining to existing methods, and the drawbacks associated with them, on the measurement of ion diffusion in wood. The objective of this thesis is stated in Chapter 3, with the focus being on the diffusive mass transport of ions in wood. The materials and methods used in this work, together with the theoretical model formulated for the determination of transport properties (effective diffusion coefficient and tortuosity factors) of ions in wood, are described in Chapter 4. This is followed by Chapter 5, which provides the important findings of the various investigations carried out, along with some discussion and interpretation of the results. Chapter 6 discusses the contributions made to this field of research. Finally, Chapter 7 summarizes the outcomes of the various investigations and the conclusions that are drawn.

3

2. Background2. Background2. Background2. Background

2.1 2.1 2.1 2.1 The The The The structure and compositionstructure and compositionstructure and compositionstructure and composition of woodof woodof woodof wood The classification of tree species as either “hardwoods” or “softwoods” is based on a botanical distinction. Softwoods (or “gymnosperms”) are conifers: they are needle-bearing trees and are often evergreen. Hardwood trees (or “angiosperms”), on the other hand, are broad-leaved and mostly deciduous. Today, more than 90% of the world’s trees are angiosperms. Gymnosperms includes species such as pine (Pines), spruce (Picea) and fir (Abies), and angiosperms include birch (Betula), beech (Fraxinus), oak (Quercus) and poplar (Populus). The wood species used most commonly in the Swedish forest industry are spruce (Picea abies), pine (Pinus sylvestris) and birch (Betula verrucosa) (Henrikson et al., 2008).

The structure of wThe structure of wThe structure of wThe structure of wood ood ood ood

Wood is a heterogeneous, hygroscopic, anisotropic and porous material. There are large variations in structure both between and within individual trees. Many studies on the structure of wood can be found in the literature, e.g. Fengel and Wegener, 1984; Sjöström, 1993; Miller, 1999; Rowell, 2005; Wiedenhoeft, 2010. An understanding of the nature of the various capillary components of wood and the extent to which each influences the transport of liquids, dissolved materials and chemicals through wood is imperative. Wood is composed of different kinds of capillaries of variable size: from structures observable to the naked eye down to capillaries that approach molecular dimensions. In general, softwoods have a more uniform structure than hardwoods, with a fewer number of cell types. Since this study focuses on Norway spruce, only structural details of softwood are described.

Levels of structureLevels of structureLevels of structureLevels of structure in woodin woodin woodin wood

The fundamental structure of wood, from the molecular to the cellular or anatomical level, determines its properties and behaviour. Four different levels of structure can be identified in wood:

Macroscopic structureMacroscopic structureMacroscopic structureMacroscopic structure

When a tree stem is cut transversely, a surface appears that is comprised of various annual growth rings presented in the form of concentric bands (Figure 2.1). A tree consists of several layers which, from the outside in, are known as the outer bark, inner bark, vascular cambium, sapwood (Sw), heartwood (Hw) and pith. The outer bark is a corky dead part, whereas the inner bark (phloem) is a thin living part through which sugars produced by photosynthesis are translocated from the leaves to the roots or parts of the tree that are growing. The outer bark not only provides mechanical support and protection to the softer inner bark, it also decreases the evaporative water loss (Wiedenhoeft, 2010).The cambium layer, located with the inner bark, forms cells of wood and bark; it can be seen only with the aid of a microscope. Growth in the thickness of the bark and wood is the result of cell division in the cambium.

4

Figure 2.1 Macroscopic view of a transverse section of a tree stem.

Sapwood, which is next to the cambium, contains both living and dead cells. Its primary functions are to store photosynthate, provide mechanical support and handle the transport of sap, which consisting mainly of sugars and mineral salts dissolved in water. Heartwood is comprised of dead cells that have no function in either conducting water or storing nutrients/photosynthate. Hw has a lower moisture content than Sw. The transition from Sw to Hw is accompanied by an increase in the content of extractives; it is often these which darken the colour of the Hw. The extractives in Hw may also affect wood by reducing its permeability, thus making Hw slower to dry and more difficult to impregnate with chemical substances. The pith is a small core of tissue located at the centre (inner layer) of a tree stem, and comprises remnants of the early growth of the tree stem before the formation of wood. The inner part of the growth ring, formed first in the growing season, is called earlywood and the outer part, formed later on in the growing season, is called latewood (see Figure 2.1).

Microscopic structureMicroscopic structureMicroscopic structureMicroscopic structure

Softwoods consist mainly of longitudinal tracheids (90-95 % of cell volume) and a small number of ray cells (5-10 % of cell volume) (Fengel and Wegener, 1984; Brändström 2001). For the sake of simplicity, the longitudinal tracheids are hereafter referred to simply as “fibres”. All fibres are hollow, elongated tubes that are tapered and closed at both ends, with a cross-section that is something between rectangular and elliptical in shape. The hollow void in the fibres is called the lumen, and it is this which enables water to be transported within them. The distribution of fluids between adjacent cells is achieved through a system of pits that allows intercellular connection. The average length of Scandinavian softwood (Norwegian spruce and Scots pine) fibres is 2-4 mm and the average diameter is around 0.02-0.04 mm. The main function of earlywood, which is characterized by cells with relatively large cavities and thin walls, is to transport water and nutrition. Latewood cells serve more as mechanical support, as they have smaller cavities, thicker walls and are longer in length than earlywood cells. The average content of latewood in Norway spruce has been reported to be 20-25% on a volume basis. When growth rings are prominent, as is the case in most softwoods, earlywood differs markedly from latewood in its physical properties in that it is lighter in weight, softer and weaker than latewood (Brändström, 2001; Miller, 1999). The average thickness of cell walls of earlywood fibres is 2-4 µm whereas that of latewood fibres is 4-8 µm (Sjöström, 1993). The

5

radial system is made up of ray cells, which run perpendicular to the fibres and across the void volume of the wood, and are known as wood rays (see Figure 2.1). Ray cells have an average length of 0.01-0.16 mm and diameter of 2-50 µm. The main functions of wood rays are to redistribute and store, for example, starch. A number of softwoods also contain vertical and horizontal resin ducts (channels), but these are usually filled with resin; their contribution to diffusion and flow is minor (Burr and Stamm, 1947). The other types of cells present in wood are epithelial cells (resin canals) and parenchyma (ray canals), the latter of which are involved in the transport of liquids in the horizontal direction (Rowell, 2005).

UltraUltraUltraUltra----structurestructurestructurestructure

The wall of a mature wood cell consists of different layers (see Figure 2.2 A): a thin outer layer called the primary wall (P) and a thicker inner layer towards the cell cavity called the secondary wall (S). Three sub layers make up the secondary wall: the outer, or transition, layer (S1), the central layer (S2) and the inner layer or tertiary wall (S3). Individual cells are connected together by the intercellular middle lamella region (ML).

Pits are characteristic to the structure of wood (see Figure 2.2 B). They represent canals that facilitate the transport of liquids both horizontally and vertically through the cell walls. Earlywood tracheids have more pits than latewood tracheids, which makes it easier to impregnate earlywood than latewood. Earlywood has approximately 50 to 300 pits per tracheid whereas latewood has only 10 to 50 (Stamm, 1946).

Figure 2.2 A. The different layers in the cell wall of wood. ML: middle lamella, P: primary cell wall, S1, S2, S3: secondary walls (from Saltberg, 2009). B. The radial surface of a softwood tracheid, showing pits on the side walls and the lumen within the hollow cell.

Pits of adjacent cells are normally paired off and form pit pairs, of which there are three common types: simple (link parenchyma cells together), bordered (in the radial walls of the tracheids and link tracheids to tracheids) and half-bordered (connect the vertical tracheids with horizontal ray parenchyma cells) (see Figure 2.3).

6

Figure 2.3 The three basic types of pit pairs: a simple, b bordered and c half-bordered. A aperture, C chamber, M middle lamella-primary wall, S secondary wall and T torus.

All pits have essentially two main components: a cavity and a membrane. The pit membrane consists of the primary wall and the middle lamella and, since pits occur in pairs, it is therefore composed of two primary walls and the middle lamella. Pits are localized exclusively on radial cell walls and are more frequent at the ends of tracheids than in the middle. The pit membranes of bordered pit pairs have a central thickened portion called the torus (T). This may be pushed against the opening of a cell wall and possibly become sealed in this position, thereby reducing the rate of material transport significantly: such pits are known as aspirated pits. Bordered pits on the tangential walls of longitudinal tracheids are always smaller than those on the radial walls (Kazi, 1996). The pit chamber is roughly in the form of a truncated cone, with the smaller end opening into the fibre cavity and the larger end closed by the pit membrane. The pit membrane is three to four times greater in diameter than the pit opening.

Molecular structureMolecular structureMolecular structureMolecular structure

Wood polymers can be classified into three major types: cellulose, hemicelluloses and lignin (Sjöström, 1993). The proportion of these three polymers varies between species. A typical composition of softwood is 40-50% cellulose, 20-30% hemicelluloses and 20-30% lignin, based on the weight of the extractive-free wood (Kazi, 1996). Cellulose is a polysaccharide; its monomers, β-D-glucose units, are linked together by (1→4)-glycosidic bonds. The equatorial orientation of the hydroxyl groups in the structure of cellulose provides many opportunities for strong intra and intermolecular interactions, such as hydrogen bonds and hydrophobic interactions, to form. The cellulose chains show a tendency to self-order: they arrange themselves into so-called elementary fibrils that have regions with varying degrees of order. The elementary fibrils result in highly ordered crystalline structures, which have great tensile strength and low accessibility. Structures that are less ordered, on the other hand, are amorphous and easily degradable. Hemicelluloses form a diverse group of polysaccharides that are substantially smaller in size than cellulose. The five main sugar monomers that can compose the chains of hemicelluloses are: D-glucose, D-mannose, D-galactose, D-xylose and L-arabinose (Sjöström 1993). Of these, L-arabinose and D-xylose are pentose (C5) sugars while the rest are hexose (C6) sugars. Hemicelluloses may also contain small amounts (as side chains) of L-rhamnose, D-glucuronic acid, 4-O-methyl-D-glucoronic acid and D-galacturonic acid. Three main types of

7

hemicellulose occur in softwoods: (galacto) glucomannan, arabinoglucuronoxylan and arabinogalactans, although the latter is found primarily in larch wood. Lignin is an amorphous and highly branched irregular complex polymer, predominantly constituted of three phenyl propane units that form the major building blocks, namely p-coumaryl alcohol, coniferyl alcohol and sinapyl alcohol, which give rise to a random sequence of p-hydroxyphenyl (H-lignin), guaicyl (G-lignin) and syringyl (S-lignin) sub-units, respectively, in the polymer. It has been identified that softwood lignins are mostly composed of G units, with the remainder being H units. The main function of lignin is to act as an adhesive between the cellulose fibrils and hemicelluloses, thereby providing the fibres and their structure with strength, hydrophobicity and resistance against microorganisms (Sjöström 1993). There are other chemical compounds in wood, but these usually account for only a few percent of its total composition. One class of such compounds is called extractives; they include both water-soluble and organic-soluble extractives and form, on average, up to 10% of softwood. There are also inorganic compounds present in wood; as a group, they are generally known as ash. Metal ions such as calcium (Ca), potassium (K), magnesium (Mg), manganese (Mn) and silicon (Si) are elements commonly found in wood (Saltberg, 2009).

Water in woodWater in woodWater in woodWater in wood

Water has an inevitable effect on the properties of wood material (Engelund, 2011). The cell walls of wood are made of polymers that are more or less hydrophilic and can thus accommodate a substantial amount of water within the structure itself. Functional groups capable of attracting water molecules are termed as sorption sites. In wood, most of the sorption sites are found in the hemicelluloses, followed by cellulose and then lignin.

Shrinking and swelling Shrinking and swelling Shrinking and swelling Shrinking and swelling in the in the in the in the structurestructurestructurestructure of woodof woodof woodof wood

Swelling and shrinking in wood occur as a result of changes in the moisture content. The moisture present in wood has two forms: bound or hygroscopic water, and free or capillary water. Bound water is found in the cell wall: it is believed to be hydrogen-bonded to the hydroxyl groups of primarily the cellulose and hemicelluloses and, to a lesser extent, to the hydroxyl groups of lignin too. The moisture content at which the cell walls are saturated with bound water, with no free water in the lumen cavities, is called the “fibre saturation point” (Siau, 1984). At saturated conditions, liquid water is available only in the cell walls: once the cell walls have been saturated, however, liquid water may also enter the cell cavities. The fibres in green (i.e. living) wood are always swollen. The moisture content in its structure may decrease when wood is cut; when it is reduced to below the fibre saturation point, the fibres shrink. This shrinking causes the cracking in the fibre walls that affects the mass transport of chemicals in wood, as an increased number of paths and/or smaller lumen develop. In most cases, wood pieces used in diffusion measurements have a moisture content that is higher than the fibre saturation point, which then leads to tangential swelling. Tangential swelling can easily be seen in the form of an increase in volume (Skaar, 1972); longitudinal swelling is much smaller than transversal swelling. Some electrolytes increase swelling by softening the cell walls (Wallström and Lindberg, 2000); they usually have a very high pH, e.g. NaOH (aq) and white liquor. This phenomenon may occur because the solution can penetrate the wood from the lumen, through the secondary wall layer S3, and into the middle lamella.

8

2.2 Impregnati2.2 Impregnati2.2 Impregnati2.2 Impregnatingngngng wood with chemicalswood with chemicalswood with chemicalswood with chemicals Impregnation is a method commonly used in the treatment of wood material whereby a variety of chemicals are employed to improve properties such as dimension stability, chemical and biological resistance, weathering and mechanical properties and resistance to fire (Kazi, 1996). Impregnation is also the first step in conventional pulping or wood conversion processes, followed by cooking, steam explosion and other processes. Proper impregnation of wood with chemicals is of great importance in attaining economically viable and efficient chemical pulping and biorefinery operations, because homogeneous impregnation not only increases the uniformity of the treatments but also reduces reaction times. During impregnation, chemicals are transferred into the core of the wood by two entirely different primary mechanisms (Stone and Förderreuther, 1956). The first of these, liquor penetration, is the flow of liquor into the gas/vapour-filled voids of the wood material under the influence of a pressure gradient. The second, diffusion, refers to the movement of ions or other soluble matter through liquid present within the wood material under the influence of a concentration gradient. Both mechanisms occur simultaneously in fresh wood but, in certain cases (such as in Kraft cooking), most of the chemicals are transported into wood by diffusion. The role of penetration is to fill the fibre cavities with liquid, which enables a faster and more even diffusion of chemicals/ions into the wood (Määttänen and Tikka, 2012). Penetration occurs very rapidly initially if there are any gas/vapour-filled pores, whereas diffusion is the mechanism involved if the pores are completely filled with liquid, and is much slower. In the conventional pulping process, a major portion of the cooking chemicals (white liquor) required may be added at the beginning of the process. The wood material contains water in the form of moisture from the beginning of the pulping process, which contributes to the dilution of the cooking liquor that penetrates into the interior of the wood chips. A concentration gradient (the driving force for the diffusion mechanism) is thus established between the bulk of the cooking liquor surrounding the wood chips and the liquor inside the fibre lumina. Following this initial penetration stage, the distribution of chemicals within the chips occurs mainly by diffusion, due to the chemicals (causing a decrease in concentration) present in the cooking liquor being consumed by various reactions that occur between lignin and polysaccharides, e.g. the neutralization of acidic groups present in wood or produced by the peeling of polysaccharides.

Diffusion in woodDiffusion in woodDiffusion in woodDiffusion in wood

Diffusion is a phenomenon involving the movement of constituents from one zone, with a high concentration, to another with a lower concentration (Stamm, 1967; Siau, 1984). The basic mechanism behind diffusion is random molecular motion. Diffusion in its simplest form is uni-dimensional diffusion, and can be expressed using Fick’s first law of diffusion thus:

J = −�� (2.1)

where J = mass flux, c = concentration of the solute in the liquid phase, x = direction of transport and Df = diffusion coefficient of the “free-solution”. The diffusive transport of solutes in wood is influenced by tortuous pathways. There are lesser mass fluxes in wood than in free solution because some of the cross-sectional area in wood is occupied by solid portions. Equation (2.1) must therefore be modified for diffusion in wood as follows:

9

J = −�� (2.2)

where c = concentration in the liquid phase of the wood pores and = volumetric water content, which is defined as:

= �� (2.3)

where n = total wood porosity and S = degree of saturation of the wood (0-1). The maximum flux for liquid phase diffusion will therefore occur when the wood is saturated (S = 1), all other conditions in Equation (2.2) being equal. The tortuosity of the wood is usually accounted for by the inclusion of a tortuosity factor, τ, in the transport equation; in most of the cases, tortuosity factors cannot be measured independently. It is thus convenient to define an effective diffusion coefficient (Saripalli et al. 2002), De as:

� = ��ɸƟ� (2.4)

where ɸ = effective, transport-through-porosity (porosity that is accessible to diffusion), Ɵ = a constrictivity factor that accounts for the constricted transport paths caused by the small pores and pore throats in a porous medium (usually assumed ~ 1) and τ = tortuosity factor that accounts for the reduction in the diffusive flux caused by the tortuous path lengths traced by the solute molecules (in contrast to the straight paths in an unrestricted aqueous medium). For time-dependent (transient) transport of non-reactive solutes in wood (i.e. solutes that are non-interactive with wood), Fick’s second law is assumed to apply

�� = � ��

�� (2.5)

However, diffusion into, or from, a porous material such as wood is much more complicated than diffusion under steady-state conditions through a porous, inert material. The boundary conditions, for example, change continuously. Whilst wood is being impregnated with chemicals, it becomes swollen to varying degrees: in alkaline solutions, for example, the structure of the wood changes due to swelling. Reactions with various wood components also consume alkali and organic matter becomes dissolved which, in turn, affects the mass transport rate due to the increased permeability of the cell walls (Gustafson, 1988). Thus, for a given concentration of agent in solution, such as NaOH, the swollen wood matrix can be considered as having an irregular geometry with time.

A solute migrates faster through an open area (lumen) than through a partially-closed area (pit or pit membrane). Moreover, wood material has a variety of functional groups in its cell walls; the solutes are often subjected to chemical interactions with the components of cell walls, which also affects the rate of transport. Mass transport in wood is thus affected by both physical and chemical conditions. Under these dynamic conditions (e.g. to account for sorption interactions), Equation (2.5) may be modified thus:

�� = � ��

�� − �� (2.6)

where �� = sorbed concentration of the chemical species expressed in terms of the mass of sorbed species per unit volume of voids (i.e. occupied by the liquid phase), or thus:

10

�� = �� (2.7)

where q = sorbed concentration expressed as the mass of solute sorbed per mass of dry wood; ρb = bulk density of the wood; and = volumetric water content.

The partition coefficient (k) is defined as:

� = �� (2.8)

where q = sorbed concentration (mass of solute sorbed per mass of dry wood) and c = concentration of solute in the liquid phase of the wood pores. When the relationship of q versus c is linear, k is termed as the “distribution coefficient, kd.” Otherwise, k is a function of the equilibrium concentration of the liquid in the wood’s pores. A plot of the mass of the solute sorbed per mass of wood, q, versus the concentration of the solute in solution, c, is called a “sorption isotherm”.

Important factors that affect the mass transport of ions in wood

Properties of wProperties of wProperties of wProperties of wood:ood:ood:ood: Different types of wood (within the tree: Sw vs. Hw, juvenile vs. mature wood, reaction vs. compression wood, as well as different species) have different capillary structures, which influence the diffusion paths within the wood. This affects the rate of mass transport through wood, at least in the early stage of impregnation, where its permeability remains unchanged. Törnqvist et al., (2001b) found significant differences in diffusion between different wood species: birch had a diffusion coefficient that was significantly higher than spruce and pine in the transversal (radial) direction due to their different capillary structures, but the differences were negligible in the longitudinal direction. The grain orientation also influences the rate of diffusion, and the rate of diffusion in different directions in wood generally follows the rule: longitudinal >> radial > tangential. Degree of saturation: Degree of saturation: Degree of saturation: Degree of saturation: It is generally true that a higher moisture content facilitates diffusion, and that the diffusion rate varies depending on the uniformity of the water throughout the wood’s matrix. In unsaturated pieces of wood, local barriers to diffusion are found in zones of gas or vapour that break up the water phase. This leads to a decrease in the number of pathways available for diffusion. The diffusion rate is at its maximum when the pore structure of a piece of wood is filled with water. Temperature:Temperature:Temperature:Temperature: The diffusion rate generally increases with an increase in temperature. Temperature affects the mobility of ions and thereby also diffusion and, consequently, the diffusion coefficient. Properties of Properties of Properties of Properties of the the the the chemical solution:chemical solution:chemical solution:chemical solution: The important factors that affect the mass transport rate of chemicals in wood with respect to the properties of the chemical solution are sorption, temperature, pH, concentration and viscosity (Stone and Förderreuther, 1956).

Sorption in woodSorption in woodSorption in woodSorption in wood

When a metal ion in solution interacts with a wood tissue, it can be sorbed by physisorption or chemisorption. An ion exchange takes place between the incoming cation and either the sorbed metal ions or the hydrogen ions of the functional groups at the surface of the wood material. The sorption of metal ions from the solution onto the wood’s surface may also occur via the formation of surface complexes in which site-specific interactions take place between metal ions and functional groups at the wood surface (Hubbe et al., 2011). Based on the behaviour

11

of heavy metal sorption on sawdust, it has been speculated that an ion exchange and hydrogen bonding may be the principal mechanisms in the sorption of heavy metals (Shukla et al., 2002). The sorption equilibrium is dependent on the pH of the aqueous solution. At pH values greater than the pKa of most functional groups on wood surfaces, the sites are mainly in the dissociated form and can therefore bind with the ions in the solution. At pH values lower than the pKa of functional groups on wood, cations can, for example, bind to the sorption sites through the ion exchange mechanism of releasing protons (H+).

2.3 Measurement of 2.3 Measurement of 2.3 Measurement of 2.3 Measurement of the the the the diffusion of chemicals in wooddiffusion of chemicals in wooddiffusion of chemicals in wooddiffusion of chemicals in wood

MMMMethods for measuring ethods for measuring ethods for measuring ethods for measuring the the the the diffusion of diffusion of diffusion of diffusion of chemicals in woodchemicals in woodchemicals in woodchemicals in wood

Several methods are described in the literature for measuring liquid phase diffusion in wood, but there is none that is standard. A review of the literature showed that early research on molecular diffusion into wood was reported by Cady and Williams in 1935. It was inspired by experiments of diffusion across the interface between a moving liquid and a gel in a stirred bath, and was extended to conditions valid for wood. These authors studied the diffusion of urea, glycerol and lactose into several water-saturated softwoods. In 1947, another paper with the title “Diffusion in Wood” was published by Burr and Stamm. Most of the research during this period was related to developing the theoretical expression for the rates of transport of vapors, liquids and dissolved materials through softwoods in the three structural directions under various conditions of temperature and moisture content (Stamm, 1946; Burr and Stamm, 1947). In the early 1950s, the diffusion of salts, ions and organic molecules through green timbers was studied extensively (Christensen and Williams, 1951; Christensen, 1951a & b; Narayanamurti and Ratra, 1951; Narayanamurti and Kumar, 1953; Behr et al., 1953), with a quantitative theory on diffusion in wood being provided by Christensen and Williams (1951). The effect of the wood’s structure on diffusion was studied in depth using electrical conductivity measurements made on impregnated blocks of wood. These measurements were then compared with the electrical conductivity of the solution. The ratio of the two conductivities was considered as being a direct measurement of the effective capillary cross-sectional area (ECCSA), defined as the ratio between the area available for diffusion and the total area (Stone 1957). Stone and Green (1959) used this method to study the effect of pH on the diffusion of KCl in the three structural directions of aspen wood. They found that, up to a pH of 12, the ECCSA in the longitudinal direction is approx. 5 times greater than in the transverse and radial directions. At a pH above 13, however, fibre swelling causes the ECCSAs to be about the same in all three directions. Later on, in the 1980s, most of the research in this field was directed towards measurements of the diffusion of non-electrolytes (monohydric alcohols and polyethylene glycols) through wood (stika spruce Hw) filled with water using a highly sensitive, differential refractometer (Fukuyama and Urakami, 1980, 1982 and 1986). These studies emphasized the variation in the diffusion coefficient associated with the structure of wood. Tracer techniques using radioactive isotopes have been used to evaluate the impregnation of pulping chemicals during chemical and chemi-mechanical pulping (Bengtsson and Simonson 1984). After impregnation, the samples were sliced and the beta activity from the whole surface of a slice was used to provide an indication of the degree of diffusion. One of the techniques frequently employed to measure the diffusion of ions in water-filled wood is based on the rate at which ions diffuse (a) through a slice of wood, the sides of which are in contact with a solution of different concentrations, or (b) in a water bath, into which a

12

completely impregnated block of wood has been placed. Robertsen (1993) used this method to study the effects of concentration and temperature on the diffusion of NaOH in the radial direction of spruce wood. One of its limitations is that it gives only an average diffusivity, as it cannot differentiate between areas of varying diffusivity within a sample of wood. The Scanning Electron Microscopy/Energy Dispersive X-ray Analysis (SEM/EDXA) method was developed to measure liquor diffusion by measuring the concentration of a chemical at a defined area in a wood chip (Bengtsson et al., 1988; Sharareh et al., 1996). The advantage of this method is that it allows the concentration profile of the ion in the wood chip to be obtained, qualitatively and quantitatively, as the ion is transported from the edge of the chip inwards, towards its centre. Sharareh et al., (1996) used this method to measure the concentration of sulphur in the three structural directions in an impregnated wood chip simultaneously. However, this method cannot differentiate between bound and free sulphur. In order to study the effect of the diffusion mechanism on the alcohol-based organosolv pulping process, unsteady-state diffusion experiments on methanol in Douglas fir Hw were conducted at high temperatures (Meijer et al., 1996). Results from this research indicated that the diffusion mechanism was the same as for alkaline solution diffusion in kraft pulping, although a deviation from Fick’s second law was also observed at low methanol concentrations. A diffusion model for the impregnation of lignocellulosic materials was developed in the same period of time to ensure a uniform chemical distribution within the lignocellulosic matrix prior to rapid steam treatment (Kazi 1996). During their impregnation experiments with NaOH, these authors also observed that a large fraction of the inorganics (e.g. ash) present in the lignocellulosic material (i.e. straw) could be removed. The quality of the fibre was thus improved and pure cellulose was produced. A new approach called “Drift speed” was proposed with the aim of studying the diffusion of ions in wood (Törnqvist et al., 2001a). This research focused only on the effects of possible swelling during the transport of ions in the wood; this phenomenon is dependent only on the diffusion length and not the active area available for diffusion. In order to improve a number of properties of wood, such as surface hardness and weathering resistance, some researchers have used UV microscopy to investigate the mass transport of resin in the cell walls of softwood (Gindl et al., 2002; Gindl et al., 2003). Results from such work have demonstrated that UV microscopy is a suitable technique for addressing the question of the possible transport of chemicals in the cell walls of wood. Much research has been conducted in the field of wood preservation to understand the diffusion mechanism in cell walls. Some researchers have studied extensively the effect of various factors, such as temperature, moisture content and pH, on the diffusion of wood preservatives (e.g. boron and copper) in pine wood (Cooper, 1998; Ra et al., 2001). Fourier Transform Near-Infrared Transmission spectroscopy (FT-NIR) was applied to monitor the diffusion of deuterium-labeled molecules in beech wood by Tsuchikawa and Siesler (2003). Two of the important conclusions drawn from this particular work were that, independent of the wood species, the diffusion of the penetrant into the amorphous region was faster than into the semi-crystalline region, and the size effect of the diffusants (oA) plays an important role in the diffusion process in wood. Recently, tritiated water (a conservative tracer for water) was used to understand the mass transfer phenomenon in water-filled wood particles of pine and aspen wood (Jacobson and Banerjee, 2006). The authors of that study calculated the diffusion coefficient as a function of tortuosity, porosity and the self-diffusion coefficient. They concluded that mass transfer in wood was more than a simple Fickian mechanism, and introduced the concepts of size exclusion and charge exclusion. However, this method must be tested for different chemical species and wood species, and for the effects of their structures, to gain greater knowledge and insight.

13

A model that describes the leaching of calcium ions from softwood chips, and also takes into account cracks, rough surfaces and the size distribution of the wood chips, has been developed (Saltberg 2009). Leaching experiments with hand-sawn pieces of spruce wood were conducted to determine the diffusion coefficient for calcium. However, this method cannot differentiate the concentration profiles at various locations within a single wood piece. It just calculates the effective diffusion coefficient and, unfortunately, the natural content of calcium ions (which for softwoods is 600 to 1200 mg/kg wood) aggravates the measurements. Recently, the free diffusion of ions through the secondary walls and middle lamellae of wood was investigated as a function of moisture content (MC) using synchrotron-based X-ray fluorescence microscopy (XFM) (Zelinka et al., 2015). For the ions investigated in this study (i.e. K, Cl, Zn and Cu), a threshold MC was identified below which ion transport did not occur. It was also observed that the threshold for ion transport depends upon the ion species, cell wall layer and anatomical direction in the wood.

Drawbacks associated with Drawbacks associated with Drawbacks associated with Drawbacks associated with existing existing existing existing methods methods methods methods

• It is clear from the literature review that three different methodologies are commonly used to measure the diffusion of solute ions in wood. The first measures either the rate at which ions diffuse through a wood block the sides of which are in contact with solutions of different concentration, or the change in the ion concentration in a water bath in which a completely impregnated block of wood has been placed. The second method measures the electrical conductivity of the impregnated wood block and compares it with that of the impregnating chemical solution; the ratio of the two conductivities gives a direct measurement of the ECCSA. The third method employs tracer techniques using radioactive isotopes, whereby samples are sliced and the radiation activity from the whole surface of a slice is used as an indication of the degree of diffusion. It is, however, these methodologies have limitations: the first and second methods give only an average diffusivity, and cannot be used for measuring concentration profiles, and the third is limited to mapping concentration profiles in only one direction at a time, since the whole surface of each wood slice is used to measure the concentration.

• Most of the investigations involving the measurement of cation diffusion in wood have been conducted using diffusing substances such as NaCl (sodium chloride), KCl (potassium chloride) and NaOH (sodium hydroxide). If NaCl or KCl are used, there is the possibility of inaccuracy occurring in determining the diffusion of cations into the wood: Na+/ K+ ions (in the form of alkali metal ions) are present naturally in wood. In the case of NaOH, reactions between the OH- ions and the wood’s components influence the properties, and thus the mass transport, of the Na+ ions.

• Some observations indicate clearly that almost all wood pieces have micro-cracks that may be too small for direct visual detection, but they can change the capillarity of the surface layer, thereby influencing its interaction with the surrounding liquor, and/or also re-open the aspirated pits, making the layer more open in structure. These micro-cracks are the result of a damaged surface layer caused by the mechanical preparation of the pieces of wood, e.g. sawing (Salin, 2008), and influence the behaviour of the wood pieces when diffusion measurements are made.

• Wood material has a variety of chemical compositions (e.g. polymers) in its cell walls; interactions between diffusing solute ions and these components (e.g. via sorption) affect the measurement of concentration profiles, so these factors affect the determination of the diffusivities in the system.

14

• Native Norway spruce wood contains small amounts of metal ions (typically 1-5 g/kg wood), the main ones being Ca2+ , Mg2+ , Mn2+ , Na+ and K+ . Although these ions are present in small quantities, they may participate in the sorption mechanism induced by ion-exchange, and thus aggravate the measurements. Therefore, prior to the concentration profile measurements being made, the sample of wood can be treated with a dilute acid solution (e.g. sulphuric acid: pH=0.5/0.31 M H2SO4), to remove native metal ions, air and any floating fines present.

• In most of the methods, at the end of the impregnation time, the impregnated samples were either immediately processed in the diffusion measurement procedures or stored in a refrigerator (e.g. Kazi and Chornet, 1998) at freezing temperature to stop any further diffusion. These slow freezing rates may lead to incorrect measurements being made so, in order to minimize this effect, the impregnated pieces of wood may be placed in liquid nitrogen (-180oC), which has a very rapid freezing rate. The effect of inward migration of chemicals into the impregnated samples of wood can thus be minimized.

15

3. Research Objectives3. Research Objectives3. Research Objectives3. Research Objectives The aim of this thesis is to develop knowledge of the diffusive mass transport of ions in wood. With the focus placed on developing a methodology for investigating the diffusive mass transport of ions in wood, using both experimental and modelling approaches, the objectives of this thesis are:

• To develop a method for measuring the concentration profiles of ions passing through the porous structure of the wood matrix.

• To develop a method that is suitable for differentiating the amount of ions that are

sorbed onto the surface of the cell walls of solid wood from those that are dissolved in the solution in its pores.

• To develop a model that describes the diffusion of ions into a liquid-filled porous

medium and quantifies the transport properties, such as the effective diffusion coefficients and tortuosity factors of the ions in a piece of wood.

• To investigate potential interactions during the sorption of ions from an aqueous bulk

solution onto the surface of wood material.

17

4. Materials and Methods4. Materials and Methods4. Materials and Methods4. Materials and Methods

4.1 Materials and 4.1 Materials and 4.1 Materials and 4.1 Materials and chemicalschemicalschemicalschemicals

MaterialsMaterialsMaterialsMaterials

The wood materials investigated in this study were: wood flour and pieces of wood (both Sw and Hw), MCC (Avicel PH-101; particle size ~50 µm; SIGMA-ALDRICH Co., Germany), Xylan from beech wood (SIGMA-ALDRICH Co., Germany) and Kraft lignin (a softwood lignin, extracted from black liquor using the LignoBoost process at the Bäckhammar Mill, Sweden), along with isolated wood fractions from Sw flour: holocellulose fractions with varying contents of lignin and lignin carbohydrate complex (LCC).

ChemicalsChemicalsChemicalsChemicals

Dilute sulphuric acid solution (0.31 M) was prepared by dissolving reagent grade (95-97%) sulfuric acid (Scharlau, Scharlab S.L., Spain) in deionized water. Stock solutions of different Li+ ion concentrations were prepared by dissolving LiCl salt (≥ 99 %) (Merck KGaA, Darmstadt, Germany) in deionized water. 2 wt. % nitric acid solution was prepared by dissolving trace analysis grade (69 %) nitric acid (ARISTAR, VWR® PROLABO®, Leuven) in deionized water. Peracetic acid (39 wt. %), sodium (meta) periodate (≥ 99.8 %), acetone (≥ 99.8 %), cyclohexane (≥ 99.5 %), Ethylene glycol (≥ 99 %) and potassium bromide (FTIR grade; ≥ 99 %) were purchased from Sigma-Aldrich, and used as received. Sodium chloride (min. 99.5 %) were purchased from Scharlau, and used as received. NaOH (ACS, Reag, Ph Eur) and NaAc (Solution 50%) purchased from Merck, and used as received.

Preparation of wood pieces and wood flourPreparation of wood pieces and wood flourPreparation of wood pieces and wood flourPreparation of wood pieces and wood flour

A stemwood disc 23 cm in diameter (free from bark) taken from a 31±1 year-old Norway spruce (Picea abies L) was investigated. Samples of both Sw and Hw were prepared carefully using a vertical band saw (Mossner Rekord, August Mössner KG, D-7075 Mutlangen, Germany and a metal cutting band saw : L.S. Starrett Company Ltd., Jedburgh, Scotland with 14 teeth per inch) and were kept preliminarily in an airtight polyethylene (PE) bag at 1oC. It was assumed that a stemwood disc contains 50% Sw, 30% Hw and that the remaining 20% is an intermediate between Sw and Hw (Sandberg and Sterley, 2009). In this study, both Sw and Hw were investigated. Only samples free from rot and other deformations were selected and cut into rectangular prototype pieces using the same vertical band saw mentioned above. The dimensions of the prepared pieces of wood were 100 × 25 × 7.5 mm3 (L × R × T). The radial dimension was chosen to be approx. 3 times larger than the tangential dimension in order to minimize the effect of mass transport through the radial direction on the measurements of mass transport in the tangential direction. The material was then stored in an airtight PE bag in a freezer at -18oC. Defrosting the pieces took 24 h at ambient temperature. For the study in which the concentration profiles were measured, surface damage to the wood caused by rough sawing was minimized by planing all four vertical surfaces of the pieces carefully (parallel to the longitudinal fibre direction) with a hand plane (Stanley Hand Tools, Stanley Canada, Mississauga, ON) to remove a surface layer of approx. 1-1.5 mm in thickness thus wood pieces used in the actual experiments had dimensions of 100 × 23 × 6.5 mm3 (L × R × T). A similar approach of minimizing the surface damage to the wood caused by rough

18

sawing, was not possible for the two surfaces perpendicular to the longitudinal direction of the fibres, so they can be expected to have a substantial number of micro-cracks e.g.in middle lamella. In the case of the equilibrium sorption studies, the pieces of wood were ground in a Wiley-type mill (< 1 mm). The resulting wood flour was stored in an airtight PE bag until further use.

IsolatIsolatIsolatIsolating theing theing theing the components of the components of the components of the components of the wood wood wood wood

Peracetic acid (PAA) delignificationPeracetic acid (PAA) delignificationPeracetic acid (PAA) delignificationPeracetic acid (PAA) delignification: Prepar: Prepar: Prepar: Preparinginginging holocellulose fractions with varying contents of holocellulose fractions with varying contents of holocellulose fractions with varying contents of holocellulose fractions with varying contents of

lilililiggggnin nin nin nin from Sw flour from Sw flour from Sw flour from Sw flour

Holocellulose fractions with varying contents of lignin were prepared from the Sw flour using PAA. The delignifying procedure, reported in detail by Kumar et al., (2013), was performed at room temperature with 5 wt. % Sw flour samples and reaction times between 6 and 72 h (namely 6, 24, 51 and 72 h). PAA loadings of 5.5 g/g dry Sw flour were used.

Periodate oxidation: Periodate oxidation: Periodate oxidation: Periodate oxidation: Preparation of lignin/LCC fraction from Sw flourPreparation of lignin/LCC fraction from Sw flourPreparation of lignin/LCC fraction from Sw flourPreparation of lignin/LCC fraction from Sw flour

The lignin/LCC fraction was prepared by oxidizing the Sw flour with sodium periodate, employing a procedure reported in detail by Alam and Van De Ven, (2014). The oxidation procedure was carried out in a glass beaker equipped with an overhead stirrer, and the reaction mixture (Sw flour, sodium metaperiodate and sodium chloride) was stirred continuously in the dark, at room temperature, for 6 days. At the end of the reaction, ethylene glycol (~ 4 ml) was added to the reaction mixture to quench the residual periodate and the mixture was washed repeatedly with deionized water. A purified and oxidized suspension of Sw flour was then stirred gently at 80-90 oC in an oil bath for 6 hours. The sample was cooled to room temperature and the non-dissolved, brown, solid material (lignin/LCC fraction) was separated off by vacuum filtration.

Analysis Analysis Analysis Analysis ofofofof the cthe cthe cthe compositionompositionompositionomposition of the woodof the woodof the woodof the wood

The untreated and delignified solids were dried in an oven at 50 oC for several days. The following methods were then used to analyse their contents of structural carbohydrates and acid-soluble and insoluble (Klason) lignin.

Analysis of Analysis of Analysis of Analysis of Klason lignin Klason lignin Klason lignin Klason lignin

The method used was based on the procedure presented by Theander and Westerlund (1986), whereby Klason lignin is measured gravimetrically after complete acid hydrolysis with 72% H2SO4. In the present study, a 200 mg oven-dried sample was weighed before 3 ml of 72% H2SO4 was added. The sample was evacuated for 15 min and placed in a water bath at 30 oC for 1 h. Then 84 g of deionized water was added before it was heated to 125 oC in an autoclave for 1 h. Afterwards the sample was filtered off, and the solid residue collected was referred to as Klason lignin. The filtrate from the hydrolysis process was used later to measure the content of acid-soluble lignin and for analysis of carbohydrates.

AAAAnalysis of analysis of analysis of analysis of acidcidcidcid----soluble ligninsoluble ligninsoluble ligninsoluble lignin

The content of acid-soluble lignin was calculated in relation to the absorbance value measured with UV at a wavelength of 205 nm in a Specord 205 (Analytik Jena); an absorptivity constant of 110 dm3g-1cm-1 was assumed (Lin and Dence, 1992).

19

Analysis of cAnalysis of cAnalysis of cAnalysis of carbohydratearbohydratearbohydratearbohydratessss

The filtrate resulting from the acid hydrolysis was diluted to 100 ml in a volumetric flask. This solution was diluted a further 5 times before being filtered through a 0.45 µm PVDF filter. The monomeric sugars were analysed using the Dionex ICS-5000 HPLC system equipped with CarboPac PA1 columns and run using NaOH, NaOH/NaAc (0.2 M) as the eluents; fucose was used as the internal standard. An electrochemical detector was used for detection, and the software used was Chromeleon 7, Chromatography Data System, Version 7.1.0.898. The amounts of cellulose, (galacto) glucomannan and xylan were calculated from the carbohydrate analysis using the assumptions and corrections described by Jedvert et al., (2012).

4.2 4.2 4.2 4.2 Measurement of dMeasurement of dMeasurement of dMeasurement of density and porosity ensity and porosity ensity and porosity ensity and porosity The mass of air-dried (at normal conditions i.e. room temperature (R T)) samples divided by their volume gives the bulk density (ρb). The volume of the samples was determined by simply measuring their dimensions. The AccuPyc 1330 helium pycnometer (MICROMETRICS) was used to determine the volume of the cell walls of the solid wood. Helium gas was used as the displacement medium to fill as small pores as possible; for these measurements, wood flour (< 1 mm sieve) was used to have as many pores as possible open to the displacement medium. The density of the cell wall (ρp) was then calculated with the solid volume that was measured for the cell walls and the weight of the sample. Using the normal bulk density (ρb) and the solid density of the cell wall (ρp), the total porosity can be calculated as:

�� !"�#�$%�& = 1 − ��( (4.1)

This total porosity includes all of the pores, regardless of whether they are open or closed. Hg porosimetry was carried out with a Micrometrics AutoPore IV Mercury Porosimeter and analysed using the software AutoPore IV 9500, Version 1.09. The porosity determined by mercury intrusion porosimetry (Hg porosimetry) only determines the percentage of pores that are accessible to Hg. Specimens of about 0.2 g in size were cut perpendicular in the dimensions of about 2.5 (L) x 24 (R) x 7.5 (T) mm. Specimens approx. 2.5 mm in length were the smallest practical length that could be prepared. This length is nonetheless small enough to cut most of the fibres (the average fibre length is about 3 mm) and means that, in most cases, mercury can enter the lumen of the fibre without passing through a pit pore. Consequently, the intrusion volume and the diameters of the lumen could be measured in a satisfactory way. Several duplicates were prepared from two different pieces of wood and they were all analyzed. The porosity (Hg porosimetry) can be calculated from the intrusion pore volume, determined using Hg porosimetry, and the solid volume of the cell wall, determined using helium pycnometer, thus:

"�#�$%�&)*+,�#�$%-.�#&/ = 01�234561762 8693: :;44⁄01�234561762 8693: :;44⁄ =� 99>;994695?8693: :;44⁄ (4.2)

In this study, the intrusion volume of a pore was measured using Hg-porosimetry divided into 2 parts: (i) the total intrusion volume of a pore, which applies to the whole range that the instrument is able to measure (i.e. diameters 0 - 330 µm), and (ii) the effective transport-through-intrusion (diffusion-accessible) volume of a pore (i.e. diameters 0 - 41 µm).

20

4.3 4.3 4.3 4.3 Studies of eStudies of eStudies of eStudies of equilibrium sorption quilibrium sorption quilibrium sorption quilibrium sorption

Sorption experimentsSorption experimentsSorption experimentsSorption experiments

Batch sorption experiments were conducted on flour samples of both Sw and Hw at various temperatures and using different concentrations of bulk aqueous LiCl solution in order to determine the amount of Li+ ions that are sorbed onto the surfaces of solid wood at equilibrium conditions (qwood,e). The experimental procedure for studying sorption is reported in detail in Paper III. Wood flour was placed in polypropylene (PP) beakers containing a bulk aqueous LiCl solution of the desired concentration, with a ratio of liquor to wood flour of 100:1(400:4 (g/g)). This mixture was shaken sporadically during the course of the experiment and, at the end of the specified residence time for each particular temperature, it was centrifuged (SERIE 130; F.I.E.M.M.E., Establishments Couprie 7, Quai Clande Bernard, Lyon, France) at approx. 2960 rpm for 5 minutes to separate the wood flour from the solution. Samples of wood flour treated with LiCl were then lyophilized (Labconco, Kanas City, MO, USA) for 5 days. The freeze-dried sample of wood flour treated with LiCl was then acid leached at room temperature with 2 wt. % HNO3 for ~ 24 h. The leaching liquor was analyzed for its concentration of Li+ ions using Flame Atomic Emission Spectroscopy (FAES).

4.4 4.4 4.4 4.4 Measurement of the cMeasurement of the cMeasurement of the cMeasurement of the concentration profileoncentration profileoncentration profileoncentration profilessss Figure 4.1 illustrates the steps in the experimental procedure followed in this study, a detailed description of which is reported in Papers II and V. The wood pieces were pre-impregnated with either water or dilute acid solution in a vacuum-pressure cycle, which removed air and any metal ions that may present in the wood. During the preliminary study (Papers I and II), in which experiments were carried out for treatment times of 4 and 12 h at the specified temperature (R T, 40 and 60oC), the pieces of wood were pre-impregnated with water. In the later chemical impregnation experiments (Paper V) at 40 oC and for treatment times of 36, 72 and 168 h, the pieces were pre-impregnated with a dilute acid solution (pH=0.5/0.31 M H2SO4). In the experiments using samples of wood flour, only traces of major metal ions (Ca2+, Mg2+, Mn2+, Na+ and K+) were observed after dilute acid treatment (Paper III). The effect of these metal ions on mass transport of Li+ in wood is thus negligible.

Impregnation experimentsImpregnation experimentsImpregnation experimentsImpregnation experiments ---- Impregnation with LiCl Impregnation with LiCl Impregnation with LiCl Impregnation with LiCl

The pieces of wood pre-impregnated with either water or dilute acid were first dabbed with filter paper before being immersed in a LiCl solution. During the preliminary study (Paper II), with samples of both Sw and Hw, the pieces were immersed in a 1 M LiCl solution with a wood-to-liquor ratio of 1:125. In the later experiments, the samples of Sw (Paper V) were immersed in a 1 M LiCl solution and Hw in a 0.5 M, at a wood-to-liquor ratio of 1:75. After chemical impregnation at the time intervals specified and at each particular temperature, the pieces were removed and placed in liquid N2 (-180oC) to minimize the further migration of any ions. Finally, these frozen pieces were lyophilized (LABCONCO®, Kansas City, MO, USA) for about 2 weeks to avoid secondary thermal diffusion.

In preparation for measuring the concentration of Li+ ions, each of the impregnated pieces of wood was cut into small cubes (Figure 4.2) with dimensions of approx. 6 x 6 x 6.5 mm3. Each cube was microtomed in the tangential direction into slices ~ 0.3 mm thick and then oven-dried at 105oC for 1 h to remove any moisture accumulated during the microtoming process. The slices of wood were placed in desiccators containing blue gel salt and left at R T until constant mass was reached. Each slice was then acid leached at R T with 2 wt. % HNO3 for ~ 24 h. The

21

leaching liquor was analysed for its concentration of Li+ ions using Flame Atomic Emission Spectroscopy (FAES).

Figure 4.1 The steps in the experimental procedure used to measure the concentration profiles.

Figure 4.2 (A) Locations of the small cube samples in an idealized piece of impregnated wood. (B) Microtoming a small cube into slices of thickness ~0.3 mm for use in measuring the concentration profile of Li+ ions. DL and DT (= DR) are the diffusion coefficients of the Li+ ions from the aqueous bulk solution to the wood in the longitudinal, tangential and radial directions, respectively.

22

4.5 4.5 4.5 4.5 CharacterizationCharacterizationCharacterizationCharacterization of the mof the mof the mof the material aterial aterial aterial

Flame Atomic Emission Spectroscopy Flame Atomic Emission Spectroscopy Flame Atomic Emission Spectroscopy Flame Atomic Emission Spectroscopy (FAES)(FAES)(FAES)(FAES)

The leaching liquor was analyzed for Li+ ion concentration at the end of the leaching period using Flame Atomic Emission Spectroscopy (Thermo Scientific, iCE 3000 series, AA spectrometer, Cambridge, United Kingdom). An air-acetylene torch was used as the source of the flame; the emission was measured at a wavelength of 670.8 nm. The concentration of Li+

ions was kept at an optimum working concentration range of 0.02 - 5 µg/ml. The average deviation from the mean values of these measurements was ~8%.

Fourier Transform Infrared Spectroscopy (FTIR) Fourier Transform Infrared Spectroscopy (FTIR) Fourier Transform Infrared Spectroscopy (FTIR) Fourier Transform Infrared Spectroscopy (FTIR)

FTIR spectra were recorded in a Fourier Transform Infrared Spectrophotometer (Perkin Elmer, Spectrum One, FTIR-Spectrometer) using the KBr discs technique. Each spectrum was collected 16 times in the range of 4000 - 400 cm-1 with a resolution of 4 cm-1 and corrected for background noise. Each spectrum was collected 16 times in the range of 4000-400 cm-1 with a resolution of 4 cm-1 and corrected for background noise. The spectra were normalized using a peak at 1424 cm-1, which corresponds to the CH2 contributed by the cellulose.

XXXX----ray photoelectron spectroscopy (XPS) ray photoelectron spectroscopy (XPS) ray photoelectron spectroscopy (XPS) ray photoelectron spectroscopy (XPS)

X-ray photoelectron spectra were collected using a Quantum 2000 scanning Electron Spectroscopy for Chemical Analysis (ESCA) microprobe from Physical Electronics (Chanhassen, Mn, USA) with a monochromatic Al Kα source (1486.6 eV). The samples were analyzed at a take-off angle of 45o, with the area under analysis being about 400 x 500 µm in size and 4-5 nm in depth. The pressure in the chamber during XPS analysis was lower than 10-

9 torr, making it an ultra-high vacuum regime. An overall survey spectrum (i.e. low-resolution) from 0 to 1100 eV binding energy, and a high resolution spectrum of the C1s region from 280 to 300 eV and O1s region from 525 to 535 eV, were collected. Analyses of the chemical bonds of carbon and oxygen were made by fitting the curves of the C1s and O1s peaks from the high resolution spectra and deconvoluting them into sub-peaks using ESCA tools (MultiPak Vers. 6.1).

BET measurements of surface areaBET measurements of surface areaBET measurements of surface areaBET measurements of surface area

The specific surface area of both untreated and LiCl treated samples were measured according to the BET theory (Brunauer–Emmett–Teller) from measurements of nitrogen adsorption using a Micrometrics TriStar 3000 instrument. Hornification whilst the samples are being dried is avoided by performing a solvent exchange and replacing water with cyclohexane. This was done through a solvent exchange procedure in which the water was replaced initially with dry acetone which, in turn, was replaced with cyclohexane (Palme et al., 2014). The samples were then dried overnight in a stream of nitrogen.

4.4.4.4.6666 Theoretical model formulationTheoretical model formulationTheoretical model formulationTheoretical model formulation: E: E: E: Estimatistimatistimatistimating theng theng theng the transport properties transport properties transport properties transport properties

of of of of ionionionions in woods in woods in woods in wood The model used to determine the transport properties of Norway spruce wood, such as the effective diffusion coefficients and tortuosity factors of Li+ ions, is based on the theory of the transport of diluted species in porous media (Paper V). The model is solved using COMSOL® multiphysics modelling software (Vers. 5.0, COMSOL, Stockholm, Sweden). Equation (4.3)

23

describes the transport of solutes in a porous media saturated with liquid (The Transport of Diluted Species in Porous Media Interface, COMSOL®, Vers, 5.0) thus:

)ɸ + AB�/ �∁D�� + E∁5 − AF∁F,5H �ɸ�� = ∇ ∙ KE� ,5H∇∁5L (4.3)

where Ci is the concentration of the species i in the liquid located in the porous media (g/l), CP,

i is the amount sorbed onto the solid particles (g/kg), k is the wood-Li+ ion equilibrium partition coefficient (l/kg), De,i is the effective diffusion coefficient of species in the porous medium (m2/s), ɸ is the effective transport-through-porosity calculated using Equation (4.2), and ρb and ρp are the densities of the bulk and solid phases, respectively, of the wood (kg/m3). The two terms on the left-hand side of Equation (4.3) describe the accumulation of the species within the liquid and the solid, respectively, whilst the term the right-hand side introduces the transport of the species due to diffusion.

The initial condition (I.C) and boundary condition (B.C) are:

I.C: time t = 0, Concentration Ci = 0 g/l for all domains of the geometry

B.C: for all times t, Concentration Ci = 1 M = 6.941 g/l for all surfaces of the geometry.

The effective diffusion coefficient (De) for the transport of a solute in a porous medium saturated with a liquid is significantly lower than the free-water diffusion coefficient (Df) because of the constricted and elongated (tortuous) transport paths. In a saturated porous medium, De is related to Df as (Saripalli et al., 2002):

� = ��ɸƟ� (4.4)

where ɸ is the porosity of the piece of wood that is accessible to diffusion, Ɵ is a constrictivity factor to account for the constricted transport paths (caused by the small pores and pore throats in a porous medium and usually assumed to be ~1) and τ is a tortuosity factor. The latter accounts for the reduction in diffusive flux caused by the tortuous path lengths formed by the solute molecules, which can be compared to the straight paths in an unrestricted aqueous medium. The estimation procedure used to determine the self-diffusion coefficient of Li+ ions in a free solution (Df) can be found in Appendix A1.

25

5555.... Results and DiscussionResults and DiscussionResults and DiscussionResults and Discussion 5.1. Development of 5.1. Development of 5.1. Development of 5.1. Development of an an an an experimental procedure: experimental procedure: experimental procedure: experimental procedure: CCCConcentration profiles oncentration profiles oncentration profiles oncentration profiles

of ions inof ions inof ions inof ions in piecepiecepiecepiecessss of of of of wood wood wood wood One finding of this work is that the preparation of samples is of the utmost importance. Our results indicate that even if the surface seems to be smooth from visual inspection, this is not in fact the case. The following abbreviations are used in the text below: NWP (Normal Wood Pieces) refers to pieces of wood that were prepared carefully from a disc of stemwood using a vertical band sawing machine and SsWP (Surface smoothed Wood Pieces) refers to pieces that were planed carefully on all 4 vertical surfaces (four surfaces parallel to the longitudinal fibre direction) to a depth of about 1 - 1.5 mm with a hand plane in order to peel off the damaged surface layer from the NWP. The concentration of ions (i.e. moles of Li+ /kg dried wood) is plotted against the thickness of the impregnated pieces of wood to describe the effects of various factors on the concentration profiles. It should be kept in mind that this concentration, which is measured experimentally, is the total concentration of the Li+ ions: it is the sum of the Li+ ions that are sorbed onto the surfaces of the wood’s cell walls and dissolved in the solution located in the pores. Figure 5.1 provides an example of how damaged surface layers influence the concentration profiles of Li+ ions in samples of Norway spruce. The figure shows that there is a higher concentration of Li+ ions in the outer surface layers (< 1.5 mm depth) of the NWP compared with SsWP. This is a clear indication that the rough surface layers, which were formed during the rough sawing process, may have affected the mass transport of Li+ ions in wood. It therefore seems that the formation and behaviour of the damaged surface layers depends on the method of preparation the pieces of wood have been subjected to. So, in the sections discussed below, SsWP are used in order to minimize the effects of such damaged surface layers (micro-cracks) on the measurement of concentration profiles of Li+ ions in wood.

Figure 5.1 Concentration profiles of Li+ ions measured in a piece of Norway spruce Sw treated with 1 M LiCl solution at room temperature for 4 h residence time: NWP vs. SsWP for Cube a (see Figure 4.2 (A)).

26

Concentration profiles at various locations in a piece of woodConcentration profiles at various locations in a piece of woodConcentration profiles at various locations in a piece of woodConcentration profiles at various locations in a piece of wood

Figure 5.2 shows the concentration profiles at different locations in a single piece of LiCl (aq.) impregnated wood. The differences between the profiles are due to the fact that mass transport properties are strongly influenced by the anisotropy of the samples of wood. It is evident here that the concentration profiles of Cubes a (taken from the middle of the piece) and c (taken from the edge) (see Figure 4.2 (A)) differ. In the case of Cube a, this is because the diffusion of Li+ ions was mainly influenced by mass transport in the transversal direction alone, whereas in the case of Cube c, diffusion was influenced by mass transport in both the longitudinal and the transversal directions. Moreover, as shown in Figure 5.2, the concentration profiles observed for Cubes a and b were within the experimental error: it is therefore reasonable to assume that, in the case of Cube a, the concentration is not affected by longitudinal mass transport due to the diffusion distance, which is relatively long compared to mass transport in the radial and tangential directions. However, the behaviour seen at the outer layers (< 1 mm depth) and the edge (Cube c) may be attributed to micro-cracks caused by the mechanical preparation of the pieces (e.g. sawing). The four surfaces parallel to the longitudinal fibre direction were planed using a hand plane, which eliminated a large fraction of the micro-cracks. It was not possible to use a similar approach with the two surfaces perpendicular to the longitudinal fibre directions, however, so a substantial amount of micro-cracks can be expected at these two surfaces, e.g. in the middle lamella.

Figure 5.2 Concentration profiles of Li+ ions measured in a piece of Norway spruce Sw treated with 1 M bulk LiCl solution at 60 oC for 4 h residence time. Cubes a, b and c refer to the position within the piece of wood from where the samples originate (see Figure 4.2 (A)).

For the experimental conditions studied (i.e. relatively neutral or low pH), it was the carboxylic sites (pKa 3-4) that were predominantly ionized (Sjöström 1993): they therefore exist in anionic forms, and the Li+ cations can bind to the carboxylic sites. However, the concentrations of Li+ ions observed at the outer layers (< 1 mm depth) and the edge (Cube c), were much higher than those of the carboxylic acid groups (0.086±0.007 mol/kg) (Werkelin et al., 2010) present in the samples of native Norway spruce. This result suggests that, in addition to the carboxylic sites, other functional groups of the components of the wood’s cell walls (such as carbonyl and hydroxyl) might also be involved in the interactions with the Li+/Cl- ions. It is therefore evident that the transport of Li+ ions in wood may be influenced by other phenomena, such as surface diffusion, adsorption/desorption and Donnan equilibrium.

27

5.2. 5.2. 5.2. 5.2. InvestigaInvestigaInvestigaInvestigation of possible interactions between Norway spruce tion of possible interactions between Norway spruce tion of possible interactions between Norway spruce tion of possible interactions between Norway spruce

wood and aqueous LiCl solutionwood and aqueous LiCl solutionwood and aqueous LiCl solutionwood and aqueous LiCl solution

XPS XPS XPS XPS characterizationcharacterizationcharacterizationcharacterization

The XPS survey spectra of the samples of Sw flour before and after treatment with aqueous LiCl solution reveal the presence of Li and Cl ions in the samples treated with LiCl, thus indicating sorption onto their surfaces. Wide-scan XPS spectra are presented in Figure 5.3, and the composition of the surface of the elements detected, estimated using the atomic sensitivity factor and the area under each peak, is presented in Table 5.1. In general, the theoretical value of the O/C ratio for carbohydrates (0.83 for cellulose and 0.8 for hemicelluloses) is higher than the corresponding value for lignin (0.33) (Sernek, 2002), which reflects the abundance of the hydroxyl groups present in carbohydrates. In this case, although the O/C values measured were significantly lower than those estimated from the chemical composition and elemental analysis, they are in good agreement with values reported previously for wood material by Sinn et al., 2001 and Inari et al., 2006. Whilst this difference reflects the restriction of the XPS analysis to the surface layer, it may also be an effect of the adaptation of cellulose surfaces which, upon drying, expose more hydrophobic structural motifs, as reported previously (Johansson et al., (2012)). According to previous findings, the high vacuum released during XPS analysis, which allows for vaporization of the residual water contained in the material, can also contribute to the low O/C values observed (Inari et al., 2006). A slight decrease in the O/C ratio was observed for Sw flour upon LiCl treatment, as shown in Table 5.1. As mentioned above, this result is likely to be associated with the rearrangement of cellulose fibrils during drying, leading to the reduced accessibility of –OH groups and a more pronounced hydrophobic surface. The indication here is that this rearrangement seems to be promoted by LiCl treatment.

Figure 5.3 Survey spectra of Norway spruce Sw flour (a) before and (b) after treatment with 1 M LiCl aqueous solution for 3 h at room temperature (R T).

Table 5.1

Surface composition of the elements on samples of Sw flour before and after treatment with 1M LiCl for a residence time of 3 h at R T, as detected by XPS. Sample Element

C1s O1s Li1s Cl2p O/C Untreated Sw flour 73.96 26.04 - - 0.35 LiCl treated Sw flour 73.12 24.29 2.06 0.53 0.33

28

Deconvolution of the carbon (C1s) and oxygen (O1s) signals observed could provide further insight into these changes. The C1s signal was deconvoluted into four sub-peaks corresponding to (i) C1- carbon atoms bonded only to hydrogen or other carbon atoms (C-C or C-H), appearing at a binding energy (BE) of 283.4±0.11 eV and arising predominately from the lignin and extractive constituents of the wood (Nzokou and Pascal Kamdem, 2005), (ii) C2 - carbon atoms bonded to a single, non-carbonyl, oxygen atom (C-O or C-OH), at a BE of 285.0±0.13 eV, arising from both the carbohydrate and lignin constituents (Sernek, 2002), (iii) C3 - carbon atoms bonded to one carbonyl (mostly lignin) or two non-carbonyl oxygen atoms of polysaccharides (C=O or O-C-O), with a BE of 286.5±0.18 eV, and (iv) C4 - carbon bonded to one carbonyl and one non-carbonyl oxygen (O-C=O) at a BE of 287.7±0.17 eV (Dorris and Gray, 1978). The O1s signals were deconvoluted into two sub-peaks: O1 (at a BE of 531.3±0.13 eV), corresponding to oxygen atoms bonded to carbon with a double bond (O-C=O), and the O2 (at a BE of 532.4±0.11 eV) arising from oxygen atoms bonded to carbon with a single bond (C-O) (Inari et al., 2006; Nzokou and Pascal Kamdem, 2005). Figure 5.4 shows the high resolution and deconvoluted spectra of the C1s and O1s signals from the Sw samples before and after treatment with LiCl.

Figure 5.4 Spectra of carbon (C1s) (A) and oxygen (O1s) (B) signals from Norway spruce Sw flour (a) at high resolution and (b) deconvoluted. (i) Before and (ii) after treatment with 1 M LiCl aqueous solution for 3 h residence time at R T (as indicated).

It is evident from Figure 5.4 (A)-(a) that the changes in the C1s signal observed upon LiCl treatment are associated with a slight increase in the contribution from the C1 (C-C/C-H) component as well as a decrease in that from the C2 (C-O/C-OH) and C3 (C=O/O-C-O) components. This indicates, once again, an enrichment of the more hydrophobic C-C/C-H structures on the surface that are a result of the rearrangement of the wood matrix discussed above. Corresponding analyses of the O1s signals confirmed the same trend: the contribution of the O2 component (C-O structures) decreased, indicating the withdrawal of C-O structures

29

from the surface and thus giving rise to a simultaneous relative increase in the O1 (C=O) signal. More importantly, a new signal (O0) at 530.1±0.01 eV was observed (Figure 5.4 (B)-(b)). This signal, which is generally assigned to metal oxides (Chastain and King, 1992), can be attributed to interactions between the functional groups of the wood material that contain oxygen and the Li+ ions (Li-O interactions). Harilal et al. (2009) assigned this signal to lithium oxides and peroxides (Li2O, Li2O2) and other species containing lithium and oxygen (Li2CO3, LiOH). In order to explore these observations further, samples of Sw flour were fractionated into its constituent fractions: holocelluloses with high and low lignin contents and lignin/LCC fractions. The resulting fractions were treated individually with LiCl and analysed using XPS (Paper IV). The same decreasing trend observed for the Sw samples was also observed for holocelluloses in the samples with high and low contents of lignin in the values of the O/C ratios upon LiCl treatment, which indicates the rearrangement of the cellulose fibrillary structures (as mentioned above). It is evident that this rearrangement is probably facilitated by the LiCl treatment step: it could possibly be explained by the fact that LiCl disrupts the stabilizing H-bonding and thereby promotes mobility within the fibre’s matrix. Enhanced mobility in the more delignified fraction makes this rearrangement even more pronounced and results in the O/C ratio being decreased further. The deconvoluted C1s spectra of the holocellulose fractions, in line with the observations above, show an accumulation of C1 class carbon on the surface, with a simultaneous decrease in the C2 contribution upon LiCl treatment; the effect is much more pronounced for the more delignified samples (holocelluloses with low lignin content) with enhanced matrix mobility. This indicates that the rearrangement at the surface (leading to the accumulation of the C-C/C-H motifs at the surface) is facilitated by the action of the LiCl. The action of LiCl is more noticeable in the more delignified samples with enhanced accessibility: matrix mobility could therefore be attributed to the disruption of the H-bonds. The corresponding analysis of the deconvoluted O1s spectra indicates, once again, the occurrence of Li-O interactions (the appearance of an O0 signal), with increased intensity upon lignin removal, when compared to the original Sw samples. Furthermore, the deconvoluted oxygen spectra reveal a decrease in the O1 signal upon LiCl treatment, being more pronounced for more delignified samples. This could indicate the withdrawal of the carbonyl functionalities (C=O) from the surface of the holocellulose samples as a part of the rearrangement promoted by the LiCl, and facilitated by the improved mobility in the more delignified samples. On the other hand, it could also indicate the removal of hemicellulose acetyl groups during LiCl treatment. Moreover, changes in carbonyl signals may also be associated with the removal of oxidized shortened chains formed as side products during PAA delignification. However, no significant oxidative effects on carbohydrates are expected under the mild conditions applied in this study (Jääskeläinen et al., 2000), so the losses observed in carboxyl signals are indicative instead of the enhanced removal of hemicellulose carbonyls (deacetylation) from the more accessible matrix. The variations observed in the O2 signal (C-O) also concur with this explanation. In the original Sw sample, the O2 portion on the surface was reduced significantly upon LiCl treatment, which can be understood as being the surface hydrophobization mentioned above that occurs upon rearrangement. As delignification proceeds, water-soluble oxidized carbohydrates become more accessible. Some of these are washed out during LiCl treatment, giving rise to a reduction of O1 and the simultaneous relative increase of O2, as observed. It should be pointed out here that, when changes in O2 are considered, the gradual removal of lignin (in Sw, from holocelluloses with a high lignin content to those with a low lignin content)

30

makes it increasingly difficult to “hide” C-O motifs from the surface upon rearrangement, which may affect the variations observed. The lignin fraction in the Sw sample was isolated and included in the XPS analysis for the purpose of comparison (Paper IV). No rearrangement leading to the accumulation of C-C on the surface could be observed. Here, it is likely that the structure created during drying after LiCl treatment is determined by hydrophobic interactions of the lignin motifs, leading to the exposure of more hydrophilic groups on the surface. It was even more interesting to observe that there were no indications of Li-O interactions either. It is obvious that both of these phenomena seem to be characteristic of the carbohydrate fibrillar network that is capable of causing surface adaptation and interactions with Li+/Cl- ions. Nevertheless, the lignin fraction seems to retain a considerable amount of LiCl, probably through non-specific interactions between the surfaces of the lignin/LCC and the aqueous LiCl solution.

BET BET BET BET measurements of measurements of measurements of measurements of surface area surface area surface area surface area

In order to investigate the rearrangements that were observed as being promoted by LiCl further, variations were monitored in the surface area of samples with holocelluloses with high and low contents of lignin before and after LiCl treatment. Upon LiCl treatment, the BET specific surface area increased from 1.5±0.0 m2/g to 2.4±0.1 m2/g for holocelluloses with a high lignin content, and from 1.6±0.03 m2/g to 2.7±0.6 m2/g for those with low a lignin content. LiCl treatment promoted an increase in surface area, with the effect being more pronounced as the content of carbohydrate increased, i.e. there is an increase in both the matrix mobility and the accessibility of the carbohydrate components. The increase observed in the surface area may be an effect of the aforementioned disruption of the stabilizing H-bonding, which promotes rearrangement of the fibrillar structure on the surface caused by the action of the Li+/Cl- ions. The literature (Nada et al., 2009) reports that the crystallinity of the cellulose materials decreases upon LiCl treatment, allowing an increased surface area to be observed. However, the increased surface area of the cellulose-rich samples can also be affected by the removal of soluble components during LiCl treatment.

FTIR FTIR FTIR FTIR characterizationcharacterizationcharacterizationcharacterization FTIR characterization of Sw flour samples (untreated, acid treated, and after acid treatment, followed by LiCl treatment) was carried out to identify the functional groups that are affected by the interactions between the Sw flour and the Li+/Cl- ions, and thus probably play an important role in the sorption of ions (Figure 5.5). The relevant functional groups and their tentative band assignments are in agreement with those reported in the literature for softwoods (Schwanninger et al., 2004; Pandey, 1999). Comparing the spectra of the original wood with those of wood treated with LiCl, it is clear that there are differences in the absorbance signals originating from the groups containing –OH (e.g. alcohols, phenols and absorbed water) and carbonyl (C=O) and C-H functionalities. The differences observed in the absorption bands suggest that these functional groups are affected by the interactions of Li+/Cl- ions with Norway spruce Sw flour. Changes in the stretching vibrations of the –OH detected within a broad range of frequencies (3700-3100 cm-1) indicate interactions between the Li+/Cl- ions and the –OH groups. The changes in the signals corresponding to the C=O stretch in carbonyls (e.g. acetyl groups) present in hemicelluloses (1738-1709 cm-1) and C-H deformations in carbohydrates (1470-1300 cm-1) were also observed upon LiCl treatment (Figure 5.5 (ii)). Corresponding bands shown a clear decrease in intensity, thus indicating the partial removal of acetyl groups during the treatment,

31

which is in agreement with the XPS findings. Furthermore, signals originating from the sorbed water (1650-1635 cm-1) were enhanced significantly in the LiCl treated samples, implying that the enhanced interaction with water is an effect of LiCl treatment.

Figure 5.5 FTIR spectra of the Sw flour samples: untreated (native), treated with dilute acid and treated with 3 M LiCl after acid treatment at room temperature: (i) Overall spectra and (ii) magnified spectra between 1800 and 1300 cm-1.

It was not possible to draw any conclusions from the results described above as to which functional groups of the individual wood constituents are involved in the sorption of ions. Therefore FTIR spectroscopy was also used to study the behaviour of the individual model components of wood: cellulose (Micro Crystalline Cellulose (MCC)), hemicelluloses (beechwood xylan) and lignin (Kraft lignin) (Paper III) as well as isolated fractions of Sw flour: holocelluloses with high and low contents of lignin and lignin/LCC fractions (Paper IV), and before and after 3 M LiCl treatment at room temperature. It is evident from this study that the bands assigned to the –OH groups of cellulose, xylan and lignin show changes upon LiCl-treatment, thus indicating their participation. The partial removal of acetyl groups assumed during aqueous LiCl treatment was supported further by a decrease in the intensity of the signals corresponding to C=O and the C-H of carbohydrates. As the lignin was removed gradually from the Sw flour, and upon LiCl treatment, the spectral changes in these wave number intervals were even more pronounced, implying that the removal of acetyl groups was even more extensive. Furthermore, signals originating from the sorbed water (1652-1630 cm-1) were enhanced significantly in delignified samples, implying an enhanced interaction with water during LiCl treatment. Moreover, significant changes in the -OH signals were also detected in the more delignified samples, implying the participation of these functional groups in interactions with LiCl: this might be due to changes in the H-bonding patterns, as already indicated by the XPS results. The lignin/LCC samples showed even more pronounced changes in the signals originating from these groups (i.e. -OH, C=O and sorbed water). According to Alam and Van De Ven, (2014), the softwood lignin isolated by the procedure employed in this study is rich in both phenolic groups and carbonyls originating from the oxidized carbohydrate residues in lignin-carbohydrate complexes. Drastic changes in the -OH region of this sample are indicative of the altered interactions of the phenolic groups due to the presence of LiCl and/or enhanced interactions with water.

32

FAES measurementsFAES measurementsFAES measurementsFAES measurements The effect the composition of the wood had on the amount of Li+ ions sorbed onto samples of Norway spruce was measured quantitatively using Flame Atomic Emission Spectroscopy (FAES). The results revealed that the amount of Li+ ions sorbed increased as the amount of lignin decreased in the sample (Paper IV). The highest concentration of Li+ ion was observed in the carbohydrate-rich portion (i.e. holocelluloses with a low lignin content) and the lowest concentration in the lignin/LCC fraction. It is evident from this study that samples with more OH groups show higher amounts of Li+ ions sorbed, suggesting that OH functional groups have significant interactions with Li+ ions, as was observed in the FTIR studies. Moreover, this result also indicates that there is an increase in the number of interactions the functional groups make with the LiCl solution, which is due to the increased accessibility of the carbohydrate network upon delignification.

Effect of pH during the sorption experimentsEffect of pH during the sorption experimentsEffect of pH during the sorption experimentsEffect of pH during the sorption experiments

The sorption experiments were performed in a closed-batch system in this study, without any adjustment of the pH. It was observed that the pH of the liquid phase decreased, and also that the drop in pH increased with increasing concentration of the Li+ ion in the bulk solution (Figure 5.6). Similar results were observed for the samples of wood flour that were not acid treated prior to LiCl treatment. The pH drop observed was, however, more pronounced for acid treated than for non-acid treated samples prior to LiCl treatment (Figure 5.6). It was estimated that a drop in pH of approximately 1.9 units was due to acid liquor in the wood pores: the remaining drop must therefore have been caused by either the exchange of ions between the wood components and the Li+/Cl- ions or other structural changes in the wood that might have occurred upon LiCl treatment (such as deacetylation). Only a small part of the remaining decrease in pH can be explained by carboxylic groups; at these low pH values it is unlikely that phenolic groups are involved. The remaining drop in pH may therefore be due to deacetylation of carbohydrates upon treatment of wood samples with LiCl, as observed above in the FTIR and XPS studies.

Figure 5.6 Decrease in pH vs concentration of Li+ ions in bulk LiCl aqueous solution during the sorption experiments with Sw flour at room temperature.

33

5.3. Density and porosity 5.3. Density and porosity 5.3. Density and porosity 5.3. Density and porosity measuremeasuremeasuremeasuremememementntntntssss of Norway spruce of Norway spruce of Norway spruce of Norway spruce woodwoodwoodwood Data pertaining to the porosity and volume of pores that are accessible for the diffusion of ions in samples of wood pieces and wood flour is required in order to (a) measure the concentration profiles of the ions in the porous structure of wood and (b) estimate the amount of ions sorbed onto the surfaces of wood cell walls at equilibrium conditions. Measurements of density and porosity are therefore described first. The densities measured (i.e. of the normal bulk and the cell walls) and the total porosities of Norway spruce wood are consistent with previous findings (Plötze and Niemz, 2011) (Table

5.2). In this table, the values of the intrusion volume and porosity of the pore determined by mercury intrusion porosimetry are also reported. Figure 5.7 shows the incremental intrusion volume and pore size distribution of pores of diameter 0 - 41 µm (these diameters refer to pores accessible to diffusion) for untreated samples of both Sw and Hw of Norway spruce. In softwood, pores with a diameter larger than 20 µm may represent the resin canals and earlywood tracheids, and those between 20 and 6 µm represent the latewood tracheids (Zauer et al., 2014), although there is a relatively large overlap between earlywood and latewood tracheids. Both of these pore classes are termed as macrovoids: pore diameters smaller than 6 µm (microvoids) represent the pointed ends of all cell elements as well as of the pit pores (Thygesen et al., 2010).

Sw vs. HwSw vs. HwSw vs. HwSw vs. Hw

Significant differences in porosity and intrusion pore volume were observed between Sw and Hw pieces of Norway spruce (Figure 5.7 (B)). The results show that Hw piece has lower porosity and intrusion pore volume than Sw piece. These differences in Sw and Hw piece are most likely due to differences in the structure of Sw and Hw, e.g. Hw is usually much less permeable than Sw due to pit aspiration and incrustation. But in the case of wood flour samples, similar intrusion pore volume and porosity were observed for both Sw and Hw flour samples (Figure 5.7 (A)). In the determination of concentration profiles and estimation of transport properties discussed in the following sections, constant porosity (57.8±8.9 %) and intrusion pore volume (0.99±0.3 ml/g) are used for both Sw and Hw pieces.

Figure 5.7

Table 5.2

34

Figure 5.7 Pore size distribution of untreated Norway spruce (a) Sw and (b) Hw samples of wood flour (A) and wood piece (B) samples (as indicated); sample 1, sample 2…are replicates that were prepared from 2 different wood pieces.

35

Table 5.2 Normal bulk density, specific cell wall density, total porosity, intrusion pore volume and porosity (Hg porosimetry) of Norway spruce wood.

This study Sample Cell wall density

(ρp) (g/cm3)

Normal bulk

density (ρb)

(g/cm3)

Total porosity

n = 1-(ρb/ ρp) %

Total

intrusion

pore volume

(mL/g)

(0 to 330 µm)

Total porosity

(Hg

Porosimetry)

%

(0 to 330 µm)

Effective

transport-

through

intrusion pore

volume (mL/g)

(0 to 41 µm)

Effective

transport-

through

porosity (Hg

Porosimetry)

% (0 to 41 µm)

Sw flour 1.48±0.002(n*=3) 2.9±0.1(n=3) 81.5±0.7(n=3) 0.45±0.0(n=3) 39.9±1(n=3)

Hw flour 1.48±0.006(n=3) 2.6±0.3(n=3) 79.5±1.6(n=3) 0.41±0.06(n=3) 37.7±3.5(n=3)

Sw piece 0.41±0.01(n=3) 72.4±0.34(n=3) 1.1±0.4(n=10) 59.7±8.6(n=10) 0.99±0.3(n=10) 57.8±8.9(n=10)

Hw piece 0.49±0.01(n=3) 66.7±0.97(n=3) 0.4±0.1(n=5) 38.3±4(n=5) 0.4±0.1(n=5) 36.1±4(n=5)

Reference

(Plötze and

Niemz,

2011)

Norway

spruce

Wood

piece

1.524 0.401 73.68 68.39

*n = number of samples prepared from the two different pieces of wood

36

5.4. Sorption equilibrium of ions onto s5.4. Sorption equilibrium of ions onto s5.4. Sorption equilibrium of ions onto s5.4. Sorption equilibrium of ions onto sololololid portions of wood flourid portions of wood flourid portions of wood flourid portions of wood flour The amount of Li+ ions sorbed per unit weight of (dry) Norway spruce wood flour that is measured experimentally is the total amount of Li+ ions sorbed by the wood (qtotal), i.e. the quantity of Li+ ions that is sorbed onto the surfaces of the cell walls of the solid wood (qwood) and dissolved in the solution within the wood pores (qpores). The measurements are aimed at steady state, so it is reasonable to assume that the concentration of Li+ ions in the liquid in the wood pores is equal to that of the bulk liquor. Using Equations (5.1) and (5.2) below, it is thus possible to estimate the amount of Li+ ions sorbed onto the solid wood. All of the results reported in this section were based on the amount of Li+ ions sorbed onto the solid wood (qwood). qtotal = qwood + qpores (5.1)

qpores = CLi+

, bulk * Vpores (5.2)

where CLi+

, bulk and Vpores are the concentrations of Li+ ions in the bulk solution (mg/ml) and the volume of the pores in the wood flour (ml/g), respectively. The volume of the pores in the samples of wood flour was determined by the mercury porosimetry technique mentioned previously. Vpores was calculated as the total incremental intrusion volume of the pores in the wood flour, with the pores having a diameter of between 0 and 41 µm. The wood-Li+ ion equilibrium partition coefficient (k), is defined as:

� = �NOOP,QRSDT (5.3)

where qwood, e is the amount of Li+ ions sorbed onto the solid wood (mg/g) and CLi+ is the Li+

ion concentration in the solution located in the porous structure of the wood matrix (mg/ml) at an apparent equilibrium condition. In this study, the surface area (sorption sites) available in the wood flour and samples of wood piece was also assumed to be constant. In order to determine the residence time needed to obtain an apparent equilibrium (i.e. the time at which there is no change in the amount of Li+ ion sorbed to wood flour), the experimental data (qwood) was measured at several contact times ranging between 30 minutes and 30 days. The changes observed in the qwood values as the residence time was varied were small, and indeed almost within the experimental error. Furthermore, the results indicated that, for all the experiments, there is a rapid initial rise in the sorption of Li+ ions and steady state is almost reached within 8 h (Figure 5.8). However, the experimental data measured at 24 h was used as the amount of ions sorbed at a condition of apparent equilibrium (qwood,e). Similar results were also observed for Hw flour samples.

37

Figure 5.8 Variations in the amount of Li+ sorbed onto Norway spruce Sw flour with time, at R T. The contact time varied between (a) 30 min and 30 days, and (b) 0.5 and 48 h.

The wood-Li+ ion equilibrium partition coefficients (k) for both Sw and Hw, calculated using Equation (5.3) for experiments with 1 M LiCl aqueous solution at different temperatures, are reported in Table 5.3. The plot of qwood, e versus the concentration of Li+ in bulk LiCl solution (Ce) is called sorption isotherm. In this study, at lower concentrations of bulk LiCl solutions (up to 1 M i.e. 6.941 mg/ml), the relationship qwood, e vs. Ce observed is linear (Figure 5.9): the slope of the linear curve is thus considered as being the distribution coefficient. The values of k obtained using Equation (5.3) (Table 5.3) and the value of the slope obtained from the linear sorption isotherm (Figure 5.9) agreed well. The results indicate that the equilibrium partition coefficient is almost constant with respect to temperature and concentration up to 1 M bulk LiCl solution for samples of both Sw and Hw flours. However, as the temperature increased from R T to 60 oC, the value of k decreased slightly for Sw flour, while the value of k increased slightly for Hw flour. This may be due to the structural and/or chemical changes that might have occurred in the wood at higher temperatures, although the differences are within the estimated experimental errors.

Figure 5.9 Experimental sorption isotherm of Li+ ions onto Norway spruce wood in experiments with lower concentration of bulk LiCl solution (Upto 1 M i.e. 6.941 mg/ml) at different temperatures: (a) Sapwood (Sw) and (b) Heartwood (Hw) (as indicated).

38

Table 5.3

The values of qwood, e and k measured for samples of Norway spruce Sw and Hw in experiments using 1 M LiCl aqueous solution at various temperatures.

Temperature Sw Hw qwood, e (mg/g) k (ml/g) qwood, e (mg/g) k (ml/g)

R T 5.9±0.2 0.8±0.0 5.7±0.4 0.8±0.0 40 oC 5.6±0.5 0.8±0.1 5.6±0.0 0.8±0.0 60 oC 5.3±0.7 0.7±0.1 6.5±1.2 0.9±0.2

EffectEffectEffectEffectssss of temperature and wood structure on equilibrium sorption isothermsof temperature and wood structure on equilibrium sorption isothermsof temperature and wood structure on equilibrium sorption isothermsof temperature and wood structure on equilibrium sorption isotherms

In Figure 5.10, qwood, e is plotted against the concentration of the Li+ ions in the bulk liquid (Ce) at R T and 60 oC for samples of both Sw and Hw flour. Only a weak dependence on temperature can be observed for both samples. Approximately the same quantities were sorbed for bulk LiCl solutions of concentrations up to 1 M; above 1 M, these quantities increased slightly with temperature, which may be due to changes in the properties of the wood (e.g. a large surface becomes accessible) that, in turn, may result in slight changes in the interactions with Li+ ions. Furthermore, for concentrations above 1 M, the sorbed quantities increased more for Hw than for Sw. These differences in the sorbed quantities of Sw and Hw may be attributed to differences in the structural and chemical compositions of Sw and Hw: the extractive content in Hw is, for example, higher than in Sw, and has a broader variety of functional groups available for Li+ ion interactions (Siau, 1984). However, the differences are small: in most cases they are within, or almost within, the experimental errors estimated.

Figure 5.10 Effect of wood type (Sw and Hw) and temperature on the equilibrium sorption isotherms of Li+ ions onto Norway spruce wood flour at different temperatures. (a) Sw and (b) Hw.

Study of sStudy of sStudy of sStudy of sorption isotherms orption isotherms orption isotherms orption isotherms

An initial investigation was made into the possibility of fitting the Langmuir sorption isotherm model to the individual sorption isotherms described above (Paper III). It was, however, found that there was a very strong correlation between the two parameters (qwood,max and b) fitted and no stable results were obtained. Furthermore, as discussed above, there were only small differences between the four different sorption isotherms (Sw at R T, Sw at 60 oC, Hw at R T

39

and Hw at 60 oC). It was decided therefore that all the data from the 4 sorption isotherms should be used to estimate the parameters in both the Langmuir and Freundlich models: both of these models fitted the experimental data reasonably good (R2 > 0.9). The parameters estimated are thus valid for both Sw and Hw for the temperature range between R T and 60 oC, and for the concentrations range of Li+ ions between 0 and 6 M.

Langmuir isothermLangmuir isothermLangmuir isothermLangmuir isotherm

The qwood,max and b values for Li+ ions on Norway spruce wood were found to be approx. 132 mg/g and 0.05 l/mol, respectively. The estimated qwood,max value (approx. 132 mg/g or 19 mmol/g) is much higher than the concentration of the carboxylic acid groups (COOH) (0.086±0.007 mmol/g) (Werkelin et al., 2010) and the theoretical OH groups (12.7 mmol/g) (see Appendix A2) for the estimation procedure) in wood samples of native Norway spruce. This result therefore suggests that, apart from the interactions of carboxyl, hydroxyl and other groups (e.g. C-H), as shown in the FTIR study and in XPS study (visible as Li-O interactions), a non-site-specific retention of ions might also have occurred during the sorption of Li+ ions onto Norway spruce wood flour under the experimental conditions studied.

Freundlich isothermFreundlich isothermFreundlich isothermFreundlich isotherm

The Freundlich isotherm constants kF and 1/n were found to be approx. 6.5 (mg/g) (l/mol)1/n

and 0.87 (dimensionless), respectively. A value for 1/n below one implies chemisorption processes, whereas above one indicates co-operative sorption (Foo and Hameed, 2010). In this study, the value of 1/n was below one (1/n =0.87), so it is most likely that the sorption process for the experimental conditions studied includes chemisorption.

5.5. Concentration profiles of ions in 5.5. Concentration profiles of ions in 5.5. Concentration profiles of ions in 5.5. Concentration profiles of ions in the the the the porous structure of porous structure of porous structure of porous structure of piecepiecepiecepiecessss of of of of

wood wood wood wood Native Norway spruce wood contains small amounts of metal ions (typically 1-5 g/kg wood). These may participate in the sorption mechanism induced by the exchange of ions and thus may aggravate the measurement of concentration profiles. So, prior to the LiCl impregnation experiments being carried out, the samples of wood pieces were treated in a dilute sulphuric acid solution (pH=0.5/0.31 M H2SO4) until no floating wood could be observed, in order to remove native metal ions and the air present in the wood material (Paper V). The concentration of Li+ ions measured experimentally was the total concentration of Li+ ions (qtotal), i.e. the sum of the Li+ ions sorbed onto the surfaces of the cell walls of the wood and those present in the solution located in the pores. The concentration of Li+ ions in the solution in the pores (CLi

+) was calculated using Equation (5.4), which is based on mass balance.

For each slice (i.e. thickness level), the qtotal is defined as:

��6�;9 = )UV5T ∗ �/ + EUV5T ∗ X>66?75 � H (5.4)

where, CLi+ is the concentration of Li+ ions available in the solution located in the porous

structure of the wood matrix (g/l) and k is the wood-Li+ ion equilibrium partition coefficient (l/kg). Vwood piece is the effective pore volume of the transport-through-intrusion (diffusion-accessible) of the piece of wood (0.99±0.27 l/kg). From this section onwards, all the results reported were based on the concentration of Li+ ions in the solution located in the porous structure of the wood matrix, which was calculated by using Equation (5.4).

40

Effect of impregnation Effect of impregnation Effect of impregnation Effect of impregnation (residence) time(residence) time(residence) time(residence) time

An increase in the concentration of Li+ ions was observed at the corresponding positions in the pieces of wood as the impregnation time increased. In Figure 5.11, which shows an example at 40oC, it can be observed that a shift occurred in the concentration profiles of the Li+ ions towards the centre of the piece of wood as the duration of the treatment increased from 36 to 168 h. In the eluate of slices taken from the outermost surface layer, a concentration of Li+ ions greater than in the bulk solution (6.941 g/l) was observed at extended residence times (i.e. 168 h). It appears that the micro-cracks continue to affect the concentration profiles to some extent, and thereby contribute to the different shapes of the concentration profiles at extended residence times.

Figure 5.11 Concentration profiles measured for Li+ ions (present in the wood pore solution) in SsWP Norway spruce Sw after residence times of 36, 72 and 168 h, at 40 oC for Cube a.

5.65.65.65.6. . . . EEEEffective diffusiffective diffusiffective diffusiffective diffusion on on on coefficientscoefficientscoefficientscoefficients and tortuosity factors of ions in and tortuosity factors of ions in and tortuosity factors of ions in and tortuosity factors of ions in

woodwoodwoodwood The symbols and lines in Figure 5.12 represent the experimental concentration profiles and the best fit of the model, respectively. Although the model fits these profiles reasonably well for Cube a (Figure 5.12 (A)), as can be seen, it was not satisfactory for Cube c, especially at residence times of 36 and 72 h (Figure 5.12 (B)-(a) and (b)). This behaviour is most likely due to micro-cracks caused by the mechanical preparation of the pieces of wood (e.g. during sawing), as mentioned previously. Regardless of the fact that it was not possible to verify the concentration profiles at shorter times (36 and 72 h) for the edge portions (Cube c) of the piece of wood in this work, rough estimations of the longitudinal diffusion coefficients and tortuosity factors have been made. Similar results were also observed for the samples of Hw (Figure

5.13). The estimated De,T(=R) and De,L are listed in Table 5.4, where it can be seen that, for samples of both Sw and Hw and for the experimental conditions studied, the predicted De,T(=R) is about two orders of magnitude less than the De,L. This result is reasonable since the diffusion of chemicals/ions in the transversal direction takes place mainly through the cell walls and pits whereas in the longitudinal direction, it is mainly through the liquid in the fibre lumen. Tortuosity factors for both samples (Sw and Hw) were calculated using Equation (4.4) in two diffusional directions, i.e. the longitudinal and transversal, and are given in Table 5.4. As expected, the tortuousness is smaller in the longitudinal direction than in the transversal

41

direction for samples of both Sw and Hw. This result is understandable because the transport of ions in the longitudinal direction occurs mainly through the lumen openings in the tracheids and the distances through the pit pores are short. Sawing may also have caused some fibre separation in longitudinal direction at the edge part of the piece of wood, causing the formation of some micro-cracks between the fibres which may, in turn, enhance longitudinal mass transport. In the transverse direction, however, the ions are transported mainly through pit pores, which have rather small areas and their positions along the fibre wall cause the tortuosity.

Figure 5.12 Measured and calculated concentration profiles of Li+ ions in Norway spruce Sw at 40 oC for (A) Cube a (middle portion) and (B) Cube c (edge portion) of the impregnated piece of wood in the experiments with 1 M bulk LiCl solution after the residence times (a) 36, (b) 72 and (c) 168 h.

42

Figure 5.13 Measured and calculated concentration profiles of Li+ ions in Norway spruce Hw at 40 oC for (A) Cube a (middle portion) and (B) Cube c (edge portion) of the impregnated piece of wood in the experiments with 0.5 M bulk LiCl solution after the residence times (a) 36, (b) 72 and (c) 168 h.

Table 5.4

Diffusive resistance and tortuosity factors of Li+ ions in samples of Norway spruce Sw in the longitudinal and transversal (Radial & Tangential) directions at 40 oC. Sample Concentration of

bulk LiCl solution (M) Direction of diffusion

Effective diffusion coefficient (m2/s)

Tortuosity factor

Sw 1 Longitudinal Approx.1x10-9 1.2 Transversal 1.7 ± 0.7x10-11 76 ± 35

Hw 0.5 Longitudinal Approx. 6.5x10-10 1.9

Transversal 4.7 ± 2x10-12 305 ± 159

43

6. 6. 6. 6. CCCContributions ontributions ontributions ontributions made to thismade to thismade to thismade to this field of researchfield of researchfield of researchfield of research

• The methods employed to prepare the samples used in this study are suitable for elucidating reasonably good measurements of the concentration profiles of ions in the porous structure of wood in the transversal direction. The diffusion of ions in the central part of a piece of wood is affected mainly by mass transport in the transversal direction, which has much shorter diffusional distances than the longitudinal direction. However, the method proposed is unable to explain the concentration profiles at the edge part of the wood piece, where the diffusion of ions is affected by mass transport in both longitudinal and transversal directions. This is probably due to the structural damage that might have occurred during the mechanical preparation of the wood pieces at the edge part of the wood piece; sawing, for example, may also cause some separation of fibres in the longitudinal direction, thus allowing micro-cracks between the fibres may to be formed.

• The chemical groups involved in the interactions between the Li+/Cl- ions and wood are partly identified. A decrease in the carbonyl content of the wood samples upon LiCl treatment indicates the deacetylation of hemicelluloses.

• The method employed to measure the sorption equilibrium is suitable for differentiating the amount of ions that is sorbed onto the surfaces of the cell walls of the wood from that dissolved in the solution within its pores. The sorption equilibrium of Norway spruce wood-Li+ ions from LiCl solution is partially understood. At lower concentrations of LiCl solution, the major part of the sorption of Li+ ions onto the wood can be explained by the ion exchange that occurs between the carboxylic groups in the wood’s components and the Li+/Cl-. However, at higher concentration of LiCl, other interactions/bondings might be involved in the sorption process.

• Considering both the diffusive mass transport of ions through liquid-saturated wood pores and the sorption of ions onto the matrix of the solid wood, the transport model adopted provided reasonably good results for the transport properties (i.e. effective diffusion coefficients and tortuosity factors) in the transversal direction.

45

7. 7. 7. 7. ConclusionsConclusionsConclusionsConclusions 7.1. Sorption equilibrium experiments: development of the method

• The experimental methodology proposed was found to be suitable for determining the concentration of the ions sorbed onto the matrix of the solid wood.

• The results show that Norway spruce wood has a considerable affinity towards Li+ ions. 7.2. Sorption isotherm models

• Both the Langmuir and Freundlich sorption isotherm models described the sorption of Li+ ions onto Norway spruce wood flour reasonably good for the conditions investigated.

7.4. Changes in surface chemistry during sorption

• The introduction of Li+ and Cl- ions from an aqueous LiCl solution into the matrix of Norway spruce Sw is associated with the disruption of H-bonding and the accompanying enhanced mobility, which were demonstrated by an enhanced adaptation of the surface of the fibrillar structure upon drying and an enlarged surface area.

• Spectroscopic analysis of the wood material before and after LiCl treatment indicated clearly that the elements Li and Cl participated in interactions with the constituents of wood: the hydroxyl, carboxylic and carbonyl groups in the wood involved in the interactions and the amounts of Li+ ions sorbed onto carbohydrate-rich fractions were higher than for the lignin fractions.

7.5. Measurement of the concentration profiles: development of the method

• An experimental methodology for determining concentration profiles of chemicals in porous structure has been adapted to wood.

• The method gave reasonable results, although defects (i.e. micro-cracks) in the pieces of wood remained and were detectable.

7.6. Transport properties of ions in wood

• The effective diffusion coefficient of Li+ ions in samples of Norway spruce Sw and Hw is about two orders of magnitude less in the transversal direction than in the longitudinal direction for the conditions used in this study.

• The tortuousness is longer in the transversal direction than in the longitudinal direction for both the Sw and Hw: the tortuousness was calculated as being 63±29 and 160±83 times longer in the transversal direction for Sw and Hw, respectively, for the conditions used in this study.

47

8888. Acknowledgements. Acknowledgements. Acknowledgements. Acknowledgements I would like to thank:

My examiner and main supervisor Professor Hans Theliander for giving me this opportunity to be a Ph.D. student in his research group. I am grateful for his contributions of time, tremendous support, ideas and valuable guidance which were vital for the outcome of the research work presented in this thesis. Dr. Merima Hasani, my advisor and co-author, for excellent collaboration. I am grateful for her support, commitment and guidance. The Chalmers Energy Initiative (CEI) for financial support. Kurt Löfgren and Torbjörn Jönsson for their help with carpentry: preparing samples of wood pieces from logs and then small cubes from the pieces. Ms. Anne Wendel for her help with XPS measurements, and for answering many of my immature questions and providing input to help me understand the results obtained. Mr. Tommy Friberg for his skillful help with smoothing the wood pieces. Ms. Joanna Wojtasz for her help with analysing Klason lignin and carbohydrates and Mr.

Jonas Wetterling for his help with BET surface area and Density (Pycnometer) measurements. Professor Per Lincoln, for interesting discussions. Ms. Eva Kristenson, Ms. Malin Larsson and Ms. Carina Pettersson for their assistance with administrative tasks. Linguistic advisors: Ms. Deborah Fronko, for linguistic review of Paper II and Licentiate thesis, and Ms. Maureen Sondell for a much valued linguistic review of the three manuscripts (Papers III, IV and V) and the Ph.D. thesis. Dr. Seshendra Karamchedu for listening to, and providing support for, many of my thoughts concerning my personal as well as my professional life over the course of these years. My former and present colleagues and friends at the Division of Forest Products and Chemical Engineering and Chemical Environmental Science for a fascinating and enjoyable working environment with people from different cultures from all over the world. All of my friends from India who reside in Gothenburg region, some of whom also work at Chalmers University, for sharing many enjoyable weekend breaks during my time here. All of my teachers, past and present, for their guidance, support and encouragement. I would especially like to thank Dr. P. Dinesh Sankar Reddy, Prof. Pallab Ghosh and Dr. Tamal

Banerjee for their valuable suggestions and support. I also thank a special teachers from my school, Mr. B. Rajasekhar Rahul and his wife N. K. Saraswathi, for it was them who also encouraged me to continue on to higher studies. This teacher’s family is special to me, since I married to their daughter Sumahitha Rahul. Their support has been essential to me during these years in taking care of my wife and son (Rituparan), who reside in India. My dear family, for all their constant support. Many thanks for understanding, and being with me through some tough times, and especially to my parents: mother (Kolavali Subba

Lakshmi) and father (Kolavali Aswartha Narayana), who understand the importance of education and worked hard to provide the many things I needed to reach this level in my life. My mother never went to school at all and my father had only basic education, having had only 4 years of schooling during childhood. I am truly greatful for all their great support and encouragement, and for allowing me to realize my own potential.

49

9. References9. References9. References9. References Akhtaruzzaman, A.F.M. and Virkola, N-E. (1979). Influence of chip dimensions in kraft pulping. Paperi ja Puu 9:578–580. Alam, M.N. and Van De Ven, T.G. (2014). An aqueous process for the separation of cellulose from biomass. Journal of Science & Technology for Forest Products and Processes 4(3):22-26. Behr, E.A., Briggs, D.R. and Kaufert, F.H. (1953). Diffusion of dissolved materials through wood. J. Phys. Chem.57:476-480. Bengtsson, G. and Simonson, R. (1984). Chemi-mechanical pulping of birch wood chip. Paperi ja Puu 64(3): 187-193. Bengtsson, G., Simonson, R., Heitner, C., Beatson, R. and Ferguson, C. (1988). Chemi-mechanical pulping of birch wood chips. Part 2. Studies on impregnation of wood blocks using scanning electron microscopy and energy dispersive X-ray analysis. Nordic Pulp & Paper Research Journal 3:132-137. Brändström, J. (2001). Micro and ultrastructural aspects of Norway spruce tracheids: A review. IAWA journal 22(4): 333-353. Burr, H.K. and Stamm, A.J. (1947). Diffusion in wood. J. Phys. Chem. A. 51:240-261. Cady, L.C. and Williams, J.W. (1935). Molecular diffusion into wood. The Journal of Physical Chemistry A 39(1): 87-102. Chastain, J. and King, R. C. (Eds.). (1992). Handbook of X-ray photoelectron spectroscopy: a reference book of standard spectra for identification and interpretation of XPS data (p. 261). Eden Prairie, MN: Perkin-Elmer. Cooper, P.A. (1998). Diffusion of copper in wood cell walls following vacuum treatment. Wood Fibre Sci. 30(4): 382-395. Colom, X., Carrillo, F., Nogués, F. and Garriga, P. (2003). Structural analysis of photodegraded wood by means of FTIR spectroscopy. Polymer Degradation and Stability 80(3):543-549. Christensen, G.N. (1951a). Diffusion in wood. II. The temperature coefficient of diffusion through wood. Austral. J. Appl. Sci. 2(4): 430-439. Christensen, G.N. (1951b). Diffusion in wood. III. Ion selection and its effect on the diffusion of electrolytes. Austral. J. Appl. Sci. 2(4): 440-453. Christensen, G.N. and Williams, E.J. (1951). Diffusion in wood. I. A quantitative theory of diffusion in porous media and its application to wood. Austral. J. Appl. Sci. 2(4): 411-429. Dorris, G.M. and Gray, D.G. (1978). The surface analysis of paper and wood fibres by ESCA (Electron Spectroscopy for Chemical Analysis). I Application to cellulose and lignin. Cellulose Chem. Technol. 12:9-23. Engelund, E.T. (2011). Wood-water interactions: Linking molecular level mechanisms with macroscopic performance. Ph.D. thesis, Technical University of Denmark, Copenhagen, Denmark. Fengel, D. and Wegener, G. (1984). Wood-Chemistry, Ultrastructure, Reactions. Walter de Gruyter, New York. Fernando, S., Adhikari, S., Chandrapal, C. and Murali, N. (2006). Biorefineries: current status, challenges and future direction. Energy Fuels 20: 1727-1737. Foo, K.Y. and Hameed, B.H. (2010). Insights into the modeling of adsorption isotherm systems, Chemical Engineering Journal 156:2-10. Fukuyama, M. and Urakami, H. (1980). Diffusion of nonelectrolytes through wood saturated with water I. Measurements of the diffusion rates by highly sensitive differential refractometer. Mokuzai Gakkaishi 26:587-594.

50

Fukuyama, M. and Urakami, H. (1982). Diffusion of nonelectrolytes through wood saturated with water II. Diffusion rates of the monohydric alcohols. Mokuzai Gakkaishi 28:17-24. Fukuyama, M. and Urakami, H. (1986). Diffusion of nonelectrolytes through wood saturated with water III. Diffusion rates of polyethylene glycols. Mokuzai Gakkaishi 32:147-154. Gamstedt, E. K., Bader, T. K. and de Borst, K. (2013). Mixed numerical experimental methods in wood micromechanics. Wood Science and Technology 47(1):183-202. Gindl, W., Dessipri, E. and Wimmer, R. (2002). Using UV-microscopy to study diffusion of melamine-urea-formaldehyde resin in cell walls of Spruce wood. Holzforschung 56:103-107. Gindl, W., Zargar-yaghubi, F. and Wimmer, R. (2003). Impregnation of softwood cell walls with melamine-formaldehyde resin. Bioresource Technology 87: 325-330. Gustafson, R. (1988). The role of diffusion during initial delignification of alkaline pulping. Tappi Journal 71(4): 145. Harilal, S.S., Allain, J.P., Hassanein, A., Hendricks, M R. and Nieto-Perez, M. (2009). Reactivity of a lithium-exposed graphite surface. Applied Surface Science 255:8539-8543. Henrikson, G., Brännvall, E. and Lennholm, H. (2008). Chapter 2 The Trees, Wood Chemistry, The Ljungberg Textbook, Gothenburg, Forest products and Chemical Engineering, Chalmers University of Technology. Copyright: Fibre and Polymer Technology, KTH Sweden. Hubbe, M. A., Hasan, S. H. and Ducoste, J. J. (2011). Cellulosic substrates for removal of pollutants from aqueous systems: A review. 1. Metals. BioResources 6(2):2161-2287. Inari, G.N., Petrissans, M., Lambert, J., Ehrhardt, J.J. and Gérardin, P. (2006). XPS characterization of wood chemical composition after heat‐treatment. Surface and Interface Analysis 38(10):1336-1342. Jacobson, A.J. and Banerjee, S. (2006). Diffusion of tritiated water into water-saturated wood particles. Holzforschung 60: 59-63. Jedvert, K., Saltberg, A., Lindström, M. E. and Theliander, H. (2012). Mild steam explosion and chemical pre-treatment of Norway spruce. Bioresources 7(2): 2051-2074. Johansson, L.S., Campbell, J.M., Hännien, T., Ganne-Chedeville, C., Vuorinen, T., Hughes, M. and Laine, J. (2012). XPS and the medium-dependent surface adaptation of cellulose in wood. Surface and Interface Analysis 44:899-903. Jääskeläinen, A-S., Tapanila, T. and Poppius-Levlin, K. (2000). Carbohydrate reactions in peroxyacetic acid bleaching. Journal of Wood Chemistry and Technology 20(1):43-59. Kazi, K. M. F. (1996) Impregnation: A key step of biomass conversion processes. Ph.D. thesis, University of Sherbrooke, Quebec, Canada. Kazi, K.M.F. and Chornet, E. A. (1998). Diffusion Model for the Chemical Impregnation of Hardwoods and its Significance for Rapid Steam Treatments. Paperi ja Puu-Paper and Timber 80(1):41. Kumar, R., Hu, F., Hubbell, C. A., Ragauskas, A. J. and Wyman, C. E. (2013). Comparison of laboratory delignification methods, their selectivity, and impacts on physiochemical characteristics of cellulosic biomass. Bioresource Technology 130: 372-381. Lin, S. Y. and Dence, C. W., (1992). Methods in lignin chemistry. Springer Verlag, Berlin, Germany.

51

Liu, S., Amidon, T.E., Francis, R, C., Ramarao, B.V., Lai, Y-Z. and Scott, G, M. (2006). From forest biomass to chemicals and energy – biorefinery initiative in New York State. Feature Commentary, Industrial Biotechnology 2(2):113-120. Määttänen, M. and Tikka, P. (2012). Determination of phenomena involved in impregnation of softwood chips. Part 2: alkali uptake, alkali consumption and impregnation yield. Nordic Pulp & Paper Research Journal 27(3):559-567. Meijer, M., Zwan, R.P. and Militz, H. (1996) Unsteady-state diffusion of methanol in Douglas-fir heartwood at high temperatures. Holzforschung 50:135-143. Miller, R. B. (1999). Structure of wood. Wood handbook: wood as an engineering material. Madison, WI: USDA Forest Service, Forest Products Laboratory, General technical report FPL; GTR-113: Pages 2.1-2.4. Nada, A.M.A., Abd El-Mongy, S. and Abd El-Sayed, E.S. (2009). Effect of different treatments on cellulose toward carboxylation and its application for metal ion absorption. Bioresources, 4(1), 80-93. Narayanamurti, D. and Kumar, V.B. (1953). Diffusion of organic molecules through wood. Journal of Polymer Science 10(6):515-524. Narayanamurti, D. and Ratra, R.S. (1951). Diffusion of ions through some Indian timbers. Proceedings of the Indian Academy of Sciences, Section A 33(6): 349-359. Nzokou, P. and Pascal Kamdem, D. (2005). X‐ray photoelectron spectroscopy study of red oak (Quercus rubra), black cherry (Prunus serotina) and red pine (Pinus resinosa) extracted wood surfaces. Surface and Interface Analysis 37(8):689-694. Pandey, K. K., (1999). A study of chemical structure of soft and hardwood and wood polymers by FTIR spectroscopy. Journal of Applied Polymer Science 71(12):1969-1975. Palme, A., Idström, A., Nordstierna, L. and Brelid, H. (2014). Chemical and ultrastructural changes in cotton cellulose induced by laundering and textile use. Cellulose 21(6):4681–4691. Pettersen, R. C. (1984). The chemical composition of wood. In: Rowell, Roger M., ed. The chemistry of solid wood. Advances in Chemistry Series 207. Washington, DC: American Chemical Society; Chapter 2. Plötze, M. and Niemz, P. (2011). Porosity and Pore Size Distribution of Different Wood Types as Determined by Mercury Intrusion Porosimetry. Eur. J. Wood Prod 69:649. Ra, J. B., Barnes, H.M. and Conners, T.E. (2001). Determination of boron diffusion coefficients in wood. Wood Fibre Sci. 33:90-103. Ra, J. B., Barnes, H. M. and Conners, T. E. (2002). Predicting boron diffusion in wood from surface sorption. Forest Products Journal 52(10):67. Robertsen, L. (1993). Diffusion in wood. Ph.D. thesis. Department of Pulping Technology, Faculty of Chemical Engineering, Åbo Academy, Finland. Rosu, D., Teaca, C. A., Bodirlau, R. and Rosu, L., (2010). FTIR and color change of the modified wood as a result of artificial light irradiation. Journal of Photochemistry and Photobiology B: Biology. 99(3):144-149. Rowell, R.M. (ed.) (2005). Handbook of wood chemistry and wood composites: Boca Raton, Fla: Taylor & Francis: CRC Press, Corp. Salin, J.G. (2008). Almost all wooden pieces have a damaged surface layer: impact on some properties and quality. Proceedings Conference COST E53, Delft, The Netherlands, 135-143. Saltberg, A. (2009) Inorganic ions in wood chips - Leaching prior to Kraft pulping and the effect of calcium on kraft delignification. Ph.D. thesis, Chalmers University of Technology, Gothenburg, Sweden.

52

Sandberg, K. and Sterley, M., (2009). Separating Norway Spruce Heartwood and Sapwood in Dried Conditions with Near-Infrared Spectroscopy and Multivariate Data Analysis. European Journal of Forestry Research, 128:475-481. Saripalli, K. P., Serne, R. J., Meyer, P. D. and McGrail, B. P. (2002). Prediction of Diffusion Coefficients in Porous Media Using Tortuosity Factors Based on Interfacial Areas. Ground Water. 40(4):346. Schwanninger, M., Rodrigues, J. C., Pereira, H. and Hinterstoisser, B. (2004). Effects of short-time vibratory ball milling on the shape of FT-IR spectra of wood and cellulose. Vibrational Spectroscopy. 36(1):23-40. Sernek, M. (2002). Comparative analysis of inactivated wood surfaces (Ph.D. thesis, Virginia Polytechnic Institute and State University, USA). Sharareh, S., Tessier, P. and Lee, C-L. (1996). Penetration of sodium sulphite into Black Spruce and Aspen wood chips using SEM/EDXA. J. Pulp and Paper Science 22(3):J71-J77. Shukla, A., Zhang, Y. H., Dubey, P., Margrave, J. L. and Shukla, S. S. (2002). The role of sawdust in the removal of unwanted materials from water. J. Hazardous Materials 95(1):137-152. Siau, J.F. (1984). Transport processes in wood, Springer Verlag, Berlin, Germany. Sinn, G., Reiterer, A. and Stanzl-Tschegg, S. E. (2001). Surface analysis of different wood species using X-ray photoelectron spectroscopy (XPS). J. Materials Science 36(19):4673-4680. Sjöström, E. (1993) Wood Chemistry-Fundamentals and Applications. Academic Press, San Diego, USA. Skaar, C. (1972). Water in wood. Syracuse University Press, Syracuse, New York. Skaar, J. and Siau, J.F. (1981). Thermal diffusion of bound water in wood. Wood Science Technology 15:105-112. Stamm, A.J. (1946). Passage of liquids, vapors and dissolved materials through softwoods. Tech. Bull. 929: 1-79. U.S. Dept. Agr. Washington, DC. Stamm, A.J. (1967). Movement of fluids in wood: Part II. Diffusion. Wood Science and Technology 1:205-230. Stone, J. E. (1957). The effective capillary cross-sectional area of wood as a function of pH. Tappi Journal 40(7): 539-541. Stone, J.E. and Förderreuther, C. (1956). Studies of penetration and diffusion into wood. Tappi Journal 39(10): 679-683. Stone, J.E. and Green, H.V. (1959). Penetration and diffusion into hardwoods. Tappi 42(8): 700-709. Tan, I.A.W. and Hameed, B.H. (2007). Equilibrium and kinetic studies on basic dye adsorption by oil palm fibre-activated carbon, Chemical Engineering Journal 127: 111-119. Theander, O. and Westerlund, E. A. (1986). Studies on dietary fibre. 3. Improved procedures for analysis of dietary fibre. Journal of Agricultural and Food Chemistry 34(2):330-336. Thygesen, L. G., Engelund, E.T. and Hoffmeyer, P. (2010). Water Sorption in Wood and Modified Wood at High Values of Relative Humidity. Part I: Results for Untreated, Acetylated, and Furfurylated Norway spruce. Holzforschung 64: 315. Törnqvist, M., Hurme, T. and Rosenholm, J.B. (2001a). Drift speed: a way of measuring diffusion and tortuosity of porous materials. Colloids and surfaces A: Physicochem. Eng. Aspects 180:23-31. Törnqvist, M., Hurme, T. and Rosenholm, J.B. (2001b). The concentration dependence of the diffusion coefficient in Pine, Birch, and Spruce. Paperi ja Puu 83(3):204.

53

Tsuchikawa, S. and Siesler, H.W. (2003). Near-infrared spectroscopic monitoring of the diffusion process of deuterium-labeled molecules in wood. Applied Spectroscopy 57:675-681. Vinden, P. (1984). The effect of raw material variables on preservative treatment of wood by diffusion processes. Journal of the Institute of Wood Science 10(1):31-41. Wallström, L. and Lindberg, K.A.H. (2000). The diffusion, size and location of added silver grains in the cell walls of Swedish pine, Pinus sylvestris. Wood Sci.Tech 34:403-415. Werkelin, J., Skrifvars, B. J., Zevenhoven, M., Holmbom, B. and Hupa, M. (2010). Chemical forms of ash-forming elements in woody biomass fuels. Fuel 89(2):481-493. Wiedenhoeft, A. (2010). Structure and function of wood, Wood handbook-Wood as an engineering material. General Technical Report FPL-GTR-190. Madison, WI: U.S. Dept. Agr, Forest Service, Forest Products Laboratory. Pages 3.1-3.17. Yantasee, W. (2001). Kinetic and equilibrium analysis of metal ion adsorption onto bleached and unbleached kraft pulps. Ph.D. thesis, Chemical Engineering, Oregon State University, USA. Zauer, M., Hempel, S., Pfriem, A., Mechtcherine, V. and Wagenfuhr, A. (2014). Investigations of the Pore-Size Distribution of Wood in the Dry and Wet State by Means of Mercury Intrusion Porosimetry. Wood Sci. Technol 48: 1229. Zelinka, S. L., Gleber, S. C., Vogt, S., Rodríguez López, G. M. and Jakes, J. E. (2015). Threshold for ion movements in wood cell walls below fibre saturation observed by X-ray fluorescence microscopy (XFM). Holzforschung, 69(4):441-448.

55

AppendicesAppendicesAppendicesAppendices Appendix A1

Estimation of the diffusion coefficient of Li+ ion in an aqueous medium (free solution)

The diffusion coefficient of Li+ ion in an aqueous medium (free water) (Df) with a concentration of Caq (M) and at temperature T (oC) can be estimated from the following correlation, using viscosity data of the aqueous medium at a corresponding temperature and concentration (Poling et al., 2001), Equation (A1):

��,V5TZ℃ = ��,V5T\]℃ ∗ ^ Z℃\]℃_ ∗ `

abcd�e℃

abcdf℃ g

(A1)

where ��,V5TZ℃ and ��,V5T\]℃ are the diffusion coefficients of Li+ ions in an aqueous medium at temperature

T (oC) and 25 oC, respectively, and hRcdZOb and hRcd\]Ob are the viscosities of an aqueous solution at

concentration C, temperature T (oC) and 25 oC, respectively. Df,Li+ in an aqueous solution of

concentration C at 40 oC has been estimated using Equation A1 and the literature data presented in Table 2. Df,Li

+ at 40 oC in 1 M and 0.5 M aqueous solutions was found to be 2.0x10-9 m2/s and 2.1x10-9 m2/s, respectively.

Table A1

Viscosity of a 1 M LiCl solution and the self-diffusivity of Li+ ions in 1 M LiCl solution at

different temperatures.

Bulk LiCl solution concentration (M)

Temperature (oC)

Viscosity (Melinder 2007) (µ) (mPa.s)

Self-diffusivity (Braun and Weingärtner 1988) (Df,Li

+) (m2/s)

1 25 1.023 0.924x10-9 40 0.753 -

0.5 25 0.965 0.95x10-9 40 0.69 -

References

Braun, B. M. and Weingärtner, H. Accurate Self-Diffusion Coefficients of Li+, Na+ and Cs+ Ions in Aqueous Alkali Metal Halide Solutions from NMR Spin-Echo Experiments. J. Phys. Chem. 1988, 92, 1342.

Melinder, Å. Thermophysical Properties of Aqueous Solutions Used as Secondary Working Fluids. Ph.D. thesis. Dept. of Energy Technology. Royal Institute of Technology (KTH), Stockholm, Sweden. 2007.

Poling, B. E., Prausnitz, J. M. and O’Connell, J. P. The Properties of Gases and Liquids. Fifth Edition, McGraw-Hill. 2001.

Welty, J. R., Wicks, C. E., Wilson, R. E. and Rorrer, G. 24.2. The Diffusion Coefficient. In: Fundamentals of Momentum, Heat and Mass Transfer. Third Edition. John Wiley & Sons, New York, USA. 1984.

56

Appendix A2

Estimation of the number of theoretical OH groups in the sample of Norway spruce Sw

wood flour

The number of theoretical OH groups in moles per gram (mol/g) can be estimated using Equation A2, as proposed by Rowell [1] and Hill [2]:

ijkl\+ im

kl\+ \Rki\+ �

kno (A2)

where A is cellulose, B the hemicellulose hexosan, C the hemicellulose pentosan and D lignin. Analyses of the composition of Norway spruce Sw flour were performed using the procedure described by Jedvert et al., 2012 [3] and found to be: 36.9% cellulose (A), 17.2% the hemicellulose hexosan (B) and 5.8% pentosan (C), and 33% lignin (both soluble and insoluble in acid). Based on this composition, the number of theoretical OH groups that were estimated was found to be 12.72 mmol/g. Since hydroxyl groups made inaccessible through binding between wood polymers or steric hindrance are not considered, this estimate expresses a theoretical upper limit. Also, two-thirds of the hydroxyl groups of cellulose are known to be inaccessible to water due to the crystalline part of cellulose. Moreover, the accessibility of the hydroxyl groups in a sample (e.g. drying) is also influenced by its history.

[1] Rowell, R. M. Wood Sci. (1980) 13:102.

[2] Hill, C. A. S. Wood modification- chemical, thermal and other processes. Wiley, Chichester, UK. (2006).

[3] Jedvert, K., Saltberg, A., Lindström, M.E. and Theliander, H. Mild steam explosion and chemical pre-treatment of Norway spruce. Bioresources. 7(2) (2012) 2051-2074.

Paper I

Determination of the diffusion of monovalent cations into wood under isothermal

conditions based on LiCl impregnation of Norway spruce


Holzforschung 2013; 67(5):559-565

Copyright © 2013 Walter de Gruyter GmbH

DOI 10.1515/hf-2012-0182 Holzforschung 2013; 67(5): 559–565

Reddysuresh Kolavali and Hans Theliander *

Determination of the diffusion of monovalent cations into wood under isothermal conditions based on LiCl impregnation of Norway spruce Abstract: The impregnation of wood pieces in the course

of pulping can be divided into primary (advective mass

transport) and secondary (diffusive mass transport) pen-

etration. Little is known about the latter partly because

of the difficulties in the determination of the relevant dif-

fusivities for this system. In the present article, a precise

experimental methodology has been developed to meas-

ure the concentration profiles of cations as a function of

wood piece dimensions, impregnation time, temperature,

and wood structure. The cation concentration can be

measured at any position in the impregnated wood piece.

The impregnation of Norway spruce wood samples with

LiCl was investigated. The impregnated wood pieces were

cut mechanically into cubes, which were cut in slices by

means of a microtome, and the eluate of the 0.4-mm-thick

slices in HNO 3 was analyzed by flame atomic emission

spectroscopy. The method gave reasonable results, and

defects in the wood piece (microcracks) were detectable.

The preliminary results presented here have to be verified

with more replicates due to the heterogeneity of wood.

Keywords: diffusion, experimental methodology, impreg-

nation, lithium chloride, local concentration, Norway

spruce

*Corresponding author: Hans Theliander , Division of Forest

Products and Chemical Engineering, Department of Chemical

and Biological Engineering, Chalmers University of Technology,

SE-41296 Gothenburg, Sweden, Phone: + 46317722992,

Fax: + 46317722995 , e-mail: [email protected]

Reddysuresh Kolavali: Division of Forest Products and Chemical

Engineering , Department of Chemical and Biological Engineering,

Chalmers University of Technology, SE-41296 Gothenburg, Sweden

Introduction Lignocellulosic biomass is the most abundant renewable

material, accounting for 50 % of all the biomass in the

world, and its rational utilization helps preserve fossil

resources (Classen et al. 1999 ; Fernando et al. 2006 ). The

most popular keyword in this context is “ biorefinery ” . It

refers to petroleum refinery as a model for perfect utili-

zation, which produces various types of fuels and mate-

rials from petroleum. The pulping industry is the most

developed chemical technology of wood, and biorefinery

intends to establish further improvements in this area to

produce fuels, power, heat, and value-added chemicals

from biomass. It is believed that biorefinery will play a

significant role in forest clusters worldwide ( N ä yh ä and

Pesonen 2012 ). There are huge research efforts in the field

of biorefinery aiming at the production of value-added

products (Ragauskas et al. 2006 ; Dautzenberg et al. 2011 ;

G ü tsch and Sixta 2011 ; H ö rhammer et al. 2011 ; Kirsch et al.

2011 ; L ó pez et al. 2011; Martin -Sampedro et al. 2011 ; Sch ü tt

et al. 2011 ; Testova et al. 2011 ). For example, the extraction

of a portion of hemicelluloses from wood before pulping

– by acid hydrolysis, autohydrolysis, steam explosion, or

alkali extraction – seems to be promising in this context.

In any case, the impregnation of wood with chemicals will

be of great importance for the economically viable biore-

finery. The goal is a uniform distribution of moisture and

chemicals within the wood pieces at the very beginning

of the main reaction. This is the reason why the complex

chemical transport phenomena of reactants via lumina

and voids to the solid matrix must be understood better.

Most of the investigations are based on either average

flux measurement of the diffusing substance or electri-

cal conductivity measurement of the impregnated wood

samples. These methods have limitations and are not

suitable to determine the concentration profiles. The

most relevant investigations in this context have been

identified (Cady and Williams 1935 ; Stamm 1946 ; Burr

and Stamm 1947 ; Christensen 1951a,b ; Christensen and

Williams 1951 ; Narayanamurti and Ratra 1951 ; Behr et al.

1953 ; Narayanamurti and Kumar 1953 ; Stone and Green

1959 ; Fukuyama and Urakami 1980, 1982, 1986 ; Siau 1984 ;

Bengtsson and Simonson 1988 ; Robertsen 1993 ; Meijer

et al. 1996 ; Sharareh et al. 1996 ; Kazi et al. 1997 ; Cooper

1998 ; Ra et al. 2001 ; T ö rnqvist et al. 2001 ; Gindl et al. 2002 ;

Tsuchikawa and Siesler 2003 ; Jacobson et al. 2006 ). There

is no standard method for measuring diffusion into wood.

There are many aspects of diffusion of chemicals

into wood that have not yet been extensively studied. For

Brought to you by | Chalmers University of TechnologyAuthenticated

Download Date | 5/9/16 4:57 PM

560 R. Kolavali and H. Theliander: Diffusion of monovalent cations

example, most of the research on cation diffusion meas-

urements into wood was conducted with NaCl, KCl, and

NaOH. In the case of NaCl/KCl as diffusing substances, the

high natural contents of Na + /K + ions in wood aggravate the

measurements. In the case of NaOH, the reactions of OH -

ions with wood components are masking the pure diffu-

sion effects of Na + .

The present work is focusing on the diffusional mass

transfer of cations into wood by considering some of these

factors mentioned above. The intention is to determine

the concentration profiles of cations within the wood

pieces as a function of wood piece dimensions, impreg-

nation time, wood structure, and temperature. Lithium

chloride (LiCl) was chosen as agent and the Li + ion con-

centration was measured by flame atomic emission spec-

troscopy (FAES). The effects of genuine Li + and Cl - in the

cell wall are negligibly small for the impregnation results.

Su et al. (2012) demonstrated that Li + has a low affinity

toward wood components.

Materials and methods

Samples The stemwood disc with 23 cm thickness (without bark) of a 31 ± 1-year-

old Norway spruce ( Picea abies L.) was investigated. Both sapwood

(sW) and heartwood (hW) samples were carefully prepared using a

vertical band sawing machine (Mossner Rekord, August M ö ssner KG,

Mutlangen, Germany; metal cutting band saw: L.S. Starrett Co., Ltd.,

Jedburgh, UK, with 14 teeth per inch) and kept preliminarily in an

airtight polyethylene (PE) bag at 1 ° C. It was assumed that the stem-

wood disc contains 50 % sW and 30 % hW, and the remaining 20 % is

an intermediate between sW and hW (Sandberg and Sterley 2009 ).

Only sW and hW were investigated and not the intermediate mate-

rial. Only rot-free and other deformation-free samples were selected

and cut into a rectangular prototype pieces using the same vertical

band sawing machine mentioned above. Two diff erent dimensions

were prepared: 100 × 25 × 8 and 100 × 50 × 4 mm 3 (L × R × T). Then, the

material was stored in an airtight PE bag in a freezer at -18 ° C. Figure

1 illustrates the further procedure including the impregnation with

LiCl (impregnation times: 1, 4, and 12 h) and AES.

Water impregnation cycles The defrosting of the wood pieces took 24 h. Then, the pieces were

water impregnated in a vacuum-pressure cycle. This was performed

in a small polypropylene (PP) beaker placed in an autoclave fi lled

with deionized water (at ∼ 22 ° C; i.e., at room temperature). Then,

vacuum was applied for 30 min and the autoclave was pressurized

with N 2 at 0.5 MPa for 1 h. This procedure was repeated until no fl oat-

ing pieces were observed aft er the complete water impregnation (a

maximum of 3 – 5 cycles for sW and 5 – 10 cycles for hW were needed).

Norway spruce chips,sW and hW,

100 x 25 x 8 mm3

100 x 50 x 4 mm3

(-18°C)

Defrosting

Vacuum-pressureimpregnation with water

Cutting cubes withdimensions of

10 x 10 x 8 mm3

10 x 10 x 4 mm3

from different locations of the chips

Microtoming of the cubes

Slices ,0.4 mm thick

Oven drying, leaching with 2%

HNO3 for 24 h

Leachingliquor

Atomic emissionspectroscopy

Plots Li+ conc. vs. deepness in transverse

direction (mm)

,

Immersionin 1 M LiClsolution for 1, 4, and

12 h, frosen to -180°C,liophylization

Figure 1 Scheme of the experimental design of this study.

Liquor mixerwith 3 bladeimpeller stirrer

Sampleholderthreads

Woodpieces

1 M LiCIsolution

PPvessel

Figure 2 Experimental setup for impregnation with LiCl solution.

To eliminate the small bubbles left on the surfaces (possible barriers

against chemical diff usion), the pieces were kept in fi nal vacuum for

30 min before the subsequent experimental steps.

Impregnation with LiCl The pieces were dabbed with fi lter paper and immersed in a solution

of 1 m LiCl (Merck KGaA, Darmstadt, Germany) at a wood-to-liquor

ratio of 1:50 (Figure 2 ). The impregnation vessel was made of PP and

equipped with a liquor mixer (IKA, Staufen, Germany) fi tted with a

three-blade impeller (IKA, Staufen, Germany) to avoid the concen-

tration gradients in the impregnation vessel. Aft er 1, 4, and 12 h of



R. Kolavali and H. Theliander: Diffusion of monovalent cations 561

impregnation times, the pieces were removed and placed into liquid

N 2 (-180 ° C) to stop (minimize) the further migration of Li + . The fro-

zen pieces were lyophilized (instrument: Labconco, Kansas City, MO,

USA) for ∼ 2 weeks. The drying procedure was controlled with parallel

samples, the weights of which were measured.

Flame atomic emission spectroscopy Each of the impregnated wood pieces was cut into small cubes

( Figure 3 ) by means of a vertical band sawing machine (Mossner

Rekord, August M ö ssner KG, Mutlangen, Germany; metal cutting

band saw: L.S. Starrett Co., Ltd., Jedburgh, UK, with 14 teeth per

inch). Cube dimensions: 10 × 10 × 8 and 10 × 10 × 4 mm 3 for 1 h impreg-

nation experiments and 5 × 5 × 8 and 5 × 5 × 4 mm 3 in 4 and 12 h impreg-

nation experiments. Each cube was microtomed in transverse direc-

tion to slices of ∼ 0.4 mm thickness, which were oven dried at 105 ° C

for 1 h. The dried slices were kept in desiccators containing blue gel

salt and at room temperature. The slices were acid leached at room

temperature with 2 % HNO 3 (Mallinckrodt Baker, Inc., Phillipsburg,

e

g

d a

f

h

Longitudinal

Radial

Transverse

b c

i

Figure 3 Sampling of small cubes with dimensions of 10 × 10 × 4 mm 3 from an idealized wood piece (size ∼ 100 × 50 × 4 mm 3 ), which were

microtomed for Li + concentration profile measurements.

1.1

1 h impregnation

4 h impregnation

12 h impregnation

Bul

k Li

CI s

olut

ion

conc

entra

tion

(M)

1

0.90 100 200 300 400 500

Time (min)600 700 800

Figure 4 Variation of bulk LiCl solution concentration with time

during the impregnation experiments at room temperature.

NJ, USA) for ∼ 24 h. A set of leaching experiments was also conducted

for 72 h and there was no diff erence in the release of Li + ion between

the 24 and 72 h leaching experiments. At the end of the leaching

period, leaching liquor was collected with a syringe connected to

0.45 μ m polyvinylidene fl uoride membrane fi lter (Pall Life Sciences,

Ann Arbor, MI, USA). The leaching liquor was analyzed for Li + ion

0.8a

b

0.6

0.4

0.2

0

Den

sity

of w

ood

slic

e (g

cm

-3)

0.6

0.8

0.4

0.2

0

0 1 2 3

0 1

100×50×4

100×25×8

Center of the cube

Center of the cube

Sapwood Heartwood

2Position in transverse direction (mm)

3

4 5 6

Figure 5 Density profiles for cube a (cf. Figure 3) of both sW and hW

samples:

(a) 100 × 25 × 8 mm 3 piece and (b) 100 × 50 × 4 mm 3 piece.




7

6

5

sW, 100×25×8 mm3

hW, 100×25×8 mm3

sW, 100×50×4 mm3

hW, 100×50×4 mm34

Con

cent

ratio

n (g

l-1)

3

2

1

00 1 20.5 1.5

Position in transverse direction (mm)

2.5

Figure 6 Measured concentration profiles of Li + ion in Norway spruce

wood pieces at room temperature: sW vs. hW, 1 h impregnation time.

7

6

5

a) Sapwood

a 100×50×4 mm3 chips b 100×25×8 mm3 chips

1 h impr. 4 h impr. 12 h impr.

b) Heartwood b) Heartwood

a) Sapwood

Center of the cube

Center of the cube

Center of the cube

Center of the cube

4

3

2

1

0

7

6

5

4

3

2

1

0

Con

cent

ratio

n (g

l-1)

7

6

5

4

3

2

1

0

7

6

5

4

3

2

1

0

0 1 2 30.5 1.5 2.5

0 1 2 3 40.5 1.5 2.5Position in transverse direction (mm)

3.5 4.5

0 1 2 3 4 50.5 1.5 2.5 3.5 4.5 5.5

0 1 20.5 1.5Position in transverse direction (mm)

2.5

Figure 7 Measured concentration profiles of Li + ion in Norway spruce wood pieces at room temperature for two dimensions (as indicated)

as a function of impregnation time.

concentration by FAES (iCE 3000 series, AA spectrometer; Thermo

Scientifi c, Cambridge, UK). Air-acetylene was the fl ame source, and

the emission was measured at 670.8 nm. The Li + ion concentration

was adjusted to an optimal working concentration range of 0.02 – 5

μ g ml -1 .

Results and discussion The uniformity of the bulk LiCl concentration in the

impregnation vessel was measured at different time

intervals during the chemical impregnation experiments

(Figure 4 ). This figure illustrates that, in all experiments

with 1, 4, and 12 h impregnation times, the wood pieces

were uniformly exposed to the constant bulk concentra-

tion of 1 m LiCl.

The concentration of the chemical agent at the center

of the wood piece is a measure of the completeness of

wood treatment. Therefore, most of the data reported in




the present article are taken from cube a , which is from

the center portion of the piece (Figure 3). In the course

of the experiments, the density profiles were also deter-

mined in addition to the main concentration profiles. The

parameter density is important because earlywood and

latewood have different densities, which affect the trans-

port of ions. The density profiles were measured for cube

a (Figure 3) of both sW and hW samples with two dimen-

sions (Figure 5 ). As expected, earlywood and latewood

show lower and higher densities, respectively (Jyske et al.

2008 ). However, the density variation in hW is not con-

sistent compared with sW due to its different chemical

composition; for example, the extractive content in hW

is higher.

In all the concentration profiles in this section, the

concentration (g of Li + L -1 , where L is free volume of dried

wood slice) is plotted against the position in the trans-

verse direction of the specimens. These data describe the

diffusional mass transport, that is, the assumption was

that the mass transfer occurs through lumens and pits

filled with water.

Effect of impregnation time

A very similar Li + ion concentration profile was observed in

both sW and hW (Figure 6 ) during 1 h impregnation time.

Indeed, there is a diffusional mass transport of Li + ions into

the wood pieces. The Li + concentration difference between

sW and hW is within the experimental error for FAES in the

eluates. However, it should be kept in mind that the pen-

etration depth observed in this study was < 1 mm. Thus, the

similarities between the concentration profiles in Figure 6

are the situation in layers close to the surface. Probably,

the surface is modified in the course of sample prepara-

tion; for example, many microcracks may have formed,

which affect the diffusion of Li + ions into deeper layers.

In Figure 7 a and b, a shift of Li + ion concentration pro-

files is visible after 12 h impregnation toward the center in

sW and hW pieces with 4 and 8 mm thicknesses. For the

4-mm-thick piece (Figure 7a), the shapes of the concen-

tration profiles are similar to that, which can be expected

in a diffusional operation. Also for the 8-mm-thick piece

(Figure 7b), the 1 h impregnated wood piece seems to be

reasonable, but after 4 and 12 h impregnation times the

concentration profiles have quite different shapes. One

possible reason is that microcracks present in the wood

pieces may have contributed to a better accessibility for Li +

ions to diffuse further into the interior of wood pieces, with

increased impregnation time. This observation is much

more pronounced in 4-mm-thick pieces compared with

8-mm-thick pieces for the same experimental conditions.

This example demonstrates the relevance of local concen-

tration measurements in comparison with data based on

average properties. The spatial resolution of penetration is

even indicative for microcracks otherwise not visible.

Concentration profiles at various locations

In Figure 8 , the concentration profiles for three different

positions ( a , b , and c in Figure 3) are shown. Here, it is

evident that the concentration profiles for cube a (middle)

and cube b (between the middle and the end of the piece)

were quite similar, and in these two cases, the diffusion of

Li + ion was only influenced by mass transport in transver-

sal direction. However, in the case of cube c (at the edge of

the piece), the concentration profile is different because of

the influence of mass transport in longitudinal direction

in addition to transversal direction. This effect is much

7

6

5

4

3

2

Con

cent

ratio

n (g

l-1) 1

0

7

6

5

4

3

2

1

0

0 0.5

12 h impregnation

4 h impregnation

a

b

Cube a cf Fig.3 Cube b cf Fig.3

Center of the cube

Center of the cube

Cube c cf Fig.3

1.5 2.5 3.51 2 3

0 0.5 1.5

Position in transverse direction (mm)

2.51 2 3

Figure 8 Measured concentration profiles of Li + ion in Norway

spruce sW specimen at room temperature for 100 × 50 × 4 mm 3 piece

for 4 and 12 h impregnation times. a , b , and c refer to the sampling

within the wood pieces (cf. Figure 3).




more pronounced with elevated impregnation time from 4

to 12 h (Figure 8b).

Conclusion The proposed methodology is time consuming, but the tra-

ditional FAES determination of Li + in the eluate of slices

taken from different depths is precise. The local concentra-

tion profiles can be obtained, which may contribute to a

deeper understanding of the mass transport in wood. The

method could be useful for the calibration of more rapid

cation concentration measurements within the wood

piece (energy-dispersive X-ray analysis, X-ray fluorescence

analysis, secondary ion mass spectrometry, etc.). The

method is sensitive and able to detect cracks and other

defects influencing the mass transport. All these factors

should be considered in the case of precise diffusivity

experiments. However, the preliminary results from this

study must be further verified by means of more repeti-

tions and experiments performed at higher temperatures.

Acknowledgements: The authors are thankful to the Chal-

mers Energy Initiative (CEI) program for their financial

support.

Received October 31, 2012; accepted February 5, 2013; previously

published online March 6, 2013

References Behr, E.A., Briggs, D.R., Kaufert, F.H. (1953) Diffusion of dissolved

materials through wood. J. Phys. Chem. 57:476 – 480.

Bengtsson, G., Simonson, R. (1988) Chemimechanical pulping of

birch wood chips. Part 2. Studies on impregnation of wood

blocks using scanning electron microscopy and energy

dispersive X-ray analysis. Nord. Pulp Pap. Res. J. 3:132 – 137.

Burr, H.K., Stamm, A.J. (1947) Diffusion in wood. J. Phys. Chem. A

51:240 – 261.

Cady, L.C., Williams, J.W. (1935) Molecular diffusion into wood.

J. Phys. Chem. A 39:87 – 102.

Christensen, G.N. (1951a) Diffusion in wood. II. The temperature

coefficient of diffusion through wood. Austral. J. Appl. Sci.

2:430 – 439.

Christensen, G.N. (1951b) Diffusion in wood. III. Ion selection and

its effect on the diffusion of electrolytes. Austral. J. Appl. Sci.

2:440 – 453.

Christensen, G.N., Williams, E.J. (1951) Diffusion in wood. I. A

quantitative theory of diffusion in porous media and its

application to wood. Austral. J. Appl. Sci. 2:411 – 429.

Classen, P.A.M., van Lier, J.B., Lopez Contreras, A.M., van Niel,

E.W.J., Sijtsma, L., Stams, A.J.M., de Vries, S.S., Weusthuis,

R.A. (1999) Utilization of biomass for the supply of energy

carriers. Appl. Microbiol. Biotechnol. 52:741 – 755.

Cooper, P.A. (1998) Diffusion of copper in wood cell walls following

vacuum treatment. Wood Fiber Sci. 30:382 – 395.

Dautzenberg, G., Gerhardt, M., Kamm, B. (2011) Bio based fuels and

fuel additives from lignocellulose feedstock via the production

of levulinic acid and furfural. Holzforschung 65:439 – 451.

Fernando, S., Adhikari, S., Chandrapal, C., Murali, N. (2006)

Biorefineries: current status, challenges, and future direction.

Energy Fuels 20:1727 – 1737.

Fukuyama, M., Urakami, H. (1980) Diffusion of nonelectrolytes

through wood saturated with water I. Measurements of the

diffusion rates by highly sensitive differential refractometer.

Mokuzai Gakkaishi 26:587 – 594.


through wood saturated with water II. Diffusion rates of the

monohydric alcohols. Mokuzai Gakkaishi 28:17 – 24.


through wood saturated with water III. Diffusion rates

of polyethylene glycols. Mokuzai Gakkaishi 32:

147 – 154.

Gindl, W., Dessipri, E., Wimmer, R. (2002) Using UV-microscopy to

study diffusion of melamine-urea-formaldehyde resin in cell

walls of spruce wood. Holzforschung 56:103 – 107.

G ü tsch, J.S., Sixta, H. (2011) Purification of Eucalyptus globulus water prehydrolyzates using the HiTAC process

(high-temperature adsorption on activated charcoal).

Holzforschung 65:511 – 518.

H ö rhammer, H., Walton, S., van Heiningen, A. (2011) A larch based

biorefinery: pre-extraction and extract fermentation to lactic

acid. Holzforschung 65:491 – 496.

Jacobson, A.J., Smith, G.D., Yang, R., Banerjee, S. (2006) Diffusion

of sulfide into southern pine ( Pinus taeda L.) and sweetgum

( Liquidambar styraciflua L.) particles and chips. Holzforschung

60:498 – 502.

Jyske, T., M ä kinen, H., Saranp ä ä , P. (2008) Wood density within

Norway spruce stems. Silva Fennica 42:439 – 455.

Kazi, K.M.F., Gauvin, H., Jollez, P., Chornet, E. (1997) A diffusion

model for the impregnation of lignocellulosic materials. Tappi

J. 80:209 – 219.

Kirsch, C., Zetzl, C., Smirnova, I. (2011) Development of an

integrated thermal and enzymatic hydrolysis for lignocellulosic

biomass in fixed-bed reactors. Holzforschung 65:483 – 489.

L ó pez, Y., Gull ó n, B., Puls, J., Paraj ó , J.C., Mart í n, C. (2011) Dilute

acid pretreatment of starch-containing rice hulls for ethanol

production. Holzforschung 65:467 – 473.

Martin-Sampedro, R., Eugenio, M.E., Revilla, E., Martin, J.A.,

Villar, J.C. (2011) Integration of Kraft pulping on a forest

biorefinery by the addition of a steam explosion pretreatment.

Bioresources 6:513 – 528.

Meijer, M., Zwan, R.P., Militz, H. (1996) Unsteady-state diffusion

of methanol in Douglas-fir heartwood at high temperatures.

Holzforschung 50:135 – 143.

Narayanamurti, D., Kumar, V.B. (1953) Diffusion of organic

molecules through wood. J. Polym. Sci. 10:515 – 524.




Narayanamurti, D., Ratra, R.S. (1951) Diffusion of ions through some

Indian timbers. Proc. Indian Acad. Sci. Sect. A 33:349 – 359.

N ä yh ä , A., Pesonen, H.-L. (2012) Diffusion of forest biorefineries

in Scandinavia and North America. Technol. Forecasting Soc.

Change 79:1111 – 1120.

Ra, J.B., Barnes, H.M., Conners, T.E. (2001) Determination of boron

diffusion coefficients in wood. Wood Fiber Sci. 33:90 – 103.

Ragauskas, A.J., Nagy, M., Kim, D.H., Eckert, C.A., Hallett, J.P.,

Liotta, C.L. (2006) From wood to fuels: integrating biofuels and

pulp production. Ind. Biotech. 2:55 – 65.

Robertsen, L. Diffusion in wood. Ph.D. thesis. Department of

Pulping Technology, Faculty of Chemical Engineering, Å bo

Academy, Finland, 1993.

Sandberg, K., Sterley, M. (2009) Separating Norway spruce

heartwood and sapwood in dried condition with near-infrared

spectroscopy and multivariate data analysis. Eur. For. Res.

128:475 – 481.

Sch ü tt, F., Puls, J., Saake, B. (2011) Optimization of steam

pretreatment conditions for enzymatic hydrolysis of poplar

wood. Holzforschung 65:453 – 459.

Sharareh, S., Tessier, P., Lee, C.-L. (1996) Penetration of sodium

sulphite into black spruce and aspen wood chips using SEM/

EDXA. J. Pulp Pap. Sci. 22: J71 – J77.

Siau, J.F. Transport processes in wood. Springer Series in Wood

Science, vol. 2. Springer-Verlag, Syracuse, 1984.

Stamm, A.J. (1946) Passage of liquids, vapors, and dissolved

materials through softwoods. U.S. Dept. Agr. Tech. Bull. 929.

Stone, J.E., Green, H.V. (1959) Penetration and diffusion into

hardwoods. Tappi 42:700 – 709.

Su, P., Granholm, K., Pranovich, A., Harju, L., Holmbom, B.,

Ivaska, A. (2012) Metal ion sorption to birch and spruce wood.

Bioresources 7:2141 – 2155.

Testova, L., Chong, S.-L., Tenkanen, M., Sixta, H. (2011) Autohy-

drolysis of birch wood. Holzforschung 65:535 – 542.

T ö rnqvist, M., Hurme, T., Rosenholm, J.B. (2001) Drift speed: a way

of measuring diffusion and tortuosity of porous materials.

Colloids Surf. A Physicochem. Eng. Aspects 180:23 – 31.

Tsuchikawa, S., Siesler, H.W. (2003) Near-infrared spectroscopic

monitoring of the diffusion process of deuterium-labeled

molecules in wood. Appl. Spectrosc. 57:675 – 681.



Paper II

Experimental determination of the diffusion of monovalent cation into wood: Effects

of micro-cracks, wood structure, impregnation time and temperature on local

concentration profiles


J-FOR, Journal of Science & Technology for Forest Products and Processes 2014; 4(2):

29-35

Copyright © 2014 PAPTAC

ABST

RACT

29J-FOR Journal of Science & Technology for Forest Products and Processes: VOL.4, NO.2, 2014

SPECIAL BIOREFINERY ISSUE

A thorough understanding of the key phenomena that take place during the chemical transport of reactants into a wood matrix is critical for the success of today’s chemical pulp mills as well as future biorefinery operations. In the present article, our previous experimental meth-odology has been further developed to minimize the effect of micro-cracks present in wood pieces. The effects of wood structure (sapwood and heartwood), impregnation time, and temperature on local concentration profiles in Norway spruce wood were investigated using LiCl as a tracer substance. For experimental conditions of neutral/low pH, Li+ ion concentrations in the eluate of slices taken from the outer surface layers of an impregnated wood piece were found to be higher than the native anionic group content (i.e., the cationic exchange capacity) of the wood piece. The method gave reasonable results, but defects in the wood piece (micro-cracks) still existed and were detectable.

REDDYSURESH KOLAVALI, HANS THELIANDER*

EXPERIMENTAL DETERMINATION OF THE DIFFUSION OF MONOVALENT CATION INTO WOOD: EFFECTS OF MICRO-CRACKS, WOOD STRUCTURE, IMPREGNATION TIME AND TEMPERATURE ON LOCAL CONCENTRATION PROFILES

Increased global energy demands and the limited availability of fossil resources cou-pled with increased social concern about greenhouse gas emissions have prompted humankind to look for sustainable alterna-tive resources. Petroleum or fossil energy resources are considered to be nonrenew-able carbon resources because of their long geologic production cycles (~200 million years) as compared with biomass feedstocks (~<1 to 80 years), which are renewable and carbon-neutral resources [1]. In recent years, several types of poten-tial biomass feedstocks have been identi-fied, and it has been proposed that each biomass type be processed in a specific way to produce green materials, chemicals, and energy [2]. Wood lignocelluloses are the only type of biomass that is seasonally independent, and lignocelluloses are the most abundant biomass on earth, account-ing for an annual production of ~170 bil-lion metric tonnes in the biosphere. The “biorefinery” is the most popular con-cept in this context for converting forest

biomass to different forms of energy, materials, and chemicals. The biorefin-ery (with renewable biomass as feed-stock) is analogous to the petroleum re-finery (with nonrenewable petroleum as feedstock) whereby a single feedstock is fractionated into a multitude of com-modity products depending on economic and social requirements. Today, the ma-jor uses of wood include direct use as a building material, making paper products, and generating heat through combustion. The pulping industry is the most devel-oped chemical technology for processing wood, and to some extent it is already a biorefinery in which heat, electric power, and cellulosic fibres are produced from wood. The biorefinery concept establish-es additional improvements in this area: the decreasing competitiveness of tradi-tional pulp and paper mills has increased the opportunities for and the urgency of transforming chemical pulp mills into integrated forest biorefineries (IFBR) to produce higher value-added products

INTRODUCTION

*Contact: [email protected]

REDDYSURESH KOLAVALIDivision of Forest Products and Chemical Engineering, Department of Chemical and Biological Engineering, Chalmers University of Technology, Gothenburg, Sweden

HANS THELIANDERDivision of Forest Products and Chemical Engineering, Department of Chemical and Biological Engineering, Chalmers University of Technology, Gothenburg, Sweden

30 J-FOR Journal of Science & Technology for Forest Products and Processes: VOL.4, NO.2, 2014

such as ethanol, polymers, carbon fibres, and diesel fuel in addition to pulp [3]. Therefore, it is believed that the biore-finery will play a significant role in forest clusters worldwide. Enormous research efforts in the biorefinery field are under-way around the world today, targeting the production of value-added products [4–14]. An illustration of this is the extrac-tion of hemicelluloses such as xylan and glucomannans from wood chips before chemical pulping and their conversion into barrier films through acid hydroly-sis, autohydrolysis, hot-water extraction, steam explosion, or alkali extraction, lead-ing to a potentially profitable use.

In many cases, chemicals must be transported into wood before reaction. Therefore, the wood must be impreg-nated with the reactants. The phenom-ena involved can be divided into primary (penetration) and secondary (diffusion) transport mechanisms. Penetration is de-fined as the flow of liquor into gas- or steam-filled voids in the wood samples under a pressure gradient, and diffusion is defined as the movement of ions or mol-ecules through a liquid with concentration gradients as the driving force. Penetration occurs very fast in the beginning if there are any gas- or steam-filled pores, but if the pores are completely filled with liquid, diffusion is the mechanism involved and is much slower. The role of penetration is to fill fibre cavities with liquid, which enables faster and more uniform diffusion of chemical ions into wood [15]. For these reasons, proper impregnation of wood with chemicals will be of great importance for an economically viable and efficient biorefinery process because homogeneous impregnation increases treatment unifor-mity and reduces reaction times. There-fore, the complex chemical transport phe-nomena of reactants into the wood matrix are of critical importance and must be better understood to ensure the success of biorefinery operations. The most relevant investigations have been conducted in the context of measuring chemical diffusion into wood, but no standard method exists. For a complete list of literature referenc-es, the reader is referred to our previous

article [16].From the literature review, most in-

vestigations were based on either an av-erage flux measurement of the diffusing substance or an electrical conductivity measurement of the impregnated wood samples. These methods have limita-tions, e.g., these methods cannot be used to determine the concentration profile at various separate locations within a single wood piece. Many aspects of chemical dif-fusion into wood have not yet been exten-sively studied. For example, most research on cation diffusion measurements into wood has been conducted with substances such as NaCl (sodium chloride), KCl (po-tassium chloride), and NaOH (sodium hydroxide). In cases in which NaCl/KCl is used as the diffusing substance, deter-minations of cation diffusion into wood may be inaccurate due to the high original concentration of Na+/ K+ ions in wood. In cases in which NaOH is used as the diffusing substance, the OH- (hydroxyl) ions react with wood components and mask the pure diffusional effects of Na+ ions. Moreover, some observations clearly indicate that almost all wood pieces have micro-cracks, which may be too small for direct visual detection, but which can change surface-layer capillarity and thus influence the interaction with the sur-rounding liquor and/or re-open aspirated pits and thus make the layer structure more open. These micro-cracks are due to surface-layer damage resulting from me-chanical preparation of the wood pieces (e.g., sawing) [17]. These differences influ-ence the behaviour of a wood piece in dif-fusion measurements.

For these reasons, in our previous work [16], an attempt was made to de-velop an experimental methodology to determine the concentration profiles of cations in a wood piece as a function of wood piece dimensions, impregnation time, wood structure, and temperature. To overcome some of the difficulties mentioned above, LiCl (lithium chloride) was chosen as the diffusing substance, and a direct method consisting of slic-ing the wood piece and measuring the Li+ ion concentration with flame atomic

emission spectroscopy (FAES) was used. Even though wood contains very small amounts of naturally occurring Li+ and Cl- ions, the effect of the original presence of these ions on cation diffusion into wood can be considered as negligibly small. Su et al. [18] have also demonstrated that Li+ has low affinity towards wood compo-nents. Furthermore, in a previous study, we observed that the micro-cracks pres-ent on the surface layers of wood are best indicated through variation in the spatial resolution of penetration (concentration profiles) and are otherwise not visible. Therefore, in the present paper, previous experimental methodology [16] has been further developed to minimize the effect of micro-cracks on measurements of lo-cal cation concentration profiles within a wood piece. The effects of impregnation time and temperature on Li+ ion concen-tration profiles in sapwood (Sw) and heart-wood (Hw) samples are also discussed.

MATERIALS AND METHODS

Samples A stemwood disc 23 cm in diameter (with-out bark) from a 31±1-year-old Norway spruce (Picea abies L) was investigated. Both Sw and Hw samples were carefully prepared using a vertical band saw and were kept preliminarily in an airtight poly-ethylene (PE) bag at 1oC. It was assumed that a stemwood disc contains 50% Sw and 30% Hw and that the remaining 20% is an intermediate between Sw and Hw [19]. Only Sw and Hw were investigated and not the intermediate material. Only rot- and other deformation-free samples were selected and cut into rectangular prototype pieces using the same vertical band saw mentioned above. The dimen-sions of the prepared wood pieces were 100 × 25 × 8 mm3 (L × W × T). The material was then stored in an airtight PE bag in a freezer at -18oC. Defrosting the wood pieces took 24 h. To minimize sur-face damage to the wood pieces caused by rough sawing, the wood pieces were care-fully planed on all four vertical surfaces to peel off surface layers approximately 0.5 to 1.5 mm in thickness using a hand plane



(Stanley Hand Tools, Stanley Canada, Mis-sissauga, ON L5N 7K6). Further steps in the experimental procedure, equipment details such as cutting machine specifica-tions, and information about the chemi-cals used can be found in our earlier paper [16].

Water Impregnation Cycles The wood pieces were water-impregnated in a vacuum-pressure cycle in a small poly-propylene (PP) beaker placed in an auto-clave filled with deionized water (at ~22oC, i.e., at r.t.). Then a vacuum was applied for 30 min, and the autoclave was pressurized with N2 at 0.5 MPa for 1 h. This proce-dure was repeated until no floating pieces were observed after complete water im-pregnation (maximum 3–5 cycles for Sw and 5–15 cycles for Hw). To eliminate the small bubbles left on the surfaces (which constitute possible barriers against chemi-cal diffusion), the pieces were kept in the final vacuum for 30 min before the subse-quent experimental steps.

Impregnation with LiCl The pieces were then dabbed with filter paper and immersed in a solution of 1 M LiCl at a wood-to-liquor ratio of 1:125 (Fig. 1). The impregnation vessel was made of PP and placed in an unstirred wa-ter bath (Sub Aqua 26 Plus, Grant Scien-tific, Grant Instruments (Cambridge) Ltd., SG8 6GB, England). The impregnation vessel was equipped with a liquor mixer fitted with a three-blade impeller to avoid concentration gradients in the impregna-tion vessel and with a digital thermometer

(Traceable® Digital Thermometer, Con-trol Company, Friendswood, Texas 77546 USA) to monitor and control temperature in the impregnation vessel. After impreg-nation at time intervals of 4 and 12 h, at each of the specified temperatures of 40oC and 60oC, the pieces were removed and placed in liquid N2 (-180oC) to stop (minimize) further migration of Li+ ions. The frozen pieces were lyophilized for ap-proximately two weeks. The drying proce-dure was controlled by means of parallel samples, the weights of which were mea-sured.

Flame Atomic Emission Spectroscopy (FAES) Each impregnated wood piece was cut into small cubes (Fig. 2) using the verti-cal band saw. The target cube dimensions were approximately 6 x 6 x 6 mm3 in all the impregnation experiments. Each cube was microtomed in the tangential direction to slices of ~0.3 mm thickness, which were

oven-dried at 105oC for 1 h. The dried slices were kept in desiccators containing blue gel salt and at r.t. The slices were acid-leached at r.t. with 2% HNO3 for approxi-mately 24 h. A set of leaching experiments was also conducted for 72 h, and there was no difference in the release of Li+ ion be-tween the 24-h and 72-h leaching experi-ments. At the end of the leaching period, the leaching liquor was collected with a syringe connected to a 0.45 µm PVDF (polyvinylidenefluoride) membrane filter. The leaching liquor was analyzed for Li+ ion concentration using FAES. An air-acetylene torch was the flame source, and the emission was measured at 670.8 nm. The Li+ ion concentration was adjusted to an optimal working concentration range of 0.02 to 5 µg ml-1.

RESULTS AND DISCUSSION

The uniformity of the bulk LiCl concen-tration in the impregnation vessel was

Fig. 1 - Experimental setup for impregnation with LiCl solution at different temperatures.

Fig. 2 - (a) Sampling of small cube samples (6×6×6 mm3) from an idealized impregnated wood piece (100×22×6 mm3); (b) microtoming of small cube into slices of ~0.3 mm thickness for Li+

ion concentration profile measurements.


measured at four time intervals during all the chemical impregnation experiments. The results indicated that, in all experi-ments with 4 and 12 h treatment times and for experiments both at 40oC and 60oC, the wood pieces were exposed to a bulk concentration of 7 ±1 (g Li+ L-1) [1 M LiCl = 6.98 g Li+ L-1]. In a similar way, the temperature inside the impregnation ves-sel was also continuously monitored and recorded at four time intervals during the chemical impregnation experiments. Ob-servations indicated that the temperature inside the impregnation vessel was main-tained at 40 ±1oC and 60 ±1oC for the ex-periments at 40oC and 60oC respectively.

Because the concentration of chemi-cal agent at the centre of the wood piece is a reasonable measure of the completeness of the wood treatment, most of the data reported in the present article are taken from cube a, which is from the center por-tion of the piece (Fig. 2). To understand the precise effects of various factors on local concentration profiles, in all the con-centration profiles presented in this study, the concentration mol/Kg (i.e., moles of Li+ (kg dried wood)-1) is plotted against the thickness of the impregnated wood piece. Indeed, this experimentally measured concentration is the total concentration of Li+ ion on wood and in the solution in the pores. The following abbreviations are used in the following text: NWP (nor-mal wood pieces) means wood pieces that were carefully prepared from a stemwood disc, using a vertical band saw; SsWP (surface-smoothed wood pieces) means wood pieces that were carefully planed on all four vertical surfaces to peel off the

damaged surface layer of NWP to a depth of approximately 0.5 to 1.5 mm using a hand plane. In all the following sections, except for that on the effect sof micro-cracks, SsWP are used to minimize the ef-fect of micro-cracks.

Error analysis of the experimental methodologyTo verify the reproducibility of the experi-mental methodology used, concentration profiles were measured for three Sw piec-es with similar dimensional specifications (~100×22×6 mm3) after 12 h of chemical impregnation experiments at 40oC. Figure 3 shows a good example of one advan-tage of local measurements because it is obviously easy to detect various defects, such as cracks in the wood sample. Clearly, two of the three experimental series show good similarities in concentration pro-files, with an experimental error of ap-proximately 11%. The third series (wood piece 1), however, has several deviant data points, which were attributed to cracks in the wood structure. If all data points (wood pieces 1, 2, and 3) are considered when evaluating the experimental error, it becomes quite large, approximately 21%. However, if some data points that may be considered erroneous due to cracks in all three wood pieces are excluded from the error calculations, the error decreases to approximately 8%.

Effect of micro-cracksThis section presents several illustrations of the influence of damaged surface lay-ers on Li+ ion concentration profiles in Norway spruce wood samples. In Fig. 4,

a higher Li+ ion concentration in the outermost surface layers can be seen in NWP than in SsWP. This clearly indicates that the rough surface layers, which were formed during rough sawing, may have al-tered the influence of mass transfer on Li+ ion diffusion in NWP more than in SsWP. Figure 4 shows that the outermost surface layer, which is 1–2 mm thick, may have different properties from the bulk of the wood piece. Therefore, it seems that the formation and behaviour of the damaged surface layers depends on the method of preparing the wood pieces. Moreover, the effect of the damaged surface layer on Li+ ion diffusion in Norway spruce Sw was more pronounced than in Hw.

Effect of wood structureDifferences in Li+ ion concentration pro-file were observed between Sw and Hw samples of Norway spruce at different temperatures and for different treatment times (Fig. 5). Higher Li+ ion concentra-tions were observed in Sw than in Hw for eluates of slices taken at similar depths. This observation became more pro-nounced as the treatment time increased from 4 to 12 h and the temperature in-creased from 40oC to 60oC. These differ-ences in Sw and Hw concentration pro-files are due to differences in the structural and chemical compositions of Sw and Hw, e.g., a) Hw is usually much less permeable than Sw due to pit aspiration and incrusta-tion [20], and b) the extractives content in Hw is higher than in Sw. However, differ-ent concentration profile shapes were ob-served for extended treatment times (Fig. 5 A) b)) as well as at higher temperatures

Fig. 3 - Measured concentration profiles of Li+ ion in Norway spruce Sw of three different wood pieces of similar specifications after 12 h treatment time at 40oC.

Fig. 4 - Measured Li+ ion concentration profiles in Norway spruce Sw and Hw samples, at room temperature and for 4 h treatment time (as indicated): NWP vs. SsWP.



(Fig. 5 B) b)). This is probably due to the remaining effect of micro-cracks on Li+ ion diffusion in the wood samples.

Effect of Impregnation TimeIn Figs. 6 A) and 6 B), a shift in concen-tration profiles towards the centre of the wood piece is visible as the treatment du-ration increases from 4 to 12 h. This ob-servation is more pronounced in Sw than in Hw and for increased temperature from

40oC to 60oC. In the case of Hw, at 40oC, the concentration profiles remained within experimental error as the treatment time increased from 4 to 12 h. This difference in behavior occurred because Hw is less permeable than Sw. On the other hand, at 60oC, as the impregnation time increased from 4 to 12 h, higher Li+ ion concentra-tions were observed. In addition, different concentration profile shapes were ob-served for both Sw and Hw as treatment

time increased at 60oC. Moreover, the ef-fect of micro-cracks on concentration profiles apparently still exists to some ex-tent and contributes to the different con-centration profile shapes at higher tem-peratures and extended treatment-time intervals.

Effect of TemperatureMinor differences in Li+ ion concentra-tion profiles were found as the tempera-ture increased from 40oC to 60oC (Fig. 7). As expected, the Li+ ion concentra-tion in a given wood piece increased with temperature. This observation was more pronounced in Sw than in Hw after a 4 h treatment period. These differences may have been due to differences in chemical composition, e.g., the extractives content in Hw is higher than in Sw. As the temper-ature increased, the observed penetration depth also became deeper in Sw than in Hw. However, after a 12 h treatment pe-riod, the differences in concentration pro-files at 40oC and 60oC were appreciable, and different concentration profile shapes were observed.

As has been noted in this experi-mental study, the Li+ ion concentration in a wood piece increased as the treat-ment time increased from 4 to 12 h and as the temperature increased from 40oC to 60oC. The effects of extended treatment time and increased temperature not only occurred in a predictable manner as pre-dicted by diffusion theory, but also differ-ent concentration profile shapes were ob-served as well. Therefore, it is evident that Li+ ion transport in wood material may be influenced by other phenomena such as surface diffusion, adsorption/desorp-tion, and Donnan equilibrium. More im-portantly, for the experimental conditions studied (i.e., relatively neutral or low pH), in all concentration profiles, the observed Li+ ion concentrations were much higher than the concentrations of carboxylic acid groups (0.086 ± 0.007 mol/Kg) (Werkelin et al., 2010; Shoulaifar et al., 2012) present in the native Norway spruce wood sam-ples. In wood samples, at neutral or weakly acidic conditions, mainly carboxylic sites are predominantly ionized, and therefore

Fig. 5 - Measured Li+ ion concentration profiles in Norway spruce SsWP at A) 40oC and B) 60oC: Sw vs. Hw; for treatment times of a) 4 h, and b) 12 h.

Fig. 6 - Measured concentration profiles of Li+ ion in Norway spruce after different treatment-time intervals: 4 h vs. 12 h, and for different temperatures (as indicated) for A) Sw and B) Hw samples.


they exist in anionic forms, and the Li+ cations can bind to the carboxylic sites.

Concentration Profiles at Various LocationsFigure 8 shows concentration profiles for different locations in the wood piece (a, b, c, d, and e in Fig. 2). It is evident that the concentration profiles for cube a and cube c (cube c is symmetric to cube e) are

different. This phenomenon occurs be-cause in the case of cube a, Li+ ion dif-fusion was influenced only by mass trans-port in the transverse direction. However, in the case of cube c (cube e), Li+ ion dif-fusion was influenced by mass transport in both longitudinal and transverse direc-tions. This effect is evident in both Sw and Hw samples even at extended treatment-time intervals and temperatures, but with

different concentration profile shapes (Figs. 8 A) & B) b)).

CONCLUSIONS

Our previous experimental methodology was further developed to minimize the ef-fect of micro-cracks on Li+ ion concen-tration profile measurements in Norway spruce wood pieces. The proposed meth-odology was also verified with three rep-licates, thus confirming that the method provides a good prediction of concentra-tion profiles and showing that the effect of micro-cracks varies among wood piec-es. Effects of treatment-time duration and temperature on Li+ ion concentration pro-files in both Sw and Hw pieces were inves-tigated. For the experimental conditions chosen, concentrations of Li+ ion much higher than the cationic exchange capacity of Norway spruce wood were observed in the slices taken from the outer surface of the wood pieces. The improved method gave reasonable results, but defects in the wood pieces (micro-cracks) remained and were detectable.

REFERENCESLiu, S., Amidon, T.E., Francis, R.-C., Ra-marao, B.V., Lai, Y.-Z., and Scott, G.M., “From Forest Biomass to Chemicals and Energy—Biorefinery Initiative in New York State”, Feature Commentary, Industrial Biotechnology, 2(2):113–120 (Summer 2006).Schlosser, S. and Blahusiak, M., “Bio-refinery for Production of Chemicals, Energy, and Fuels”, Elektroenergetika, 4(2):8–16 (2011).Van Heiningen, A., “Converting a Kraft Pulp Mill into an Integrated For-est Products Biorefinery”, Technical Articles, Technical Association of the Pulp and Paper Industry of South Africa (TAPPSA), 1–9 (May 2007).Ragauskas, A.J., Nagy, M., Kim, D.H., Eckert, C.A., Hallett, J.P., and Liotta, C.L., “From Wood to Fuels: Integrating Biofuels and Pulp Production”, Indus-trial Biotechnology, 2: 55–65 (2006).Dautzenberg, G., Gerhardt, M., and Kamm, B., “Bio-Based Fuels and Fuel Additives from Lignocellulose Feedstock

1.

2.

3.

4.

5.

Fig. 7 - Measured Li+ ion concentration profiles in Norway spruce at different temperatures: 40oC vs. 60oC, and after treatment intervals of 4 h and 12 h (as indicated) for both A) Sw and B) Hw samples.

Fig. 8 - Measured concentration profiles of Li+ ion in Norway spruce wood pieces at various locations within a single wood piece of A) Sw sample, after intervals of a) 4 h and b) 12 h treatment at 40oC, and B) Hw sample, for a) 40oC and b) 60oC after 12 h treatment time. Cubes a, b, c, and e refer to sampling within the wood piece (as shown in Fig. 2).



WHY JOIN PAPTAC’s TECHNICAL COMMUNITIES

Sharing information on speci�c topics & challenges facing the Canadian pulp and paper industry. Accessing an exclusive Canadian technical pulp and paper network.

Being aware of the latest technological advancements and innovations. Greater value derived from participating in PAPTAC events (PaperWeek, PACWEST, conferences, webinars, etc.).

Continuing to learn from your peers, identifying and developping new problem-solving solutions.

The technical communities have been formed among PAPTAC members who are interested in a particular aspect

of the industry. Members discuss practical and scienti�c developments in their area of specialty. They work on

projects to learn more about the industry, and then disseminate this information to the industry through

technical presentations and reports.

DOI 10.1515/hf-2012-0182 (2013).Salin, J.G., “Almost All Wooden Pieces Have a Damaged Surface Layer—Im-pact on Some Properties and Quality”, Proceedings, Conference COST E53, Delft, The Netherlands, 135–143 (2008).Su, P., Granholm, K., Pranovich, A., Harju, L., Holmbom, B., and Ivaska, A., “Metal Ion Sorption to Birch and Spruce Wood”, Bioresources, 7:2141–2155 (2012).Sandberg, K. and Sterley, M., “Separat-ing Norway Spruce Heartwood and Sapwood in Dried Condition with Near-Infrared Spectroscopy and Multivariate Data Analysis”, European Journal of Forestry Research, 128:475–481 (2009).Siau, J.F., Transport Processes in Wood. Springer-Verlag. Berlin, New York (1984).Werkelin, J., Skrifvars, B-J., Zeven-hoven, M., Holmbom, B., and Hupa, M., “Chemical Forms of Ash-Forming Ele-ments in Woody Biomass Fuels”, Fuel, 89:481–493 (2010).Shoulaifar, T.K., DeMartini, N., Ivaska, A., Fardim, P., and Hupa, M., “Mea-suring the Concentration of Carbox-ylic Acid Groups in Torrefied Spruce Wood”, Bioresource Technology, 123: 338–343 (2012).

17.

18.

19.

20.

21.

22.

Way to Activate Wood for Enzymatic Treatment, Chemical Pulping, and Biore-finery Processes”, Nordic Pulp & Paper Research Journal, 27(5):828–835 (2012).Hansen, N.M.L. and Plackett, D., “Sus-tainable Films and Coatings from Hemi-celluloses: A Review”, Biomacromol-ecules, 9(6):1493–1505 (2008).Escalante, A., Goncalves, A., Bodin, A., Stepan, A., Sandström, C., Toriz, G. and Gatenholm, P., “Flexible Oxygen Bar-rier Films from Spruce Xylan”, Carbo-hydrate Polymers, 87:2381–2387 (2012).Alekhina, M., Mikkonen, K.S., Alen, R., Tenkanen, M., and Sixta, H., “Carboxy-methylation of Alkali Extracted Xylan for Preparation of Bio-Based Packaging Films”, Carbohydrate Polymers, http://dx.doi.org/j.carbpol.2012.03.048 (2013).Määttänen, M. and Tikka, P., “Determi-nation of Phenomena Involved in Im-pregnation of Softwood Chips, Part 2: Alkali Uptake, Alkali Consumption, and Impregnation Yield”, Nordic Pulp & Paper Research Journal, 27(3):559–567 (2012).Kolavali, R. and Theliander, H., “Deter-mination of the Diffusion of Monova-lent Cations into Wood under Isothermal Conditions Based on LiCl Impregnation of Norway Spruce”, Holzforschung,

12.

13.

14.

15.

16.

via the Production of Levulinic Acid and Furfural”, Holzforschung, 65(4):439–451 (2011).Hörhammer, H., Walton, S., and van Heiningen, A., “A Larch-Based Biorefin-ery: Pre-Extraction and Extract Fermen-tation to Lactic Acid”, Holzforschung, 65(4):491–496 (2011).López, Y., Gullón, B., Puls, J., Parajó, J.C., and Martín, C., “Dilute Acid Pretreat-ment of Starch-Containing Rice Hulls for Ethanol Production”, Holzforsc-hung, 65(4):467–473 (2011).Martin-Sampedro, R., Eugenio, M.E., Revilla, E., Martin, J.A., and Villar, J.C., “Integration of Kraft Pulping in a Forest Biorefinery by the Addition of a Steam Explosion Pretreatment”, Bioresources 6:513–528 (2011).Schütt, F., Puls, J., and Saake, B., “Opti-mization of Steam Pretreatment Condi-tions for Enzymatic Hydrolysis of Poplar Wood”, Holzforschung, 65(4):453–459 (2011).Testova, L., Chong, S.-L., Tenkanen, M., and Sixta, H., “Autohydrolysis of Birch Wood”, Holzforschung, 65(4):535–542 (2011).Jedvert, K., Wang, Y., Saltberg, A., Hen-riksson, G., Lindstrom, M.E., and The-liander, H., “Mild Steam Explosion: A

6.

7.

8.

9.

10.

11.

Paper III

The sorption of monovalent cations onto Norway spruce: model studies using wood

flour and LiCl solution


In Manuscript

1

The sorption of monovalent cations onto Norway spruce: The sorption of monovalent cations onto Norway spruce: The sorption of monovalent cations onto Norway spruce: The sorption of monovalent cations onto Norway spruce:

model studies using wood flour and LiCl solutionmodel studies using wood flour and LiCl solutionmodel studies using wood flour and LiCl solutionmodel studies using wood flour and LiCl solution

Reddysuresh Kolavalia, Merima Hasania and Hans Theliandera* a Division of Forest Products and Chemical Engineering, Dept. of Chemistry & Chemical Engineering, Chalmers University of Technology, SE-41296 Gothenburg, Sweden. *Corresponding author: Professor Hans Theliander Tel: +46 31 772 2992 Fax: +46 31 772 2995 E-mail: [email protected]

Abstract The transport of chemicals in a porous material such as wood is very complex and involves several processes: the diffusion of chemicals in the cell pores (lumen and pit pores), through the cell walls at certain conditions, and sorption at solid surfaces. In the present study, batch sorption experiments were performed to examine the sorption of Li+ ions from an aqueous LiCl solution onto Norway spruce wood flour samples. The experimental methodology employed is suitable for differentiating the amount of ions sorbed onto the surfaces of wood cell walls/ into the cell walls, and dissolved in the solution in wood pores. The apparent equilibrium sorption data was analyzed using two widely applied isotherm models: Langmuir and Freundlich. The results suggest that the sorption was spontaneous and, for the experimental conditions studied, probably involved several interaction types between the different functional groups of the wood and the Li+/Cl- ions.

Keywords: diffusion, impregnation, sorption equilibrium, metal ion, MCC, Xylan, Kraft lignin, Wood

1. Introduction Impregnating wood with chemicals is often of great importance when it is to be processed. The uniform distribution of chemicals into the wood prior to a reaction increases the uniformity of treatment to which it is then subjected, reduces reaction times and may, as a consequence, increase the yield of the products. The mass transport involved when wood is impregnated with chemicals can be divided into two kinds: advective (penetration) and diffusive (diffusion) [1]. Penetration is defined as the flow of liquor into the gas/vapour-filled voids in the wood under a pressure gradient, and diffusion as the movement of solute matter through the liquid present inside the wood caused by concentration gradients. In fresh wood, both mechanisms occur at the same time: in certain cases (e.g. pulping), however, the majority of the chemicals are transferred into the wood by diffusion, since chemicals are consumed via reactions inside the wood and thus create concentration gradients. Diffusion is therefore the controlling mechanism in most (but not all) wood treatment processes. Interactions between the components of the cell wall and the diffusing solutes mean that the solutes may also be sorbed onto the cell wall, so there may also be a mass transport at the cell wall itself. The overall mass transport rate may thus be influenced by several different mass transport phenomena as well as by sorption phenomena. Many researchers have investigated the mass transport of chemicals/ions in wood (e.g. [2-24]) but, to the best knowledge of the authors, none have considered the sorption phenomenon in wood, when the diffusion coefficients of ions in wood have been determined. However, to predict the diffusion coefficient of ions in wood accurately, an understanding of the sorption of ions in wood tissue is required. In our previous study [25], concentration profiles of Li+ ions in Norway spruce wood were measured at different impregnation times and temperatures. The shapes of the concentration profiles observed were as expected for a process controlled by

2

diffusion but, for the experimental condition of neutral/low pH, the concentrations of Li+ ions on the outer surface layers of the wood were found to be higher than the content of the native anionic group (i.e. the cationic exchange capacity) of the Norway spruce wood. Thus it is possible that the transport of Li+ ions in wood material may also be influenced by other phenomena, such as surface diffusion, adsorption/desorption and the Donnan equilibrium. There are quite a few investigations in the literature regarding the sorption properties of wood (sorbent). However, most of them deal with the removal of various contaminants (sorbate) such as dyes and heavy metals from wastewater streams emanating from dyeing, textiles, paint, rubber, leather, paper, plastics, ceramics, pharmaceutical, food and cosmetics industries [26-39]. The experimental methodology used in these investigations to calculate the amount of sorbate sorbed by the unit weight of wood (mg sorbate/g (dry) sorbent) is normally based on the mass balance of the sorption system: the sorbate that “disappears” from the solution must be in the wood (both in the pores and on the cell wall). This method is not suitable if it is necessary to differentiate the amount of sorbate sorbed onto the surfaces of the wood’s cell walls from that which is dissolved in the solution in the pores at equilibrium conditions. This type of information is required if the mass transport of chemicals in wood, and the mechanism involved in the sorption of chemicals in wood material, are to be understood. When a metal ion in solution interacts with a wood surface [40], it can be sorbed by the process of “ion exchange”: this may take place between the incoming cation and either a sorbed metal ion or a hydrogen ion available in the functional groups of wood material (Figure 1). On the other hand, metal ions can also be sorbed by chemical complexation, in which site-specific interactions take place between metal ions and functional groups of wood. Based on the behaviour of heavy metal sorption on sawdust, it has been speculated that ion exchange and hydrogen bonding may be the principal mechanism in the sorption of heavy metals [27]. The main objective of this study was to develop a methodology to examine the sorption of ions onto wood material. The experimental study involves model wood constituents (Micro Crystalline Cellulose (MCC), hemicellulose (Beechwood Xylan) and lignin (Kraft Lignin)) as well as samples of wood flour (Norway spruce). Batch sorption studies, using lithium chloride (LiCl) as a model component, investigated samples of both sapwood (Sw) and heartwood (Hw) at different temperatures and concentrations of bulk LiCl solution. The wood material was characterized by measuring pore size and volume (using mercury porosimetry) [41-42]. FTIR was used to investigate the samples before and after LiCl treatment to obtain information of the functional groups that might be involved in the sorption of metal ions. Langmuir and Freundlich equilibrium isotherm models were fitted to the experimental equilibrium data.

2. Materials and Methods 2.1. Materials and chemicals

The wood materials studied were wood flour (both Sw and Hw), MCC (Avicel PH-101; particle size ~50 µm; SIGMA-ALDRICH Co., Germany), Xylan from beechwood (SIGMA-ALDRICH Co., Germany), and Kraft lignin (a softwood lignin extracted from black liquor using the LignoBoost process at the Bäckhammar Mill, Sweden). A dilute solution of sulphuric acid (0.31 M) was prepared using reagent grade (95-97%) sulphuric acid (Scharlau, Scharlab S.L., Spain) and deionized water. Stock solutions of different Li+ ion concentrations were prepared by dissolving LiCl salt (≥ 99 %) (Merck KGaA, Darmstadt, Germany) in deionized water. 2 wt. % nitric acid (HNO3) solution was prepared using trace analysis grade (69 %) nitric acid (ARISTAR, VWR® PROLABO®, Leuven) and deionized water. Potassium Bromide (KBr) (≥ 99 %, FTIR grade) purchased from Sigma Aldrich. All chemicals were used as received. 2.2. Preparation of wood flour

Samples of both Sw and Hw were prepared from a stem wood disc, 23 cm in diameter (excluding bark), taken from a 31±1 year old Norway spruce (Picea abies L). Only samples free

3

from rot and deformations were selected; they were cut into rectangular prototype pieces using a vertical band saw machine. The material was then stored in an airtight polyethylene (PE) bag in a freezer at -18 oC. Defrosting the wood pieces took ~24 h at ambient temperature, after which they were ground in a Wiley-type mill (< 1 mm). The resulting wood flour was stored in an airtight PE bag for further use. 2.3. Pre-impregnation of wood flour with dilute sulphuric acid solution

Native Norway spruce wood contains small amounts of metal ions (typically 1-5 g/kg wood), mainly Ca2+, Mg2+, Mn2+, Na+ and K+. Although these ions are only present in small quantities, they may participate in the ion-exchange induced sorption mechanism, and thus aggravate the measurements made. Therefore, prior to the sorption experiments, the sample of wood flour was vacuum-impregnated in a dilute sulphuric acid solution (pH=0.5/0.31 M H2SO4) for ~18 h at room temperature (R T i.e. ~22 oC) in order to remove native metal ions, air and any floating fines present in the wood material. The concentrations of metal ions in the samples of both Sw and Hw before and after acid treatment are reported in Table A1 (see Appendix A1). Only traces of these metal ions were observed in the samples of wood flour after acid treatment. The effect of these metal ions on mass transport of Li+ ions in Norway spruce wood is thus negligible. The suspension was then vacuum-filtered using filter paper (00A Grade filter paper, Munktell Filter AB, Falun, Sweden) to separate acid solution from wood flour. The acid-treated wood flour was then used immediately in the subsequent experimental steps. 2.4. Sorption experiments

Wood flour was added to polypropylene (PP) beakers containing an aqueous LiCl solution of the desired concentration with a liquor-to-wood flour ratio of 100:1(400:4(g/g)). The mixture was shaken sporadically during the experiment and, at the end of the specified residence time, it was centrifuged (SERIE 130; F.I.E.M.M.E; Establishments Couprie; 7, Quai Clande Bernard; Lyon. approx. 2960 rpm) for 5 min to separate the wood flour from the LiCl solution. The sample of wood flour treated with LiCl was then transferred to a PP sample bottle covered with a polyester net (40 µm mesh opening) and placed in liquid nitrogen (-180 oC). The frozen sample was then lyophilized (Labconco, Kanas City, MO, USA) for 5 days. The freeze-dried, LiCl (aq.)-treated, sample of wood flour was then acid leached at R T with 2 wt. % HNO3 for ~24 h. The leaching experiments were conducted in triplicate under identical conditions and average values were calculated. The leaching liquor was collected with a syringe connected to a 0.45 µm polyvinylidene fluoride membrane filter (Pall Life Sciences, Ann Arbor, MI, USA) and was analyzed for its Li+ ion concentration by Flame Atomic Emission Spectroscopy (FAES) (iCE 300 series, AA spectrometer; Thermo Scientific, Cambridge, UK). Air-acetylene was the flame source, and the emission was measured at 670.8 nm. The concentration of the Li+ ion was adjusted to an optimal working concentration range of 0.02-5 µg/ml. The average deviation from the mean values of these measurements was ~8%. In order to investigate the changes in pH during the sorption experiments, the pH of the liquid phase (LiCl solution) that is in contact with the wood flour was recorded for the experiments conducted at R T using Sw flour. These measurements were conducted with samples of both wood flour that were untreated and, pre impregnated with dilute acid solution, to investigate the effect of acid treatment on the changes in pH during the sorption experiments with LiCl solution. In order to study the effects of wood type and temperature on the sorption of Li+ ion onto wood flour, experiments were carried out with Sw and Hw flour at both R T and 60 oC. The residence time necessary to obtain an apparent equilibrium (the time at which there is no change in the amount of Li+ ion sorbed onto the wood flour) was determined by carrying out experiments at R T for times between 30 minutes and 30 days (0.5, 1, 4, 8, 24, 48, 96, 144, 195, 360 and 720 h). These experiments were conducted for different concentrations of aqueous LiCl solution (0.1, 0.5 and 1 M) and for both Sw and Hw flour.

4

The effect of the concentration of the aqueous LiCl solution on the sorption of Li+ ion onto Sw flour was studied by varying it between 0.1 and 12 M (0.1, 0.2, 0.5, 1, 2, 3, 4, 6, 7.5, 9 and 12 M) at R T and between 0.1 and 6 M (0.1, 0.2, 0.5, 1, 2, 3, 4 and 6 M) at 60 oC. The study was also conducted on samples of Hw flour, when the concentration of the bulk LiCl solution was varied between 0.1 and 6 M at both R T and 60 oC. 2.5. Mercury porosimetry study

All samples of wood flour were freeze-dried at -20 oC for ~15 h prior to the mercury porosimetry experiment; three replicates of samples were used in both Sw and Hw experiments. Mercury intrusion porosimetry was carried out with a Micrometrics AutoPore IV Mercury Porosimeter and analyzed with the software AutoPore IV 9500, Version 1.09. 2.6. FTIR (Fourier Transform Infrared Spectroscopy) study

Samples used in the FTIR study were obtained by extracting the material from the liquid phase after centrifugation (Thermo Fischer Scientific, Heraeus Megafuge 40 R centrifuge, Germany) and drying it for ~15 h over a heated plate at a temperature slightly higher than R T. FTIR spectra were recorded in a Fourier transform infrared spectrophotometer (Perkin Elmer, Spectrum one, FTIR-Spectrometer) using the KBr discs technique. Each spectrum was collected 16 times in the range of 4000-400 cm-1 with a resolution of 4 cm-1 and corrected for background noise. The spectra were normalized using a peak at 1424 cm-1, which corresponds to the CH2 contributed by the cellulose. 2.7. Sorption isotherms study

2.7.1. Langmuir isotherm

The Langmuir isotherm model (Equation 1) is based on the assumption that maximum sorption corresponds to a saturated monolayer of solute molecules on the sorbent surface, with no lateral interaction between the sorbed molecules/ions.

��,� =��,� ��

�� (1)

where qwood,e is the equilibrium sorbed capacity (i.e. the amount of ions sorbed per unit weight of sorbent (wood flour) at equilibrium) (in mg/g) at the corresponding bulk liquid concentration Ce (mol/l), qwood,max is the maximum sorption capacity of the sorbent (mg/g), and b is the Langmuir constant (l/mol). The essential characteristics of the Langmuir isotherm model can be expressed in terms of a dimensionless constant separation factor (RL), which predicts affinity between the sorbate and the sorbent. RL is given by the following equation:

�� =�

�� (2)

where Co is the initial concentration of sorbate (mol/l) and b (l/mol) is the Langmuir constant related to the affinity of the binding sites. The values of RL indicate the type of the isotherm as being either unfavorable (RL>1), linear (RL=1), favorable (0<RL<1) or irreversible (RL=0) [43].

2.7.2. Freundlich isotherm

Originally an empirical equation, the Freundlich isotherm (Equation 3) was later interpreted as sorption to surface sites with a non-uniform distribution of sorption enthalpy and affinities [44]. It can therefore be used to describe multilayer sorption with interaction between the sorbed ions.

��,� = �� (3)

where qwood,e and Ce are as defined above in Section 2.7.1. The Freundlich isotherm constants kF (mg/g (l/mol)1/n) and n (dimensionless) are related to the affinity of the sorbate (metal ion) to the sorbent (wood flour) i.e. sorption capacity and the sorption intensity or surface

5

heterogeneity of the sorbent, respectively. The parameters n and kF are generally temperature dependent.

2.7.3. Estimation of isotherms parameters

Non-linear fitting methods were used in this study; the solver add-in facility of Microsoft Excel was used [45]. This method minimize the sum of the squared errors (SSE) between the observed and calculated values of the dependent variable (in this case the amount of ions sorbed at equilibrium (qwood,e) is the dependent variable). The SSE is defined as:

�� = ∑ !��,�" − ��,�",$%&'

"(� (4)

where N is the number of observations, qei is the ith measured value of the dependent variable and qwood,ei,m is the ith model-predicted value of the dependent variable. The goodness-of-fit measure was determined by the coefficient of determination (R2), which is calculated as:

�& = 1 −∑ !��,�*+��,�*,�%

,-*./

∑ !��,�*+��,� 01%,-

*./ (5)

where qwood,eavg is the mean of the averaged values and all the other variables are as defined above. R2 value of 1 indicates a perfect fit to the data.

3. Results and Discussion 3.1. Mercury porosimetry measurements

Figure 2 shows the incremental intrusion volume and pore size distribution of pores with diameters between 0 and 41 µm, which are considered as being diffusion-accessible pores, for untreated flour samples of Norway spruce Sw and Hw. In softwood [42], pores with a diameter larger than 20 µm may represent the resin canals and early wood tracheids whilst those between 20 and 6 µm represent late wood tracheids. However, there is a relatively large overlap between these two kinds of tracheid, and both of these pore classes are called “macrovoids”. Pore diameters smaller than 6 µm are called “microvoids” and represent the pointed ends of all cell elements and the pit pores [46]. A comparison of the distribution curves of pore size shows small differences between the Sw and Hw flour samples, which may be due to differences in the structural and chemical composition of Sw and Hw: e.g. Hw is usually less permeable than Sw due to its pit aspiration and incrustations [47]. The average total incremental intrusion volumes observed were 0.45±0.0 ml/g and 0.41±0.1 ml/g for Sw and Hw flour samples, respectively. 3.2. FTIR characterization

FTIR characterization of the samples of Sw flour (untreated, acid-treated and after acid treatment followed by LiCl treatment) were carried out to identify which functional groups are affected by the interactions between the Sw flour and the Li+/Cl- ions, as they probably play an important role in the sorption of ions (Figure 3). The relevant functional groups and their tentative band assignments are listed in Table A2 (see Appendix A2), and are in agreement with those reported in the literature for softwoods [48-49]. Comparing the spectra of the original wood with those of the LiCl treated wood, it is clear that there are differences in the absorbance signals originating from the groups containing –OH (e.g. alcohols, phenols and absorbed water), carbonyl (C=O) and C-H functionalities. The differences observed in the frequency bands suggest that these functional groups are affected by the sorption of Li+ ions by Norway spruce Sw flour.

Changes in the stretching vibrations of –OH detected within a broad range of frequencies (3700-3100 cm-1) show the presence of both free and H-bonded hydroxyl groups of neutral carbohydrates, phenols and carboxylic acids and indicates the interactions between Li+ /Cl- ions

6

and –OH groups. The changes in the signals corresponding to the C=O stretch in carbonyls (e.g. acetyl groups) present in hemicelluloses (1738-1709 cm-1) and C-H deformations of carbohydrates (1470-1300 cm-1) were also observed upon LiCl treatment (Figure 3(ii)). Corresponding bands shown a clear decrease in intensity, thus indicating the partial removal of acetyl groups during the treatment. Further, signals originating from the sorbed water (1650-1635 cm-1) were enhanced significantly in the LiCl treated samples, implying an enhanced interaction with water as an effect of the LiCl treatment. From the results described above, it was not possible to attribute observed activity of the functional groups to individual wood constituents. Therefore FTIR spectroscopy was also used to study the behaviour of the model components of cellulose (Micro Crystalline Cellulose (MCC)), hemicelluloses (beechwood xylan) and lignin (Kraft lignin) before and after 3 M LiCl treatment at R T (Figure 4 (A)-(C)). It is evident from Figure 4 that the bands assigned to the –OH groups of cellulose, xylan and lignin show changes upon LiCl-treatment, thus indicating their participation. The assumed partial removal of acetyl groups during aqueous LiCl treatment was supported further by the decrease in intensity of the signals corresponding to C=O and C-H of carbohydrates (Figure 4 (A) & (B)-(ii)). However, it should be kept in mind that the chemical processes used to produce these model components (MCC, xylan and Kraft lignin) have altered the original structure of these components, and thereby may also affect the FTIR observations. 3.3. Sorption experiments

3.3.1. Changes in pH during the sorption experiments

In this study, the sorption experiments were performed in a closed-batch system without adjustment of the pH, and observed that the pH of the liquid phase decreased. It was also found that the drop in pH increased with increasing concentration of the Li+ ion in the bulk solution (Figure 5). Similar results were observed for the samples of wood flour that were not acid treated prior to LiCl treatment. However the observed pH drop was more pronounced for acid treated samples compared with non-acid treated prior to the LiCl treatment (Figure 5). It was estimated that a drop in pH of approximately 1.9 units was due to acid liquor in the wood pores. The remaining drop must therefore be caused by either an exchange of ions between the wood components and the Li+/Cl- ions or other structural changes in the wood that might have occurred upon LiCl treatment, such as deacetylation. Only a small part of the remaining decrease in pH can be explained by carboxylic groups; at these low pH values it is unlikely that phenolic groups are involved. The remaining drop in pH may therefore be due to deacetylation of carbohydrates upon treatment of wood samples with LiCl, as observed above in the FTIR study (Section 3.2).

3.3.2. Effects of residence time and concentration of the bulk LiCl solution on sorption

equilibrium

The experimentally measured amount of Li+ ions that are sorbed per unit weight of (dry) wood flour is the total amount of Li+ ions sorbed by the wood (qtotal in mg/g), i.e. the quantity of Li+

ions sorbed onto solid wood (i.e. onto the cell wall surfaces/into the cell walls) (qwood) and those that are dissolved into the solution in the wood pores (qpores). The measurements are aimed at steady state and the experiments were conducted at high liquor to wood ratio (100:1), so it is reasonable to assume that the concentration of Li+ ions in the liquid in the wood pores is equal to that of the bulk LiCl solution. Using Equations (6) and (7) below, it is thus possible to estimate the amount of Li+ ions sorbed onto the solid wood. All of the results reported in this study were based on the amount of Li+ ions sorbed onto the solid wood (qwood in mg/g). qtotal = qwood + qpores (6)

qpores = CLi+

, bulk * Vpores (7)

7

where CLi+

, bulk and Vpores are the concentrations of Li+ ions in the bulk solution (mg/ml) and the volume of the pores in the wood flour (ml/g), respectively. The volume of the pores in the wood flour samples was determined by the mercury porosimetry technique mentioned above (Section 3.1). Vpores was calculated as the total incremental intrusion volume of the pores in the wood flour, with the pores having a diameter of between 0 and 41 µm, which are considered as being accessible to diffusion. The effects of the residence time and the concentration of the bulk LiCl solution on the amount of Li+ ions sorbed onto the solid wood portion of the Sw flour are shown in Figure 6. The experimental data (qwood) was measured at several contact times, varying between 30 min and 30 days. The changes observed in the values of qwood with residence times were small, being almost within the experimental error: the results for the longer experimental times are not shown here. It appears, from Figure 6, that there is a rapid initial rise of the sorption of Li+ ions, with the steady state being almost reached within 8 h for all of the experiments. Therefore, the experimental data was measured at 24 h, and was used as the amount of ions sorbed at apparent equilibrium (qwood,e) in the following sections. Corresponding results were also observed for Hw flour samples (results not shown here).

3.3.3. Effects of temperature and wood structure on sorption equilibrium

Figure 7 shows qwood,e (mg/g) plotted against the concentration of Li+ ions in the bulk liquid, Ce

(M) at R T and 60 oC for samples of both Sw and Hw flour. Only a weak dependence on temperature can be observed for both samples (Figure 7 (A)). Approximately the same quantities were sorbed for bulk LiCl solutions of concentrations up to 1 M; above 1 M, these quantities increased slightly with temperature, which may be due to changes in the properties of the wood (e.g. a large surface becomes accessible) that, in turn, may result in slight changes in the interactions with Li+ ions. Furthermore, for concentrations above 1 M, the sorbed quantities increased more for Hw than for Sw (Figure 7 (B)): this observation is more pronounced at R T compared with 60 oC (Figure 7(B) - (a) vs. (b)).These differences in quantities of ions sorbed by Sw and Hw may be attributed to differences in the structural and chemical compositions of Sw and Hw: e.g. the extractive content in Hw is higher than in Sw, so it has a broader variety of functional groups available for interaction with Li+ ions [47]. However, the differences are small, in most cases within, or almost within, the estimated experimental errors.

3.4. Sorption isotherms Initially it was investigated whether it was possible to fit the Langmuir sorption isotherm model to the individual sorption isotherms described above (Section 3.3.3). It was, however, found that it was a very strong correlation between the two fitted parameters (qwood,max and b) and no stable results was obtained. Furthermore, as discussed above, there were only small differences between the four different sorption isotherms (Sw at R T; Sw at 60 oC; Hw at R T; and Hw at 60 oC). Therefore it was decided to use all the data from the 4 sorption isotherms to estimate the parameters in the Langmuir as well as Freundlich models. The result is found in Figure 8 and Table 1. In conjunction with Table 1, Figure 8 shows that the Langmuir and Freundlich models fitted the experimental data reasonably good (R2 > 0.9). The estimated parameters are thus valid for both Sw and Hw at the temperature range between R T and 60 oC, and for Li+ ion concentrations range between 0 and 6 M.

3.4.1. Langmuir isotherm

The Langmuir isotherm constants qwood,max and b values for the sorption of Li+ ions in Norway spruce wood flour were found to be approx. 132 mg/g and 0.05 l/mol respectively. The estimated qwood,max value (approx. 132 mg/g or 19 mmol/g) is much higher than the concentration of the carboxylic acid groups (COOH) (0.086±0.007 mmol/g) [50] and the

8

theoretical OH groups (12.7 mmol/g) (see Appendix A3 for the estimation procedure of theoretical [OH] groups) in wood samples of native Norway spruce. This result therefore suggests that, apart from the interactions of carboxyl, and hydroxyl groups as shown in the FTIR study (Section 3.2) and in XPS study (visible as Li-O interactions) in our recent work [51], a non-site-specific retention of ions might also have occurred during the sorption of Li+ ions onto Norway spruce wood flour under the experimental conditions studied. The RL values given in Table 2 indicate that the nature of the sorption isotherm is linear for LiCl solutions of lower concentrations, whereas for higher concentrations it is favourable, for the conditions used in this study. 3.4.2. Freundlich isotherm

The Freundlich isotherm constants kF and 1/n were found to be approx. 6.5 (mg/g) (l/mol)1/n and 0.87 (dimension less), respectively. A value of 1/n below one implies chemisorption processes, whilst a value above one is indicative of co-operative sorption [52]. In this study, the value of 1/n was below one, so it is most likely that the sorption process for the experimental conditions studied includes chemisorption (Table 1).

4. Conclusions The experimental methodology proposed was found to be suitable for determining the concentration of the ions sorbed onto the solid portions of the wood matrix. Furthermore, the results show that wood flour of Norway spruce has a considerable affinity towards Li+ ions. Both the Langmuir and Freundlich sorption isotherm models described the sorption of Li+ ions onto Norway spruce wood flour reasonably good for the conditions investigated in this study. FTIR investigation indicated that hydroxyl and carboxyl functional groups of wood material are affected by the interactions between Li+/Cl- ions and Norway spruce Sw flour.

Acknowledgement

The authors are grateful for the financial support of the Chalmers Energy Initiative (CEI) programme.

References:

[1] Akhtaruzzaman, A.F.M; Virkola, N-E. Influence of chip dimensions in Kraft pulping part I. mechanism of movement of chemicals into chips, Paperi ja Puu. 1979, 9, 578-580. [2] Wadsö, L. Studies of water vapor transport and sorption in wood, Doctoral dissertation, Report TVBM-1013, Building materials, Lund University, 1993. [3] Stamm, A.J. Bound-water diffusion into wood in the fiber direction. Forest Prod. J. 1959, 9(1), 27-32. [4] Stamm, A.J. Bound-water diffusion into wood in the across-the-fiber direction. Forest Prod. J. 1960, 10(10), 524-528. [5] Kumar, S; Jian, V.K. Diffusion through wood. Diffusion of boric acid. Holsforsch. Holzvertwert. 1973, 25(1), 21-24. [6] Yata, S; Mukudai, J; Kajita, H. Morphological studies on the movement of substances into the cell wall of wood. II. Diffusion of copper compounds into the cell wall. Mokuzai Gakkaishi. 1979, 25(3), 171-176. [7] Vinden, P. The effect of raw material variables on preservative treatment of wood by diffusion process. Journal of the Institute of Wood Science 1984, 10(1), 31-41. [8] Cooper, P. Diffusion of copper in wood cell walls following vacuum treatment. Wood Fiber Sci. 1998, 30(4), 382-395. [9] Ra, J.B; Barnes, H.M; Conners, T.E. Determination of boron diffusion coefficients in wood. Wood Fiber Sci. 2001, 33(1), 90-103.

9

[10] Jeremic, D; Quijano-Solis, C; Cooper, P. Diffusion rate of polyethylene glycol into cell walls of red pine following vacuum impregnation. Cellulose 2009, 16, 339-348. [11] Narayanamurti, D; Kumar, V.B. Diffusion of organic molecules through wood. Journal of Polymer Science 1953, 10(6), 515-524. [12] Fukuyama, M; Urakami, H. Diffusion of nonelectrolytes through wood saturated with water III. Mokuzai Gakkaishi. 1986, 32(3), 147-154. [13] Tillman, L.M; Lee, Y.Y; Torget, R. Effect of transient acid diffusion on pretreatment/hydrolysis of hardwood hemicellulose. Applied Biochemistry and Biotechnology 1990, 24(25), 103-113. [14] Meijer, M; van der Zwan, R.P; Militz, H. Unsteady-state diffusion of methanol in Douglas-fir heartwood at high temperatures. Holzforschung 1996, 50(2), 135-143. [15] Narayanamurti, D; Ratra, R.S. Diffusion of ions through some Indian timbers. Proceedings Mathematical Sciences 1951, 33(6), 349-359. [16] Christensen, G.N; Williams, E.J. Diffusion in wood I. A quantitative theory of diffusion in porous media and its application to wood. Austral. J. Appl. 1951a, 2(4), 411-429. [17] Christensen, G.N. Diffusion in wood III. Ion selection and its effect on the diffusion of electrolytes. Austral. J. Appl. 1951b, 2(4), 440-453. [18] Saltberg, A; Brelid, H; Theliander, H. Removal of metal ions from wood chips during acidic leaching 2: Modeling leaching of calcium ions from softwood chips. Nordic Pulp and Paper Research Journal 2006, 21(4), 513-519. [19] Mc Kibbins, S.W. Application of diffusion theory to the washing of Kraft cooked wood chips. Tappi 1960, 43(10), 801-805. [20] Talton, J.H; Cornell, R.H. Diffusion of sodium hydroxide in wood at high pH as a function of temperature and degree of pulping. Tappi 1987, 70(3), 115-118. [21] Robertsen, L. Diffusion in wood. PhD Thesis, Faculty of Chemical Engineering, Åbo Academy, Åbo, Finland, 1993. [22] Kazi, K.M.F; Chornet, E. A diffusion model for the chemical impregnation of hardwoods and its significance for rapid steam treatments. Paperi ja Puu. 1998, 80(1), 41-48. [23] Törnqvist, M; Hurme, T; Rosenholm, J.B. Drift speed: A way of measuring diffusion and tortuosity of porous materials. Colloids and Surfaces A: Physicochemical and Engineering Aspects 2001, 180(1-2), 23-31. [24] Ekman, K.H; Fogelberg, B.C. Impregnation of sulphate cooking studied by means of radioactive sulphur. Paperi ja Puu. 1966, 48(4), 175. [25] Kolavali, R; Theliander, H. Experimental determination of the diffusion of monovalent cations into wood: effects of micro-cracks, wood structure, impregnation time, and temperature on local concentration profiles. J-FOR; Journal of Science & Technology for forest products and processes 2014, 4(2), 29-35. [26] Ho, Y.S; Mckay, G. Kinetic models for the sorption of dye from aqueous solution by wood. Trans I Chem E. 1998, 76 (2), 183-191. [27] Shukla, A; Zhang, Y-H; Dubey, P; Margrave, J.L; Shukla, S. The role of sawdust in the removal of unwanted materials from water. Journal of Hazardous Materials 2002, 95(1-2), 137-152. [28] Lister, S.K.; Line, M.A. Potential utilization of sewage sludge and paper mill waste for biosorption of metals from polluted waterways. Bioresource Technology 2001, 79(1), 35-39. [29] Özacar, M; Sengil, I.A. Adsorption of metal complex dyes from aqueous solutions by pine sawdust. Bioresource Technology 2005, 96(7), 791-795. [30] Özacar, M; Sengil, I.A. A Kinetic study of metal complex dye sorption onto pine sawdust. Process Biochemistry 2005, 40(2), 565-572. [31] Sciban, M; Klasnja, M; Skrbic, B. Modified softwood sawdust as adsorbent of heavy metal ions from water. Journal of Hazardous Materials 2006, 136(2), 266-271.

10

[32] Batzias, F.A; Sidiras, D.K. Dye adsorption by prehydrolysed beech sawdust in batch and fixed-bed systems. Bioresource Technology 2007, 98(6), 208-1217. [33] Ferrero, F. Dye removal by low cost adsorbents: Hazelnut shells in comparison with wood sawdust. Journal of Hazardous Materials 2007, 142 (1-2), 144-152. [34] Argun, M.E; Dursun, S; Ozdemir, C; Karatas, M. Heavy metal adsorption by modified oak sawdust: Thermodynamics and kinetics. Journal of Hazardous Materials 2007, 141(1), 77-85. [35] Ahluwalia, S.S; Goyal, D. Microbial and plant derived biomass for removal of heavy metals from waste water. Bioresource Technology 2007, 98(12), 2243-2257. [36] Ofomaja, A.E. Kinetic study and sorption mechanism of methylene blue and methyl violet onto mansonia (Mansonia altissima) wood sawdust. Chemical Engineering Journal 2008, 143(1-3), 85-95. [37] Ofomaja, A.E; Ho, Y-S. Effect of temperature and pH on methyl violet biosorption by Mansonia wood sawdust. Bioresource Technology 2008, 99 (1-3), 5411-5417. [38] Dulman, V; Cucu-Man, S.M. Sorption of some textile dyes by beech wood sawdust. Journal of Hazardous Materials 2009, 162 (2-3), 1457-1464. [39] Duong, T.D; Nguyen, K.L; Hoang, M. Competitive sorption of Na+ and Ca2+ ions on unbleached kraft fibers- A kinetics and equilibrium study. Journal of Colloid and Interface Science 2006, 301(2), 446-451. [40] Hubbe, M.E; Hasan, S.H; Ducoste, J.J. Cellulosic substrates for removal of pollutants from aqueous systems: a review.1. Metals. Bioresources 2011, 6(2), 2161-2287. [41] Plötze, M; Niemz, P. Porosity and pore size distribution of different wood types as determined by mercury intrusion porosimetry. Eur. J. Wood Prod. 2011, 69 (4), 649-657. [42] Zauer, M; Hempel, S; Pfriem, A; Mechtcherine, V; Wagenfuhr, A. Investigations of the pore-size distribution of wood in the dry and wet state by means of mercury intrusion porosimetry. Wood Sci. Technol. 2014, 48 (6), 1229-1240. [43] Tan, I.A.W; Hameed, B.H; Ahmad, A.L. Equilibrium and kinetic studies on basic dye adsorption by oil palm fibre activated carbon. Chemical Engineering Journal 2007, 127 (1-3), 111-119. [44] Yantasee, W. Kinetic and equilibrium analysis of metal ion adsorption onto bleached and unbleached kraft pulps, Doctoral dissertation, Chemical engineering, Oregon State University, USA, 2001. [45] Bolster, C.H; Horberger, G.M. On the use of linearized Langmuir equations. Soil Sci. Soc. Am. J. 2007, 71(6), 1796-1806. [46] Thygesen, L.G; Engelund, E.T; Hoffmeyer, P. Water sorption in wood and modified wood at high values of relative humidity. Part I: results for untreated, acetylated, and furfurylated Norway spruce. Holzforschung 2010, 64(3), 315-323. [47] Siau, J.F. Transport processes in wood, Spinger-Verlag Berlin Heidelberg, 1984. [48] Schwanninger, M; Rodrigues, J.C; Pereira, H; Hinterstoisser, B. Effect of short-time vibratory ball milling on the shape of FT-IR spectra of wood and cellulose. Vibrational Spectroscopy 2004, 36(1), 23-40. [49] Pandey, K.K; Pitman, A.J. FTIR studies of the changes in wood chemistry following decay by brown-rot and white-rot fungi. International Biodeterioration & Biodegradation 2003, 52(3) 151-160. [50] Werkelin, J; Skrifvars, b-J; Zevenhoven, M; Holmbom, B; Huppa, M. Chemical forms of ash-forming elements in woody biomass fuels. Fuel 2010, 89(2), 481-493. [51] Kolavali, R; Hasani, M. The sorption of monovalent cations onto Norway spruce wood flour: molecular interactions behind the LiCl impregnation. (Manuscript submitted to Holzforschung 2016). [52] Foo, K.Y; Hameed, B.H. Insights into the modeling of adsorption isotherm systems, Chemical Engineering Journal. 2010, 156(1), 2-10.

11



Reddysuresh Kolavalia, Merima Hasania, and Hans Theliandera* a Division of Forest Products and Chemical Engineering, Dept. of Chemistry & Chemical Engineering, Chalmers University of Technology, SE-41296 Gothenburg, Sweden. *Corresponding author: Professor Hans Theliander Tel: +46 31 772 2992 Fax: +46 31 772 2995 E-mail: [email protected]

Table 1

Characteristic parameters obtained through a non-linear method for the sorption of Li+ ions onto Norway spruce wood flour.

Isotherm Characteristic parameters Langmuir

qwood, max, mg/g 132

b, l/mol 0.05 R2 0.93 Freundlich kF,

(mg/g)(l/mol)1/n

6.5

1/n 0.87 R2 0.92

Table 2

Langmuir isotherm dimensionless constant separation factor (RL) values for the sorption of Li+ ions onto Norway spruce wood flour.

Bulk LiCl Concentration (M) RL 0.1 0.99 0.2 0.99 0.5 0.98 1 0.95 2 0.91 3 0.87 4 0.83 6 0.77

12

Figure 1. Schematic diagram of the ion exchange and chemical complexation likely to occur during the sorption of metal ions onto wood material. (Mn+ M: cation, n: valance)

Figure 2. Pore size distribution of 3 replicates (Samples 1-3) of native Norway spruce wood flour: (a) Sw (left) and (b) Hw (right) samples had pore sizes with a diameter of 0 - 41 µm.

13

Figure 3. FTIR spectra of the Sw flour samples: untreated (native), dilute acid treated and 3 M LiCl treated after dilute acid treatment at room temperature: (i) Overall spectra and (ii) magnified spectra between 1800 and 1300 cm-1.

14

Figure 4. FTIR spectra of samples of (A) MCC, (B) beechwood xylan, (C) Kraft lignin untreated (native), dilute-acid treated and 3 M LiCl treated after acid treatment at room temperature (as indicated). (i) Overall spectra and (ii) magnified spectra 1800-1300 cm-1.

15

Figure 5. Decrease in pH versus concentration of Li+ ions in bulk LiCl aqueous solution during the sorption experiments using Sw flour at room temperature.

Figure 6. Variations in the amount of Li+ ions sorbed onto the solid wood portions (qwood) of Sapwood (Sw) flour with time: in experiments with various concentrations of aqueous LiCl solution (as indicated) at room temperature.

16

Figure 7. Amount of Li+ ions sorbed onto Norway spruce wood flour at the apparent equilibrium (qe) with varying concentrations of bulk LiCl solution. (A) Effect of temperature (room temperature (R T) vs. 60 oC) and (B) Effect of wood type (Sw vs. Hw).

17

Figure 8. Equilibrium sorption isotherms of Li+ ions onto wood flour of Norway spruce at different temperatures (room temperature (R T) and 60 oC) for both Sw as well as Hw samples (as indicated).

18




Appendix

A1. Concentration of ash and major elements

Table A1

Concentration* of ash and metal ions in samples of Sw and Hw before and after treatment with dilute sulphuric acid solution (pH=0.5/0.31 M H2SO4) for ~18 h at room temperature (R T).

Sapwood (Sw) Heartwood (Hw) Before acid

treatment After acid treatment

Before acid treatment

After acid treatment

On dry matter

Ash, weight-% 0.3 <0.1 0.3 <0.1 Calcium ion (Ca2+), weight-% 0.060 <0.002 0.077 <0.002 Magnesium ion (Mg2+), weight-% 0.0087 <0.0004 0.0096 <0.0004 Manganese ion (Mn2+), weight-% 0.0038 <0.0001 0.0041 <0.0001 Sodium ion (Na+), weight-% 0.0021 <0.001 0.0016 <0.001 Potassium ion (K+), weight-% 0.053 <0.001 0.022 <0.001

*Concentration of ash and major elements (Mn, Mg, Ca, Na, and K) estimated using methods SS-EN ISO 18122 and SS-EN ISO 16967 respectively, by SP Technical Research Institute of Sweden.

19

A2. FTIR absorption bands and their assignment

Table A2

FTIR absorption bands from spectra obtained for samples shown in Section 3.5 and their

assignment

Range of Wave Number (cm-1) Functional Group/Band Assignment

3700-3100 O-H stretching

2960-2850 C-H stretching vibration in methyl and methylene

groups

1738-1709 C=O stretching vibration in unconjugated ketens and in

free aldehyde present in lignin and hemicelluloses

1650-1635 Sorbed water

1510-1500 C=C stretching vibration in the aromatic structure of

lignin

1460-1470 C-H deformations; asymmetric bending vibration of

–CH3 and –CH2 groups from lignin

1430-1416 CH2 bending vibrations related to the structure of

cellulose; aromatic skeletal vibrations

1375-1374 C-H deformation vibration in cellulose and

hemicellulose; O-H bending vibrations in phenols

(lignin)

1317-1315 C-H vibration in cellulose and C1-O vibration in syringyl

derivatives

1270-1266 Guiacyl ring breathing, C-O stretch in lignin and for

C-O linkage in guiacyl aromatic methoxyl groups

1162-1125 C-O-C group asymmetric bridge stretching vibration in

cellulose and hemicellulose

1060-1015 C-O stretch in cellulose and hemicellulose

925-892 Glucose ring stretching, C1-H deformation;

C-H stretching out of plane of aromatic ring

20

A3. Estimation of the number of theoretical OH groups in the flour sample of Norway

spruce Sw

The number of theoretical OH groups in moles per gram (mol/g) can be estimated using Eq. A1, proposed by Rowell [1] and Hill [2]:

23

�4&+ 26

�4&+ &�

�2&+ 7

�89 (A1)

where A is cellulose, B hemicelluloses hexosan, C pentosan and D lignin. The compositional analyses of Norway spruce Sw flour were performed using the procedure described by Jedvert et al., 2012 [3] and found to be: 36.9% cellulose (A); 17.2% hemicelluloses hexosan (B) and 5.8% pentosan (C) and 33% lignin (both acid soluble and insoluble). Based on this composition, the estimated number of theoretical OH groups were found to be 12.72 mmol/g. As hydroxyl groups made inaccessible through binding between wood polymers or steric hindrance are not considered, this estimate expresses a theoretical upper limit. Also, two-thirds of the hydroxyl groups of cellulose are known to be inaccessible to water due to the crystalline portion of cellulose. Moreover, the history of the sample (e.g. drying) also influences the accessibility of the hydroxyl groups in the sample.

[1] R. M. Rowell, Wood Sci. (1980) 13:102.

[2] C. A. S. Hill, Wood modification- chemical, thermal and other processes. Wiley, Chichester. (2006).

[3] K. Jedvert, A. Saltberg, M.E. Lindström and H. Theliander, Mild steam explosion and chemical pre-treatment of Norway spruce. Bioresources. 7(2) (2012) 2051-2074.

Paper IV

The sorption of monovalent cations onto Norway spruce wood flour: molecular

interactions behind the LiCl impregnation

Reddysuresh Kolavali and Merima Hasani

Submitted for publication

For Review O

nly

The sorption of monovalent cations onto Norway spruce

wood flour based on LiCl impregnation

Journal: Holzforschung - International Journal of the Biology, Chemistry, Physics and Technology of Wood

Manuscript ID HOLZ.2016.0075

Manuscript Type: Original Article

Date Submitted by the Author: 26-Apr-2016

Complete List of Authors: Kolavali, Reddy; Chalmers University of Technology, Chemistry and Chemical Engineering Hasani, Merima; Chalmers University of Technology, Chemistry and Chemical Engineering

Section/Category: Chemistry

Keywords: diffusion, impregnation, mass transport, monovalent cation, sorption, wood, XPS, FTIR, AAS, BET

https://mc.manuscriptcentral.com/holz

Holzforschung - International Journal of the Biology, Chemistry, Physics and Technology of Wood

For Review O

nly

1

The sorption of monovalent cations onto Norway

spruce wood flour: molecular interactions behind the LiCl

impregnation

Reddysuresh Kolavali & Merima Hasani*

*Corresponding author: Merima Hasani, Assistant Professor

Division of Forest Products and Chemical Engineering, Dept. of Chemistry & Chemical

Engineering, Chalmers University of Technology, SE-41296 Gothenburg, Sweden.

Tel: +46 31 772 3065 Fax: +46 31 772 2995 E-mail: [email protected]

Reddysuresh Kolavali: Division of Forest Products and Chemical Engineering, Dept. of

Chemistry & Chemical Engineering, Chalmers University of Technology, SE-41296

Gothenburg, Sweden.

ABSTRACT

Active functional groups and interactions involved in the sorption of Li+ ions from an aqueous

LiCl solution onto the Norway spruce sapwood (Sw) flour were investigated and identified.

Sw flour fractions with varying composition (holocelluloses with varying contents of lignin

and lignin fraction containing lignin carbohydrate complexes) were prepared and treated with

aqueous (aq) LiCl solution, and in terms of sorption investigated by means of XPS, FTIR and

AAS techniques. The findings indicated that Li+/Cl- were retained on LiCl treated samples

and shown interactions with the hydroxyl, carboxylic and carbonyl groups of wood material.

Upon LiCl treatment the mobility and accessibility of the wood matrix was enhanced,

possibly by disruption of existing hydrogen bonding. The carbohydrate-rich fractions retained

higher amounts of LiCl and shown more site-specific interactions.

Keywords: diffusion, impregnation, mass transport, monovalent cation, sorption, wood, XPS.

Page 1 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

2

1. Introduction

Mass transport of chemicals in wood is often of decisive importance in wood processing, but

is at the same time a very complex process possibly involving several phenomena within the

heterogeneous hygroscopic wood tissue: such as diffusive mass transport via the liquid

available in the lumen of the wood cell or within the cell walls, surface diffusion and sorption

processes within the wood tissue. The overall mass transport rate is thus usually influenced

not only by diffusive mass transport but also by sorption. Understanding the sorption process

in detail, in terms of kinetics, sorption equilibria and molecular interactions, is therefore, an

important part of understanding the overall process of mass transport of chemicals in wood.

Many investigations acknowledged the sorption process, using the wood material as the

sorbent, as one of the most effective methods of removing contaminants (sorbate), such as

dyes and heavy metals from the wastewater streams (Ahluwalia and Goyal 2007; Argun et al.

2007; Batzias and Sidiras 2007; Dulman and Cucu-Man 2009; Ferrero 2007; Ho and McKay

1998; Ofomaja 2008; Ofomaja and Ho 2008; Özacar and Şengil 2005 (a) & (b); Shukla et al.

2002; Šćiban 2006). In the pulp and paper related research, the sorption of metal ions in wood

cell walls was generally attributed to the presence of ionisable functional groups associated

with the components of the cell wall (Duong et al. 2006). Studies of the heavy metal sorption

on sawdust speculated that ion exchange and hydrogen bonding may be the principal

interaction mechanisms involved, while formation of surface complexes based on site

specific interactions between specific metal ion and functional groups in wood may also occur

(Shukla et al. 2002).

However, most of the investigations focused on sorption equilibrium and kinetic studies,

leaving a scarce knowledge about the molecular interactions between the individual

constituents of the wood and the metal ions.

Page 2 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

3

In our previous study (Kolavali et al. 2014), the shape of the observed concentration profiles

of Li+ ions in Norway spruce wood indicated that their transport was not only as expected in a

diffusion controlled process, but also possibly influenced by other phenomena such as surface

diffusion and sorption. Moreover, for the experimental conditions of neutral/low pH, the

concentrations of Li+ ions on the outer surface layers of the wood were found to be higher

than the content of the native anionic groups (i.e. the cationic exchange capacity) in the wood.

Thus, thorough investigation of physical and chemical interactions between simple ions (such

as Li+) and wood are needed to improve understanding and predicting the types of interactions

that occur during the chemical treatment of wood material, including information about

participation of the individual wood components and their functional groups.

The objective of this study was to investigate interactions between aqueous LiCl solution and

wood in terms of surface interactions and functional groups participation. For that purpose

Norway spruce sapwood flour (here denoted as Sw flour) was separated into different

fractions: holocellulose fractions containing different amount of lignin, and a fraction

containing primarily lignin and lignin-carbohydrate complexes (here denoted as lignin/LCC

fraction). These were, together with Sw flour, treated with aqueous LiCl and investigated

with FTIR, XPS and AAS techniques, in order to study changes in surface structure, amount

of retained Li+/Cl- and behavior of functional groups as an effect of the LiCl treatment. This

study employed two novel methods to fractionate the Sw material: one using peracetic acid

(PAA) delignification, leaving holocellulose with varying contents of residual lignin

determined by reaction time (Kumar et al. 2013); and another one based on oxidative

degradation of carbohydrates (using aqueous sodium periodate) (Alam and Van De Ven

2014), leaving the lignin and hemicellulose covalently attached to it as a solid product.

Although isolated wood components do not behave the same as in the wood matrix, studying

their interactions with metal ions is relevant to the understanding of the physicochemical

Page 3 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

4

parameters determining the interactions between wood material and metal ions and the

relative importance of individual components in these interactions.

2. Materials and Methods

2.1. Materials and chemicals

The wood materials studied were wood flour (Sw) and, its constituents, i.e. holocelluloses

with varying contents of lignin and lignin/LCC (Lignin Carbohydrate Complex) fractions.

Stock solutions of aqueous LiCl were prepared by dissolving LiCl salt (≥99%) (Merck KGaA,

Darmstadt, Germany) in deionized water. 72wt.% sulphuric acid solution was prepared by

dissolving reagent grade (95-97%) sulfuric acid (Scharlau, Scharlab S.L., Spain) in deionized

water. 2wt.% nitric acid solution was prepared by dissolving trace analysis grade 69wt.%

nitric acid (ARISTAR, VWR, PROLABO Leuven) in deionized water. Peracetic acid (PAA)

(39wt.%), sodium (meta) periodate (≥99.8%), acetone (≥ 99.8%), cyclohexane (≥ 99.5%),

Ethylene glycol (≥99%) and potassium bromide (KBr) (FTIR grade; ≥ 99 %) were purchased

from Sigma-Aldrich Chemie, Steinheim, Germany and used as received. Sodium chloride

(min. 99.5%) were purchased from Scharlau, Sentmenat, Spain, and used as received. NaOH

(ACS, Reag, Ph Eur) and NaAc (Solution 50%) purchased from Merck, Darmstadt, Germany

and used as received.

2.2. Preparation of holocellulose fractions with varying contents of lignin

Holocelluloses fraction with varying lignin content was prepared through delignification of

the Norway spruce Sw flour using PAA. The procedure has been reported in detail by Kumar

et al. (2013) and was performed at room temperature on 5wt.% Sw flour loadings and PAA

loadings of 5.5g g-1 dry Sw flour in Erlenmeyer flaks where the reaction mixture was stirred

for specified reaction times (6, 24, 51, and 72h). After the finished procedure, the slurry was

vacuum filtered, and the solids were washed repeatedly with room temperature deionized

water until the filtrate pH was neutral. Solids were collected in ziploc PE bags. For mass

Page 4 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

5

balances and yield calculations, the dry content of the solids in triplicates was analyzed using

dry content analyzer (Sartorius MA 30, Sartorius GMBH Göttingen, Germany).

2.3. Preparation of the lignin/LCC fraction

The lignin/LCC fraction was isolated from the Norway spruce Sw flour through an aqueous

sodium periodate oxidation of the wood, employing a procedure reported in detail by Alam

and Van De Ven (2014). The oxidation was carried out on Sw flour (10g) in 500ml of

deionized water in a glass beaker equipped with an overhead stirrer using sodium

metaperioadate (13.2g; 61.66mmol), and sodium chloride (29.25g; 1N in the overall

solution). The reaction mixture was stirred at room temperature in the dark for 144h. At the

end of the reaction, ethylene glycol (~ 4ml) was added to the reaction mixture to quench the

residual periodate. The solid fraction was washed repeatedly with deionized water and

afterwards suspended in water (2g solid in 100ml water) and stirred at 80-90oC in an oil bath

for 6h. The sample was then cooled to room temperature and the non-dissolved, brown, solid

material was separated by vacuum filtration. The filtrate (comprising soluble cellulose and

hemicelluloses) was collected and dried at 105oC to determine the extent of recovery. The

non-dissolved brown solid (lignin/LCC fraction) collected was analyzed for structural

carbohydrates and acid soluble and insoluble (Klason) lignin as described in the following

section.

2.4. Compositional analysis

The untreated and delignified solids were dried in an oven at 50oC for several days. Analyses

for determining their contents of structural carbohydrates and lignin were then performed as

follows:

Page 5 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

6

2.4.1. Klason lignin

The method used was based on the procedure presented by Theander and Westerlund 1986,

whereby Klason lignin is measured gravimetrically after complete acid hydrolysis with 72%

(12 M) sulphuric acid (H2SO4). In the present study, a 200mg oven-dried sample was treated

with 3mL of 72% H2SO4. The sample was then evacuated for 15min and placed in a water

bath at 30oC for 1h. Then 84g of deionized water was added and it was heated to 125oC in an

autoclave for 1h. Afterwards the sample was filtered off and the solid residue collected

(Klason lignin). The filtrate from the hydrolysis process was used to measure the content of

acid-soluble lignin and for carbohydrate analysis.

2.4.2. Acid-soluble lignin

The content of acid-soluble lignin was calculated in relation to the absorbance value measured

with UV at a wavelength of 205nm in a Specord 205, Analytik Jena. It was calculated

assuming an absorptivity constant of 110 dm3 g-1 cm-1 (Lin and Dence 1992).

2.4.3. Carbohydrate analysis

The filtrate from the acid hydrolysis was firstly diluted to 100ml in a volumetric flask and

then further diluted 5 times and filtered through a 0.45µm PVDF filter prior to measurement.

Fucose was used as the internal standard. The analysis of monomeric sugars was performed

using the Dionex ICS-5000 HPLC system equipped with CarboPac PA1 columns and an

electrochemical detector and run using NaOH and NaOH/NaAc (0.2M) as the eluents. The

software used was Chromeleon 7, Chromatography Data System, Version 7.1.0.898. The

amounts of cellulose, (galacto) glucomannan and xylan were calculated from the carbohydrate

analysis using the assumptions and corrections described by Jedvert et al. (2012).

Page 6 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

7

2.5. Sorption experiments

Studied wood materials (i.e. Sw flour, holocellulose fractions and lignin/LCC) were added to

polypropylene beakers containing a Li+ ion solution with a liquor-to-solid wood material ratio

of 100:1. Each experiment was carried out for a contact time of 24h, which is more than

sufficient to reach equilibrium (Kolavali et al. 2015). The mixture was shaken sporadically for

24h, after which it was centrifuged (SERIE 130; F.I.E.M.M.E; Establishments Couprie; 7,

Quai Clande Bernard; Lyon. approx. 2960 rpm) for 5min to separate the solid wood material

from the Li+ ion solution. The treated material was transferred to an airtight PE bag and

placed in liquid nitrogen (-180oC); the frozen wood flour was then lyophilized (instrument:

Labconco, Kanas City, MO, USA) for 1 week. The freeze-dried solid material was then used

in FTIR, XPS and, BET (surface area) characterization studies. For AAS studies, in order to

measure the amount of Li+ ions sorbed on solid material, the freeze-dried solid biomass was

acid leached at room temperature with 2wt.% nitric acid solution for ~24h. Each leaching

experiment was conducted in triplicate under identical conditions, and mathematic mean

values were used in the calculations. At the end of the leaching period, the leaching liquor

was collected with a syringe connected to a 0.45µm polyvinylidene fluoride membrane filter

(Pall Life Sciences, Ann Arbor, MI, USA). It was then analyzed for its total concentration of

Li+ ions (i.e. Li+ ions sorbed onto the solid (cell wall) biomass, and dissolved in the solution

located in the pores) using Flame Atomic Emission Spectroscopy (FAES) (iCE 300 series,

AA spectrometer, Thermo Scientific, Cambridge, UK). Air-acetylene was the source of the

flame and the emission was measured at 670.8nm. The concentration of the Li+ ion was

adjusted to an optimal working concentration range of 0.02-5µg ml-1. The average deviation

from the mean values of these measurements was ~8%.

Page 7 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

8

2.6. XPS characterization

X-ray photoelectron spectra were collected using a Quantum 2000 scanning ESCA (Electron

Spectroscopy for Chemical Analysis) microprobe from Physical Electronics with a

monochromatic Al Kα source (1486.6eV). The samples were analyzed at a take-off angle of

45o, with the area analyzed being about 400 x 500µm in size and 4-5nm in depth. The

pressure in the analysis chamber during XPS analysis was lower than 10-9torr, making it an

ultra-high vacuum regime. An overall survey spectrum (i.e. low-resolution) from 0 to 1100eV

binding energy and a high resolution spectrum of the C1s region from 280 to 300eV and O1s

region from 525 to 535eV were collected. Analyses of the chemical bonds of carbon and

oxygen were made by fitting the curves of the C1s and O1s peaks from the high resolution

spectra and deconvoluting them into sub-peaks using ESCA tools (MultiPak v 6.1).

2.7. FTIR characterization

FTIR spectra were recorded in a Fourier transform infrared spectrophotometer (Perkin Elmer,

Spectrum one, FTIR-Spectrometer) using the KBr discs technique. Each spectrum was

collected 16 times in the range of 4000-400cm-1 with a resolution of 4cm-1 and corrected for

background noise. The spectra were normalized using a peak at 1424cm-1, which corresponds

to the CH2 contributed by the cellulose.

2.8. BET measurements of surface area

The specific surface area of untreated samples and those treated with LiCl was estimated

according to the BET theory (Brunauer–Emmett–Teller) from measurements of nitrogen

adsorption using a Micrometrics TriStar 3000 instrument. Hornification whilst the samples

are being dried is avoided by performing a solvent exchange and replacing water with

cyclohexane. This was done through a solvent exchange procedure in which the water was

Page 8 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

9

replaced initially with dry acetone which, in turn, was replaced with cyclohexane (Palme et al.

2014). Following this solvent exchange procedure, the samples were dried overnight in a

stream of nitrogen.

3. Results and Discussions

The composition of the studied substrates, Sw flour and its fractions isolated by PAA

delignification and periodate oxidation, are shown in Table 1. As shown, in delignified Sw

fractions between 20 and 93% of lignin content was removed. Isolated lignin/LCC fraction

contained ~8.5wt.% of monosaccharides bound to lignin.

3.1. XPS

The XPS survey spectra of the Sw samples before and after treatment with aqueous LiCl

solution reveal the presence of Li and Cl in the samples treated with LiCl, thus indicating a

sorption of these ions on their surfaces (Figure 1). The surface abundance of the detected

elements, estimated using the atomic sensitivity factor and the area under each XPS-peak, are

presented in Table 2. Small amounts of silicon, calcium and nitrogen (possibly arising from

inorganic compounds) were also detected: they were so small that their contribution is

considered as being negligible. The surface composition of lignocellulosic materials may also

be analyzed in terms of atomic ratio of oxygen to carbon (O/C). In this study, variations in

this ratio upon LiCl treatment were investigated as indicative of introduced changes. In

general, the theoretical value of the O/C ratio for carbohydrates is higher (0.83 for cellulose

and 0.8 for hemicelluloses) than the corresponding value for lignin (0.33) (Sernek 2002),

which reflects the abundance of the hydroxyl groups present in carbohydrates.

In this case, although the O/C values measured were significantly lower than those estimated

from the chemical composition and elemental analysis, they are in good agreement with

values reported previously for wood material by Sinn et al. 2001 and Inari et al. 2006. Whilst

Page 9 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

10

this difference reflects the restriction of the XPS analysis to the surface layer, it may also be

an effect of the adaptation of cellulose surfaces which, upon drying, expose more hydrophobic

structural motifs, as reported previously (Johansson et al. 2012). It was also speculated that

the high vacuum released during XPS analysis, which allows for vaporization of the residual

water contained in the material, can contribute to the low O/C values observed (Inari et al.

2006).

A slight decrease in the O/C ratio was observed for Sw flour upon LiCl treatment, as shown in

Table 2. As mentioned above, this result is likely to be associated with the rearrangement of

cellulose fibrils during drying, leading to the reduced accessibility of –OH groups, and a more

pronounced hydrophobic surface. The indication here is that this rearrangement seems to be

promoted by LiCl treatment.

Deconvolution of the carbon (C1s) and oxygen (O1s) signals observed could provide further

insight into these changes. The C1s signal was deconvoluted into four sub-peaks

corresponding to (i) C1- carbon atoms bonded only to hydrogen or other carbon atoms (C-C

or C-H), appearing at a binding energy (BE) of 283.4±0.11eV, arising predominately from the

lignin and extractive constituents of wood (Nzokou and Pascal Kamdem 2005), (ii) C2 -

carbon atoms bonded to a single, non-carbonyl, oxygen atom in addition to other carbon and

hydrogen atoms (C-O or C-OH), at BE of 285.0±0.13eV, arising from both the carbohydrate

and lignin constituents (Sernek, 2002), (iii) C3 - carbon atoms bonded to one carbonyl

(mostly lignin) or two non-carbonyl oxygen atoms of polysaccharides (C=O or O-C-O), with

BE of 286.5±0.18eV, and (iv) C4 - carbon bonded to one carbonyl and one non-carbonyl

oxygen (O-C=O) at BE of 287.7±0.17eV (Dorris and Gray 1978). O1s signals were

deconvoluted into two sub-peaks: O1 (at BE of 531.3±0.13eV) corresponding to oxygen

atoms bonded to carbon with a double bond (O-C=O), and the O2 (at BE of 532.4±0.11eV)

arising from oxygen atoms bonded to carbon with a single bond (C-O) (Inari et al. 2006;

Page 10 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

11

Nzokou and Pascal Kamdem 2005). Figure 2 shows the deconvoluted, high resolution spectra

of the C1s and O1s peaks of the Sw samples before and after treatment with LiCl. Peak area

(%), distribution of the different bonds in C1s and O1s spectra of the samples (before and

after LiCl treatment) are reported in Table S1 and Table S2 respectively, in Appendix.

It is evident from Figure 2 (A)-(a) that the changes in the C1s signal observed upon LiCl

treatment are associated with a slight increase in the contribution from the C1 (C-C/C-H)

component as well as a decrease in that from the C2 (C-O/C-OH) and C3 (C=O/O-C-O)

components. This indicates, once again, an enrichment of the more hydrophobic C-C/C-H

structures on the surface that are a result of the rearrangement of the wood matrix discussed

above. Corresponding analyses of the O1s signals confirmed the same trend: the contribution

of the O2 component (C-O structures) decreased, indicating the withdrawal of C-O structures

from the surface and thus giving rise to a simultaneous relative increase in the O1 (C=O)

signal. More importantly, a new signal (O0) at 530.1±0.01eV was observed (Figure 2 (B)-

(b)). This signal, which is generally assigned to metal oxides (Chastain and King 1992), can

be attributed to interactions between the oxygen containing functional groups of the wood

material and the Li+ ions. Harilal et al. (2009) assigned this signal to lithium oxides and

peroxides (Li2O, Li2O2) and other species containing lithium and oxygen (Li2CO3, LiOH).

In order to explore these observations further, Sw samples were to different extents

fractionated into its constituents and the resulting fractions of different composition were

individually treated with LiCl and analyzed by means of XPS. Holocelluloses with the highest

lignin (HC6) and lowest lignin (HC72) contents, and lignin/LCC samples were investigated;

their composition are reported in Table 1. When 1M concentrations of LiCl and shorter

treatment times (3h) were used (as in the case of Sw flour), the detection and analysis of the

elements of interest were restricted by the atomic sensitivity factors. These fractions (i.e.

Page 11 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

12

HC6, HC72 and lignin/LCC) were therefore treated with 3M LiCl solution for 24h at room

temperature to make sure that the effects of the treatment were easily detectable.

The XPS data of the delignified samples (HC6 and HC72) reflects a decrease in the content of

lignin (i.e. less carbon) and an increase in the matrix mobility compared to the original Sw

sample. The elemental composition of the surface and the corresponding values of the O/C

ratios from the XPS survey spectra of the samples studied are presented in Table 3.The same

decreasing trend observed for the Sw samples was observed in the values of the O/C ratios

upon LiCl treatment, which indicates the rearrangement of the cellulose fibrillary structures

(as mentioned above). It is evident that this rearrangement is likely to be facilitated by the

LiCl treatment step: it could probably be explained by the fact that LiCl disrupts the

stabilizing H-bonding, which promotes mobility within the fibre matrix. Enhanced mobility

and accessibility in the more delignified fraction (HC72) makes this rearrangement even more

pronounced and results in the O/C ratio being decreased further.

Analysis of the deconvoluted high resolution spectra confirms these findings (Figures 3)

further. In line with the observations above, the deconvoluted C1s spectra of HC6 and HC72

show an accumulation of C1 class carbon on the surface, with a simultaneous decrease in the

C2 contribution upon LiCl treatment: the effect is much more pronounced for the more

delignified samples (HC72) with enhanced matrix mobility (Figure 3 (A) vs. 3 (C)). This

indicates that the rearrangement of the surface (leading to the accumulation of the C-C/C-H

motifs at the surface) is facilitated by the action of LiCl. The action of LiCl is more noticeable

in the more delignified samples with enhanced accessibility, enhanced matrix mobility and

increased presence of OH-groups and could be attributed to the disruption of H-bonding. The

corresponding analysis of the deconvoluted oxygen signals indicates, once again, the Li-O

interactions (the appearance of an O0 signal), with increased intensity upon lignin removal,

when compared to the original Sw samples (Figure 3 (B) & (D)). Furthermore, the

Page 12 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

13

deconvoluted oxygen spectra reveal a decrease in the O1 (O-C=O) signal upon LiCl

treatment, more pronounced for more delignified samples. This could indicate the withdrawal

of the carbonyl functionalities (C=O) from the surface of the holocellulose samples as a part

of the rearrangement promoted by the LiCl, and facilitated by the improved mobility in the

more delignified samples. On the other hand, it could also indicate a removal of hemicellulose

acetyl groups during LiCl treatment. Moreover, changes in carbonyl signals may also be

associated with removal of oxidized shortened chains formed as side products during the PAA

delignification. However, under mild conditions applied in this study no significant oxidative

effects on carbohydrates are expected (jääskeläinen et al. 2000) and thus the observed losses

in carbonyl signals are rather indicative of enhanced removal of hemicellulose carbonyls

(deacetylation) from the more accessible matrix.

The variations observed in the O2 signal (C-O) also concur with this explanation. In the

original Sw sample, the O2 portion on the surface was reduced significantly upon LiCl

treatment, which can be understood as the above mentioned surface hydrophobization that

occurs upon rearrangement. As delignification proceeds, carbonyl rich hemicelluloses become

more accessible. Some of them are deacetylated or washed out during the LiCl treatment,

giving rise to a reduction of O1 (O-C=O) and the simultaneous relative increase of O2 (C-O),

as observed. It should be pointed out here that, when changes in O2 are considered, the

gradual removal of lignin (from Sw, over HC6 to HC72) makes it increasingly difficult to

“hide” C-O motifs from the surface upon rearrangement, which may affect the variations

observed.

For comparison, the lignin fraction of the Sw sample was isolated and included in the XPS

analysis (Figure 4). No rearrangement leading to the accumulation of C-C on the surface

could be observed. Here, it is likely that the structure created during drying after LiCl

Page 13 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

14

treatment is determined by hydrophobic interactions of the lignin motifs, leading to the

exposure of more hydrophilic groups on the surface.

It was even more interesting to observe that there were no indications of specific Li-O

interactions either. It is obvious that both of these phenomena seem to be characteristic of the

carbohydrate fibrillar network capable of surface adapting and specifically interacting with

Li+/Cl- ions. Nevertheless, the lignin fraction seems to retain a considerable amount of LiCl,

possibly through non-specific interactions.

3.2 BET

In order to investigate the rearrangements that were observed as being promoted by LiCl

further, variations in the surface area of the cellulose-rich samples before and after LiCl

treatment were monitored. The BET specific surface area increased from 1.5±0.0m2 g-1 to

2.4±0.1m2 g-1 for HC6, and from 1.6±0.03m2 g-1 to 2.7±0.6m2 g-1 for HC72 upon LiCl

treatment. The LiCl treatment promotes an increase in surface area, with the effect being more

pronounced as the content of carbohydrate increased, i.e. there is an increase in both the

matrix mobility and the accessibility of the carbohydrate components. The increase observed

in the surface area may be an effect of the aforementioned disruption of the stabilizing H-

bonding, which promotes the surface adapting rearrangement of the fibrillar structure by the

action of the Li+/Cl- ions. Indeed, the literature (Nada et al. 2009) reports that the crystallinity

of the cellulose materials decreases upon LiCl treatment, allowing an increased surface area to

be observed. Moreover, the increased surface area of the cellulose-rich samples can also be

affected by the removal of soluble components during LiCl treatment and deacetylation.

3.3 FTIR

The FTIR spectra of the Sw flour, HC6, HC72 and lignin/LCC before and after treatment with

LiCl are presented in Figure 5; the tentative assignments of the bands observed

Page 14 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

15

(Schwanninger et al. 2004; Rosu et al. 2010; Colom et al. 2003; Pandey 1999) are reported in

Table S3 in the Appendix. In the case of Sw flour samples, changes in signals corresponding

to C=O stretch in carbonyls (e.g. acetyl groups) present in hemicelluloses (1738-1728cm-1)

were observed upon LiCl treatment (Figure 5 (A)). Corresponding bands shown a clear

decrease in intensity, indicating t partial removal of acetyl groups during aqueous LiCl

treatment: this is in agreement with the XPS findings. As the lignin was removed gradually

from the Sw flour (i.e. for HC6), and upon LiCl treatment, the spectral changes in these wave

number intervals were even more pronounced, implying that the removal of acetyls was even

more extensive. This observation is further supported by the observed decrease in pH of the

liquid phase during the sorption experiments. It was found that the pH of the liquid phase

decreased from 5.5 to 3.6, 2.5, and 2.2 for Sw, HC6, and HC72 samples respectively, which

might be an effect of deacetylation during the LiCl treatment.

Furthermore, signals originating from the sorbed water (1652-1630cm-1) were enhanced

significantly in delignified samples, implying an enhanced interaction with water upon LiCl

treatment. Moreover, significant changes in the -OH signals were also detected in the more

delignified samples (Figure 5 (B) and (C)), implying participation of these functional groups

in interactions with LiCl; they might result in complexation and changes in the H-bonding

patterns, as already indicated by the XPS results.

The lignin/LCC sample shown even more pronounced changes in the signals originating from

these groups (-OH, C=O and sorbed water) (Figure 5 (D)). According to Alam and Van De

Ven (2014), the softwood lignin isolated by the procedure employed in this study is rich in

both phenolic groups and carbonyls originating from the oxidized carbohydrate residues in

lignin-carbohydrate complexes. Drastic changes in the –OH region of this sample are

indicative of the altered interactions of phenolic groups due to the presence of LiCl and/or

enhanced interactions with water. The reduction in the carbonyl signals implies that there is a

Page 15 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

16

reduction in the C=O content caused either by the removal of soluble C=O rich structures

during LiCl treatment or by the formation of hemiacetal linkages during the drying step

following the LiCl treatment. The increase observed in the sorbed water signal indicates an

enhanced interaction with water after LiCl treatment, which concurs not only with the

changes observed in the signals of the phenolic groups but also the increase in the O/C ratio

observed with XPS.

3. 4 The effect of the composition of the wood constituents on the amount of Li+ ions

sorbed onto Norway spruce Sw flour (AAS study)

The effect the composition of the wood constituents had on the amount of Li+ ions sorbed

onto samples of Norway spruce was measured quantitatively using Flame Atomic Emission

Spectroscopy (FAES). The results showed that the amount of Li+ ions sorbed increased as the

amount of lignin decreased in the sample (Figure 6 (a)). The highest concentration of Li+ ion

was observed in the carbohydrate-rich portion (i.e. HC72) and the lowest concentration in the

lignin/LCC fraction. The amount of Li+ ions sorbed onto the samples (Figure 6) for the

experimental conditions studied (i.e. at a relatively neutral or low pH) were much higher than

the concentration of the carboxylic acid groups (0.086±0.007 mmol/g) in the samples of

native Norway spruce wood (Werkelin et al. 2010). Thus, apart from the carboxylic acids

prone to ion-exchange with Li-cations (as being predominantly ionized under applied

conditions), other functional groups, such as OH, must also act as sorption sites. The number

of theoretical OH groups estimated for the samples studied are also shown in Figure 6, see S4

in the Appendix for the estimation procedure. These numbers increased slightly upon PAA

delignification, and were observed to be higher for holocelluloses than for the lignin fraction.

It is evident from Figure 6 that samples with more OH groups show higher amounts of Li+

ions sorbed, suggesting that OH functional groups have significant interactions with Li+ ions,

as was observed in the FTIR studies. Moreover, this result also indicates that there is an

Page 16 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

17

increase in the number of interactions the functional groups make with the LiCl solution,

which is due to the increased accessibility of the carbohydrate network upon delignification.

For instance, Gross et al. (2013) state that the small size of the Li+ ions allows for many

possible interaction sites with the glucose residues of the glucan chains, including the more

restricted regions near the oxygen atom of the glycosidic bonds. Zhang et al. (2014) report

that the chloride anions (Cl-) in the aqueous LiCl solution form hydrogen bonds with hydroxyl

protons of cellulose, while lithium cations (Li+) accompany the hydrogen bonded Cl- anions

to meet the electric balance.

4. Conclusions

The introduction of Li+ and Cl- ions from an aqueous LiCl solution into a Norway spruce Sw

matrix is associated with the disruption of H-bonding and the accompanying enhanced

mobility, which are demonstrated as an enhanced adaptation of the surface of the fibrillar

structure upon drying and an enlarged surface area. The sorption of Li+ ions onto a wood

matrix relies predominantly on interactions with oxygen-containing functional groups of

carbohydrate structures, visible as Li-O interactions in XPS studies, and changes in the OH

functional groups and sorbed water in FTIR. The decrease in the carbonyl content in the wood

material upon LiCl treatment that was observed in both the XPS and FTIR studies indicates a

deacetylation of the hemicelluloses.

Acknowledgments

We are grateful to Chalmers Energy Initiative for their financial support. Our thanks go to Ms.

Anne Wendel, Senior research engineer, Applied chemistry, Chalmers University of

Technology, for her technical assistance with the XPS characterization measurements.

Appendix A. Supplementary data

Page 17 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

18

References

Ahluwalia, S.S., Goyal, D. (2007) Microbial and plant-derived biomass for removal of heavy

metals from waste water. BIORESOURCE TECHNOL. 98:2243-2257.

Alam, M.N., Van De Ven, T.G. (2014) An aqueous process for the separation of cellulose

from biomass. J-FOR-Journal of Science & Technology for Forest Products and Processes.

4:22-26.

Akhtaruzzaman, A.F.M., Virkola, N-E. (1979) Influence of chip dimensions in Kraft pulping

Part I. Mechanism of movement of chemicals into chips. Paperi ja Puu. 9:578–580.

Argun, M.E., Dursun, S., Ozdemir, C., Karatas, M. (2007) Heavy metal adsorption by

modified oak sawdust: Thermodynamics and kinetics. J HAZARD MATER. 141:77-85.

Batzias, F.A., Sidiras, D.K. (2007) Dye adsorption by prehydrolysed beech sawdust in batch

and fixed-bed systems. BIORESOURCE TECHNOL. 98:1208-1217.

Chastain, J., King, R. C. (Eds.). Handbook of X-ray photoelectron spectroscopy: a reference

book of standard spectra for identification and interpretation of XPS data (p. 261). Eden

Prairie, MN: Perkin-Elmer. 1992.

Colom, X., Carrillo, F., Nogués, F., Garriga, P. (2003) Structural analysis of photodegraded

wood by means of FTIR spectroscopy. POLYM DEGRAD STABIL. 80:543-549.

Dorris, G.M., Gray, D.G. (1978) The surface analysis of paper and wood fibers by ESCA

(Electron Spectroscopy for Chemical Analysis). I Application to cellulose and lignin. CELL

CHEM TECHNOL12:9-23.

Dulman, V., Cucu-Man, S.M. (2009) Sorption of some textile dyes by beech wood sawdust. J

HAZARD MATER. 162:1457-1464.

Page 18 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

19

Duong, T.D., Nguyen, K.L., Hoang, M. (2006) Competitive sorption of Na+ and Ca 2+ ions on

unbleached kraft fibres—A kinetics and equilibrium study. J COLLOID INTERF SCI.

301:446-451.

Gross, A.S., Bell, A.T., Chu, J-W. (2013) Preferential interactions between lithium chloride

and glucan chains in N, N-Dimethylacetamide drive cellulose dissolution. J PHYS CHEM B.

117:3280–3286.

Ferrero, F. (2007) Dye removal by low cost adsorbents: Hazelnut shells in comparison with

wood sawdust. J HAZARD MATER. 142:144-152.

Harilal, S.S., Allain, J.P., Hassanein, A., Hendricks, M R., Nieto-Perez, M. (2009) Reactivity

of lithium exposed graphite surface. APPL SURF SCI. 255:8539-8543.

Ho, Y.S., McKay, G. (1998) Kinetic models for the sorption of dye from aqueous solution by

wood. PROCESS SAF ENVIRON. 76:183-191.

Inari, G.N., Petrissans, M., Lambert, J., Ehrhardt, J.J., Gérardin, P. (2006) XPS

characterization of wood chemical composition after heat‐treatment. SURF INTERFACE

ANAL. 38:1336–1342.

Jedvert, K., Saltberg, A., Lindström, M.E., Theliander, H. (2012) Mild steam explosion and

chemical pre-treatment of Norway spruce. BIORESOURCES. 7:2051-2074.

Johansson, L.S., Campbell, J.M., Fardim, P., Hultén, A.H., Boisvert, J.P., Ernstsson, M.

(2005) An XPS round robin investigation on analysis of wood pulp fibres and filter

paper. SURF SCI. 584:126-132.

Johansson, L.S., Campbell, J.M., Hännien, T., Ganne-Chedeville, C., Vuorinen, T., Hughes,

M., Laine, J. (2012) XPS and the medium-dependent surface adaptation of cellulose in wood.

SURF INTERFACE ANAL. 44:899-903.

Page 19 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

20

Jääskeläinen, A-S., Tapanila, T., Poppius-Levlin, K. (2000) Carbohydrate reactions in

peroxyacetic acid bleaching. J WOOD CHEM TECHNOL. 20:43-59.

Kolavali, R., and Theliander, H. (2014) Experimental determination of the diffusion of

monovalent cations into wood: Effects of micro-cracks, wood structure, impregnation time

and temperature on the local concentration profiles. J-FOR, Journal of Science & Technology

for Forest Products and Processes. 4:29-35.

Kolavali, R., Merima, H., Theliander, H. (2016) The sorption of monovalent cations onto

Norway spruce: model studies using wood flour and LiCl solution (In Manuscript)

Kumar, R., Hu, F., Hubbell, C.A., Ragauskas, A.J., Wyman, C.E. (2013) Comparison of

laboratory delignification methods, their selectivity and impacts on physiochemical

characteristics of cellulosic biomass. BIORESOURCE TECHNOL.130:372-381.

Lin, S.Y., Dence, C.W. (1992) Methods in lignin chemistry. Springer.

Nada, A.M.A., Abd El-Mongy, S., Abd El-Sayed, E.S. (2009) Effect of different treatments

on cellulose toward carboxylation and its application for metal ion absorption.

BIORESOURCES. 4:80-93.

Nzokou, P., Pascal Kamdem, D. (2005) X‐ray photoelectron spectroscopy study of red oak‐

(Quercus rubra), black cherry‐ (Prunus serotina) and red pine‐ (Pinus resinosa) extracted

wood surfaces. SURF INTERFACE ANAL. 37:689-694.

Ofomaja, A.E. (2008) Kinetic study and sorption mechanism of methylene blue and methyl

violet onto mansonia (Mansonia altissima) wood sawdust. CHEM ENG J. 143: 85-95.

Ofomaja, A.E., Ho, Y.S. (2008) Effect of temperatures and pH on methyl violet biosorption

by Mansonia wood sawdust. BIORESOURCE TECHNOL. 99:5411-5417.

Page 20 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

21

Özacar, M., Şengil, Đ.A. (2005(a)). A kinetic study of metal complex dye sorption onto pine

sawdust. PROCESS BIOCHEM. 40:565-572.

Özacar, M., Şengil, Đ.A. (2005(b)). Adsorption of metal complex dyes from aqueous solutions

by pine sawdust. BIORESOURCE TECHNOL. 96:791-795.

Palme, A., Idström, A., Nordstierna, L., Brelid, H. (2014) Chemical and ultrastructural

changes in cotton cellulose induced by laundering and textile use. CELLULOSE. 21:4681–

4691.

Pandey, K.K. (1999) A study of chemical structure of soft and hardwood and wood polymers

by FTIR spectroscopy. J APPL POLYM SCI. 71:1969-1975.

Rosu, D., Teaca, C.A., Bodirlau, R., Rosu, L. (2010) FTIR and color change of the modified

wood as a result of artificial light irradiation. J PHOTOCH PHOTOBIO B. 99:144-149.

Schwanninger, M., Rodrigues, J.C., Pereira, H., Hinterstoisser, B. (2004) Effects of short-time

vibratory ball milling on the shape of FT-IR spectra of wood and cellulose. VIB

SPECTROSC. 36:23-40.

Šćiban, M., Klašnja, M., Škrbić, B. (2006) Modified softwood sawdust as adsorbent of heavy

metal ions from water. J HAZARD MATER. 136:266-271.

Sernek, M. (2002) Comparative analysis of inactivated wood surfaces. Ph.D. thesis, Virginia

Polytechnic Institute and State University, USA.

Shukla, A., Zhang, Y.H., Dubey, P., Margrave, J.L., Shukla, S.S. (2002) The role of sawdust

in the removal of unwanted materials from water. J HAZARD MATER. 95:137-152.

Sinn, G., Reiterer, A., Stanzl-Tschegg, S.E. (2001) Surface analysis of different wood species

using X-ray photoelectron spectroscopy (XPS). J MATER SCI. 36:4673-4680.

Page 21 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

22

Theander, O., Westerlund, E.A. (1986) Studies on dietary fiber. 3. Improved procedures for

analysis of dietary fiber. J AGR FOOD CHEM. 34:330-336.

Werkelin, J., Skrifvars, B-J., Zevenhoven, M., Holmbom, B., Huppa, M. (2010) Chemical

forms of ash-forming elements in woody biomass fuels. FUEL. 89:481-493.

Page 22 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

23

Table 1.

Compositional analysis of untreated and delignified biomass following PAA delignification

and periodate oxidation of Norway spruce Sw flour.

Sample Description Carbohydrate fraction

(wt.%)

Lignin fraction

(wt.%)

Cellulose GGM AGX Klason Acid

soluble

Sw Untreat Sapwood

flour

36.9 17.2 5.8 28.8 4.2

HC6 Holocelluloses after

6h of PAA

delignification

39.5 17.4 5.8 24.4 9.1


24h of PAA

delignification

42.6 18.4 6.2 6.1 6.0


51h of PAA

delignification

43.2 18.6 6.1 4.2 6.0


72h of PAA

delignification

54.1 19.1 6.7 2.7 6.6

Lignin/LCC Lignin with 0.8 5.6 2.0 63.6 5.3

Page 23 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

24

chemically bound

hemicelluloses

Table 2.

Surface composition of the elements on Norway spruce Sw flour before and after treatment

with 1M LiCl for 3h residence time, at room temperature, detected by XPS.

Sample Element

C1s O1s Li1s Cl2p O/C

Untreated Sw flour 73.96 26.04 - - 0.35

LiCl treated Sw flour 73.12 24.29 2.06 0.53 0.33

Table 3.

Surface composition of the elements on isolated wood components of Norway spruce, before

and after treatment with 3M LiCl for 24h residence time, at room temperature, detected by

XPS

Sample Element

C1s O1s Li1s Cl2p O/C

Untreated HC6 69.28 30.72 - - 0.44

LiCl treated HC6 64.70 27.65 4.58 3.07 0.43

Page 24 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

25

Untreated HC72 68.08 31.92 - - 0.47

LiCl treated HC72 61.88 26.42 6.17 5.13 0.43

Untreated Lignin/LCC 74.35 24.86 - - 0.33

LiCl treated Lignin/LCC 55.01 19.45 12.38 12.66 0.35

Figure 1. Survey spectra of Norway spruce Sw flour (a) before and (b) after treatment with

1M LiCl aqueous solution for 3h at room temperature.

Page 25 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

26

Figure 2. High resolution (a) and deconvoluted (b) spectra of carbon (C1s) (A) and oxygen

(O1s (B)) signals on Norway spruce Sw flour (i) before and (ii) after treatment with 1M LiCl

aqueous solution for 3h residence time at room temperature (as indicated).

Page 26 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

27

Page 27 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

28

Figure 3. High resolution (a) and deconvoluted (b) spectra of (A) HC6-C1s, (B) HC6-O1s, (C)

HC72-C1s and (D) HC72-O1s (i) before and (ii) after treatment with 3M LiCl aqueous

solution for 24h residence time at room temperature (as indicated).

Page 28 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

29

Figure 4. High resolution (a) and deconvoluted (b) spectra of carbon (C1s) (A) and oxygen

(O1s (B)) signals on lignin/LCC samples of Norway spruce (i) before and (ii) after treatment

with 1M LiCl aqueous solution for 24h residence time at room temperature (as indicated).

Page 29 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

30

Figure 5. FTIR spectra of Norway spruce Sw flour (A), HC6 (B), HC72 (C) and lignin/LCC

(D) before and after treatment with 3M LiCl aqueous solution at room temperature for 24h

residence time.

Page 30 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review O

nly

31

Figure 6. Variation of amount of Li+ ion sorbed on samples of different compositions of wood

constituents (as indicated); after 24h of residence time with 3M LiCl aqueous solution at room

temperature; variation of theoretical [OH] groups available on samples of different

compositions of wood constituents are also indicated.

Page 31 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review Only

1

Supporting information

The sorption of monovalent cations onto Norway spruce wood flour: molecular interactions behind the LiCl impregnation

Reddysuresh Kolavalia, Merima Hasani

a*

aDivision of Forest Products and Chemical Engineering, Department of Chemistry and Chemical Engineering, Chalmers University of Technology, SE-41296 Gothenburg,

Sweden.

*Corresponding author: Tel.: +46 31 772 3065; fax: +46 31 772 2995.

E-mail address: [email protected] (Merima Hasani).

This supporting information has six pages. It includes three tables (namely Table S1, Table S2 and Table S3) and S4: procedure to estimation of number of theoretical OH

groups in wood samples.

Table S1.

Peak area (%) distribution of the different bonds in C1s spectra of the samples (before and after LiCl treatment) studied

Bond

Type

(%)

Sample

Sw HC6 HC72 Lignin/LCC

untreat LiCl

treated

untreat LiCl

treated

untreat LiCl

treated

untreat LiCl

treated

C1 52.0 53.1 41.0 44.8 35.2 51.1 63.3 57.3

C2 35.7 34.2 41.3 40.3 50.5 36.0 23.8 29.8

C3 9.0 8.6 13.4 11.8 9.2 9.4 5.8 9.7

C4 3.4 4.2 4.4 3.2 5.1 3.6 7.1 3.1

Page 32 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review Only

2

Table S2.

Peak area (%) distribution of the different bonds in O1s spectra of the samples (before and after LiCl treatment) studied

Bond

Type

(%)

Sample

Sw HC6 HC72 Lignin/LCC

untreat LiCl

treated

untreat LiCl

treated

untreat LiCl

treated

untreat LiCl

treated

O1 - 5.8 - 7.9 - 10.5 - -

O2 70.0 72.8 72.3 70.5 78.1 58.2 55.4 66.1

O3 30.0 21.4 27.7 21.6 21.9 31.4 44.6 33.9

Table S3.

Tentative band assignments of the samples studied in this work before and after LiCl treatment

Wave number

(cm-1)

Sw flour HHLC HLLC Lignin/ LCC Sw flour HHLC HLLC Lignin/ LCC

Before LiCl treatment After LiCl treatment

3645-3620

Free OH

3570-3450

Valance

vibration of H

bonded OH

groups

3500-300

O-H

3368 3464-3306 3480-3306 3464 3368 3192-3496 3118-3516 3640-3038

Page 33 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review Only

3

stretching

C-H

stretching

2996-2902

C-H stretch

2904 2902 2902 2938 2904 2904 2906 2996,2960,

2912

New peaks 2132 2064 2128 2086

1738-1728

C=O stretch

in COOH

1734 1738 1738 1738 1728 1732 1728 -

Adsorbed

water

1652 1650 1646 1650 1638 1640 1640 1630,1654

1512-1510

Aromatic

skeletal

vibration

1510 1512 - 1510 1510 1512 - 1510,1532

1464-1462

C-H

deformation

(methyl (-

CH3) and

methylene (-

CH2))(lignin)

1462 - - 1462 - - - 1464

1428-1424

Aromatic

skeletal

vibrations

combined

with C-H in

plane

deformation

(cellulose)

1428 1428 1428 1424 1426 1428 1428 1426

Page 34 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review Only

4

1374

C-H

deformation

(symmetric)

1374 1374 1374 - 1374 1374 1374 -

1322-1320

CH2 rocking

vibration

1322 - - - 1322 1320 1320 -

1266-1262

C-O of

guaiacyl ring

plus C=O

stretch

1266 1266 - 1266 1266 1266 1262 1264

1230

C-C plus C-O

plus C=O

stretch

- - - 1230 - - - -

1156-1152

C-O-C

asymmetric

valance

vibration

1156 1154 1154 1152 1158 1158 1158 -

1146-1138

Aromatic C-

H in plane

deformation;

typical for

guaiacyl units

- - - 1138 - - - 1146

1122-1120

C-C and C-O

stretching

- 1120 1120 1122 - - - -

1106-1104

Glucose ring

- 1106 1104 1104 - - 1104 -

Page 35 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review Only

5

stretch

1088

C-O

deformation

in secondary

alcohols

- 1088 1088 1088 - - - -

1076-1072

- 1072 1072 1076 - - - -

1062

C-O stretch

mainly from

C3-O3H

- - - 1062 - - - 1062

1058-1050

O-H bending

1058 1056 1054 1050 1058 1056 1058 -

1032

C-O of

primary

alcohol,

guaiacyl C-H

- - - 1032 - - - 1032

1026-1024

Aromatic C-

H in plane

deformation;

symmetrical

C-O

stretching

- 1026 1024 - - - 1026 -

898

Glucose ring

stretching,

C1-H

deformation

898 898 - - - 896 898 -

Page 36 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Review Only

6

S4. Estimation of number of theoretical OH groups in Norway spruce Sw flour and its isolated components

The number of theoretical OH groups in moles per gram (mol g-1) can be estimated Eq. A1, proposed by Rowell [1] and Hill [2]

��

��+

��

��+

��

��+

�� (A1)

where composition- % of cellulose (A), hemicelluloses hexosan (B) and pentosan (C) and lignin (D). The compositional analyses of Norway

spruce Sw flour and its isolated components are reported in Table 1 in the main article. Following this composition, the estimated number of

theoretical OH groups were found to be slightly increased upon delignification. As hydroxyl groups made inaccessible through binding between

wood polymers or steric hindrance are not considered, this estimate expresses a theoretical upper limit. And also two-thirds of the hydroxyl

groups of cellulose are known to be inaccessible to water due to crystalline portion of cellulose. In addition to this, history of the sample e.g.

drying also influence the accessibility of hydroxyl groups in the sample.

[1] R M Rowell, Wood Sci. (1980) 13:102.

[2] C A S Hill, Wood modification- chemical, thermal and other processes. Wiley, Chichester. (2006).

Page 37 of 37



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Paper V

Solute sorption and diffusion in wood based on the LiCl impregnation of Norway

spruce


In manuscript

1

Solute sorption and diffusion in woodSolute sorption and diffusion in woodSolute sorption and diffusion in woodSolute sorption and diffusion in wood bbbbased on ased on ased on ased on thethethethe

LiCl impregnation of Norway spruceLiCl impregnation of Norway spruceLiCl impregnation of Norway spruceLiCl impregnation of Norway spruce


Abstract

Many applications require that chemicals be transported into wood: not only in the production of paper pulp and in the fractionation of different wood components in wood-based biorefineries but also in operations in which solid wood is treated with different kinds of chemicals, e.g. preservatives, fire retardants and dimension-stabilizing chemicals. The transport of chemicals in a heterogeneous porous hygroscopic material such as wood is a complex mass transfer phenomenon, often involving several processes: the diffusion of chemicals in the cell pores (lumen and pit pores), through the cell walls at certain conditions, and sorption at solid surfaces. In the present study, an experimental methodology was adapted to measure the concentration profiles of cations in the porous structure of the wood matrix. Using a transport model that took into consideration both the diffusive mass transport and the sorption of ions onto the wood material, effective diffusion coefficients and tortuosity factors were estimated. This transport model was developed using COMSOL® multiphysics modeling software. Keywords: diffusion, sorption, impregnation, model, experiment, wood

Introduction

Understanding the conditions that control the mass transfer of various chemicals/ions in wood is of great importance to wood processing: homogeneous impregnation, for example, increases the uniformity of the treatment, reduces reaction times and possibly increases the yield of the final products. The impregnation of wood material with chemicals can occur via advective (penetration) or diffusive (diffusion) mass transport1. Penetration is defined as the flow of liquor into the gas/vapour-filled voids of the wood pieces under a pressure gradient, and diffusion as the movement of ions or other solute matter through the liquid within the wood pieces caused by concentration gradients. Penetration occurs rapidly whereas diffusion is normally much slower, making it the controlling mechanism in most (but not all) wood impregnation processes. The diffusive mass transport of chemicals in wood material is a complex mass transfer. A wood cell (fibre) may be described as a pipe with both ends closed. The cells overlap each other and, at normal conditions, the mass transport between the fibres is via so-called “pit pores”. The cell wall may be described as being more or less a swollen gel. If the fibres and the pit pores become filled with a liquid, the main diffusive mass transport of the solutes (ions/chemicals) occurs through the liquid. The solutes may also be sorbed onto the cell wall and so there may also be a mass transport at the cell wall (surface diffusion) and/or in the gel-like structure of the cell wall. The overall mass transport rate may thus be influenced by several different mass transport phenomena as well as by sorption phenomena. A vast number of studies regarding mass transfer in wood can be found in the literature, e.g. estimation of the diffusion coefficients of water 2-3, preservative treating chemicals4-10, dimension-stabilizing chemicals11, nonelectrolytes12-17, inorganic ions18-21 and pulping chemicals22-27 in wood material. In the case of wood, it has been found that diffusion in the

2

longitudinal direction in its structure is faster than in the radial and transversal directions. It has been reported that28 the effective capillary cross-sectional area (ECCSA), which is defined as the ratio between the area available for diffusion and the total area, differs in the three structural directions: it is about 50% in the longitudinal direction and 5-10% in the tangential-radial directions in untreated wood for all except alkaline solutions. In alkaline solutions the cell walls of the wood swell considerably, resulting in diffusion having approximately equal rates in all three structural directions. In the literature, three different methodologies employed to measure the diffusion of solute ions into water-filled pores in wood are common29. The first method either measures the rate at which ions diffuse through a wood block whose sides are in contact with solutions of different concentration or the change in ion concentration in a water bath in which a completely impregnated wood block has been placede.g.8, 17, 19, 24, 26. The second method measures the electrical conductivity of the impregnated wood block compared with that of the impregnating chemical solutione.g.28; the ratio of the two conductivities gives a direct measurement of the ECCSA. The third method employs tracer techniques using radioactive isotopes, in which the samples are sliced and the radiation activity from the whole surface is used as an indication of the degree of diffusion e.g.27. However, these methodologies have some limitations: the first and second method gives only an average diffusivity, and cannot be used for measuring concentration profiles and the third is limited to mapping concentration profiles in only one direction at a time, since the whole surface of each wood slice is used to measure the concentration. Moreover, most of the investigations involving measuring cation diffusion in wood have been conducted using substances such as NaCl (sodium chloride), KCl (potassium chloride) and NaOH (sodium hydroxide). When NaCl/KCl are used as the diffusing substance, the Na+/ K+ ions already present in the wood (in the form of alkali metal ions) mean there is a possibility that inaccurate measurements are made. In the case of NaOH, the reactions of OH- ions with the wood components influence the structure of the wood which, in turn, influences the mass transport of Na+ ions. Furthermore, these methodologies ignore the effects of both micro-cracks and the interactions between the diffusing solute ions and the components of the cell wall on the measurement of concentration profiles and thereby these factors affects the determination of the diffusivities in the system. In order to overcome some of the difficulties stated above, LiCl (lithium chloride) was chosen as the diffusing substance; a direct method of slicing the piece of wood and measuring the Li+ ion concentration with Flame Atomic Emission Spectroscopy (FAES) were decided upon. In our previous studies30-31, concentration profiles of Li+ ions in wood samples of Norway spruce were measured at different impregnation times and temperatures. The objective of the current study was to further develop a methodology for studying the diffusion of ions through the porous structure of the wood matrix by considering the effect of the sorption of ions onto wood matrix surfaces. The effects of impregnation time on the concentration profiles of Li+ ions in Norway spruce wood at 40 oC were therefore investigated. A model describing the diffusion of solutes into a water-filled porous medium was developed to quantify the transport properties, such as effective diffusion coefficients and tortuosity factors of ions (Li+) in the Norway spruce wood sample in radial/tangential as well as longitudinal directions.

Materials and Methods

MaterialsMaterialsMaterialsMaterials A stemwood disc 23 cm in diameter (free from bark) taken from a 31±1-year old Norway spruce (Picea abies L) was investigated. The samples of both Sw and Hw were prepared carefully using a vertical band saw (Mossner Rekord, August Mössner KG, Mutlangen, Germany; cutting band saw: L.S. Starrett Company LTD., Jedburgh, Scotland with 14 teeth per inch) and were placed preliminarily in an airtight polyethylene (PE) bag at 1oC. It was assumed that a stemwood disc contains 50% Sw and 30% Hw, with remaining 20% being an intermediate of

3

Sw and Hw. In this study, only Sw was investigated. Samples free from rot and deformations was selected. These were cut into rectangular prototype pieces, using the same vertical band saw mentioned above, and had dimensions of 100 × 25 × 7.5 mm3 (L × R × T). The radial dimension was chosen to be larger (approx. 3 times) than the tangential dimension, in order to minimize the effect of mass transport through radial direction on measurements of mass transport in tangential direction. The material was then stored in an airtight PE bag in a freezer at -18oC. Defrosting the wood pieces took 24 h at ambient temperature. For the studies of the concentration profiles measurement, the wood pieces were carefully planed on all four vertical surfaces to peel off the surface layers: approx. 1 to1.5 mm were removed using a hand plane (Stanley Hand Tools, Stanley Canada, Mississauga, Canada) in order to minimize the damage caused to the surface of the wood pieces by rough sawing. However, similar approach was not possible to do with the two surfaces perpendicular to the longitudinal fiber directions and thus wood pieces used in the actual experiments had dimensions of 100 × 23 × 6.5 mm3 (L × R × T). Further steps in the experimental procedure, along with details of the equipment used, can be found in our earlier papers30-31. For the equilibrium sorption studies, the wood pieces were ground into flour in a Wiley-type mill (< 1 mm) and then stored in an airtight PE bag, ready for further use. A dilute acid solution was prepared using reagent grade (95-97%) sulphuric acid (Scharlau, Scharlab S.L., Spain) and deionized water. A bulk LiCl aqueous solution was prepared by dissolving LiCl salt (≥ 99%) (Merck KGaA, Darmstadt, Germany) in deionized water. A 2 wt. % nitric acid solution (HNO3) was prepared using trace analysis grade (69%) nitric acid (ARISTAR, VWR® PROLABO®, Leuven, Germany) and deionized water. Density deDensity deDensity deDensity determinationterminationterminationtermination The mass of the air-dried (normal conditions at room temperature) samples divided by their volume gives the bulk density (ρb). The volume of the samples was determined by simply measuring their dimensions. Specimens with a weight of about 3 g and dimensions of 30x25x8 mm3 were used. The AccuPyc 1330 helium pycnometer (MICROMETRICS) was used to determine the solid volume of the wood cell wall. Helium gas was used as the displacement medium to fill as many of the small pores as possible: wood flour (< 1 mm sieve) was used for these measurements so that as many pores as possible were open to the displacement medium. The density of the cell wall (ρp) was then calculated using the solid volume of the cell wall measured and the weight of the sample. Porosity measurementsPorosity measurementsPorosity measurementsPorosity measurements Based on the normal bulk density (ρb) and the solid density of the cell wall (ρp), the total porosity can be calculated as follows:

�� = 1 − ��

(1)

This total porosity includes all of the pores, both open and closed. Mercury (Hg) porosimetry was carried out with a Micrometrics AutoPore IV Mercury Porosimeter and analyzed with the software AutoPore IV 9500, Version 1.09. The porosity determined by mercury intrusion porosimetry only determines the percentage of pores that are accessible to Hg. Specimens of about 0.2 g were cut perpendicular in dimensions of about 2.5 (L) x 24 (R) x 7.5 (T) mm, with the smallest length that could be prepared practically being around 2.5 mm. This is nevertheless small enough to cut most of the fibres (average fibre length is about 3 mm), and means that, in most cases at least, the mercury is able to enter the lumen of the fibre without passing through a pit pore. Consequently, the intrusion volume and the diameters of the lumen could be measured in a satisfactory way. Several duplicates were prepared from two different pieces of wood and analyzed.

4

The porosity (Hg porosimetry) can be calculated from the intrusion pore volume determined using Hg porosimetry and the cell wall solid volume determined using helium pycnometer:

�� = �� ! �"# $�%" %&��⁄

�� ! �"# $�%" %&��⁄ ()"$$*&$$� $�+# $�%" %&��⁄ (2)

In this study, the volume of intrusion pores that was measured using Hg-porosimetry is classified into 2 parts: (i) the total volume of intrusion pores, which applies to the whole range of pore diameters that the instrument can measure (i.e. between 0 and 330 µm) and (ii) the intrusion pore volume of the wood piece that is accessible to diffusion (i.e. pore diameters between 0 and 41 µm). Equilibrium sorption Equilibrium sorption Equilibrium sorption Equilibrium sorption measurementmeasurementmeasurementmeasurementssss Batch sorption experiments were conducted on flour samples of Sw at 40 oC, using 1 M bulk aqueous LiCl solution, in order to determine the amount of Li+ ions that are sorbed onto solid wood surfaces at equilibrium conditions (qwood,e). The experimental procedure for sorption study has been reported in detail in our recent paper32. Wood-Li+ ion equilibrium partition coefficient (k) is defined as: , =

-.//0,23456

(3)

where qwood,e is the amount of Li+ ions sorbed onto solid wood (g/kg) and CLi+ is the

concentration of Li+ ions in the solution located in the porous structure of the wood matrix (g/l) at an apparent equilibrium conditions. Since the experiments were conducted at high liquor-to-wood ratios (100:1) and for longer residence times, it is reasonable to assume that the concentration of Li+ ions in the liquid in the wood pores is equal to that of the bulk solution. In addition, the surface area (sorption sites) available in the wood flour and wood piece samples are also assumed to be constant. Concentration profile Concentration profile Concentration profile Concentration profile measurementmeasurementmeasurementmeasurementssss The experimental procedures were identical to those reported in our earlier paper31 with the exception that, prior to the LiCl impregnation experiments, the wood piece samples were impregnated with dilute acid solution (pH=0.5/0.3 M H2SO4) to remove both air and the native metal ions present in the wood material. In our recent study32, in experiments with wood flour samples, only traces of major metal ions (Ca2+, Mg2+, Mn2+, Na+ and K+) were observed after dilute acid treatment. The effect of these metal ions on mass transport of Li+ in wood is thus negligible. The impregnation of wood piece samples with dilute acid solution was carried out in a vacuum-pressure (N2 gas; 0.5 MPa) cycle in an autoclave at room temperature until no floating pieces of wood remained. These pieces were then dabbed with filter paper before being immersed in a solution of 1 M LiCl (6.94 g/l), maintained at a temperature of 40 oC, with a wood-to-liquor ratio of 1:75. After impregnation time intervals of 36, 72 and 168 h, the pieces were removed and placed in liquid N2 (-180 oC) to minimize further migration of Li+ ions. The frozen pieces were freeze-dried for about two weeks and then cut into small cubes at various locations (Figure 1 (A)) in order to measure their concentration profiles. The target dimensions of the cubes were approx. 6 x 6 x 6.5 (L x R x T) mm3 in all the impregnation experiments. Each cube was microtomed in the tangential direction to slices ~0.3 mm in thickness (Figure 1 (B)), which were then acid leached at room temperature with 2 wt. % HNO3 for approx. 24 h. The obtained leaching liquor was then analyzed for its concentration of Li+ ions using Flame Atomic Emission Spectroscopy (FAES), which used air-acetylene as the flame source. The emission was measured at 670.8 nm.

5

Figure 1. (A) Locations of the small cube samples taken from an idealized piece of impregnated wood, (B) Microtoming a small cube into slices of thickness ~0.3 mm for use in measuring the concentration profile of Li+ ions. DL and DT (= DR) are the diffusion coefficients of the Li+ ions from the aqueous LiCl solution to the wood in the longitudinal, tangential and radial directions, respectively.

The concentration of Li+ ions measured experimentally (g Li+/kg freeze-dried wood) was the total concentration of Li+ ions (qtotal), i.e. the sum of the Li+ ions sorbed onto the surfaces of the wood cell wall/into cell wall and that is in the solution located in the wood pores. In this study, all the results were reported based on the concentration of Li+ ions available in the solution located in the porous structure of wood matrix (CLi

+), which was calculated using Equation (4) based on mass balance. At each slice (i.e. thickness level), the qtotal is defined as:

7� �&$ = �89�6 ∗ ,� ; <89�6 ∗ =* +!�")"> (4)

where CLi+ is the concentration of Li+ ions available in the solution located in the porous

structure of the wood matrix (g/l), k is the wood-Li+ ion equilibrium partition coefficient (l/kg) and Vwood piece is the intrusion pore volume of the wood piece that is accessible to diffusion (l/kg).

EstimationEstimationEstimationEstimation of transport propertiesof transport propertiesof transport propertiesof transport properties ---- Model formulationModel formulationModel formulationModel formulation The model used to determine the transport properties in Norway spruce wood, such as effective diffusion coefficients and tortuosity factors of Li+ ions, is based on the theory of the transport of diluted species in a porous medium. The model was solved using COMSOL® multiphysics modeling software (Version 5.0; COMSOL, Stockholm, Sweden). Equation 5 describes the transport of solutes in a liquid-saturated porous medium (The Transport of Diluted Species in Porous Media Interface, COMSOL®, and Version 5.0). The two terms on the left-hand side of this equation describe the accumulation of the species within the liquid and solid, respectively. On the right-hand side, the term for the transport of species due to diffusion is introduced.

�ɸ ; @A,�B∁5B�; <∁� − @D∁D,�>

Bɸ

B�� E ∙ G<H",�>E∁�I (5)

6

where Ci is the concentration of species i in the liquid located in the porous media (g/l); CP, i is the amount sorbed onto the solid particles (g/kg); k is the equilibrium partition coefficient (l/kg); De,i is the effective diffusion coefficient of species i in the porous media (m2/s); ɸ is the porosity

of the wood piece that is accessible to diffusion, calculated by using equation (2); ρb and ρp are the bulk and solid phase density of wood (kg/m3), respectively.

The initial conditions (I.C) and boundary conditions (B.C) are:

I.C: for time t = 0 the concentration Ci = 0 g/l for all domains of the geometry

B.C: for all times t, the concentration Ci = 1 M = 6.94 g/l for all surfaces of the geometry.

Wood piece modelWood piece modelWood piece modelWood piece model

Rectangular geometry was selected to determine the effective diffusion coefficient in both the longitudinal and transversal directions. The dimensions chosen were 100 x 23 x 6.5 (L x R x T) mm3, i.e. the same as the wood pieces used in the actual experiments. The mesh consists of approx. 2501539 elements, which was found to suffice in obtaining a solution independent of the mesh. The simulations were run for 0 to 10080 min with a time interval of 1 min; both the absolute and relative tolerances were kept at 10-3. The computational time was ~ 4 h. In the case of Cube a (i.e. the middle part of the impregnated wood piece in Figure 1 (A)), it is assumed that for the conditions used in this study, the concentration is not affected by longitudinal mass transport due to the diffusion distance, which is long when compared to the radial and tangential directions. This assumption is reasonable, since the concentration profiles observed for Cube a and Cube b, were within the experimental error (the results of which are not shown here). Furthermore, the effective diffusion coefficient in the radial direction (De,R) is assumed as being equal to that in the tangential direction (De,T). De,R (=De,T) was estimated by fitting the experimental concentration profiles of Cube a. The range of initial guess values for De,R (=De,T) was obtained from the cube model (see Appendix A1); the guess values were varied until a minimum in the differences between the experimental and calculated concentration profiles was found. In the case of Cube c (i.e. close to the edge part of the impregnated wood piece in Figure 1 (A)), the diffusion of ions was influenced by mass transport in both the longitudinal (De,L) and transversal directions (De,R (=De,T)). It was assumed that the value of De,R (=De,T) determined in the case of Cube a was also valid here, and was therefore used in the estimation of De,L. De,L was estimated by fitting the experimental concentration profiles of Cube c. The initial guess values for De,L were varied, starting with the self-diffusion coefficient of Li+ ions in free solution, Df (see Appendix A2). This value was then reduced until a minimum in the differences between the experimental and calculated concentration profiles was found. The effective diffusion coefficient (De) for the transport of a solute in a liquid-saturated porous medium is significantly lower than the free-water diffusion coefficient (Df) because of the constricted and elongated (tortuous) transport paths. In a saturated porous medium, De is related to the Df as: 33

H" =JKɸƟM (6)

where ɸ is the porosity of the wood piece that is accessible to diffusion, Ɵ is a constrictivity factor to account for the constricted transport paths caused by the small pores and pore throats in a porous medium (usually assumed to be ~ 1) and τ is a tortuosity factor. The latter accounts for the reduction in diffusive flux caused by the tortuous path lengths formed by the solute molecules, which can be compared to the straight paths in an unrestricted aqueous medium.

7

Results and discussions

Density and porosity measurementsDensity and porosity measurementsDensity and porosity measurementsDensity and porosity measurements The average normal bulk density and the cell wall density of the Sw samples were found to be 0.41±0.0 g/cm3 and 1.5±0.0 g/cm3, respectively. The average total porosity, calculated using the values of normal bulk density and cell wall density of the Sw wood piece samples, was 72.4±0.3 %. These densities and total porosities are consistent with previous findings34 (Table 1). Making porosity measurements of wood material using mercury intrusion methods is quite difficult due to the heterogeneous structure of the wood, and that virtually all of the fibres must be cut to avoid errors caused by pit pores. In this study, measurements were made on 10 different samples and, as shown in Figure 2, six of them (Samples 1, 2, 3, 7, 8 and 9) gave consistent results. The pores in untreated spruce wood with a diameter above 20 µm may represent resin canals and earlywood tracheids, whilst those between 20 and 6 µm mostly represent latewood tracheids35. However, there is a relatively large overlap between earlywood and latewood tracheids. Pore diameters smaller than 6 µm represent the pointed ends of all cell elements, pits and the voids of the pit membrane36. In Table 1, the values of the intrusion pore volume and porosity, determined using mercury intrusion porosimetry, are shown. The average diffusion accessible intrusion pore volume and the porosity measured for the piece of Sw were 0.99±0.3 l/kg and 57.8±8.9 %, respectively.

Figure 2. Pore size distribution of 10 samples prepared from 2 untreated Sw pieces of Norway spruce. Samples 1-3 and 7-9 came from one piece of wood and Samples 4-6 and 10 from the other.

Table 1

8

Table 1 Normal bulk density, specific cell wall density, total porosity, intrusion pore volume, Porosity (Hg-Porosimetry) of Norway spruce wood

This study Sample Cell wall

density (ρp)

(g/cm3)

Normal bulk

density (ρb)

(g/cm3)

Total porosity

n = 1-(ρb/ ρp)

%

Total

intrusion pore

volume

(mL/g)

(0 to 330 µm)

Total porosity

(Hg

Porosimetry)

%

(0 to 330 µm)

Effective

transport-

through

intrusion pore

volume (mL/g)

(0 to 41 µm)

Effective

transport-

through

porosity (Hg

Porosimetry)

%

(0 to 41 µm)

Sw flour 1.5±0.0(n*=3) 2.9±0.1(n=3) 81.5±0.7(n=3) 0.45±0.0(n=3) 39.9±1.0(n=3) Sw piece

0.41±0.0(n=3) 72.4±0.3(n=3) 1.1±0.4(n=10) 59.7±8.6(n=10) 0.99±0.3(n=10) 57.8±8.9(n=10)

Reference34

(Plötze, M

& Niemz,

p. (2011))

Norway spruce

1.524 0.401 73.68 68.39

*n = number of samples that are prepared from two different wood pieces.

9

Equilibrium sorption studiesEquilibrium sorption studiesEquilibrium sorption studiesEquilibrium sorption studies Figure 3 shows the amount of Li+ ions sorbed onto solid wood portions of Sw flour samples (qwood) at 40 oC, in sorption experiments with 1 M LiCl aqueous solution, for contact times between 0.5 and 48 h. It was observed that an apparent equilibrium was reached quite quickly. However, the experimental data (qwood) measured at 8, 24 and 48 h was used to calculate the average value (qwood,e) that was used in the calculation of equilibrium partition coefficient (k), which was then used in the model for estimating the transport properties. For the conditions used in this study, the value of qwood,e and k of Sw flour was 5.7±0.6 g/kg and 0.8±0.1 l/kg, respectively.

Figure 3. Variation in the amount of Li+ ions sorbed onto Sw flour from 1 M LiCl bulk solution with time at 40 oC.

Estimation of transport propertiesEstimation of transport propertiesEstimation of transport propertiesEstimation of transport properties Figure 4 shows the experimental concentration profiles obtained for Cube a (in the middle) and Cube c (close to the edge) in the piece of Sw after different treatment times at 40 oC; it is evident that the concentration profiles differ. This is because the diffusion of Li+ ions is mainly influenced by mass transport in the transversal direction in Cube a, but in both the transversal and the longitudinal directions in Cube c. In the case of Cube a, a shift in the concentration profiles of the Li+ ions was observed with an increase in concentration in the centre of the wood piece as the residence time increased from 36 to 168 h (Figure 4 (A)). The results for Cube a seems to be reasonable, with the exception of the outer layers. In the case of Cube c, very high concentrations were obtained already after 36 h of residence time (Figure 4 (B)). One possible reason for this, is the methods that were used in this study for the preparation of wood pieces: The four surfaces parallel to the longitudinal fiber direction has been planed by using a hand plane and thus a large fraction of the micro-cracks has been eliminated. Similar approach was not possible to do with the two surfaces perpendicular to the longitudinal fiber directions and thus at these two surfaces it can be expected to exist a substantial amount of micro-cracks e.g. cracks in the middle lamella. However, at longer residence times, i.e. after 168 h, the results of Cube c seemed to be very close to an apparent equilibrium (Figure 4 (B)-(c)). The behaviour observed at the outer layers (e.g. at longer residence times, i.e. 168 h) and at the edge (Cube c) has also been observed in

10

earlier studies31, and is thought to be due to micro-cracks caused by the mechanical preparation of the wood pieces (e.g. during sawing).

Figure 4. Measured and calculated concentration profiles of Li+ ions in Norway spruce Sw at 40 oC for (A) Cube a (middle) and (B) Cube c (edge) taken from the impregnated piece of wood after the treatment times of (a) 36, (b) 72 and (c) 168 h.

The lines in Figure 4 represent the best fit of the model. As can be seen, the model fits the experimental concentration profiles reasonably well for Cube a (Figure 4 (A)) but the fit was not satisfactory for Cube c, especially at residence times of 36 and 72 h (Figure 4 (B)-(a) and (b)). Although it has not been possible to verify the concentration profiles at shorter times for the edge portions of the piece of wood in this work, a rough estimation of the longitudinal diffusion coefficients and tortuosity factors have nevertheless been made. The estimated values of De,T(=R) and De,L are 1.7±0.7x10-11 m2/s and ~1x10-9 m2/s, respectively at 40 oC. For the Sw samples, the predicted De,T(=R) is about two orders of magnitude less than for De,L. This result is reasonable since the diffusion of chemicals/ions in the transversal direction takes place mainly through cell walls and pits whereas in the longitudinal direction, it is mainly through liquid in the fibre lumen. Tortuosity factors for Sw samples of Norway spruce were calculated using Equation 6 in two diffusional directions, i.e. longitudinal and transversal, and found to be ~1.2 and 76±35 respectively. The tortuousness is smaller in the longitudinal than in the transversal direction, as expected. This result is understandable because the transport of Li+ ions in the longitudinal direction occurs mainly through the lumen openings in tracheids and the distances through the pit pores are short. At the edge part of the wood piece, the sawing may also cause some fiber separation in longitudinal direction, thus some micro cracks between the fibers may have been formed which may enhance longitudinal mass transport. In the transverse direction, however, the Li+ ions are transported mainly through pit pores, which have rather small areas and their positions along the fiber wall causes the tortuosity.

11

Case study: Case study: Case study: Case study: Testing the DTesting the DTesting the DTesting the De,T(=R) e,T(=R) e,T(=R) e,T(=R) and Dand Dand Dand De,Le,Le,Le,L predicted by the predicted by the predicted by the predicted by the model at shorter model at shorter model at shorter model at shorter

impregnation timesimpregnation timesimpregnation timesimpregnation times The estimated values of De,T(=R) and De,L were tested by using them to calculate the concentration profiles at shorter residence times (4 and 12 h), which were compared with the experimental data from an earlier test series31 (Figure 5). The experimental procedures employed and conditions used to measure the concentration profiles at shorter residence times31 were almost identical to those in this paper. The difference is that the pieces of wood used in the present work were pre-impregnated with dilute acid solution, which was not the case in the experiments at shorter residence times. For the case of Cube a, Figure 5 shows that, the calculated and experimental concentration profiles agree reasonably well only at the outer layer of the piece of wood. In the inner structure of the piece of wood (>1 mm in thickness), on the other hand, the differences between the calculated and experimental concentration profiles becomes larger. For the case of Cube c, the calculated and experimental concentration profiles do not agree well (results not shown here) and, the differences observed between the calculated and experimental concentration profiles are much larger compared with Cube a. One reason for the differences may be due to structural changes that might have occurred in the wood due to treatment with dilute acid prior to impregnation by the LiCl. Another is that two pieces of wood never have exactly the same properties, which was observed in our earlier studies31.

Figure 5. Measured and calculated concentration profiles of Li+ ions in Norway spruce Sw at 40 oC for Cube a (middle portion of the impregnated piece of wood) after treatment times of (a) 4 and (b) 12 h.

Conclusions

• A methodology (both an experimental procedure and a model) for the diffusive mass transport of chemicals into a porous structure has been adapted to wood.

• Reasonable results were obtained, although defects (micro-cracks) in the pieces of wood remained and were detectable.

• The transversal effective diffusion coefficient for Li+ ion in Norway spruce Sw for the conditions used in this study was calculated as being 1.7±0.7x10-11 m2/s at 40 oC. The longitudinal effective diffusion coefficient was found to be approx. 60 times the transversal.

ACKNOWLEDGEMENT

The authors are thankful to the Chalmers Energy Initiative (CEI) programme for their financial support.

12

References

1. Akhtaruzzaman, A.F.M. and Virkola, N-E. Influence of Chip Dimensions in Kraft Pulping Part I. Mechanism of Movement of Chemicals into Chips. Paperi ja Puu. 1979, 9, 578.

2. Stamm, A.J. Bound-water Diffusion into Wood in the Fiber Direction. Forest Prod. J. 1959, 9(1), 27.

3. Stamm, A.J. Bound-water Diffusion into Wood in the across-the-Fiber Direction. Forest Prod. J. 1960, 10(10), 524.

4. Kumar, S. and Jian, V.K. Diffusion through Wood. Diffusion of Boric Acid. Holsforsch. Holzvertwert. 1973, 25, 21.

5. Yata, S., Mukudai, J. and Kajita, H. Morphological Studies on the Movement of Substances into the Cell Wall of Wood. II. Diffusion of Copper Compounds into the Cell Wall. Mokuzai Gakkaishi. 1979, 25(3), 171.

6. Yata, S., Mukudai, J. and Kajita, H. Morphological Studies on the Movement of Substances into the Cell Wall of Wood. III. Diffusion of Zinc Compounds into the Cell Wall. Mokuzai Gakkaishi. 1981a, 27(11), 761.

7. Yata, S., Mukudai, J. and Kajita, H. Morphological Studies on the Movement of Substances into the Cell Wall of Wood. II. Diffusion of Hexavalent Chromium into the Cell Wall. Mokuzai Gakkaishi. 1981b, 27(12), 821.

8. Vinden, P. The Effect of Raw Material Variables on Preservative Treatment of Wood by Diffusion Process. Journal of the Institute of Wood Science. 1984, 10(1), 31.

9. Cooper, P. A. Diffusion of Copper in Wood Cell Walls following Vacuum Treatment. Wood Fiber Sci., 1998, 30(4), 382.

10. Ra, J. B., Barnes, H. M. and Conners. T. E. Determination of Boron Diffusion Coefficients in Wood. Wood Fiber Sci. 2001, 33(1), 90.

11. Jeremic, D., Quijano-Solis, C. and Cooper, P. Diffusion Rate of Polyethylene Glycol into Cell Walls of Red Pine following Vacuum Impregnation. Cellulose. 2009, 16, 339.

12. Narayanamurti, D. and Kumar, V. B. Diffusion of Organic Molecules through Wood. Journal of Polymer Science. 1953, 10(6), 515.

13. Fukuyama, M. and Urakami, H. Diffusion of Nonelectrolytes through Wood Saturated with Water I. Mokuzai Gakkaishi. 1980, 26, 587.

14. Fukuyama, M. and Urakami, H. Diffusion of Nonelectrolytes through Wood Saturated with Water II. Mokuzai Gakkaishi. 1982, 28, 17.

15. Fukuyama, M. and Urakami, H. Diffusion of Nonelectrolytes through Wood Saturated with Water III. Mokuzai Gakkaishi. 1986, 32, 147.

16. Tillman, L.M., Lee, Y.Y. and Torget, R. Effect of Transient Acid Diffusion on Pretreatment/Hydrolysis of Hardwood Hemicellulose. Applied Biochemistry and Biotechnology. 1990, 24(25), 103.

17. Meijer, M., van der Zwan, R. P. and Militz, H. Unsteady-State Diffusion of Methanol in Douglas-fir Heartwood at High Temperatures. Holzforschung. 1996, 50, 135.

18. Narayanamurti, D. and Ratra, R.S. Diffusion of Ions through Some Indian Timbers. Proceedings Mathematical Sciences. 1951, 33(6), 349.

19. Christensen, G.N. and Williams, E.J. Diffusion in Wood I. A Quantitative Theory of Diffusion in Porous Media and its Application to Wood. Austral. J. Appl. 1951a, 2(4), 411.

20. Christensen, G.N. Diffusion in Wood III. Ion Selection and its Effect on the Diffusion of Electrolytes. Austral. J. Appl. 1951b, 2(4), 440.

21. Saltberg, A., Brelid, H. and Theliander, H. Removal of Metal Ions from Wood Chips during Acidic Leaching 2: Modeling Leaching of Calcium Ions from Softwood Chips. Nordic Pulp and Paper Research Journal. 2006, 21(4), 513.

13

22. Mc Kibbins, S.W. Application of Diffusion Theory to the Washing of Kraft Cooked Wood Chips. Tappi. 1960, 43, 801.

23. Talton, J.H. and Cornell, R.H. Diffusion of Sodium Hydroxide in Wood at High pH as a Function of Temperature and Degree of Pulping. Tappi. 1987, 115.

24. Robertsen, L. and Lönnberg, B. Chemical Diffusion in Wood. Proc. 6th Int. Symp. on Wood and Pulping Chemistry; Appita, Australia. 1991, 179.

25. Kazi, K.M.F. and Chornet, E. A Diffusion Model for the Chemical Impregnation of Hardwoods and its Significance for Rapid Steam Treatments. Paperi ja Puu-Paper and Timber. 1998, 80(1), 41.

26. Törnqvist, M., Hurme, T. and Rosenholm, J.B. Drift Speed: A Way of Measuring Diffusion and Tortuosity of Porous Materials. Colloids and Surfaces A: Physicochemical and Engineering Aspects. 2001, 180, 23.

27. Ekman, K. H. and Fogelberg, B. C. Impregnation of sulphate cooking studied by means of radioactive sulphur. Paperi ja Puu. 1966, 48(4), 175.

28. Stone, J.E. The Effective Capillary Cross-Sectional Area of Wood as a Function of pH. Tappi. 1957, 40(7), 539.

29. Sharareh, S., Tessier, P. and Lee, C-L. Penetration of Sodium Sulphite into Black Spruce and Aspen Wood Chips using SEM/EDXA. Journal of Pulp and Paper Science. 1996, 22(3), J71.

30. Kolavali, R. and Theliander, H. Determination of the Diffusion of Monovalent Cations into Wood under Isothermal Conditions based on LiCl Impregnation of Norway spruce. Holzforschung. 2013, 67(5), 559.

31. Kolavali, R. and Theliander, H. Experimental Determination of the Diffusion of Monovalent Cations into Wood: Effects of Micro-Cracks, Wood Structure, Impregnation Time, and Temperature on Local Concentration Profiles. J-FOR; Journal

of science & technology for forest products and processes. 2014, 4(2), 29. 32. Kolavali, R., Merima, H. and Theliander, H. The sorption of monovalent cations onto

Norway spruce: model studies using wood flour and LiCl solution (Manuscript ready to submit).

33. Saripalli, K. P., Serne, R. J., Meyer, P. D. and McGrail, B. P. Prediction of Diffusion Coefficients in Porous Media Using Tortuosity Factors Based on Interfacial Areas. Ground Water. 2002, 40(4), 346.

34. Plötze, M. and Niemz, P. Porosity and Pore Size Distribution of Different Wood Types as Determined by Mercury Intrusion Porosimetry. Eur. J. Wood Prod. 2011, 69, 649.

35. Zauer, M., Hempel, S., Pfriem, A., Mechtcherine, V. and Wagenfuhr, A. Investigations of the Pore-Size Distribution of Wood in the Dry and Wet State by Means of Mercury Intrusion Porosimetry. Wood Sci Technol. 2014, 48, 1229.

36. Thygesen, L. G., Engelund, E.T. and Hoffmeyer, P. Water Sorption in Wood and Modified Wood at High Values of Relative Humidity. Part I: Results for Untreated, Acetylated and Furfurylated Norway spruce. Holzforschung.2010, 64, 315.

14

Appendix A1 Cube model - Estimation of the initial guess of effective diffusion coefficients

in the transversal direction Cubical geometry was selected to determine the range of the initial guess of the effective diffusion coefficient in the transversal (De, R = De, T) direction. In the case of Cube a (i.e. the middle part of the impregnated piece of wood, Figure 1 (A)), it is assumed that concentration is not affected by mass transport in the longitudinal direction because the distances involved are much longer than in the radial and tangential directions. Furthermore, the effective diffusion coefficient in the radial direction is assumed as being equal to the effective diffusion coefficient in the tangential direction. Consequently, Cube a is used to determine the effective diffusivity in both the radial and tangential directions.

The dimensions of the geometry were 6 x 6 x 6.5 (L x R x T) mm3. The mesh consists of approximately 32105 elements, which was sufficient to obtain a solution that was independent of the mesh. The simulations were run for 0 to 10,080 min (i.e. 0 to 168 h) with a time interval of 1 min; both absolute and relative tolerances were kept to 10-3. The computational time was 2 to 3 min. The anisotropic diffusion coefficient matrix is defined as:

O0 0 00 H",Q 00 0 H",RS (A1)

The value of H",Q(TR) was provided as the intial guess in the model to calculate the concentration profiles that correspond to the centre portion of the wood piece (i.e. Cube a) in the actual experimental data. The initial guess values were varied within the broad range of diffusion coefficients expected for solid material (Welty et al. 1984), i.e. from 10-14 to 10-10 m2/s, until the differences between the experimental and calculated concentration profiles were minimal.

Appendix A2 Estimation of the diffusion coefficient of Li+ ions in an aqueous medium (free

solution)

The diffusion coefficient of Li+ ions in an aqueous medium (free water) (Df) at temperature T (oC) and for a concentration of the aqueous medium Caq (M) can be estimated from the following correlation, using viscosity data of the aqueous medium at a corresponding temperature and concentration (Poling et al. 2001):

HU,9�6R℃ = HU,9�6WX℃ ∗ Y R℃WX℃Z ∗ [\]^_`a℃\]^_b℃ c (A2)

where HU,9�6R℃ and HU,9�6WX℃ are the diffusion coefficients of the Li+ ion in an aqueous medium at

temperature T (oC) and 25 oC, respectively and d3^_R/] and d3^_WX/] are the viscosities of an aqueous solution at concentration C, at temperature T (oC) and 25 oC, respectively.

Df,Li+ in an aqueous solution of 1 M LiCl at 40 oC was estimated using Equation A2 and the

data from the literature presented in Table A1. Df,Li+ at 40 oC was found to be 2.0x10-9 m2/s.

15

Table A1

Viscosity of 1 M LiCl solution and self-diffusivity of Li+ ions in 1 M LiCl solution at two

different temperatures

Temperature (oC)

Viscosity (Melinder 2007)

(µ) (mPa.s)

Self-diffusivity (Braun and Weingärtner 1988)

(Df,Li+) (m2/s)

25 1.023 0.924x10-9

40 0.753 -

References

Braun, B. M. and Weingärtner, H. Accurate Self-Diffusion Coefficients of Li+, Na+, and Cs+ Ions in Aqueous Alkali Metal Halide Solutions from NMR Spin-Echo Experiments. J. Phys. Chem. 1988, 92, 1342. Melinder, Å. Thermophysical Properties of Aqueous Solutions Used as Secondary Working Fluids. Doctoral Thesis. Dept. of Energy Technology. Royal Institute of Technology (KTH), Stockholm, Sweden. 2007. Poling, B. E., Prausnitz, J. M. and O’Connell, J. P. The Properties of Gases and Liquids. Fifth Edition, McGraw-Hill. 2001. Welty, J. R., Wicks, C. E., Wilson, R. E. and Rorrer, G. 24.2. The Diffusion Coefficient. In: Fundamentals of Momentum, Heat and Mass Transfer. Third Edition. John Wiley & Sons: New York. 1984.

diffusive masstransport of ions in wood

Documents