Forestcluster Fubio Biorefinery programme report

Future BiorefineryProgramme Report 2009–2011

Programme Report 2009-2011

Future Biorefinery

4

Content

Foreword

Introduction

Novel ionic liquids for wood processing

Hot water treatment of lignocellulose

Modification of xylan and galactoglucomannan

Hemicellulose and cellulose-based films and barriers

Co-polymerisation of three hydroxy acids

Dry-jet wet fiber spinning – Creating a new cellulose regeneration infrastructure

Functional cellulose beads

Cellulose/polymer blends

Improving the extensibility and formability of paper and board

A novel process for the production of dialdehyde cellulose microfibers

Stimuli-responsive materials and their applications

Papermaking with hemi-lean pulp

Selected extractives in protection of wood products and human health

Immunomodulatory effects of bark and knot extract and compounds

Anti-carcinogenic and metabolic effects of wood-derived extracts

5

6

12

24

48

56

64

74

84

96

110

122

130

142

152

162

168

Copyright Forestcluster Ltd 2011. All rights reserved.This publication includes materials protected under copyright law, the copyright for which is held by Forestcluster Ltd or a third party. The materials appearing in publications may not be used for commercial purposes. The contents of publications are the opinion of the writers and do not represent the official position of Forestcluster Ltd. Forestcluster Ltd bears no responsibility for any possible damages arising from their use. The original source must be mentioned when quoting from the materials.

ISBN 978-952-92-9718-4 (paperback)ISBN 978-952-92-9719-1 (PDF)

5

Foreword

Foreword

Introduction





Co-polymerisation of three hydroxy acids











Wood is one of the most versatile biological raw materials that is available today in large, renewable reserves around the world. Wood products have countless impor-tant industrial applications, such as in design, furniture and construction. These ap-plications have a bright future ahead. At the same time, chemical and mechanical wood processing provides the basis for a growing range of globally significant fibre-based tissue, paper and packaging applications and solutions. At the sharp end, ad-vances in the use of the individual chemicals and polymers that make up wood are creating the foundation for future biorefineries and helping change the shape of so-ciety for the better.

Human use of wood reaches back thousands of years. Today’s escalating global pop-ulation and limited natural resources, however, call for new ways of improving the ef-ficiency of our use of this vital natural resource. This opens up significant opportuni-ties for products based on renewable, non-food materials (‘non-food bio-products’). In the face of current oil price and sustainability challenges, the bio-economy concept is fast winning ground. Could an increasing share of our consumer products be pro-duced using renewable raw materials, like wood, instead of non-renewables, like oil?

The forest-based industry sees this opportunity and believes that the industry will play a decisive role in the development towards a bio-economy. With this goal in sight, diversifying the product output of the primary wood refining process – pulping – is a rational strategic starting point. The pulp mills of today are being redefined as the biorefineries of tomorrow.

The change requires new competencies. This is happening on several levels. In Fin-land, the foundation for this development is being laid by the joint research company Forestcluster Ltd (www.forestcluster.fi) and especially its Future Biorefinery strategic area (FuBio). FuBio is a 5-year undertaking with a volume of around EUR 50 million. This programme report covers most of the main results of the first two years of FuBio, i.e. the ‘FuBio Joint Research 1’ research programme. Most of the themes of Fu-Bio Joint Research 1 are now being carried forward either in the continuation of FuBio or as company-lead development projects.

Mikael HannusStora Enso Oyj

Chairperson of the FuBio Joint Research 1 Research Programme Management Group

6

Introduction

1. Background The forest-based industry is a pillar of the global bio-economy. In Europe, the industry, including forestry, is estimat-ed to have employed approximately 4.8 million persons in 2009 (CEPI and CEI-BOIS). The Public-Private Partnership For-estcluster Ltd was established in 2007 to accelerate the development of the Euro-pean knowledge-based bio-economy from a Finnish perspective. The initial strategy process of Forestcluster identified three focus areas for R&D: ‘Forward Custom-er Solutions’ strives to develop new busi-ness models and market ideas for exist-ing products of the pulp and paper indus-try. The focus of ‘Intelligent, Resource-Efficient Production Technologies’ is on improving the resource efficiency of pulp and paper mills and modifying them for the production of novel bio-products. ‘Fu-ture Biorefinery’, or FuBio, is developing the scientific and technological basis for transitioning from mills’ existing markets to markets where wood-based products currently have no or only a minimal pres-ence. If successful, such new bio-prod-ucts would replace existing petroleum-based products and thus bring consider-able sustainability benefits. FuBio is also developing pioneering processes for the fractionation of wood, thus generating en-tirely new bio-materials for further pro-cessing and a range of new high-poten-tial applications.

The Forestcluster strategy process was continued, and in terms of FuBio, po-tential target markets for the diversifica-tion process were identified. The outcome is summarised in Figure 1.

FuBio is planned to run for 5 years. It was initiated by a 2-year research pro-

gramme, FuBio Joint Research 1 (March 2009–May 2011), with a total volume of about EUR 19 million financed by the owners of Forestcluster and the Finnish Funding Agency for Technology and Inno-vation (Tekes). The owners of Forestclus-ter include (August 2011) nine compa-nies (Andritz, Kemira, Metso, Metsä-Bot-nia, Metsäliitto Cooperative, M-Real, Myl-lykoski, Stora Enso and UPM-Kymmene), two research organisations (the Finnish Forest Research Institute (Metla) and VTT Technical Research Centre of Finland), as well as eight universities (Aalto Universi-ty, Lappeenranta University of Technolo-gy, University of Eastern Finland, Univer-sity of Helsinki, University of Jyväskylä, University of Oulu, Tampere University of Technology and Åbo Akademi University).

2. Structure and participants

FuBio Joint Research 1 comprised five Themes (see Figure 2). A sixth Theme (T4) was also planned, but it did not be-gin during FuBio Joint Research 1.

The programme partners included the owners of Forestcluster Ltd. Additional re-search was also subcontracted to external partners, primarily Danisco Sweeteners, Finex, GloCell, Pharmatest Services (Or-thotopix), Pöyry Management Consulting, Separation Research, University of Tam-pere and University of Turku. Leading in-ternational expertise was also included in the programme. Key collaborators in-cluded Karlsruhe Institute of Technolo-gy (Germany), RWTH Aachen (Germany) and FPInnovations (Canada).

7

Figure 1. The six target markets of FuBio.

Figure 2. Structure of the FuBio Joint Research 1 research programme (March 2009-May 2011).

8

3. Management

FuBio Joint Research 1 was managed by a Management Group (MG). The execu-tion of the programme was headed by the Programme Manager together with theme-specific leaders and tutors. The MG had the following members:• Mikael Hannus, Stora Enso,

Chairman of the MG• Lars Gädda, Forestcluster• Eeva Jernström, UPM-Kymmene• Annaleena Kokko, VTT Technical

Research Centre of Finland, Theme 3 Leader

• Jukka Leppälahti, Tekes• Paterson McKeough, Andritz• Jussi Mäntyniemi, Metso• Klaus Niemelä, VTT Technical

Research Centre of Finland, Theme 1 Leader

• Erkki Peltonen, Myllykoski• Ismo Reilama, Metsä-Botnia• Kari Saari, Kemira• Pekka Saranpää, Finnish Forest

Research Institute (Metla), Theme 5 Leader

• Anna Suurnäkki, VTT Technical Research Centre of Finland, Theme 2 Leader

• Niklas von Weymarn, VTT Technical Research Centre of Finland, Programme Manager

The tutors were Herbert Sixta, Aalto University (Theme 1), Lars Gädda (Theme 2), Ali Harlin, VTT Technical Re-search Centre of Finland and Maija Ten-kanen, University of Helsinki (Theme 3), as well as Tiina Nakari-Setälä, VTT Tech-nical Research Centre of Finland and Bjarne Holmbom, Åbo Akademi Univer-sity (Theme 5).

4. Specific research areas

The specific research areas of FuBio Joint Research 1 are on the following pages portrayed as five concepts. The yellow boxes illustrate existing industrial oper-ations, the green boxes are the process-es that FuBio Joint Research 1 aims to in-tegrate into the existing operations. The fourth box could also be ‘Paper produc-tion’ (now ‘Board production’).

5. Future plans Two new research programmes were launched on June 1, 2011. The first of these, FuBio Joint Research 2, is pursu-ing four of the target markets outlined in Figure 1, namely ‘Structural compos-ites’, ‘Novel packaging and filtration ma-terials’, ‘Polymers, resins and chemicals’, and ‘Health-promoting products’. The ‘Regenerated fibre and chemicals’ target market is the focus of the second new research programme, ‘FuBio Products from dissolved cellulose’. The sixth target market, ‘Wood preservatives and glues’, is currently in the planning stage (as of September 2011).

This report summarises the results of the first two years of research activity within FuBio. The structure of the report mirrors the structure of the research pro-gramme itself.

9

Figure 4. The concept has two goals: i) to develop novel bio-based materials to replace current petroleum-based materials (paper chemicals and coating), and ii) to study ways to add mouldability to the fibre web.

Figure 3. The concept comprises the separation of either lignin or organic acids from black liquor and the consequent upgrading of the said components to novel bio-products. The acids could alternatively be produced directly from wood-based sugars (‘sugar platform’).

Wood harvest

Wood handling

Pulp production

Board production

Converting & printing

Chemical recovery

Water treatment

Additives

Energy By-products

Residues Bark Knotwood

Additives

Lignin sep.

Acid sep. Moulding

Fibre Additives

Composites

Upgrading

Chemicals Resins Polymers

Fermen- tation

Wood-based sugar

Wood harvest

Wood handling

Pulp production

Board production


Chemical recovery

Water treatment New bio-

barriers

Energy By-products


New paper chemicals

Fibre modification for improved mouldability

10

Figure 6. The concept comprises the development of new processes for cellulose dissolution and regeneration.

Figure 5. The concept comprises the extraction of compounds from wood and using them in different applications (from health promotion to the protection of wood products).

Wood harvest

Wood handling

Pulp production

Board production


Chemical recovery

Water treatment

Additives

Energy By-products


Additives

Tall oil GGM

Product protection

(high volume)

Health products, (food and medicine;

high value)

Wood harvest

Wood handling

Pulp production

Dissolution & Regeneration Converting

Chemical recovery

Water treatment

Energy By-products

Residues Bark Knotwood Derivatisation

Thermally formed products Cellulose beads (Textiles) (Nonwovens)

Chemicals

11

Figure 7. The concept comprises the development of completely new processes for the fractionation of wood, thus generating novel materials for further upgrading. The processes are based on hot water treatment and/or novel ionic liquids.

On behalf of Forestcluster Ltd and myself, I extend my sincere gratitude to all partici-pants from industry and the research community for their active efforts in getting this programme successfully off the ground. Together, we have taken the first key step to-wards shaping the biorefineries of the future.

Niklas von Weymarn

VTT Technical Research Centre of FinlandProgramme Manager, FuBio Joint Research 1 Research Programme

Wood/ saw dust

New fraction.

Cellulose/fibre (incl. Theme 2 applications)

Hemicelluloses: •  Barriers •  Films

Lignin

12


Editor

Partners:

Aalto University

Åbo Akademi University

University of Helsinki

Alistair W. T. King

Key researchers:

Herbert Sixta, Michael Hummel, Lauri K. J. Hauru, Anne Michud

Jyri-Pekka Mikkola, Päivi Mäki-Arvela, Pasi Virtanen, Ikenna Anugwom

Ilkka Kilpeläinen, Alistair W. T. King, Pirkko Karhunen, Jorma Matikainen, Arno Parviainen, Timo Leskinen, Paula Järvi

13

AbstractNew concepts in wood processing with novel ionic liquids, as potentially environmentally benign

reaction media, are presented. These include a refinement in our understanding about the ef-

fects that these unique solvators have on woody material and examples of the new structures

under development, which offer increased process sustainability over the initial generations.

In addition to selective dissolution of purified wood biopolymers, wood can now be fibrillated

using ionic liquids, under mild treatment conditions. This can be achieved without chemical mod-

ification of the material or extracts. The process is accompanied by removal of pectins, including

small amounts of mobile lignin and hemicelluloses, which are extracted from the middle lamella.

As lignin is not extracted completely and the fibres maintain their strength, the new material is,

in a way, similar to thermo-mechanical pulp (TMP). Surprisingly from a processing point of view,

it seems that fibrillation of wood does not require chemical fragmentation of lignin and can be

achieved under relatively mild conditions using ionic liquids. In consideration to wood fraction-

ation, different ionic liquids and treatment conditions are also available that will allow for dissolu-

tion of all wood biopolymers, including wood itself and selective extraction of wood biopolymers

from the wood matrix. Systematic screening and literature review has resulted in our improved

understanding of the factors that affect these effects and phenomena.

Ionic liquid recyclability has been foremost on our minds during this research period. The re-

sult is the discovery that new classes of ‘switchable’ and ‘distillable’ ionic liquids can effective-

ly lignocellulose. This takes advantage of the electron density afforded by superbases such as

1,8 - diazabicycloundec - 7 - ene (DBU) and 1,1,3,3 - tetramethylguanidine (TMG) as bulk

chemicals. Whereas previous structures have low recyclabilities, the new structures can be con-

verted from their ‘ionic’ form into neutral species, which allows for distillation of the materials in

high yields and recovery. This increases the overall sustainability of the prospective processes,

beyond what was capable before and offers significant energy savings.

Overall we have developed understanding that makes us highly competitive, on the interna-

tional scale in the area of bioprocessing with ionic liquids.

14

1. Background

Ionic liquid bioprocess sustainabilityWood is regarded as a leading sustainable resource, to replace fossil fuels. However, challenges will exist in the incorporation of wood-based feedstocks and process-es into, traditionally, fossil-based value chains. Therefore, new ‘tuneable’ meth-ods offering more varied and improved selectivities, for the fractionation and pro-cessing of wood biopolymers, into materi-als, chemicals and energy are necessitat-ed. The reported high efficiency in the dis-solution of cellulose by ionic liquids (ILs), has thus afforded new processing oppor-tunities, whereby wood itself can be ef-fectively solvated and processed accord-ingly. In addition, room temperature ion-ic liquids (RTILs), such as 1-ethyl-3-me-thylimidazolium acetate ([Emim][OAc], m.p. -45°C), offer so effective cellulose solvation capabilities that they are now considered to be industrially viable me-dia for existing and novel cellulose pro-cessing applications. An important ex-ample is in the replacement of N-methyl-morpholine-N-oxide hydrate (NMMO·H2O) in a ‘Lyocell’ process, circumventing haz-ardous thermal stability issues. [Emim][OAc] has been so successful in cellulose solvation that BASF is now producing it on a ton-scale. Publications are appear-ing, however, that highlights the insta-bility of [Emim][OAc], in the presence of lignocellulosic solutes. Basic ILs such as [Emim][OAc] are also known to have re-duced thermal stabilities. This effectively prevents the recovery of the IL on an in-dustrial-scale by distillation. As such, oth-er methods of recycling or more recycla-ble structures/systems are sought.

2. Objectives

The main application objectives of this work was in the development of process-es that utilise ionic liquids:• Assess and develop the potential

for fractionation of wood into its components using ionic liquids as solvating media

• Assess and develop the potential of ionic liquids in the dissolution and regeneration of cellulose-rich pulp to novel materials; in particular, develop new fibre-spinning processes, based on the Lyocell process, by dissolution of pulp into ionic liquids and subsequent regeneration by air-gap spinning into water

• Develop recyclable and low toxicity ionic liquid systems to maintain sustainability of processing

On the whole, the objectives called for both assessments of existing structures for their efficiency in the fractionation of woody material and also for the develop-ment of new ionic liquids. This was nec-essary due to the poor stability and re-cyclability of existing structures. A gen-eral pre-requisite for new ionic liquids for the above application was that they be effective at dissolving both lignin and wood polysaccharides such as cellulose or hemicelluloses.

3. Research approachDue to the infancy of ionic liquids re-search and the structural complexity of ionic liquids, in comparison to molecular solvents, this work package demanded a more academic approach. This was con-ducted alongside assessment ionic liquids for their efficacy for the processing of lig-nocellulosics. Three main areas were fo-cused on (Figure 1).

15

The research strategy outlined in Figure 1 can be summarised as follows:• The starting point was the synthesis

of existing imidazolium-based ionic liquids

• The next step was to further our understanding of different effects and phenomena that occur when lignocellulosics are contacted with ionic liquids (e.g. dissolution capabilities, fractionation efficiency, fibrillation, chemical reaction).

• Develop quantitative structure property relationships in regard to the physical properties of the ionic liquids (e.g. thermal stabilities) and the observed effects and phenomena (mainly biopolymer solubilities). This was achieved through a process of parameterization of effective and non-effective ionic liquids. Computational methods were also

used to predict physiochemical properties of structures.

• Through this understanding of effects, parameters and the generation of new structures, it was possible to develop hypotheses about which structural features would allow for more advantageous effects (e.g. lower viscosities or more recyclable structures)

• New series of structures were synthesised with an improved understanding of physical properties and recycling issues, in particular

Although the above strategy requires a large set of structures and data to work most efficiently, thorough literature re-view and the development of our own da-ta resulted in the development of highly novel structures and associated intellec-tual property.

Figure 1. Strategy for academic development of novel and effective ionic liquids.

16

4. Results

4.1 Initial ionic liquids synthesis & testingA thorough literature review was per-formed looking at the solubility of ligno-cellulosics in current ionic liquids. This combined with additional screening of a range of structures (Table 1) allowed us to identify potential opportunities with re-gard to biopolymer solubilities.

A new concept, called wood chip ‘fi-brillation’, was identified for ionic liquids, which are known to be highly effective at solvating cellulose. Despite the abili-ty of ionic liquids to dissolve wood itself, under more harsh dissolution conditions, these ionic liquids were capable of reduc-ing wood chips into fibres. The products were similar to TMP (Figure 2) and fibril-lation was achieved under mild treatment conditions (110°C, 3 d, without stirring).

Table 1. Efficiency of cellulose dissolution, lignin dissolution, and wood fibrillation for in-house ionic liquid structures.

Ionic Liquid Preparation Cellulosea Lignina Fibrillationa

[mmim][Me2PO4] Synthesized Dissolves +++++ +++ (Softwood) [amim]Cl Synthesized Dissolves +++ ++ (gels) (Softwood) [amim]Br Synthesized Dissolves - (Softwood) [amim][Me2PO4] Synthesized Dissolves +++++ +++ (Softwood) [emim]Cl Merck Dissolves - (Softwood) [emim][Me2PO4] Synthesized Dissolves +++++ +++++ (Softwood) [emim][Et2PO4] Synthesized Dissolves +++++ ++++ (Softwood) [emim][SCN] Merck Degrades ++++ - (Softwood) [emim][MeHPO3] Synthesized Dissolves +++++ +++++ (Softwoods) [emim][EtHPO3] Synthesized Dissolves +++++ +++ (Softwood) [emim][HSO4] Merck Fragments - (Softwood) [emim][MeSO4] Iolitec Fragments +++++ - (Softwood) [emim][OTs] Iolitec Fragments +++++ - (Softwood) [emim][OAc] Iolitec Dissolves ++++ (Hard and Softwoods) [emim][Me2PO3] Synthesized Dissolves +++++ - (Softwood) [emim][OCOCF3] Iolitec - - (Softwood) [emim][NTf2] Merck - - - (Softwood) [eeim][Et2PO4] Synthesized Dissolves ++ (Softwood) [mmmim][Me2PO4] Synthesized ++ (Softwood) [emmim]Cl Synthesized - (Softwood) [emmim][Et2PO4] Synthesized +++ (Softwood) [prmim][Me2PO4] Synthesized +++ (Softwood) [iprmim][Pri

2PO4] Synthesized + (Softwood) [bmim][Me2PO4] Synthesized ++ (Softwood) [bmim][HSO4] Merck Fragments - (Softwood) [omim][OctSO4] Merck - - (Softwood) [hemim]Cl Iolitec - - (Softwood) [Hmim]Cl BASF - - (Softwood) [P4444]Cl Iolitec - ++++ - (Softwood) [P14444]Cl Iolitec - ++++ - (Softwood) [P14666]Cl Iolitec - ++++ - (Softwood) [P4442][Et2PO4] Iolitec ++++ - (Softwood) [P4441][OTs] Iolitec - (Softwood) [mPyr][MeHPO3] Synthesized (impure) Degrades ++ (Softwood) [DBUH][OCOC2H5] Synthesized Dissolves

a efficiency of dissolution or fibrillation: +++++ (strong effect), - (no effect)

17

Figure 2. SEM of Norway spruce chip ‘fibrillated’ from 1-ethyl-3-methylimidazolium dimethylphosphate ([emim][Me2PO4]) at 110°C for 3 d.

Of these structures, [Emim][Me2PO4] was found to be most effective for soft-woods such as spruce and pine. [Emim][OAc] was found to be highly effective for both softwoods and hardwoods such as birch and aspen. Wood chips were about 2 * 2 * 0.25 cm and pre-dried by solvent exchange with acetone. Sugar analysis and NMR analysis of the chips and resid-ual ionic liquid showed that mainly pec-tins were removed from the middle lamel-la. Small quantities of lignin and hemicel-luloses were also removed but essential-ly the fibrillation occurred, without bulk fragmentation of removal of cellulose, hemicellulose or lignin. 1-Ethyl-3-methy-limidazolium methylphosphonate ([Emim][MeHPO3]) was also found to be very ef-fective at fibrillating softwoods but it be-came apparent that the ionic liquid was reacting with the fibres. No reaction or complexation with fibres was observed with [Emim][Me2PO4] or [Emim][OAc]. The discovery resulted in the filing of a patent application.

4.2 Ionic liquid parameterizationWhile basic imidazolium-based ionic liq-uids (e.g. [Emim][OAc]) are known to be highly effective at dissolving cellulose, several reports have appeared, including those from FuBio, indicating that they are capable of reacting with various lignocel-lulosic functionalities. To learn more about the various parameters that affect both dissolution and reactivity, two methods of parameterisation were chosen:• Kamlet-Taft solvatochromic

parameterisation; prediction of α (H-bond acidity), β (H-bond basicity) and π* (dipolarity/polarisability). This is the most comprehensive parameterization strategy to date for ionic liquids and should allow for quantitative understanding of how the ionic liquids structural features effect solvation through the development of linear solvation energy relationships. An initial manuscript is ready for submission, which more

18

comprehensively demonstrates the effect of the ionic liquids H-bond basicity, as a requirement for dissolution. This was applied to existing ionic liquids, novel TMG-based ionic liquids and traditional molecular solvents such as LiCl/DMAc and NMMO•H2O (Figure 3). Reduction in basicity was also observed to be the major factor in the regeneration of pulp from ionic liquid, upon addition of water.

• Ab initio proton affinity computational calculations on anions and neutral bases, as a measure of the enthalpy of deprotonation in the gas-phase (i.e. measure of basicity or nucleophilicity); Allows for prediction of relative inherent basicity, acidity and nucleophilicity of ILs as properties predicting cellulose solubility and thermal

stability. From a process of measuring the thermal stability of imidazolium-based ionic liquids and comparison with the proton affinities for those anions, it was possible to see an approximate correlation (Figure 4). It is also known the main mechanism of decomposition of basic imidazolium-based ionic liquids is through nucleophilic attack of the anion on the imidazolium alpha-positions. This indicates that as the ability of ionic liquids to dissolve cellulose increases (increasing basicity), their thermal stability decreases (increasing nucleophilicity). This is consistent with the fact that [emim][OAc], as a basic ionic liquid, is now known to react with lignocellulosics. This also has additional implications for the recyclability sustainability of processes using imidazolium-based ionic liquids as many of

Figure 3. The difference β-α (aggregate basicity - essentially basicity of the anion minus the acidity of cation) plotted against β (basicity); full symbols are effective cellulose solvents, empty symbols non-solvents, half-empty poor solvents; LiCl/DMAc data shown in the range 20–100°C.

19

them decompose under the practical processing conditions that available for distillation of these materials. Under some circumstances this is not such an issue, e.g. cellulose dissolution/regeneration but for more complex systems, including those involving wood, sustainabilities can drop quickly under the wrong processing conditions.

4.3 ‘Distillable’ ionicliquidsThe calculation of proton affinities for neutral bases also had the benifit of al-lowing us to predict which combination of acids and bases would be effective at dis-solving cellulose (Figure 5).

For example, [Emim][OAc] can be thought of a combination of acetic ac-id and the 1-ethyl-3-methylimidazol-2-ylidene carbene ([Emim]:), as the C2-H is the most acidic position on the im-idazolium ring. The proton affinity for this neutral species is -262.9 kcal mol-1 at the MP2/6-311+G(d,p)//MP2/6-311+G(d,p) level of ab initio theory. Im-ino-tris(dimethylamino)phosphorane (PhosP1), 1,8-diazabicycloundec-7-ene (DBU) and 1,1,3,3-tetramethylguanidine

Figure 4. The correlation of proton affinities vs decomposition temperatures.

(TMG) with proton affinities of -253.9, -244.9 and -248.9 kcal mol-1 respective-ly, all dissolve cellulose, in combination with acetic acid. When the proton affin-ity values of neutral bases rise above ~ -240 kcal mol-1 the resulting ionic liquids formed, by stoichiometric combination with acetic or propionic acid, no longer dissolve cellulose. Moreover, as the acid-ity of the cation is only just higher than the acidity of carboxylic acids, in the neat ionic liquid, it was possible to distil some of the mixtures (Figure 6). Essentially by heating the sample we could dissociate the conjugated acid and base, thus af-fording a vapour pressure to the system and allowing for distillation and recom-bination.

We recently reported a new class of ILs that both dissolved cellulose and were distillable at much lower tempera-tures and higher pressures than for imid-azolium-based ILs, such as [Emim][OAc]. In our hands we achieved > 99% recov-ery and 99% purity of distillate which is a considerable improvement over the distill-ability of archetypical structures (imidazo-liums). Undoubtedly the electron density provided by the more basic neutral bas-es allowed for formation of the ionic liq-

Figure 5. Molten acid-base conjugates, capable of dissolving cellulose, as the acetate salts.

20

uids in the first place, but also preserva-tion of that basicity, in the media, which makes them effective solvators of cellu-lose. A patent application was filed.

4.4. ‘Switchable’ ionic liquids (SILs)As in the case of the distillable ionic liq-uids, superbases such as DBU allowed for the conjugation of alcohols and gases, such as carbon or sulphur dioxide. When these materials were combined in stoi-chiometric quantities, they formed sta-

Figure 6. Distillation and X-ray crystallographic structure of [TMGH][CO2Et].

ble ionic liquids. The general structures of which were protonated DBU cations with carbonate or sulphite anions. These materials were observed to dissociate again upon heating to higher tempera-tures or upon bubbling inert gas through the mixtures. In effect the solvents could be ‘switched’ between ionic and neutral species when perturbed. This has impor-tant implications concerning the proces-sibility and recyclability of the system (Figure 7).

Several combinations of alcohols and gases with DBU were tested and wood

Figure 7. Synthesis, use and regeneration of novel switchable ionic liquids.

21

4.5 Novel imidazolium-based ionic liquidsSince both, the fractionation trials and spinning experiments demand substan-tial amounts of [Emim][OAc] of high pu-rity, cost-effective syntheses were inves-tigated. [Emim][OAc] can be prepared in excellent yields and high purity (Figure 9). This was possible through a novel me-tathesis step. Dimethylsulfite was used as alkylating agent. The methylsulfite anion

was treated with the most processable structures. The result was that it was pos-sible to extract hemicelluloses and selec-tively lignin, in some cases by treatment under mild conditions (100°C). Some of the structures even had the capability of fibrillating wood chips under mild treat-ment conditions (Figure 8). The results of the research resulted in the filing of a patent and preparation of several re-search articles.

Figure 8. Birch wood chip treated with different switchable ionic liquids. (A) Treated with (SIL #1) for 5 days at 100°C, (B) treated with (SIL #2) for 5 days at 100°C and (C) is the untreated wood. No mechanical stirring was applied during the experiments.

Figure 9. Preparation of [Emim][OAc] via methylation of imidazole with dimethylsulfite.

22

formed can be hydrolyzed and expelled as sulphur dioxide upon addition of acetic acid, leaving acetate remaining. Accurate control of the reaction conditions is re-quired to suppress the formation of meth-anesulfonate during the alkylation step.

Even though [Emim][OAc] is the working horse for the spinning and frac-tionation trials, new cellulose dissolving ionic liquids are always of interest. In co-operation with Professor Herwig Schot-tenberger from the University of Inns-bruck, Austria we synthesised sever-al new ionic liquids with O,S-phosphoro-thioate or O, Se-phosphoroselenoate an-ions (Figure 10). Their structural similari-ty to [Emim][Me2PO4] explains their abil-ity to dissolve cellulose. As such, are also potential candidates for the fibrillation of wood chips. An article was published this year concerning the work.

5. Future business potential

On the whole important progress by inter-national standards has been made, in the area of the processing of lignocellulosics using ionic liquids, through FuBio - Phase 1. Moderate advances have been realized with the overall application-based targets. 3 Patents have been filed and are at the

PCT assessment stage. Cellulose regener-ation activities are well under way. The in-frastructure is available for continuing the research and optimizing processes. The major advances have been in the devel-opment of recyclable structures allowing for increased sustainability. This is neces-sary for future developments. Overall un-derstanding of the area and chemistry is at a very high standard and teamwork be-tween partners is now at a distinct level.

6. Key development needs and future plans

Increased recyclability of novel ionic liq-uids and ensuing intellectual property are always of interest. Refinement of ionic liq-uid structures developed in FuBio - Phase 1 will be continued. More thorough ap-plication testing is also necessary with scale-up and feasibility analysis of the more defined concepts. Physiochemical properties of the ionic liquids and biopoly-mer solutions need more comprehensive study for application development. From an academic perspective, a better un-derstanding of the factors that affect the fractionation of lignin in particular from wood is necessary to allow for more tu-neable fractionation processes. Ionic liq-uids are highly effective at solvating wood

Figure 10. Synthesis of imidazolium O,S-dimethylphosphorothioates via salt metathesis using sodium O,S-dimethylphosphorothioate.

23

and wood biopolymers but this academic challenge persists and will always be an obstacle to fractionation efficiency. Sup-porting academic projects are intended to improve this situation.

7. Publications and reports

Mäki-Arvela, P., Anugwom, I., Vir-tanen, P., Sjöholm, R., Mikkola, J.-P. 2010. Dissolution of lignocellulosic mate-rials and its constituents using ionic liq-uids-A review. Industrial Crops and Prod-ucts 32(3): 175-201.

Anugwom, I., Mäki-Arvela, P., Vir-tanen, P., Damlin, P., Sjöholm, R., Mikkola, J.-P. 2011. Switchable ionic liquids (SILs) based on glycerol and acid gases. RSC Advances 1, 452-457.

Anugwom, I., Mäki-Arvela, P., Vir-tanen, P., Willför, S., Sjöholm, R., Mikkola , J.-P. 2011. Dissolution of wood using novel switchable ionic liquids based on 1,8-diazabicyclo-[5.4.0]-undec-7-ene - glycerol applying acid gases as trig-gers. Manuscript Submitted – Carbohy-drate Polymers Anugwom, I., Mäki-Arvela, P., Vir-tanen, P., Willför, S., Sjöholm, R., Mikkola, J.-P. 2011. Selective extraction of hemicellulose from spruce with switch-able ionic liquids. Carbohydrate Polymers, in press

Mikkola, J.-P., Anugwom, I., Mäki-Arvela, P., Virtanen, P. 2010. Dissolu-tion, fractionation and processing of ligno-cellulosic materials and polymers with bi-carbonate ionic solvents formed from am-ides, alcohols and carbon dioxide. Finnish Patent Application 2010/6142

King, A. W. T., Asikkala, J., Mutikain-en, I., Kilpeläinen. I. 2011 Distillable Acid-Base Conjugate Ionic Liquids for Cel-lulose Dissolution and Processing. Ange-wandte Chemie International Edition 50, 6301-6305

King, A. W. T., Parviainen, A., Kar-hunen, P., Matikainen, J., Hauru, L. K. J., Sixta, H., Kilpeläinen, I. 2011. Rel-ative and inherent nucleophilicity/basic-ity/diffusivity of imidazolium-based ion-ic liquids – the implications for lignocel-lulose processing applications. Submitted to Green Chemistry

Hauru, L. K. J., Hummel, M., King, A. W. T., Kilpeläinen, I., Sixta, H. 2011. New insights into the requirements for dissolution of cellulose into ionic liquids. Submitted - Journal of the American Chemical Society

Hummel, M., Froschauer, C., Laus, G., Röder, T., Kopacka, H., Hauru, L. K. J., Weber, H. K., Sixta H., Schotten-berger H. 2011. Dimethyl phosphoro-thioate and phosphoroselenoate ionic liq-uids as solvent media for cellulosic ma-terials. Green Chemistry 13, 2507-2517

Karhunen, P., Matikainen, J., King, A. W. T., Kyllonen, L., Willför, S., Kil-peläinen, I. 2011. Ionic liquids for fibril-lation of wood chips. Manuscript under preparation

King, A. W. T., Karhunen, P., Mati-kainen, J., Kilpeläinen, I. 2010. Pro-cess for fibrillating lignocellulosic materi-al, fibres and their use, Finnish Patent Ap-plication 2010/5272

King, A. W. T., Kilpeläinen, I. 2010. Method of dissolving lignocellulos-ic materials, Finnish Patent Application 2010/5727

24


Editors

Partners:

Aalto University

Metla


University of Jyväskylä

VTT Technical Research Centre of Finland

Herbert Sixta, Marc Borrega and Lasse Tolonen

Key researchers:

Herbert Sixta, Marc Borrega, Lasse Tolonen, Kaarlo Nieminen, Ville Alopaeus, Juha Visuri, Susanna Kuitunen

Hannu Ilvesniemi, Kaisu Leppänen, Veikko Kitunen, Peter Spetz, Risto Korpinen

Stefan Willför, Bjarne Holmbom, Andrey Pranovich, Tao Song, Jens Krogell, Henrik Grenman

Raimo Alén, Joni Lehto

Marjatta Kleen

25

AbstractTreating wood, e.g. wood chips or saw dust, with hot water results in a partial deconstruction

of certain wood components into the water. Hot water treatment (HWT) processes can thus be

used for the removal of hemicelluloses from wood prior to subsequent delignification for the

production of pulp, or as a mean to extract valuable hemicelluloses for a wide range of applica-

tions. The extraction efficiency depends on several process variables, such as extraction inten-

sity (time-temperature), wood species, wood particle size, process mode (batch, flow-through)

and pH. The use of mild extraction conditions favours the recovery of water-soluble, medium-

and high-molar mass hemicelluloses in quasi-intact (close-to-native) form, but at low yield. The

water-solubility was a prerequisite for further ease of handling and usability. Increasing the ex-

traction severity increases the yield of extracted hemicelluloses, which are then recovered large-

ly as oligo- and monosaccharides. Under severe extraction conditions (above 200 °C), signifi-

cant amounts of acetic acid and degradation products, such as furfural and HMF, were recovered

from the water extracts. In this chapter, results from two different HWT sub-studies are present-

ed: i) HWT of soft- and hardwoods at temperatures up to about 200 °C, and ii) cellulose disso-

lution in near- and supercritical water treatment.

In regard to HWT, similar results were obtained by using batch and flow-through process

modes. However, the particle size (chips vs. saw dust) and the liquid-to-wood ratio had signifi-

cant effects on the extraction efficiency, presumably related to mass transfer and solubility limi-

tations, respectively. Furthermore, the influence of pH and extraction time was also found to be

critical in terms of the molar mass of the hemicelluloses obtained. Without pH adjustment, the

end-pH dropped below 3.5. This also led to autohydrolysis, which caused severe depolymeri-

sation, especially at longer extraction times. The removal of lignin increased considerably with

increasing extraction temperatures. The combination of intense HWT and mild alkaline pulping

may allow for the production of high-purity cellulosic pulps, due to the quasi-quantitative remov-

al of hemicelluloses.

In regard to near- and supercritical water treatment, literature sources suggest that cellulose

can be dissolved in near- and supercritical water as a polymer that precipitates upon cooling.

To test the concept, two reactors were built and operated at Karlsruhe Institute of Technology,

Germany. The conversion of microcrystalline cellulose under various reaction conditions was in-

vestigated. An extensive conversion occurs in ten seconds in nearcritical water at 300 °C, and in

supercritical water the conversion is complete in a fraction of a second. For instance, at 360 °C,

54 % of cellulose was converted in 0.25 seconds. Of the converted fraction, 44 % was found as

cellulose precipitate, 35 % as DP2-5 oligomers, and 5 % as glucose. The structural changes in

the cellulose residues were investigated to elucidate the conversion mechanism. Cellulose was

found to be depolymerized significantly in all trials. There were certain indications that supercrit-

ical water swells or otherwise damages cellulose crystallites.

26

1. Background

The use of hot water treatment (HWT; in literature also known as ‘hot water ex-traction’, ‘autohydrolysis’ or ‘hydrother-molysis’) for wood fractionation is cur-rently receiving considerable attention. The hemicelluloses dissolved in the aque-ous extract may be recovered and con-verted into products of high added val-ue. The wood residue after the extraction, composed mostly of cellulose and lignin, can be further subjected to pulping, using less chemicals and shorter pulping times than in the case of untreated wood. At an industrial scale, however, HWT is solely utilised in the form of steam in pre-hydro-lysis-Kraft processes for the production of dissolving pulps.

HWT is often applied to hardwoods. The higher amount of acetyl groups bound to the hemicelluloses and a good delig-nification efficiency, with lesser tenden-cy for lignin condensation, make hard-woods more suitable to water extraction than softwoods. The temperatures used usually range between 130°C and 240°C. The particle size of the raw material stud-ied varies from industrial-size chips to fine wood meal. Due to mass and heat transfer limitations, the larger the particle size, the lower is the yield of extracted products. The extraction efficiency may also depend on the liquid-to-wood (L:W) ratio of the process, owing to solubility limitations.

Temperatures closer and over 300°C offers interesting new possibilities. Cer-tain solvent properties of hot liquid wa-ter are shifted as a function of temper-ature and pressure. For instance, densi-ty, viscosity and surface tension are de-creased as temperature increases. Dielec-tric constant decreases concomitantly, which makes water to behave like a less polar solvent at higher temperatures. Ion product is increased in subcritical temper-ature range. The effect of pressure is pro-nounced under supercritical conditions at temperatures above 374°C, enabling tai-

loring the solvent properties by shifting the pressure at a given temperature.

Given the higher ion product, and therefore a higher H+ concentration, hy-drothermal treatment in subcritical water has been proposed as a promising meth-od to hydrolyse biomass and cellulose. As an additional advantage, no acid neutral-isation is required because the H+ con-centration decreases again when temper-ature is reduced.

2. ObjectivesIn terms of HWT:- Exploitation of the full potential

of HWT for the separation of hemicelluloses (130–240°C)

- Effect of wood species (birch vs. spruce/pine)

- Effect of particle dimensions - Effect of liquor-to-wood ratio

(dilution)- Effect of pH- Comparison of batch and flow-

through systems- Relationship between molar mass

of isolated carbohydrates (xylan or GGM) and yield.

- Fractionation of wood polymers as part of a biorefinery concept

- Lignin removal and activation to facilitate subsequent delignification for the manufacture of pure cellulose pulps

- Kinetic modelling of reactions- Purification of water extracts- Upscaling of the HWT process

In terms of treatment at near- and super-critical conditions:- Establish an international

collaboration with Karlsruhe Institute of Technology (KIT) in Karlsruhe, Germany, which has a long experience with high pressure high temperature processes for biomass conversion.

27

- Together with KIT, to build and evaluate a reactor system capable to operate under subcritical and supercritical conditions using reasonable short reaction times.

- Study the effect of raw material, as well as process time and temperature on the conversion of cellulose and formation of reaction products.

- Confirm the formation of cellulose precipitate, and evaluate its potential for material applications.

- Characterization of cellulose residue prior and after the treatment in order to increase knowledge about the conversion mechanism.

3. Research approachHot water treatmentFive research groups from Aalto Univer-sity (AALTO), Åbo Akademi University (ÅBO), University of Jyväskylä (JYV), Met-la (METLA) and VTT (VTT) carried out ex-periments on water HWT, using different wood species (birch, spruce and in one case pine), wood particle dimensions (fine and coarse wood meal, wood chips), pro-cess modes (batch, continuous batch and flow-through) in different scales (from

0.05 L to 300 L) and different conditions (L:S ratio, temperature, time, pH, etc.).

Naturally, the question arises how these results compare with each other? The definition of an intensity factor re-lated to the quantitative extraction of hemicelluloses (in particular xylan) into the aqueous phase was considered to be a reasonable basis for the comparative evaluation of the different processes. Fur-ther, some simplified kinetics was com-puted to determine the rate of xylan or mannan extraction into the water phase as a function of temperature, time and acidity. The experimental set-ups and re-action conditions of the different research groups are summarized in Tables 1 (for birch) and 2 (for spruce and pine). Fur-thermore, the influence of pH was thor-oughly studied.

These tables show that the majority of experiments were done with birch and spruce in batch mode. The precise condi-tions and experimental set-ups are com-prehensively described in the reports of the individual research groups.

Near- and supercritical treatmentTwo reactors were used for the experi-ments in Karlsruhe. The first reactor, Zyk-lon, was capable to operate only under subcritical conditions only, limited by the

Table 1. Overview on the different HWT reaction conditions of birch wood.

28

Table 2. Overview on the different HWT reaction conditions of spruce (S) and pine (P).

reaction time that was too long for super-critical temperatures. The second reac-tor, Mikki-reactor, reached reaction times down to 0.20 seconds, enabling experi-ments to be carried out also under super-critical conditions.

In all the experiments, no catalyst was added, and water alone was used as reaction medium. Commercial microcrys-talline cellulose from Merck was used as cellulose substrate. In the second set of experiments with Mikki-reactor, two oth-er cellulose powders from Rettenmeyer GmbH were used in addition to MCC from Merck.

4. Results

PART1: HOT WATER TREATMENT OF SOFT- AND HARDWOODS

4.1. Comparative evaluation of different experimental concepts: Common basis of comparison

4.1.1. Kinetic studyIn an effort to compare the results origi-nating from the participating laboratories using different reactors, reaction condi-tions and wood particle dimensions, the rates of xylan and mannan extraction in-to the water phase have been comput-ed as a function of the pH. Owing to the shortcomings of the available data base for kinetic evaluations a rather simple ap-

proach was applied. The simplified reac-tion scheme assumes a consecutive pseu-do-first order reaction as illustrated in scheme 1:

Scheme 1.

in which CS denotes the concentration of the carbohydrates in the solid residue, CL the concentration of the dissolved carbo-hydrates and D the concentration of the degraded carbohydrates. Other, more complex reaction schemes have been tested, such as scheme 2:

Scheme 2.

however, with limited success. Thus, the following differential equations derived from scheme 1 have been solved (Sci-entist®):

.[ ]

.[ ]

29

where A1 and A2 are the pre-exponen-tial factors of reactions 1 and 2, EA1 and EA2, the activation energies, R the uni-versal gas constant, T the absolute tem-perature in Kelvin and k1 and k2 the rate constants.

Reaction rate k2, which describes the velocity of the degradation or the dehy-dration reactions of dissolved carbohy-drates, is insignificant at low intensity HWT and/or high acidity. Thus, reliable results on k2 have been received only in a few cases (Table 3). Overall, it was pos-sible to simulate the experimental results sufficiently well as demonstrated in Fig-ure 1 and Figure 2.

Figure 1a reveals the differences be-tween the process modes: in the batch mode the xylan dissolution starts right away, while it undergoes a time lag in the flow-through mode which can be eas-ily explained by the different pH profiles as illustrated in Figure 1b. In the perco-lation mode performed on birch sawdust (which has been applied in the case of METLA), the hydrolyzed carboxylic acids (mainly acetic acid) are not accumulated in the reactor as in the case of the batch mode, but are eluted into the receiving tank. However, after a certain period of time the acidity drops much faster than in the batch reactor and, accordingly, the

Figure 1. a (left): Experimental and calculated total xylan concentrations on oven dry birch wood dissolved in water as a function of time, temperature and pH. The data points reveal the experimental results, while the lines derive from the kinetic model. b (right): Development of the pH during the dissolution of the carbohydrates from birch wood.

rate of xylan hydrolysis increases rapidly. The pH reaches lower values in the flow-through mode, presumably owing to the elution of cations from the wood. Conse-quently, less buffering capacity is avail-able and thus the pH of the hydrolysate drops below the value in the batch hydro-lysate where all cations accumulate and form a buffer system with the released acids. The use of sawdust in HWT batch operations results in relatively flat pH pro-files of generally low pH level (Fig 1b). Lower mass transfer restrictions allow a faster release of acetic acid as compared to the case of wood chip hot water extrac-tions and thus accelerate the HWT pro-cess. Consequently, in batch mode, the pH profiles of wood chip hot water treat-ments are steeper and the minimum pH level is reached after longer treatment times (or higher temperatures) as com-pared to the hot water treatments of saw-dust. As expected, the xylan concentra-tion increases faster when wood meal (dash-dot, VTT) was used as compared to the use of chips (solid line, VTT) or compressed chips (dash, VTT). The per-formance of xylan hydrolysis from wood chips can be largely explained by their associated pH profiles (Fig 1b), which in turn are affected by the applied lignin-to-solid ratio (L:S).

30

k_internal (1) represents the hydro-lytic rate of the cleavage of an internal bond in a xylo-oligosaccharide in homo-geneous dilute sulphuric acid hydrolysis; k_internal (2) was derived from xylobiose hydrolysis in water at pH 4.0; k_Xylan hy-drolysis originated from results obtained from the water HWT of Eucalyptus salig-na. Italic means that the listed values are outside the experimental range of which the rate constant has been determined.

The rate constant k1 calculated from the VTT_BC_B_30 and VTT_BC*_B_30 experiments are comparatively high, but are not included in the discussion since the data set is too small to allow reliable computation of rate constants. Relative to their pH values the rates of xylan ex-traction were rather low in the case of the METLA results which might be explained by the Donnan effect which is caused by non-diffusible cations. It has been shown

that with an increasing ionic strength, as it is the case in the batch mode (see above), a general increase of the hydro-lysis rate was observed. In other words, the ratio of the proton concentration in-side the solid phase relative to that in the bulk liquid can be increased by the addi-tion of electrolytes.

The amount of hydrolysed xylan un-dergoes a maximum value of about 17 to 18% on odw which represents 80 to 85% of the total xylan present in birch. In this respect the results from METLA (flow-through) and AALTO (batch) were pretty much the same. However, in both cases very high liquor-to-wood ratios were ap-plied which favours the solubility of poly-meric xylan and counteract degradation reactions. There are clear indications that the solubilisation of xylan is affected by the liquor-to-wood ratio. It can be as-sumed that the efficiency of xylan solu-

Table 3. Summary of the computed rate constant k2, expressing the degradation rate of the solubilised xylan from birch wood at pH 3.5.

Table 4. Summary of the computed rate constant k1, expressing the rate of xylan solubilisation from birch wood at pH 3.5.

31

bilisation is superior in a percolation than in a simple batch reactor system. The lat-ter, however, may result in higher xylan concentrations in the hydrolysate which in turn favours process economy. The com-bined recirculation-percolation process constitutes the best option since it com-bines the advantages of both batch and flow-through modes.

The extraction of mannan from soft-wood has been investigated by METLA, JVY and ÅBO. The results are depicted in Figure 2.

Again, the calculated mannan dissolu-tion rates fit very well to the experimen-tal results (Figure 2a). In the batch mode

Figure 2. a (left): Experimental and calculated total mannan concentrations on oven dry softwood dissolved in water as a function of time, temperature and pH. The data points reveal the experimental results, while the lines derive from the kinetic model. b (right): Development of the pH during the dissolution of the carbohydrates from softwood.

(ÅBO, JYV), the mannan dissolution pro-ceeds smoothly from the start of the re-action and peaks at 8 wt% on odw wood (70% of the total mannan), while in the case of the flow-through mode it starts only after a certain time lag, accelerates fast and peaks at 10 wt% on odw (86% of the total mannan in wood). The man-nan dissolution behaviour of the differ-ent reaction concepts can be explained by their corresponding pH profiles. The low and retarded mannan dissolution of METLA_SC_FT_300 as compared to that of METLA_SM_FT_0.05 both at 160°C can be nicely explained by the high pH val-ues because the rate constants k1, ex-

Table 5. Summary of the computed rate constant k1, expressing the rate of mannan solubilisation from softwood at pH 3.5.

32

trapolated to pH 3.5, are at a compara-ble level (Table 5). As expected, the use of chips compared to wood meal is ex-pressed by a later onset of mannan disso-lution and a lower rate constant k1 (MET-LA_SC_FT_300 vs. METLA_SM_FT_300).

The results from ÅBO (spruce) and JYV (pine) obtained at 150°C show a very good correspondence despite the different wood furnish. Both used a batch reactor and a relatively low liquor-to-wood ratio. Unfortunately, no experiments at high-er temperature have been executed with pine as a raw material. Thus, no compre-hensive evaluation of the HWT kinetics of this particular wood furnish is available.

The degradation of the dissolved mannan is promoted by its concentration and the applied temperature (Table 6). The results from ÅBO reveal that at low liquor-to-wood ratio the degradation of mannan starts to become significant al-ready at 160°C. At high dilution, the rate of dissolved mannan degradation remains insignificant up to 180°C as demonstrat-ed in Table 6.

Water HWT affects not only the dis-solution and degradation of the major hemicellulose components in wood, but also the minor hemicellulose components, the side chains of the hemicelluloses, the lignin, the cellulose and the whole wood morphology. It is obvious that the reac-tions of the wood components are mainly affected by the intensity of the HWT re-action. In the following chapter, intensi-

Table 6. Summary of the computed rate constant k2, expressing the rate of mannan solubilisation from softwood at pH 3.5.

ty factors are introduced which allow the prediction of quantitative changes of the wood composition triggered by the HWT reaction.

4.1.2. Intensity of HWTThe intensity of HWT is conveniently ex-pressed as P-factor using an Arrhenius-type of expression. A value of 125.6 kJ mol-1 for the fast-reacting xylan (XF), based on extensive investigations of xylan hydrolysis from Eucalyptus saligna, has been suggested for the P-factor (P-XF) calculation in a pre-hydrolysis kraft pulp mill. Overend and Chornet introduced the severity factor, R0, to quantify the inten-sity of hydrothermal biomass treatment using the following expression:

R0=t*Exp[(T-100)/14.75]

where T, the temperature, is measured in °C, t, the time in minutes. The logarith-mic plot of R0 allows the illustration of the data in a more condensed form. The relationship between log(P-XF) or logR0 with the xylan content of the wood resi-due, based on the initial wood, revealed significant scattering in the high intensi-ty HWT region, representing a xylan frac-tion of less than 0.2, which is typically de-noted as slowly-reacting or more resistant xylan (XS) (Figure 3). A reassessment of the kinetics of the HWT of xylan pres-ent in birch wood (Betula pendula) in the temperature range of 150 to 240°C yield-

33

ed an activation energy of 187 kJ/mol for the hydrolysis of the slowly-reacting xy-lan fraction XS, assuming the presence of two types of xylan that hydrolyse via par-allel first-order reactions. This value was used to calculate the P-factor, P-XS, rep-resenting the intensity of the HWT of the resistant xylan fraction. Indeed, log(P-XS) revealed a fairly well relationship with the remaining xylan content in the wood, par-ticularly for the values below 5% on od birch wood. (Figure 3).

The relationship between the HWT intensity and the amount of sugar com-ponents and their dehydration products (furfural and hydroxymethylfurfural) re-leased to the water phase is, however, more precisely described by the log(P-XF) indicating that the application of more se-vere conditions particularly promotes the degradation reactions.

4.2. HWT of Betula pendula

4.2.1. Characterization of the extract

4.2.1.1. Sugar and acetic acid balanceFollowing the principle of autocatalysed reactions, the hydrolytic cleavage of car-bohydrates and the release of water sol-uble fragments are initiated only when a certain threshold value of HWT intensity (P-XF) is exceeded. The first water solu-ble, polymeric xylan fractions appear at a log(P-XF) of about 1.6 in the hydroly-sate as demonstrated in Figure 4a. With increasing reaction intensity, the rate of the hydrolytic cleavage further acceler-ates until it reaches a fairly constant val-ue at log(P-XF) values between 1.9 and 2.4. Still, only polymeric and oligomeric xylan fractions are released to the aque-ous phase. The hydrolysis rate of solid xy-lan slightly decelerates at higher log(P-XF) because at the same time the gener-ation of xylose monomers through the hy-drolytic cleavage of oligomers starts at an

Figure 3. Relationship between the xylan content in auohydrolysed birch wood and HWT intensity factors, expressed as log(P-XF), log(P-XS) and log(R0).

appreciable rate. The release of oligomer-ic and polymeric xylan fractions reaches a maximum of 14.7% on odw at a log(P-XF) of about 2.85, while the maximum of the total hydrolysed xylan fraction is shifted to a log(P-XF) of 3.0 and amounts about 17.5% on odw. The results from AALTO, JVY, VTT(wood meal) and those of MET-LA at higher intensities show a compara-ble relationship with log(P-XF) (Fig. 4a). This is quite interesting because JVY used wood chips, while the other research labs used wood meal or sawdust. It could be expected that wood chips require a high-er log(P-XF) than wood meal to extract a certain amount of xylan as exemplified by VTT results using wood chips (VTT_BC_B_30, see Fig. 4a).

As expected, the maximum amount of bound acetyl groups in the hydroly-sate equals the maximum amount of xy-lo-oligomers (Figure 4b). A further inten-sification of HWT is answered by an in-creasing yield of monomeric xylose, peak-ing at an amount of 9.5% on odw at a log(P-XF) of 3.45 (AALTO, batch mode), and a progressive dehydration of pentos-

024 60

5

10

15

20

25

Xyla

n, %

on

odw

Autohydrolysis Intensitylog(P

Xf

): r2=0.980; log(PX

S

): r2=0.997; log( R0): r2=0.972

AALTO

34

Figure 4. a (left): Yield of xylan equivalents during HWT as a function of log(P-XF); b (right): Yield of bound acetyl groups and free acetic acid as a function of log(P-XF).

es to furfural. The latter peaks at a log(P-XF) of 4.24 reaching 10% of xylose equiv-alent on odw, which slightly exceeds the maximum monomeric xylose yield (Fig-ure 4a). The yield of monomeric xylose is lower in the flow-through than in the batch mode which is demonstrated by the results provided by METLA.

The yield of free acetic acid in the aqueous liquor is surprisingly low at moderate log(P-XF) values. Thus, HWT in combination with SAQ pulping for the production of rayon grade pulps is not a profitable source for the recovery of ace-tic acid. However, high yields of acetic acid (6.5% on odw) and furfural (10% on odw) can be obtained at high inten-sity HWT equivalent to a log(P-XF) value of about 4.25 which, succeeded by SAQ pulping, might result in the manufacture of a xylan-free dissolving pulp suitable for specialty applications.

4.2.1.2. Molar mass of sugarsAlthough HWT has shown to release main-ly oligomeric and polymeric xylan frac-tions, the autohydrolysate from batch re-actions seems to be not a very rewarding source for high molecular weight xylans.

At a P-factor 200 (log(P-XF) = 2.30), the total yield of the released xylan fragments in solution comprises about 10% on odw, which constitutes the minimum yield for a commercially attractive application.

The results can be divided into two clusters (Figure 5): The first represents the low-Mw xylans with Mw’s lower than 3.5 kDa derived from autohydolysis rang-ing between log(P-XF) 2.3 to 3.5, while the second displays the high-Mw xylans with Mw’s ranging from 4.7 to 17 kDa. The yield of the latter, however, is very low since they derive from low-intensi-ty HWT, particularly applying the flow-through reaction mode.

4.2.1.3. Formation of furanic compoundsDehydration reactions are favoured at low pH (0.05 M mineral acid) and high tem-peratures (e.g. 200°C). During moderate HWT the formation of dehydration prod-ucts remains insignificant. Thus, high in-tensity HWT is necessary to initiate the formation of furanic compounds (Figure 6).

The quantitative formation of both fu-ranic compounds can be precisely predict-

35

ed by HWT intensity expressed as log(P-XF) factor. The onset of furfural produc-tion is at log(P-XF) > 3.0, while that of HMF is connected to severe cellulose hydrolysis which starts at log (P-XF) > 3.5. The high yields of both furanic com-pounds may be attributed to the high li-quor-to-wood ratio of 40:1. Water HWT in commercial-size with liquor-to-wood ra-tios below 3.0, however, would result in significantly lower yields of furanic com-pounds owing to the onset of severe side

reactions such as resinification or conden-sation reactions.

4.2.2. Characterization of the wood residue

4.2.2.1. Yield and sugar compositionThe yield losses of wood through HWT are mainly attributed to the removal of hemicelluloses up to relatively high reac-tion intensities (Figure 7). The removal of lignin starts from the very beginning but becomes significant in the case of birch wood when HWT intensity is exceeding log(P-XS) of 3.5 (Figure 8). It is known since a long time that HWT of hardwood induces lignin solubility.

Figure 8 reveals that delignification undergoes a maximum for each reaction time in which the maximum shifts to a higher degree of delignification with ris-ing temperature. Unfortunately, the time range at each temperature gets narrower with increasing temperature owing to the competing acid-catalysed re-condensation reactions. The cellulose fraction in the birch wood stays surprisingly stable un-til very high HWT intensities are reached (log(P-XS) 5.5 – 6.0). Cellulose, however,

Figure 5. Weight average molar mass (Mw) of xylans originating from birch autohydrolysates.

Figure 6. a (left): Furfural formation in the birch autohydrolysate from dehydration of pentoses. b (right): Hydroxymethylfurfural (HMF) formation in the birch autohydrolysate from dehydration of hexoses.

36

Figure 7. Material balance of birch wood as a function of HWT intensity expressed as log(P-XF).

remains not unaffected. In fact, it is sub-jected to severe depolymerisation reac-tions through hydrolytic cleavage starting in the amorphous domains, but also af-fecting the crystalline domains depending on the HWT intensity. The newly created reducing end groups are starting points for intensive peeling reactions in subse-quent alkali treatments. The development of effective and industrially competitive stabilization measures is a subject of the continuation of FuBio (‘FuBio 2’).

Figure 8 demonstrates that the ex-tent of delignification is decreasing with increasing wood particle dimensions due to mass transfer limitations and, even more pronounced, with decreasing liquor-to-wood ratio owing to solubility limita-tions which promote condensation reac-tions inside the cell wall architecture. To overcome these limitations, at least part-ly, the combined recirculation-percolation process is recommended and is subject of investigations in one of the FuBio 2 pro-grammes.

4.2.2.2. Lignin balanceThe amount of soluble lignin (as % on original dry wood) in the liquid phase af-

Figure 8. Lignin content in the birch residue as a function of log(P-Xs), temperature, liquor-to-wood ratio and wood particle dimensions.

ter birch wood meal HWT between 180 °C and 240 °C was about 8-10% (results not shown). At any extraction temperature, the amount of soluble lignin in the hy-drolysate remained mostly constant, re-gardless of the extraction time. An insol-uble fraction, mainly related to lignin frac-tions precipitated during cooling, was al-so found in the hydrolysate, particularly after severe HWT severities. Degradation products from carbohydrates and/or oth-er degradation products may be attached to the insoluble lignin fraction.

4.3. HWT of softwood

4.3.1. Characterization of the extract

4.3.1.1. Sugar and acetic acid balancesÅBO and METLA performed HWT trials with spruce, JYV with pine. The relation-ship between log(P-XF) and the extracted amounts of C5 and C6 sugars is reason-ably well. Figure 9 shows the above de-scribed differences in the dissolution pat-terns between batch and flow-through modes. The later onset of the sugar re-

37

lease in the case of METLA’s experiments in the 300 L-reactor compared to those in the small lab-scale reactor can be ex-plained by the lower pH values in the lat-ter. The lower pH values during pine HWT compared to spruce HWT seems to be al-so the reason for the higher C5 and C6 sugar yields at given log (P-XF). As ex-pected, the flow-through mode results in higher sugar yields as compared to the batch mode. The latter shows rather low sugar yields (ÅBO) and a surprising-ly sharp yield decrease when exceeding a log (P-XF) of about 3, which is presum-ably due to the low liquor-to-wood ratio.

4.3.2. Molar mass of sugars

4.3.2.1. General overviewA major aim of softwood HWT is the re-covery of high molecular weight polysac-charides mainly consisting of galactoglu-comannan (GGM). METLA and ÅBO have determined the molar masses by SEC measurements from both the unpurified, dissolved sugars (METLA) and the poly-meric sugar recovered by ethanol precip-itation (ÅBO). The weight average molar

Figure 9. a (left): Extracted amounts of C6 sugars from softwood as a function of HWT intensity. b (right): Extracted amounts of C5 sugars from softwood as a function of HWT intensity.

masses, Mw, are related to the intensity of HWT, log(P-XF).

Figure 10 shows the expected de-crease in the Mw with progressive HWT intensity. The correspondence between the data provided by METLA and ÅBO is fairly good, which is quite surprising since both the process mode and the liquor-to-wood ratios differ significantly. The dis-tinct degradation of the dissolved sugars at log(P-XF) below 3.0 in the case of the ÅBO batch trials is by no means reflect-

Figure 10. Weight average molar mass (Mw) of xylans originating from spruce autohydrolysates.

38

ed in the Mw of the precipitated sugars. On the contrary, the Mw increases (again) with increasing reaction temperature and intensity (see encircled points in Figure 10). These (conflicting) results cannot be explained on the basis of the existing re-sults. Also quite surprising are the very high Mws at low log(P-XF) values which significantly exceed the MWs of alkaline extracted hemicelluloses following known analytical protocols.

4.3.3. Characterization of the wood residue4.3.3.1. Yield and sugar compositionThe yield decreased slightly with increas-ing the extraction intensity up to a log (P-XS) of about 3, and thereafter the yield started decreasing significantly (Figure 11). Both pine and spruce wood show a comparable behaviour. At higher temper-ature, extraction is more effective and less non-cellulosic carbohydrates are left in residuals. Therefore, by applying lower temperature it is possible to obtain poly-meric GGM with higher molar mass with-out dramatic loss in yield. Lower tem-perature is also more promising because causes less degradation in residual wood.

4.3.3.2. Lignin BalanceThe lignin content in the pine wood res-idue after the hot water extractions was rather similar after any extraction condi-tions, being about 30 % of the dry wood residue (results not shown). Most of the lignin was determined as Klason lignin, with only a minor amount of acid-solu-ble lignin. As the yield decreased with in-creasing extraction intensity, it can be ex-pected that the residual lignin decreased as well.

4.4. Purification of the hydrolysateThe concomitant release of substan-tial amounts of very reactive lignin frac-tions which partly form gluey precipi-tates avoids the commercialization of wa-ter HWT. Thus, different measures have

Figure 11. Yield of softwood as a function of HWT intensity expressed as log(P-XF).

been investigated to remove both the in-soluble and the soluble lignin.

4.4.1. XAD-4 treatmentBirch HWT hydrolysates were treated with an Amberlite XAD-4 resin (regener-ated with NH4OH). The objective of the treatment was to remove non-carbohy-drate components from hydrolysates. As seen in Table 7, XAD treatments did not have any significant effect on the content of free monosaccharides or on the total carbohydrate content. In addition, the ef-fects on the content of volatile acids were small. However, dissolved lignin and fura-noic compounds were efficiently removed from the hydrolysates (Table 7). The pH of the resin did not have any clear influ-ence on the removal efficiency.

4.5. Modelling of hot water treatmentA population balance based model was developed at Aalto to describe the depo-lymerisation reactions of hemicellulos-es taking place in HWT conditions. Kinet-ics of the cleavage of glycosidic bonds in hemicellulose was studied and two scis-sion mechanisms were tested.

39

The literature part of this modelling work consisted of exploring the phenom-ena that need to be taken into account when HWT is modelled. The previous ki-netic studies were reviewed and the re-actions taking place were identified. The modelling of depolymerisation was stud-ied to get an overview on the topic. Mass transfer and thermodynamic aspects were also considered.

P-factor

10 20 30 41 60 119 179 238

Total carbohydrates

- Untreated - XAD (neutral) - XAD (basic)

0,91

0,79

0,69

1,35

1,17

1,17

1,78

1,63

1,59

2,63

2,15

2,27

2,75

2,51

2,37

6,87

4,63

6,07

11,41

10,17

10,96

14,68

14,30

15,53

Soluble lignin


1,18

0,86

0,82

1,85

1,13

1,22

1,98

1,47

1,56

2,51

1,77

1,65

2,88

1,62

1,63

4,70

2,63

2,70

5,20

3,51

3,51

5,84

4,00

3,69

HMF


1,03

0,64

0,56

1,58

0,81

0,89

2,31

1,24

1,36

3,02

1,63

1,43

4,09

1,94

2,12

6,28

3,43

3,92

10,26

5,85

5,63

16,40

8,50

8,28

Furfural


2,78

1,01

1,20

4,78

0,98

1,40

7,62

1,69

2,25

9,74

1,98

1,82

10,98

2,16

2,35

24,98

5,59

6,79

52,52

10,67

12,73

108,58

12,11

21,60

Acetic acid


0,23

0,21

0,20

0,36

0,32

0,30

0,46

0,40

0,44

0,67

0,56

0,54

0,60

0,39

0,46

1,16

0,98

0,94

1,84

1,53

1,64

2,17

2,13

2,29

Formic acid


0,12

0,11

0,10

0,14

0,11

0,12

0,14

0,12

0,13

0,15

0,14

0,13

0,17

0,13

0,12

0,21

0,16

0,15

0,24

0,19

0,19

0,23

0,16

0,22

Table 7. Compounds content in birch hydrolysates before and after XAD-4 extraction for various extraction intensities (P-factor).

In the applied part, kinetic param-eters for the two scission mechanisms were optimized for the depolymerisation of galactoglucomannan. The effect of hy-drogen ion concentration on the scission rate was also studied. The experimental data for galactoglucomannan reactions was received from ÅBO. The obtained ki-netic parameters were compared against the acid hydrolysis data of disaccharides.

40

PART2: CELLULOSE DISSOLUTION IN SUB- AND SUPERCRITICAL WATER TREATMENT

4.6. Conversion kinetics and reaction productsThe experiments were started in autumn 2009 in Karlsruhe with the Zyklon-reactor. In these first runs the goal was to study the reaction kinetics under subcritical temperatures, and gain first-hand expe-rience of such reaction systems. The Zyk-lon-runs were limited to subcritical tem-peratures. The main results from these experiments are given in Figure 12.

These results were combined into a simple descriptive kinetic model assum-ing that cellulose dissolves as oligomers that depolymerize to monomeric glucose and further to degradation products. The model supported the assumption that a short reaction time together with a high temperature provides a high yield of oligomers and glucose whereas a long time resulted in a high amount of degra-dation products, as shown in Figure 13.

Figure 12. Left: Degree of conversion vs. reaction temperature using different reaction times. White markers for unextracted and black markers for acetone extracted samples. In ~15 s series extraction was not done. Right: Concentration of glucose in product solution for different reaction times and temperatures. Dashed lines according to the developed kinetic model.

Based on the experiences with the Zyklon-reactor, a new reactor named Mik-ki was built. Experiments were carried out using reaction times of 0.25/0.50/0.75 seconds, at temperatures of 280-390 °C. The behaviour of different celluloses was investigated by treating two additional cellulose powders in addition to micro-crystalline cellulose. The results are sum-marized in Figure 14 and Figure 15.

A complete conversion of MCC was reached at 390 °C as fast as in 0.25 sec-ond. As expected, the high yields of cel-lulose precipitate were reached with the shortest reaction times. The best yield of the precipitate was 24% on treated cellu-lose. The best combined yield of precipi-tate, oligomers, and glucose was 44% on cellulose, or 84% on the solubilized part. The yield of precipitate was in agreement with reported literature values but did not reach the best reported values: 51% in a treatment at 400°C for 0.02 second.

The effect of raw material was sig-nificant (Figure 15). The used MCC was much more reactive than that of the used

41

Figure 13. Appearance of liquor products after the removal of cellulose residue. a) 300 °C for 6 s. b) 287 °C for 2.5 seconds containing white cellulose precipitate.

Figure 14. Concentrations of different compounds in the system after hot water treatment of microcrystalline cellulose with Mikki-reactor. More complete mass balance including C2-C5 oligosugars was carried out for the 0.25s series.

Figure 15. Effect of raw material. The mass of quantified compuounds in the system using a reaction time of 0.50 s. MCC: microcrystalline cellulose from Merck, AC1: Arbocell cellulose powder, grain size 1 µm, and AC30: Arbocell cellulose powder, grain size 30 µm.

42

cellulose powders. This resulted in lower degrees of conversion under equivalent reaction condition, and consequently also as lower yields of precipitate and glucose fractions. The exact reason for the differ-ence was not solved but was not explicit-ly related to viscosity or particle size. One may speculate that the different reactivi-ty is related to the natural origin of cellu-lose or to the production method of MCC and cellulose powders.

4.6.1. Characterization of residue and precipitateUndissolved cellulose residue was depo-lymerized in hot water treatments (Fig-ure 16). The time-temperature combi-nation did not a considerable effect but the DP was largely defined by degree of conversion. There was a major difference between the Zyklon and the Mikki reac-tors. In the experiments with the Zyklon-reactor, the cellulose residues exhibited a clear level-off DP behaviour, whereas with the Mikki reactor the DPs in the res-idues reached much lower values. This is of high scientific importance and will be considered in following studies.

To have a better insight to depolymer-isation, molar mass distributions (MMD) were measured by a size-exclusion chro-matograph coupled with a MALLS detec-tor. The obtained MMDs showed a bimodal shape (Figure 17). It was possible to de-convolute the MMDs into two log-Gauss-ian-distributed populations with a high goodness of fit. The MMDs and deconvo-lution are more discussed in the related paper and conference proceeding.

Crystallinity was not considerably af-fected in the Zyklon-trials, the crystal-linity indexes remaining between 51 and 55% by NMR, and 52-57% by WAXS. In the Mikki-runs the crystallinity values var-ied more, NMR crystallinities being from 41% to 69%. Transformation to cellulose II polymorph has been considered as an evident of swelling of cellulose crystals. In

Figure 16. Viscosity of cellulose residues (MCC) vs. degree of conversion. Conditions: 0.25/0.50/0.75 seconds with the Mikki-reactor, 2.5/6 seconds with the Zyklon reactor, varying temperature.

the Zyklon-trials under subcritical condi-tions there was no indication of cellulose I to cellulose II transformation, suggest-ing that the reaction mechanism is close to that of normal acid hydrolysis of cellu-lose. In the Mikki-samples treated under supercritical conditions at 380 °C, how-ever, we found cellulose II in the residue. This is a possible indication that cellulose crystals indeed swells or are otherwise af-fected by supercritical water.

Also the formed cellulose precipitate was characterized (Figure 18). NMR and WAXS showed exceptionally high crystal-linity and a clear cellulose II structure. The degree of polymerization was low, by GPC from DP9 to DP30, and viscosity val-ues were typically 10-20 ml/g. To note, the precipitation was a slow process, con-tinuing for days. For real applications, dif-ferent ways such as concentration or ad-dition of non-solvent must be considered. These aspects were not investigated in the first phase of FuBio.

43

Figure 17. Molar mass distributions of cellulose residues treated with Zyklon-reactor (left) and Mikki-reactor (right). The MMDs are multiplied by the yield of residue.

Figure 18. Left: SEM-micrographs of cellulose precipitate. Right: WAXS-intensities of untreated microcrystalline cellulose and cellulose precipitate.

44

5. Future business potentialIn terms of HWT:- Value-added products from

monomeric-, oligomeric and polymeric hemicellulose

- High-reactivity lignin isolated from the hydrolysate (low-molecular phenols).

- High-purity lignin isolated from the cooking liquor.

- Hemicelluloses-lean pulps (dissolving pulps) of higher purity than obtained from conventional PHK pulps, or more precisely, high-purity dissolving pulps currently producible only by a combination of pre-hydrolysis and cold caustic extraction.

- The model for degradation of hemicelluloses may be further developed and set into a user-friendly process simulator, helping the operating of a plant which manufactures biopolymers for various applications.

- The model may also be used as a process design tool for future biorefineries.

In terms of near- and supercritical treat-ment:- Creation of fundamental knowledge

on the treatment of cellulose under sub-, and supercritical water conditions.

- Formation of valuable cellulose degradation products: oligomeric and monomeric sugars as well as retroaldol reaction products.

6. Key development needs and future plans In terms of HWT:- Reactor type: (a) more systematic

trials with a flow-through reactor (percolation-type of reactor). (b) Flow-through reactor with combined recirculation and percolation mode to increase the concentration of the reaction products. (c) Shrinking-bed reactor to minimize hydrolysis of dissolved sugars and condensation of dissolved lignin. (d) Cascade reactor with accurate temperature control, controllable turbulence, and flexible sampling points.

- Pre-treatment of lignocellulosic substrates: particle size vs. efficiency and selectivity of HWT reactions.

- Combination of HWT and mild sulphur-free delignification processes

- Commercially viable solution for the removal of insoluble and soluble lignin.

- Separation technology: o polymeric/oligomeric/

monomeric sugars o furanic compounds- Downstream processing:

purification, drying, etc.- Commercially viable technical

solution for the manufacture of a high-purity dissolving pulp adopting sequential HWT and alkaline cooking processes. Key steps: reinforced HWT-mild alkaline pulping with in-situ cellulose stabilization etc.

45

- The kinetic model development will continue in FuBio 2. Mass transfer models are implemented for wood meal and wood chips. Chemical reactions such as deacetylation and furfural and HMF formation will be included. The model will be extended to other wood species in addition to spruce. The depolymerisation models for cellulose and xylan will be implemented and digester level mass transfer aspects will also be looked into.

In terms of near- and supercritical treat-ment:- Continuous improvement with the

experimental plant and procedure for rapid high temperature trials.

- Better understanding the effect of cellulose characteristics and pre-treatments like acid hydrolysis on the rate of conversion.

- Elucidation of conversion/dissolution mechanism.

- Control the depolymerisation, and sugar degradation reactions, like retro-aldol reactions, in order to maximize a high yield of wanted compounds by using of pH control or radical scavengers.

7. Publications and reportsBorrega, M., Nieminen, K., Sixta, H. 2010. Delignification kinetics of birch wood HWT. 11th European Workshop on Lignocellulosics and Pulp, 16th–19th Au-gust, Hamburg, Germany.

Borrega, M., Nieminen, K., Sixta, H. 2011. Degradation kinetics of the main carbohydrates in birch wood during hot water extraction in a batch reactor at el-evated temperatures. Bioresource Tech-nology (in press).

Borrega, M., Nieminen, K., Sixta, H. 2011. Effects of hot water extraction in a batch reactor on the delignification of birch wood. BioResources 6(2), 1890-1903.

Borrega, M., Sixta, H. 2011. Fraction-ation of birch wood by a hot water treat-ment. 3rd Nordic Wood Biorefinery Con-ference, 22nd–24th March, Stockholm, Sweden.

Borrega, M., Sixta, H. 2011. Production of cellulosic pulp by subcritical water ex-traction followed by mild alkaline pulping. 16th International Symposium on Wood, Fiber and Pulping Chemistry, 8th – 10th June, Tianjin, China.

Grenman, H., Eränen, K., Krogell, J., Willför, S., Salmi, T., Murzin, D.Yu. 2011 Kinetics of aqueous extraction of hemicelluloses from spruce in an inten-sified reactor system. Industrial & En-gineering Chemistry Research, 50(7), 3818–3828.

46

Kilpeläinen, P., Leppänen, K., Spetz, P., Kitunen, V., Ilvesniemi, H., Pra-novich, A., Tenkanen, M., Willför, S. 2011. Efficient and novel extraction of xy-lan from birch. Manuscript, to be submit-ted to Holzforschung).

Kilpeläinen, P., Leppänen, K., Spetz, P., Kitunen, V., Pranovich, A., Tenk-anen, M., Ilvesniemi, H., Willför, S. 2011. Extraction of xylans from birch using pressurized hot water. 3rd Nordic Wood Biorefinery Conference, 22nd–24th March, Stockholm, Sweden.

Kleen, M., Liitiä, T.M., Tehomaa, M.M. 2011. The effect of the physical form and size of raw materials in pressurized hot water extraction of birch. 16th Interna-tional Symposium on Wood, Fiber and Pulping Chemistry, 8th–10th June, Tian-jin, China.

Kleen, M., Määttänen, M., Asikain-en, S.A., Liitiä, T.M., Tehomaa, M.M. 2011. Stepwise hot water and alkali ex-traction of birch sawdust to produce xylan and cellulose. 16th International Sympo-sium on Wood, Fiber and Pulping Chemis-try, 8th–10th June, Tianjin, China.

Lehto, J., Alén, R. 2011. HWT of carbo-hydrates from birch wood chips prior to delignification. 3rd Nordic Wood Biorefin-ery Conference, 22rd–24th March, Stock-holm, Sweden.

Pranovich, A., Song, T., Holmbom, B., Willför, S. 2010. Two-stage water extrac-tion of galactoglucomannan from spruce wood. 11th European Workshop on Lig-nocellulosics and Pulp, 16th–19th August, Hamburg, Germany.

Pranovich, A.V., Song, T., Holmbom, B., Willför, S. 2011. Structure-preserved hot-water extraction of galactoglucoman-nan from spruce wood. International Con-ference “Renewable Wood and Plant Re-sources: Chemistry, Technology, Pharma-cology, Medicine”, 21st–24th June, St. Pe-tersburg, Russia.

Sixta, H., Borrega, M., Testova, L., Costabel, L., Alekhina, M., Guetsch, J. 2011. Progress and challenges in the separation and purification of xylan from hardwood. 3rd Nordic Wood Biorefinery Conference, 22nd–24th March, Stockholm, Sweden.

Song, T., Pranovich, A., Holmbom, B. 2011. Effects of pH control with phthalate buffers on hot-water extraction of hemi-celluloses from spruce wood. Bioresource Technology (in press).

Song, T., Pranovich, A., Holmbom, B. 2011. Effects of pH control with phthalate buffers on hot-water extraction of hemi-celluloses from spruce wood. 16th Inter-national Symposium on Wood, Fiber and Pulping Chemistry, 8th–10th June, Tianjin, China.

47

Tolonen, L., Aksoy, B., Kruse, A., Six-ta, H. 2011. Depolymerization pattern of microcrystalline cellulose in sub- and supercritical water. 3rd Nordic Wood Bio-refinery Conference, 22nd–24th March, Stockholm, Sweden.

Tolonen, L., Kruse, A., Sixta, H. 2010. Hot water treatment for controlled dis-solution of cellulose. 11th European Workshop on Lignocellulosics and Pulp, 16th–19th August, Hamburg, Germany.

Tolonen, L.K., Zuckerstätter, G., Pent-tilä, P.A., Milacher, W., Habicht, W., Serimaa, R., Kruse, A., Sixta, H. 2011. Structural changes in microcrystalline cel-lulose in subcritical water treatment. Bio-macromolecules 12(7), 2544–2551.

Visuri, J., Kuitunen, S., Alopaeus, V. 2010. Modeling of Hot Water Extraction. Master’s thesis, Aalto University.

Visuri, J., Kuitunen, S., Alopaeus, V. 2011. Modeling of hot water extraction. 3rd Nordic Wood Biorefinery Conference, 22nd–24th March, Stockholm, Sweden.

Willför, S., Holmbom, B.R., Pranov-ich, A.V. 2011. Spruce galactoglucoman-nans — non-cellulosic heteropolysaccha-rides with tremendous potential for new application areas. Renewable Wood and Plant Resources: Chemistry, Technology, Pharmacology, Medicine RR 2011, June 21–24, Saint-Petersburg State Forest Technical Academy Saint-Petersburg, Rus-sia

Willför, S., Tenkanen, M. 2011. Up-grading of spruce O-acetyl-galactogluco-mannans: Progress toward new applica-tion areas. Abstracts of papers, 241st ACS National Meeting & Exposition, The An-selme Payen Award Symposium: Polysac-charides for Sustainable Chemistry, Ana-heim, CA, March 27–31

48


Editors

Partners:




Ali Harlin, Annaleena Kokko and Christiane Laine

Key researchers:

Ali Harlin, Annaleena Kokko, Christiane Laine, Adina Anghelescu-Hakala, Jonas Hartman, Sari Hyvärinen, Janne Kataja-aho, Hannu Mikkonen, Heikki Pajari, Jarmo Ropponen, Harri Setälä, Sauli Vuoti, Björn Krogerus

Ilkka Kilpeläinen, Jari Karakka

Stefan Willför, Victor Kisonen, Andrey Pranovich, Markku Auer (partly VTT), Joakim Jakobson, Anna Sundberg, Patrik Eklund

49

AbstractDevelopments in wood pulping as well as in the field of cellulosic ethanol may result in the pro-

duction of hemicelluloses in industrial quantities, which in turn would enable the rise of new

hemicellulose-based products. Such products could, for instance be applied in replacement of

plastics and technical polymers. This report demonstrates the wide opportunities of wood-based

hemicellulose derivatisation. Routes to chemicals and materials that could be useful in paper-

making, coating and packaging, especially as barrier, are also presented.

50

1. Background

Hemicelluloses have a great potential in various chemical and material appli-cations. Whereas cellulose is the most abundant plant material in nature, hemi-celluloses is the second most abundant. The global annual growth of wood and other renewable raw materials is about 170-200 billion tons as dry weight. These raw material streams contain in average 40-50% cellulose and 20-35% hemicellu-loses, which means that the annual maxi-mum availability of cellulosic and hemicel-luloses is around 70-100 billion and 35-70 billion tons, respectively, represent-ing thus, in theory, a practically unlimit-ed source. Xylans are the main hemicel-lulose in hardwood. They are also plen-tiful in other plants such as grasses, ce-reals and herbs [1,2]. Galactoglucoman-nans (GGM) are the main hemicellulose in softwoods [1,3]. The amount of hemicel-lulose, in total, outranges the global pro-duction of synthetic thermoplastic poly-mers. estimated to be about 270 million tons in 2010, by a factor of 500. Even if these raw materials are not in practise fully available for industrial production, the order of magnitude shows that both, cellulose and hemicelluloses represent a promising renewable polymeric resource for biochemicals and biomaterials.

Extraction of the xylan from wood (and agro-based material) can be achieved rather easily in alkaline con-ditions. This technology has been wide-ly studied and is also broadly patented [4,5]. Alkaline extraction typically cleaves the native acetyl substituents of xylan and affects thus markedly the solubility of xylan. Water soluble O-acetyl galactoglu-comannans (GGMs) can be extracted, for example, from process waters of thermo-mechanical pulping (TMP) of spruce (or other softwoods) [6], particle board pro-duction [7,8], or from wood directly using pressurised hot-water extraction [9]. Mi-

crowave treatment and steam explosion have also been studied for this purpose. Especially GGM recovered from the TMP process water has been thoroughly stud-ied at Åbo Akademi University.

Hemicelluloses are also formed when cellulolytic biomasses are converted (hy-drolysed) to mono-sugars. This is the cased, for instance, when the aim is to produce cellulosic ethanol. Such a pro-cess, if built around enzymatic hydrolysis, begins with a so-called pre-treatment, which open up the lignocellulose matrix for the enzymes. As a consequence hemi-cellulose polymer and/or hemicellulose-derived mono-sugars are obtained. The derivatisation of mono-sugars is not dis-cussed here.

No large-scale hemicellulose produc-tion existing yet and this is also a key bottleneck for development and applica-tion of hemicellulose-derived chemicals and materials. New sources for hemicel-luloses will be found in the pulping pro-cess, where the actual pulping will be in-tensified and accelerated by means of re-moving hemicellulose prior the actual di-gestion. Increased interest towards dis-solving pulp (alpha crystalline cellulose) will also result in industrial amounts of hemicelluloses available.

In FuBio, the focus, in terms of devel-opment of new wood fractionation meth-ods, is set on pressurised hot water ex-traction and use of ionic liquids. The re-gard to hemicellulose, the primary goal would be to extract them as polymeric molecules. A similar concept has been de-veloped and tested, e.g., by the Universi-ty of Maine, USA. The “Integrated Forest Products Refinery”(IFPR) includes ‘near neutral’ (slightly alkaline) pre-extraction of polymeric hemicelluloses in conjunction to hardwood pulp production [10].

Various possibilities for chemical mod-ification of hemicelluloses and celluloses have been reported in literature, such as esterification, etherification (e.g. alkox-

51

ylation, cationisation or carboxymethyl-ation) and methacrylation. Esterification and etherification, the most applied mod-ification methods of polysaccharides, are typically alternatives (either or), but they can also be used after each other. Both of these derivatisation methods can be performed in either slightly aqueous or in dry organic solvents. Rather broad and industrially relevant assortments of de-rivatization chemicals are available for both methods. Selected properties such as hydrophobicity-hydrophilicity balance, solubility, thermo-plasticity, film form-ing properties, etc. can be adjusted with these methods. The main difference is that the ether bond is chemically more stable than the ester bond, which is hy-drolysed in alkaline conditions. In addi-tion, etherification performed using so-called glycidyl reagents, such as propyl-ene oxide with an epoxy functionality, generates a new free hydroxyl group as a result of opening of the epoxy ring.

2. ObjectivesIn this research the main objective was to design novel value chains, in which wood-derived hemicelluloses as polymers are to be converted into novel biopolymers. Such biopolymers could potentially be used e.g. in packaging, as coatings and films, and to improve runnability.

3. Research approachThe reactivity of hemicellulose is strong-ly affected by its extortion methods and handling of the raw materials. Drying of cellulose material leads to hornification shown as decreased pore volume and re-activity as demonstrated for never-dried, freeze-dried and dried samples. Similar features are expected also with hemi-cellulose. In the research, significant at-

tention was paid on utilizing never-dried hemicelluloses, which should improve both reactivity and quality of the modi-fication and thereafter improve even the total efficiency of hypothetical process.

Chemical modification of starch for different products is well known and in-dustrially well applied. Important to notify is that hemicelluloses have no nutrition-al value as such and represent a material stream that does not interfere with pro-duction of food or feed. Now it is possi-ble to see the modified hemicelluloses as a more sustainable alternative compared to starch and possibly also a more per-forming replacement. However, we were not limiting only to typical derivatisations but also to targeted technologies beyond.

4. ResultsIt was demonstrated that one can affect the final properties such as solubility, zeta potential, and plasticization of both xylan and GGM in wide range by chemical mod-ification. The modification method should be selected based on application purpose of the derivative. Whilst some derivatives had several good properties some of the derivatives were not soluble to common solvents. This was a challenge to some of the application areas as water solutions are preferred.

Hydroxypropylated allyl/butyl xylan derivatives were prepared and tested for film formation. Two main etherification strategies were used. In the first route so-called never-dried xylan or in some of the cases also dried xylan samples pre-pared from a bleached birch pulp using alkaline extraction method was used as starting materials. The strategy after the first hydroxypropylation step was to per-form a next etherification step where – most likely - the new hydroxyl group of added hydroxypropyl substituent re-acts with added butyl and/or allyl glycid-

52

yl reagents. In the second route, in sim-ilar manner, a bleached birch pulp was used in a so-called reactive extraction method where birch pulp was first treat-ed in NaOH at ambient conditions, and then without any separation or purifica-tion steps, the derivatizing reagents were added. The xylan derivative can be eas-ily separated from cellulosic fibres using a simple filtration procedure with some washing and concentration steps. The film (thickness 0,2 mm) prepared from butyl-ated and allylated xylan (X-BA) deriva-tive was also mechanically the strongest one within all the xylan derivatives: Ten-sile strength was 44 MPa (highest value), Elongation at break 22% (highest value), and Young’s Modulus was 524 MPa (mod-erate).

Allyl/butyl derivatives of xylan were prepared using glycidyl reagents yielding water-soluble xylan derivatives with good properties for film formation and applica-tion as coating binder (Figure 1). In ad-dition ester derivatives of long and short-chain carboxylic acids were prepared from xylan and dispersed.

Water-soluble xylan derivatives and

GGM (ÅA) cross-linked with citric acid were tested for film forming and barri-er properties. They showed film forma-tion and best overall barrier in the case of coated paperboard. The best coatings had oxygen barrier better than PET, but there is still room for optimization. Also specific hydrophobic xylan deriva-tives (xylan hexanoate) provided rela-tively good water vapor barrier (similar to PET) on uncoated board. The key to good film formation and properties lies in water solubility/MW, particle dimension, additives and the structure of the base substrate.

As far as coating and printability are concerned, soluble hemicellulose bind-ers gave better surface strength for coat-ed papers than dispersions. This can be due to large particle size distribution ob-tained in the tested dispersions. The best surface strength properties were obtained with a xylan that had been both butylat-ed and allylated. Wet-pick strength with the best tested hemicellulose binders was significantly poorer than with the refer-ence latex but comparable to that with the reference starch. Brookfield viscosi-

Figure 1. A transparent and flexible film obtained from a butylated and allylated xylan (left). Reaction scheme (right): (1) hydroxypropylated (HPX), (2) hydroxypropylated and butyl-allylated (HPX-BA), and (3) butyl-allylated (X-BA) xylans.

53

Figure 2. The synthesis of cationised xylan and cationised hydroxypropyl xylan using glycydyltrimethyl ammoniumchloride.

Figure 3. Cationisation of GGM.

ties of the coating colors containing hemi-cellulose binders were significantly higher than those with latex or starch.

Cationisation of hemicelluloses was performed for xylans and GGMs (Fig-ure 2 and 3) and these derivatives were tested as wet-end additives. Re-acety-lation of cationised GGM is also possible (Figure 4). The latest cationic xylan and GGM derivatives in the project behaved quite similarly to a commercial fixative as far as turbidity and cationic demand are concerned. These improved pigment re-tention and optical and mechanical prop-

erties of paper when compared to refer-ence paper without fixative. Some early derivatives had problems when it came to e.g. solubility and MW.

The xylan and cationised xylan maleates were successfully synthesized by esterification of xylan/cationised xylan with maleic anhydride. Different degree of substitution can be achieved depending on the reaction conditions. The xylan and cationised xylan maleates derivatives with PVA are interesting for their testing as flocculants and fixatives (Figure 5).

54

Figure 5. Synthesis and structure of xylan and cationised xylan maleates as well as derivatives linked with PVA.

Figure 4. Re-acetylation of cat-GGM.

55

5. Future business potentialPlease see chapter “Hemicellulose and cellulose-based films and barriers”.


High molar mass hemicelluloses were typ-ically not obtained using the studied hot water extraction processes, which results us to conclude that high-value applica-tions must be developed also for hemicel-lulose fragments. Examples of such appli-cations include the role coordinating func-tionality on cellulose surfaces or increas-ing the efficient molar radius for floccu-lation and bonding. Fatty acid-modified hemicellulose fragments could also be ap-plied in barrier applications. On the oth-er hand, thermoplastic derivatives of cel-lulose and hemicellulose should be fur-ther studied. This could enable a total-ly new product platform to be developed. Furthermore, even more efficient modifi-cation technologies should be developed suitable for industrial use, such as reac-tive fractionation in situ for high degree of substitution.


Anghelescu-Hakala, A.G., Hyvärinen, S., Liitiä, T. Xylan derivatives as new ma-terials from natural resources. 3rd Nor-dic Wood Biorefinery Conference, NWBC 2011, 22-24 March 2011, Stockholm, Sweden.

Anghelescu-Hakala, A.G., Hyvärinen, S., Liitiä, T., Kataja-aho, J., Haavis-to, S., Asikainen, J., Harlin, Renewable materials from xylan derivatives as reten-tion chemicals. Nordic Polymer Days, NPD 2011, 15-17 June 2011, Stockholm, Swe-den.

8. References1. Sjöström E (1993) Wood chemistry

– fundamentals and applications. Academic, London

2. Ebringerová A and Heinze T (2000) Macromol. Rapid Commun. 21, 542–556.

3. Willför S, Sundberg A, Pranovich A, Holmbom B (2005) Wood Sci. Technol. 39:8, 601–617

4. US 2001/0020091 A15. US5430142A16. Willför S, Rehn P, Sundberg A,

Sundberg K, Holmbom B (2003) Tappi J., 2:11, 27–32

7. Lindblad MS et al (2007) Abstracts of Papers, 233rd ACS National Meeting

8. Lundqvist J, Jacobs A, Palm M, Zacchi G, Dahlman O, Stålbrand H (2003) Carbohydr. Polym., 51: 203–211

9. Leppänen, K, Spetz P, Pranovich, A, Hartonen, K. Kitunen, V, Ilvesniemi, H (2011) Wood Science and Technology 45:2, 223–236

10. Van Heiningen (2010) Integrated Forest Products Refinery (IFPR)FINAL REPORT, http://www.osti.gov/bridge/servlets/purl/979929-FuFfqu/979929.pdf

56


Editors

Partners:




Ali Harlin, Christiane Laine and Jonas Hartman

Key researchers:

Ali Harlin, Christiane Laine, Jonas Hartman, Annaleena Kokko, Sari Hyvärinen, Hannu Mikkonen, Heikki Pajari, Jarmo Ropponen, Harri Setälä, Riku Talja, Mika Vähä-Nissi, Björn Krogerus

Ilkka Kilpeläinen, Maija Tenkanen, Mari Granström

Stefan Willför, Victor Kisonen, Andrey Pranovich, Markku Auer (partly VTT)

57

Abstract The demand for bio-based chemicals and materials is steadily increasing. Often the motiva-

tion to develop new bio-products arises from the strive to find alternatives for petroleum-based

products. The pulp and paper industry is not yet fully bio-based. Petroleum-based materials are

still used, for instance, as barrier in paper and board products. The work reported here high-

lights some of the results obtained in FuBio 1 related to novel bio-barriers. Key materials studied

included xylan, galactoglucomannan (GGM) and cellulose. The polysaccharides were modified

with novel chemistry and conversion technologies (see previous chapter of this report), whereby

materials with improved plasticization, processability, and barrier properties were obtained. The

work focused especially on oxygen barrier, but also water vapour barrier was improved. Barrier

materials comparable with synthetic plastics were achieved.

58

1. Background

The oil/petroleum-derived plastic indus-try, in addition to struggling with contin-uous fluctuations in petroleum price, is also facing a more detrimental challenge in the term of increasing crude oil price. A major leap in the price of crude oil has not yet taken place, but several models predict very challenging times. This has resulted in increased interest towards us-age of renewable raw materials is source for chemicals, plastics, etc. Today bioplas-tics is still a niche market. The global vol-ume of bioplastics represent only about 0.5 % of the total production of plastics (PROBIB, 2009). However, the worldwide production of bioplastics is growing and the usage of bioplastics is widening to new industrial sectors.

Bioplastics are made using both prod-ucts of agriculture (e.g. starch) and wood components. Some bioplastics are based on the native, natural polymer (starch, cellulose, etc.), whereas others are pro-duced from bio-based building blocks like sugar and fatty acids. Bioplastics are usu-ally either biodegradable or non-biode-gradable. An advantage of biodegradable materials, according to EN 13432, Direc-tive 94/62/EC, and Directive 2004/12/EC, is that they reduce the landfill waste in contrast to recycled plastics (Frost & Sul-livan, 2007). Thus, the development of bio-based and biodegradable barrier ma-terials has become increasingly interest-ing due to demands for improved com-postability, sustainability and recycling of consumer packages.

Biodegradable barrier materials can, potentially, be made out of naturally oc-curring polysaccharides as well as biode-gradable polyesters. Polylactide (PLA), polyhydroxybutyrate (PHB), polyglycolide (PGA), polycaprolactone (PCL), and poly-hydroxyvalerate (PHV) are examples of polyesters that are derived from their cor-responding hydroxy acids or cyclic esters of these. Some of the monomers can be

produced by a fermentation process us-ing sugars as the staring material. Alter-natively, separation from industrial side-streams is also a possibility. Both of these concepts would results in bio-based end-products. A more detailed description of the biopolymers mentioned above, their properties and relevant market penetra-tion has been recently compiled (PROBIB, 2009).

Regarding industrial activity in terms of bio-based barriers and films, Xylo-phane is a Swedish start-up company that is commercializing xylan-based bar-rier materials for various packaging ap-plications. An example found in literature includes a material with a clearly com-petitive oxygen transfer rate (OTR) val-ues of 0.2-1.1cm3 µm/(m2 day kPa) (Xy-lophane, 2011).

Ecosynthetix, USA, has developed a starch-based biolatex, Ecosphere®, which aims to replace traditional oil-based bind-ers in the paper industry. Its properties as a barrier material are at present un-fortunately unknown. Ecosynthetix al-so has two other products: EcoMer® and EcoStix®. EcoMer® is a family of nov-el sugar-based oligomers that are made available for polymer manufacturers as a bio-based building block to create new water-borne sugar-acrylic adhesives and resins. EcoStix®, on the other hand, is a water-borne pressure sensitive adhesive (PSAs) based on the EcoMer® technolo-gy. Mater-bi® by Novamont, Italy, is also a special grade starch derivative devel-oped for extrusion coating on paper for cups and tableware. The barrier proper-ties of Mater-bi® are also unknown.

2. ObjectivesThe main objective of this work was to develop competences and technologies that would enable the use of wood-de-rived cellulose and hemicelluloses as sources for novel biopolymers. In FuBio

59

1, the research was still performed in sev-eral laboratories using several concepts, ranging from hemicellulose to dissolved cellulose. The key applications studied were stand-alone films and barrier mate-rials in paper and board products.

The belief is that some new bio-based materials and structures could provide sufficient barrier properties, be applicable in existing processes and cost structures, and finally, biodegradability and sustain-ability. Such biopolymers could potential-ly be used e.g. in packaging, as coatings and films, and to improve runnability.

3. Research approachA large part of the research was based on xylans obtained by alkaline extraction from birch pulp, galactoglucomannans (GGMs) isolated from TMP process waters and xylan and GGM obtained by pressur-ized hot water extraction (PHWE). The re-search also included some fatty acid cel-lulose derivatives, and even polymer-cel-lulose blends.

PHWE was successfully used to ex-tract both xylan and GGM from birch and spruce sawdust. The purity and low mo-lecular weights of the extracted hemicel-lulose fractions require more work and

will be a clear goal in FuBio 2. Membrane filtration was applied to recover and con-centrate polymeric hemicelluloses hav-ing, preferably, molar masses markedly over 20,000 g/mol from PHWE solutions. In practise, this was poorly achieved as the extracts comprised mostly low mo-lecular weights hemicelluloses. Separa-tion of lignin and lignin-based compounds from hemicelluloses is challenging using directly ultrafiltration. Hence, a part of the lignin is retained with the hemicellu-loses. Pressured filtration combined with hot water extraction of spruce chips was found to be on interesting option to ob-tain a polysaccharide fraction of after re-moval of primary filtrate. By diafiltration the average molar mass of hemicelluloses could be somewhat increased (e.g. from 17,000 to 22,000 g/mol).

Various chemical modifications – esters and ethers – of birch xylan and spruce GGMs have been carried out and tested (see pervious chapter). Fatty acid cellulose derivatives were also prepared and tested as films and coatings.

Chemical structure was confirmed ei-ther with FTIR spectroscopy or with NMR spectroscopy. The molar masses of the starting materials and their derivatives were determined by size exclusion chroma-tography (SEC) against pullulan standards.

Material

OP (cm3

µm/m2,day,kPa) 0% R.H.

WVP1 (g mm/

m2,day,kPa)

OTR (cm3/m2,day)

50% R.H.

OTR (cm3/m2,day)

80% R.H.

Stress2/strain at break

(MPa / %)

osAX 4.3 4.04 22.5 195 25 / 6.0

osAX + 20% S 5.9 1.18 6.0 430 14 / 14

osAX + 20% G 8.4 8.22 NA NA 8 / 10

osAX + 40% S 4.5 2.12 9.0 510 8 / 24

osAX + 40% G 19.2 21.57 NA NA 3 / 11

Table 1. Barrier performance of oat spelt arabinoxylan cast films.

60

4. Results

Oat spelt arabinoxylan (osAX) based ma-terials, bought from chemical provider TCI, were used as barrier reference. The main results obtained on free-standing films are shown in Table 1. Based on re-sults, sorbitol may be considered to be a better plasticizer than glycerol as the OTR value of the films is lower with sorbitol compared to glycerol at the same plas-ticizer contents. A reason for the poorer performance of glycerol is thought to be due to its migration out of the polymer matrix. This migratory effect is seen also in mechanical properties of the osAX films where changes as a function of glycerol content in tensile strength are more pro-nounced and less favourable with increas-ing plasticizer content. Similarly, sorbitol affects the elongation at break more. The film thicknesses varied from 21–65 µm within various experiments.

Notable is that the alkali-extracted xylans did not form self-standing films (Table 2). It was concluded that drying of hemicelluloses before modification result-

ed in practical challenges and this is not advisable (proven with xylan). Hydroxy-propyl and butyl-allyl derivatives of xylan were able form films. These functional-ized xylans exhibited high tensile strength values. On the other hand, galactogluco-mannans (GGM) with various degrees of acetyl substitution were not able to form self-standing films or continuous coatings. The addition of plasticizer (40 % sorbitol) permitted the preparation of free-stand-ing films.

The best oxygen gas barriers in coat-ings were obtained for water soluble hemicelluloses as GGM and hydroxypro-pylated xylan (HP-XYL 5% CA; Table 2). These barrier coatings are able to com-pete with oat spelt xylan barriers. Inter-nally plasticized and cross-linked xylan behaved similarly to externally plasticized and cross-linked GGM. Cross-linking im-proved the water vapour barrier for plas-ticized GGM. The grease resistance for all of the tested samples herein was excel-lent.

The longest fatty acid hemicellulose derivatives likely create a more hydro-

Table 2. Oxygen permeability (OP), normalized water vapour transmission (WVTR) and grease resistance values for GGM and xylan (XYL) coatings on pre-coated board.

G = glycerol; S = sorbitol; GX = glyoxal; CA = citric acid; HP = hydroxypropylated; BA = butyl-allyl; Ste = stearate; Hex = hexyl

Sample OP, 0% rh, 23°C (cm3μm/d,m2,kPa)

Norm. WVTR (g,μm/m2,d)

Grease barrier

KIT grease resistance

GGM 0.3 ± 0.2 379 ± 31 > 24 h > 12 GGM + 5% CA 0.4 ± 0.1 467 ± 23 > 24 h > 12 GGM + 20% G:S (1:1) + 5% CA 12 185 ± 10 HP-XYL + 5% CA 12±1 420 ± 20 > 24 h > 12 XYL-BA 193 ± 23 808 ± 83 > 24 h > 12 XYL-BA + 5% CA 118 ± 22 493 ± 47 > 24 h > 12 HP-XYL-BA+ 20% G:S 1:1 1168 ± 66 1000 ± 140 > 24 h > 12 HP-XYL Ste + 5% CA 115 ± 98 791 ± 38 XYL-Hex 1892 ± 258

61

phobic and plasticised coating layer and a more viscous solution affecting film for-mation (Figure 1). Amorphous hemicellu-loses are likely more prone for the reac-tion with fatty acids than cellulose. Cel-lulose derivatives showed poor oxygen barrier which probably was mainly due to the poorer packing density of the formed coating layer.

In addition to hemicellulose deriv-atives, also cellulose fatty acid deriva-tives were generated and coated success-fully on cupboard (Figure 2 and 3). Pin-hole- and crack-free coatings show high hydrophobicity and stability compared with commercial PLA-coated standards. Cellulose fatty acid derivatives provid-ed good grease, water and water vapour barrier, while the best hemicellulose de-rivatives were equally good grease and water vapour barriers. Some differenc-

Figure 1. Principle of hemicellulose (cellulose) fatty acid derivatisation.

es in performance can partly be due to the fact that the cellulose derivatives were applied dissolved in solvent, where-as GGM and xylan derivatives were ap-plied as aqueous solutions or dispersions. The best water and water vapour barri-er was achieved with the cellulose deriv-atives with the longest fatty acid deriv-atives together with hemicellulose poor birch kraft.

With the best derivatives the water vapour barrier was better than with PLA and talc-filled dispersion, and similar to polyester. Oxygen barrier of hemicellu-lose derivatives depended on the type of derivative and film uniformity. Therefore, pre-coated substrates are preferred in the cases of polymer solutions or mixtures of water soluble and insoluble components, while dispersions can also be used for un-coated substrates.

Figure 2. SEM images of unmodified paperboard and palmitoyl cellulose barriers.

62

5. Future business potentialKey drivers behind the current strive to move from petroleum-based materials to bio-based materials include, among oth-ers, factors such as climate change (de-creasing the CO2 footprint), increasing oil prices and changes is waste manage-ment directives in the EU. Waste man-agement in the EU is controlled by the Di-rective 2008/98/EC. It reflects the princi-ples of waste prevention, recycling & re-use and improvement on final displace-ment/monitoring. Key aspects of the di-rective are encouragement of sustainable use of resources and the principle or ‘pol-luter pays’. In Finland, a new, even strict-er law for waste management has been approved in 2011 and is expected to be in force in 2012. This will further advance the demand of bio-based (biodegradable) products, including barriers and films.

Various applications for hemicellu-loses and cellulose have been studied, ranging from low value products, such as biofuels, to high value products such as barrier materials and films. In addi-tion to using the natural polymer as the starting point for modification, the natural polymers can also be hydrolysed to smaller molecules and then re-built to new polymers. In regard to use as barrier or film, a major potential application area is

Figure 3. Application of cellulose tall oil fatty acid esters.

up-grading of paper and board products. The main results of FuBio 1 are sum-

marised in Figure 4. Strong hemicellulose films were obtained from a few hemicel-lulose derivatives. Thus, it is possible to produce mechanically strong films having water vapour barrier levels comparable to those of markedly weaker oat spelt xylan films. Furthermore, on pre-coated board the best bio-based coatings had oxygen and water vapour barrier properties in the same range as that of PET.

In terms of business potential, indus-trial availability of hemicellulose is still a major bottleneck. In current industri-al side-streams the quality and/or con-centration is typically not enough, which then reflects on the price of hemicellu-lose-based products (esp. compared to starch). This could, however, change quickly, if the demand for dissolving pulp would increase. Hemicellulose will also be available in industrial amount, when the new generation of cellulosic ethanol pro-duction starts.


In the future, attention must be put on development of the most potential appli-cations. The hemicellulose derivatives are chemically similar to starch-based deriv-atives. This gives a good starting point to develop both applications as well as busi-ness models. A key dissimilarity between hemicellulose and starch is still the differ-ence in price. This might improve signifi-cantly when industrial amounts of hemi-cellulose become available.

The work done in FuBio 1 provides a sound basis for further R&D. As with cur-rent barrier solutions, also bio-based so-lutions are most likely founded on multi-layer approached. Thus, in the future, combinations of different materials should be studied more closely.

63

Figure 4. Comparison of barrier properties of barrier materials developed in FuBio compared to commercial materials. TOFA = tall oil fatty acid; Cell = cellulose; PE = polyethylene; Hemi/Cell = other derivatives of hemicelluloses or cellulose. The chart is an adaption of material from the FP6 project SustainPack/PIRA International Ltd.


Frost & Sullivan, 2007, European mar-kets for bioplastics, M186-39.

Kosonen, V-, Eklund, P., Auer, M., Sjöholm, R., Pranovich, A., Hemming, J., Sundberg, A., Aseyev, V., Willför, S. 2010. Hydrophobication and character-ization of 0-acetyl galactoglucomanna for papermaking and barrier applicatons. Submitted to Carbohydr. Res.

Mikkonen, K.S, Heikkilä, M.I., Liljeström, V., Serimaa, R., Willför, S., Tenkanen, M., 2011, Films from gly-oxal-crosslinked spruce galactoglucoman-nans plasticized with sorbitol, submitted to Int. J. Polym. Sci., in press.

PRO-BIB, 2009, Product overview and market projection of emerging bio-based plastics, Report, University of Utrecht, The Netherlands.

Talja, R. A., Kulomaa, T. P. S., Labafzadeh, S., Kyllönen, L. E., King, A. W. T., Kilpeläinen, I. and Poppius- Levlin K. 2011, Cellulose Esters From Birch Kraft Pulps – New Biomaterials For Barrier Coating. 16th International Symposium on Wood, Fiber and Pulping Chemistry, June 08-10, 2011, Tianjin Chi-na. Proceedings manuscript submitted.

Xylophane, 2011, www.xylophane.com

LDPE

HDPE PP

PLA

PVC

PET

PVdC PA

EVOH 44%

EVOH 27%

Cellophane

LCP

Aluminum Wheat gluten

Chitosan0,01

0,1

1

10

100

1000

10000

0,01 0,11 10 100 1000

OTR

[cm

3 /m

2da

y ba

r]

WVTR [g/m2 d]

TOFA-Cell

Hemi/Cell

PE/Cellblends

64

Co-polymerisation ofthree hydroxy acids

Editors

Partners:


Jarmo Ropponen, Tuomas Mehtiö and Leena Nurmi

Key researchers:

Ali Harlin, Mika Härkönen, Jarmo Ropponen, Tuomas Mehtiö, Leena Nurmi

65

Abstract In this study, a series of co-polymers of glycolic acid (GA), D,L-lactic acid (D,L-LA) and D,L-2-hy-

droxybutyric acid (D,L-2HBA), as well as a D,L-2HBA homopolymer were prepared via melt con-

densation polymerisation. The acid monomers studied represent a mixture that could be ob-

tained by fractionation of black liquor of the kraft pulping process (i.e. the three smallest hy-

droxy acids, which are typically found in the same fraction). The co-polymers were character-

ized with 1H NMR and 13C NMR, size exclusion chromatography, differential scanning calorime-

try and thermo-gravimetric analysis. The obtained co-polymers had molecular weights between

Mw = 3500–10 000 g/mol. The reactivity of D,L-2HBA was similar to D,L-LA in the co-polymer-

isations. D,L-2HBA lowered the glass transition temperatures and the molecular weights of the

co-polymers. However, significant decrease of molecular weight was observed only when the

amount of D,L-2HBA in the feed exceeded 60 %. The results show that DL-2HBA can be co-po-

lymerised with GA and D,L-LA and therefore, indicates that such a fraction could be up-graded

by polymerisation.

66

1. BackgroundOne major side stream in the pulp & pa-per industry is the cooking liquors, “black liquor” when talking about the kraft pulp-ing process. Black liquor consists, among others, of numerous carbohydrate deg-radation products, depending on the raw material and the pulping conditions [1,2]. Among other compounds, up to 29 wt% of black liquor dry content consists of dif-ferent carboxylic acids, which are mainly hydroxy acids. Traditionally, concentrated black liquor is combusted to produce heat that is required in the pulping process, as well as to enable recycling of the pulping chemicals. However, majority of the black liquor hydroxy acids have very low heat value, and therefore it might be advanta-geous to separate them from the black li-quor to be utilised, e.g., as monomers for various polymeric products.[3]

Black liquor fractionation methods have been studied for several decades. Chromatographic methods together with membrane separation have proved to be effective in separation of high molecu-lar weight compounds from smaller hy-droxy acids.[4,5] However, separation of the three smallest α-hydroxy acids: gly-colic acid (GA), D,L-lactic acid (D,L-LA) and D,L-2-hydroxybutyric acid (D,L-2HBA) from each other is difficult due to their similar physical properties, including boil-ing points and solubilities. The possibility to utilize the mixture of these three com-ponents directly, without further separa-tion steps, as a starting material for po-lymerization would significantly increase the applicability of the compounds from the fractionation point of view.

GA and D,L-LA are already commonly used as monomers in preparation of bio-degradable polyesters, and poly(GA-co-D,L-LA) copolymers are also well studied in the literature.[6,7] However, there are hardly any reports concerning homo- or co-polymerisations of D,L-2HBA. Co-po-lymerisation of black liquor hydroxy ac-id mixture containing these three com-

pounds would therefore open utiliza-tion potential for D,L-2HBA as monomer-ic compound. The main application area of polymers of GA and D,L-LA is current-ly on medical field, e.g. in controlled drug delivery or tissue engineering, which can also be seen in the recent review articles related to the subject.[8-11]

Hydroxy acid polymers can be syn-thesized either via ring-opening polym-erization (ROP) of lactones or by direct polycondensation of hydroxy acids. Cur-rently ROP is the more commonly utilized method, as it generally provides higher molecular weights and shorter reaction times.[9] However, direct polycondensa-tion method is more easily applicable in the polymerization of hydroxy acid mix-tures. The most common way to perform polycondensation is in bulk conditions at elevated temperature and reduced pres-sure. In general it can be said of thermal properties of hydroxyl acids that copoly-mers of GA and D,L-LA are amorphous. The glass transition temperatures of the copolymers vary between 20–40°C. Crys-talline materials are obtained if GA con-tent is above 90%.[12] Homopolymers of D,L-lactic acid are soluble in various organic solvents such as tetrahydrofu-ran (THF) and chloroform. Solubilities of poly(GA-co-D,L-LA)s in tetrahydrofuran and chloroform decrease as amount of GA increases over 70% [12] and these poly-mers are soluble in only highly fluorinat-ed solvents. Biodegradability of poly(GA-co-D,L-LA)s is dependent on the compo-sition of the polymer intermediate copo-lymers being more unstable to hydrolyt-ic degradation then homopolymers.[13]

2. ObjectivesThe main objective of this study was to find out whether the mixtures containing the three smallest α-hydroxy acids (gly-colic, lactic and 2-hydroxy butanoic ac-ids) could be polymerised similarly to pre-

67

viously reported methods for LA and GA homo- and co-polymers. Since the LA and GA homo- and copolymers prepared by condensation polymerization are relative-ly widely studied and reported the goal was to see how the third added monomer, 2HBA, affects the polymerization proce-dure and properties of the resulting poly-mer (molecular weight and thermal prop-erties)

The objectives of the project were to • Polymerize three smallest hydroxy

acids obtained from black liquor using condensation polymerization

• Vary the monomer ratios of hydroxyl acids

• Study thermal properties of obtained polymers

3. Research approachThe monomers in question were pur-chased from commercial source but rep-resent mixtures that are obtainable from kraft pulp black liquor by fractionation. The structures of the monomers are pre-sented in Figure 1. The utilisation of D,L-2HBA as monomer in corresponding sys-tems has not been previously reported. Therefore, the aim of the study was to investigate how its presence affects the copolymerizations and the properties of

copolymer products, including molecular weights and thermal properties.

4. Results

4.1. Polymerization with varying monomer compositionsThe poly(α-hydroxy acid)s were synthe-sized by melt condensation polymeriza-tion. A series of polymerizations were conducted with varying monomer com-positions. For example, polymerization with following composition in feed: 45 mol% GA, 45 mol% D,L-LA and 10 mol% D,L-2HBA was conducted as follows. GA (1.901 g, 25 mmol), D,L-LA (2.502 g containing 10 wt% of water, 25 mmol), D-2HBA (0.289 g, 2.8 mmol), L-2HBA (0.289 g, 2.8 mmol) and Sn(II)oct. cata-lyst (24 mg, 0.5 wt% of monomers) were weighed into a three neck flask equipped with a magnetic stirrer and a distillation apparatus. Nitrogen flow was led under the surface of the reaction mixture. The flask was immersed into oil bath at room temperature and temperature was in-creased to 165°C within 1 h. After 5 h re-action time the pressure was reduced. The reduction was performed stepwise with-in 1.5 h to a final pressure of 250 mbar. Continuous nitrogen flow was stopped to-

Figure 1. Chemical structures of glycolic acid (1), D,L-lactic acid (2) and D,L-2-hydroxybutyric acid (3).

68

Figure 2. 1H NMR spectrum of PLGHA-HW.

Figure 3. The 13C NMR spectrum of PLGHA-HW.

69

wards the end of pressure reduction. Af-ter complete polymerization time of 10 h obtained polymer was poured out from the reaction flask and purified by dissolv-ing in chloroform and precipitating from methanol. Purification procedure was re-peated three times. Purified polymer was dried in vacuum oven at room temper-ature for 24 h before further character-ization.

4.2 Polymer compositionCompositions of prepared polymers were analysed with 1H NMR and 13C NMR. 1H NMR spectra of sample PLGHA-HW is pre-sented in Figure 2. Peaks at 1.98 ppm (4) and 1.04 ppm (6) are distinctive to meth-ylene and methyl protons of D,L-2HBA units respectively and peak at 1.58 ppm (5) to methyl protons of D,L-LA units. Overlapping signals 1-3 between 4.58-5.36 ppm are from polymer backbone carbons. 13C NMR spectra were used to calculate compositions of polymers. 13C NMR spectrum of PLGHA-HW is present-ed in Figure 3. Peaks at 60.75 ppm (6), 69.12 ppm (5) and 73.79 ppm (4) are distinctive to polymer backbone units

GA, D,L-LA and D,L-2HBA, respectively. Composition calculations were based on integral areas of peaks 6, 5 and peaks at 24.38 ppm (7) or 9.20 ppm (9), dis-tinctive to methylene and methyl car-bons of D,L-2HBA respectively. Calculat-ed compositions of polymers are present-ed in Table 1. Results show that composi-tions of polymers are similar compared to monomer feed. However, the amount of GA units in the polymer was consistent-ly slightly higher than in feed. Similar re-sults have also been reported earlier and explained to be due to higher reactivity of primary hydroxyl group of GA over sec-ondary hydroxyl groups.

4.3 Size exclusion chromatography (SEC)Molecular weights of the synthesized co-polymers were determined with size ex-clusion chromatography (SEC) and re-sults are presented in Table 1. Obtained molecular weights are comparable to molecular weights of previously report-ed PLGA polymers prepared by simi-lar procedures.[12,14] Amount of D,L-

Table 1. The conducted polymerizations and the monomer compositions, SEC results and thermal properties of the obtained polymers.

Composition in feed (mol%) Composition in polymer (mol%)

(13C NMR) SEC (g/mol) DSC (¼ C) TGA (¼ C)

GA D,L-

LA D,L-

2HBA GA D,L-LA D,L-2HBA Mw Mn Mw/Mn Tg Td(10%)

PLGA 50,0 50,0 0,0 58 42 0 9000 4200 2.12 21 224

PLGHA1 47,5 47,5 5,0 49 46 5 5900 3700 1.61 20 230

PLGHA2 45,0 45,0 10,0 47 44 9 9300 5800 1.60 23 235

PLGHA3 20,0 20,0 60,0 22 21 57 6200 4300 1.45 7 240

PLGHA4 10,0 10,0 80,0 11 10 79 3400 2700 1.26 -1 237

PHBA 0,0 0,0 100,0 0 0 100 2800 2300 1.18 -3 244

PLGHA-SW 34,6 48,6 16,8 38 46 16 10100 6800 1.50 28 243

PLGHA-HW 24,4 30,9 44,6 29 30 41 7500 4500 1.66 22 236

70

2HBA in feed can be up to 60 mol% be-fore it started to have effect on molecu-lar weight of the obtained polymer. Nota-ble decrease in molecular weight can be seen when amount of D,L-2HBA in feed is increased to 80 mol%. However, these monomer ratios are higher that can be obtain in black liquor. In addition, molec-ular weight of PHBA, the homopolymer of D,L-2HBA, was already significantly low-er than ones of copolymers. In overall, mainly due to difficult water removal from the reaction mixture obtained molecular weights by condensation polymerization remain relatively low.

4.4 Thermal analysisThermal behaviour of synthesized poly-mers was studied with differential scan-ning calorimetry (DSC) and thermogravi-metric analysis (TGA). No melting was ob-served in any of the polymers indicating that all the synthesized polymers were amorphous. This was expected for copoly-mers due to the statistical nature of their structure. Also the homopolymer of D,L-2HBA was expected to be amorphous but for different reason. Since both enantio-mers of D,L-2HBA were used in the syn-thesis the polymer was not able to form crystalline structures. Similar behaviour has been observed for poly(lactic acid) which has been reported to lose its crys-tallinity when optical purity is decreased below 66–76%.[15,16]. By contrast, the GA monomer has a regular structure. Ac-cording to literature, GA-co-D,L-LA copo-lymers can contain GA crystals in case the GA content is above 90%.[12] However, the GA content did not exceed 50 % in the current copolymer series.

Glass transition temperature (Tg) val-ues for all the prepared polymers are pre-sented in Table 1. When the amount of D,L-2HBA incorporated in the polymer was 40 mol% or lower, the measured Tg values were between 20–28°C. These val-ues are comparable to previously report-

ed values for GA-co-D,L-LA copolymers (20–40°C)[12]. When the amount of D,L-2HBA incorporated in the polymer struc-ture was higher than 40%, the Tg values started to decrease, reaching Tg = -3°C in the D,L-2HBA homopolymer sample. However, in the case of some of the sam-ples (PLGHA4 and PHBA), the lowering of Tg can also be partially due to their lower molecular weights.

TGA showed that no major differences between samples could be seen in ther-mal decomposition behaviour. Td10% val-ues for all polymers are collected in Ta-ble 1. Results show that Td10% values for all polymers were around 230°C. Ther-mo-gravimetric (TG) and differential ther-mo-gravimetric (DTG) curves of PLGHA-HW are presented in Figure 4. From these curves it can be seen that thermal de-composition started after temperature was increased over 200°C. Thermal deg-radation mechanism for poly(lactic ac-id) and poly(glycolic acid) homopolymers shows that the dominant degradation mechanism for these homopolymers was intramolecular transesterification result-ing in cyclic oligomers. Thermal degrada-tion started with cleavage of C-O bonds in the polymer structure. Due to this dom-inant degradation mechanism typical for poly(α-hydroxy acid)s no differences can be seen in thermal degradation behav-iour of α-hydroxy acid copolymers even though the monomer composition is var-ied.


Biodegradable plastics such as starch based resins and, degradable polyes-ters and polylactic acid (PLA), accounted for the vast majority (nearly 90%) of bio-plastics demand in 2008. It is esti-mated that consumption of bio-plastics

71

will arise from 2008 just above 225,000 tons up to 900,000 tons by 2013. This growth will be bred by a number of fac-tors, including consumer demand for more environmentally sustainable prod-ucts, the development of bio-based feed-stocks for commodity plastic resins and increasing restrictions on the use of plas-tic products. However, the most important will be the expected continuation of high crude oil and natural gas prices, which will allow bio-plastics to become more cost-competitive with petroleum-based resins. Looking ahead to 2018, world bio-plas-tics demand is forecast to reach nearly 2 million tons, with a market value of over US$5 billion.[17] Regulatory environment demands and increasing oil price renders growing markets also for lower molecular weight biopolymer applications.

6. Key development needs and future plansTo achieve full benefit of these results and realize the complete application potential of these hydroxyl acid copolymers further investigations is required. E.g. hydrolyt-ic degradation and mechanical properties need to be determined in future. In ad-dition, knowing the difficulties in separa-tion of the black liquor hydroxy acids from each other these results are extreme-ly promising regarding the utilization of them as a polymerisable mixture. There-fore separation techniques should be im-proved to produce hydroxy acid mixtures for polymerization. The results will be ap-plied and developed further under the Fu-Bio JR2 programme.

Figure 4. TG (solid line) and DTG (dashed line) curves of PLGHA-HW.

72

7. Publications and reportsMehtiö, T., Rämö, V., Harlin, A., Rop-ponen, J. Polyesters based on hydroxy acids obtained from black liquor. Presen-tation. NWBC2011 – Nordic Wood Biore-finery Conference 2011, Stockholm 22-24 March

Mehtiö, T., Nurmi, L., Rämö, V., Har-lin, A., Ropponen, J. Copolymers of gly-colic acid, D,L-lactic acid and D,L-2-hy-droxybutyric acid obtainable from kraft pulp black liquor. Manuscript

8.References1. Rautiainen R, Alén R.

Characterization of black liquors from kraft pulping of first-thinning Scots pine (Pinus sylvestris L.). Holzforschung 2010 01/01; 2010/12;64(1):7-12.

2. Käkölä J, Alén R, Pakkanen H, Matilainen R, Lahti K. Quantitative determination of the main aliphatic carboxylic acids in wood kraft black liquors by high-performance liquid chromatography–mass spectrometry. Journal of Chromatography A 2007 1/19;1139(2):263-70.

3. Alén R, Moilanen V, Sjöström E. Potential recovery of hydroxy acids from kraft pulping liquors. Tappi Journal 1986;69(2):76-8.

4. Alén R, Sjöström E, Suominen S. Application of Ion-Exclusion Chromatography to Alkaline Pulping Liquors; Separation of Hydroxy Carboxylic Acids from Inorganic Solids. J. Chem. Tech. 1990;51:225-233.

5. Niemi H, Lahti J, Hatakka H, Kärki S, Rovio S, Kallioinen M, et al. Fractionation of Organic and Inorganic Compounds from Black Liquor by Combining Membrane Separation and Crystallization. Chem.Eng.Technol. 2011;34(4):593-8.

6. Okada M. Chemical syntheses of biodegradable polymers. Progress in Polymer Science 2002 2;27(1):87-133.

7. Södergård A, Stolt M. Properties of lactic acid based polymers and their correlation with composition. Progress in Polymer Science 2002 7;27(6):1123-63.

8. Lipsa R, Tudorachi N, Vasile C. Poly(alpha-hydroxy acids) in biomedical applications. Synthesis and properties of lactic acid polymers. e-Polymers 2010(087).

9. Nair LS, Laurencin CT. Biodegradable polymers as biomaterials. Progress in Polymer Science 2007 9;32(8-9):762-98.

10. Gunatillake P, Mayadunne R, Adhikari R. Recent developments in biodegradable synthetic polymers. In: M. Raafat El-Gewely, editor. Biotechnology Annual Review: Elsevier; 2006, p. 301-347.

11. Gunatillake PA, Adhikari R. Biodegradable synthetic polymers for tissue engineering. Eur. Cell. Mater. 2003;5:1-16.

12. Wang Z, Zhao Y, Wang F, Wang J. Syntheses of poly(lactic acid-co-glycolic acid) serial biodegradable polymer materials via direct melt polycondensation and their characterization. J Appl Polym Sci 2005;99(1):244-52.

73

13. Chen G, Kim H, Kim E, Yoon J. Synthesis of high-molecular-weight poly(l-lactic acid) through the direct condensation polymerization of l-lactic acid in bulk state. European Polymer Journal 2006 2;42(2):468-72.

14. Zhou S, Deng X, Li X, Jia W, Liu L. Synthesis and characterization of biodegradable low molecular weight aliphatic polyesters and their use in protein-delivery systems. J Appl Polym Sci 2004;91(3):1848-56.

15. Tsuji H, Ikada Y. Crystallization from the melt of poly(lactide)s with different optical purities and their blends. Macromolecular Chemistry and Physics 1996;197(10):3483-99.

16. Tsuji H, Ikada Y. Stereocomplex formation between enantiomeric poly(lactic acid)s. 6. Binary blends from copolymers. Macromolecules 1992 10/01;25(21):5719-23.

17. http://www.sustainableplastics.org/news/bioplastic-consumption-reach-2-million-tons-2018

74


Editors

Partners:

Aalto University

RWTH Aachen, Germany

ITCF Denkendorf, Germany

Fourné Polymertechnik, Germany

Herbert Sixta and Michael Hummel

Key researchers:

Herbert Sixta, Michael Hummel

Philipp Schuster

Frank Hermanutz

-

75

AbstractControlled regeneration of cellulose in the form of value added products from non-derivatizing

solvents such as ionic liquids requires a special spinning process. For this reason, the aim of this

project was to establish the necessary infrastructure to perform high-level research and, con-

sequently, provide expertise for the production of high quality cellulosic products, such as fibers

and films. After initiating collaborations with renowned textile institutes all over Europe to trans-

fer existing knowledge to Finland, a dry-jet wet piston spinning machine was designed, manu-

factured and installed. Supplemented with customized dissolution tools and high standard an-

alytical devices, a unique basis to perform both fundamental and application oriented research

during the successive FuBio2 phase was created.

76

1. Background

First man-made cellulosic fibers date back to the beginning of the 20th century and had their highest annual output of 3.86 Mt in 1973. Thereafter, the rise of cheaper oil-based synthetic fibers caused a steady de-cline in demand for natural fibers. Howev-er, facing the depletion of fossil carbon de-posits and inexorably growing markets in rapidly industrialising nations, an opposite trend can now be observed.

The world market of man-made cellu-losic fibers is currently dominated by Vis-cose. But as a matter of fact, its production requires substantial amounts of hazard-ous chemicals. To bypass this problem, di-rect, i.e. non-derivatizing solvents attract-ed more and more scientific attention dur-ing the past few decades. Amongst many potential solvent systems, only N-methyl-morpholine N-oxid (NMMO) could be com-mercialized so far in the so called Lyocell process. This very promising technology mainly suffers from the intrinsic instabili-ty of NMMO itself, necessitating stabilisers to prevent runaway reactions. A new sub-stance class, ionic liquids (ILs), and their ability to dissolve (holo-)cellulose in pro-cess-relevant concentrations (>12%) holds great potential to develop the Lyocell tech-nology further.

When aiming for value added cellu-lose products with high structural perfor-mance the properties of the world wide es-tablished Viscose and Lyocell fibers have to be regarded as benchmarks to be ex-ceeded. This means the fibers should ex-hibit a wet-to-dry tenacity ratio superior to viscose (even that of Modal® quality), while the tendency of fibrillation upon me-chanical abrasion should not be substan-tially increased (as is the case for the clas-sical Lyocell fiber).

This ambitious goal can only be reached if in-depth knowledge of the spinning pro-cess that means fiber formation on a mi-croscopic level is generated. Mechanistic understanding and, thus, intentional high

quality design is the only way to compete with high-quantity-low-price products from Asian and South American providers.

The aim of the first phase of FuBio was to install the necessary infrastructure for dry-jet wet solution spinning and establish a network with European textile research institutes.

2. ObjectivesThe long term objective of this work is to generate novel knowledge on the mecha-nisms of cellulose fiber spinning from di-rect solvent systems. Thus, a basis for the viable production of new cellulose products shall be provided, amplifying Finland’s com-petitiveness on the world market.To reach this ambitious goal• a national network has to be

established to combine expertise and share analytical duties

• international partnerships are sought to transfer existing knowledge about fiber spinning to Finland and plan the setup of the required infrastructure

• all necessary equipment has to be designed and installed to provide the appropriate infrastructure for high level research. This includes:

- a vertical kneader system (high-shear dissolution device to prepare moisture sensitive, highly viscous solutions at elevated temperature and under

vacuum) - a high pressure hydraulic filtration unit - rheological equipment to characterize the biopolymer

solutions - a dry jet-wet piston spinning device for continuous fiber and film forming from cellulose

solutions - standard fiber analytics for inter-laboratory comparison

77

3. Research approach

In the recent past, ionic liquids (ILs) as solvents for (ligno-)cellulosic material at-tracted researchers’ interest all over the world. Many publications dealing with the dissolution properties of ILs are avail-able, indicating the huge potential of ILs for cellulose processing. However, these reports can only serve as starting point as detailed research about the regenera-tion of the respective biopolymer solute in the form of value added products such as fibers and films is still missing. To clearly demonstrate the opportunities in terms of industrial applications offered by ionic liq-uids, this project was initiated.

From available literature at the begin-ning of this programme, a few ionic liq-uids known to dissolve cellulose were se-lected and implemented. 1-Ethyl-3-me-thylimidazolium acetate was chosen as the main solvent, mainly due to its low viscosity, good cellulose dissolving prop-erties and relatively low price. Further, various pulps and cotton linters from all over the world were acquired to study the influence to pulp composition, purity and molecular weight distribution on the spin-

ning process. The preparation of the cel-lulose-IL solution (dope) at processing relevant concentration levels, i.e. 10-15 w/w-% represents already an obstacle that is hardly ever addressed in scientif-ic articles. The steep increase of viscosity upon dissolution of biopolymers requires elevated temperatures and the employ-ment of high shear forces. For this rea-son, a vertical kneader system was de-signed and installed at Aalto University (Figure 3). This device provides the possi-bility of highly efficient, reproducible dope production.

Successful spinning of a polymer so-lution requires detailed knowledge about the visco-elastic properties of the dope. Depending on the geometry and design of the spinneret, only the right combination of temperature, extrusion velocity, and pressure leads to the formation of stable filaments in the air gap. Therefore, the rheological profile of each pulp-IL com-bination has to be assessed by means of a conventional rotational rheometer pri-or to the spinning trials (Figure 1 left). However, it is known that classical oscil-lating rheology is lacking a certain sensi-tivity towards small portions of high mo-

Figure 1. Classical rotational rheometer (left) and Capillary Extensional Break-up Rheometer (CaBER, right).

78

lecular weight cellulose which has proven in practice to promote spinning stability. Further, the orientation of the biopolymer in the filament and, thus, structure for-mation of the fiber are strongly influenced by drawing, i.e. stretching of the dope fil-ament in the air gap. For this reason a Capillary Break-up Extensional Rheome-ter (CaBER, Fig 1 right) was purchased to determine elongational-rheological prop-erties of the spinning dope. This tool al-lows the assessment of the influence of high molecular weight solute fractions on the rheological state of the solution and, at the same time, can be used to mim-ic the filament draw in the air gap of the dry-jet wet spinning device.

The spinning machine clearly repre-sents the centre piece of the Cellulose Re-generation task. Aiming for a comprehen-sive understanding of the structure for-mation in cellulosic fiber and film forming from direct solvents, a controlled regener-

ation in the form of a dry-jet wet spinning unit is required. With the assistance and guidance from experts of the Institute of Textile Chemistry and Chemical Fibers (ITCF), Denkendorf, Germany and engi-neers from Fourné Polymertechnik a spin-ning line comprising a dry-jet wet piston spinning unit, a washing and drying zone, and take-up winder was planned, manu-factured, and finally installed at Aalto Uni-versity in February 2011 (Figure 2). With-in a collaboration with the Textile Insti-tute of the Rheinisch-Westfälische Tech-nische Hochschule Aachen (ITA), Ger-many, a similar spinning device could be tested and gained knowledge used to cus-tomize our machine while it was manu-factured (Figure 5). The dry-jet wet spin-ning machine is designed in such dimen-sions to serve as interface between lab-scale experiments and pilot-plant produc-tion and, thus, meets both academic and industrial interests.

Figure 2. Single stand-alone units of the dry-jet wet spinning line.

79

4. Results

4.1 Dope preparationThe vertical kneader system (purchased from b&b Gerätetechnik, Germany) en-ables the fast preparation of highly vis-cous, air bubble-free polymer-IL solu-tions on a 200-1000 ml scale (Figure 3). A powerful rotor generates high shear forces which are necessary to dissolve the cellulose in a short period of time, thus minimizing heat-induced degradation. By measuring the torque exerted to the ro-tor, the dissolution state can be moni-tored. A protocol was established to en-sure reproducible dope preparation.

4.2 Rheological characterizationSupported by Thermo Fisher Scientific, Karlsruhe, Germany, first experiments

with solutions of different pulps and cot-ton linters in 1-ethyl-3-methylimidazoli-um acetate were performed. Kier boiling allowed to degrade cotton linters in a con-trolled way, meaning that number (Mn) and weight average molecular mass (Mw) were decreased to the same extent lead-ing to an unchanged PDI. The resulting cotton linters were then mixed to obtain blends with different molecular weight distribution.These blends were dissolved in EMIM OAc at different concentration levels and sub-jected to extensional-rheological and ro-tational-rheological characterization. The extensional relaxation time increased with increasing number average molar mass Mn (Figure 4). This seems to coincide with observations regarding the spinnability of different dopes.

Figure 3. The vertical kneader system can generate high shear forces, which are necessary for the efficient preparation of lignocellulose-IL solutions. The dissolution state can be monitored by means of a torque measurement device.

80

4.3 Cellulose regeneration in the form of fibers and filmsFirst spinning relevant results were gained at the ITA, Aachen where the temperature which enables the extrusion of stable fil-aments was evaluated. Figure 5 shows mono-filament spinning at two different temperatures. Insufficiently high temper-ature leads to a phenomenon called cold shearing, resulting in an inhomogeneous, pearl-string like filament which breaks easily at the diminution sites (Figure 5 left).

Another collaboration with the ITCF Denkendorf, one of the most experienced German institutes in fiber spinning (melt and solution), was started in February 2011. After establishing the contact al-ready in April 2009, a joint master the-sis within 3 departments was initiated. Thereby, the expertise of the ITCF Denk-endorf, the Department of Forest Products Technology (Aalto) and the Department of Biotechnology and Chemical Technology (Aalto) were combined to investigate the relationship between the rheological prop-erties of the spinning dope and its behav-ior in the spinning process. A relation be-

tween the spinnability and the molecular weight distribution of the cellulosic solute could be demonstrated.

The research focus of this task al-so includes cellulose regeneration in the form of films. As there is virtually no in-formation about continuous film forming from cellulose-IL solutions, stationary film forming was performed first to gain ini-tial knowledge. In a collaboration with the Institute of Biopolymers and Chem-ical Fibres (Łódź, Poland) we studied the influence of the pulp concentration on the casting behaviour and film properties (Figure 6).

4.4 Ionic liquid synthesisEven though EMIM OAc will be our working horse for the spinning and frac-tionation trials, we are also looking for new cellulose dissolving ionic liquids. Within a cooperation with the University of Innsbruck, Austria we could synthesize several new ionic liquids with O,S-phos-phorothioate as anion (Figure 7). Their structural similarity to dimethylphosphate explains the capability to dissolve cellu-lose.

Figure 4. Relaxation time determined via CaBER for cotton linters blends with different Mn.

81

Figure 5. Mono-filament spinning at different temperatures. Left: spinning temperature of 110°C led to an instable, pearl-string like filament. Right: extrusion of the same dope at 120°C giving a homogenous mono-filament.

Figure 6. Cellulose film samples prepared from cotton linters-EMIM OAc solutions.

Figure 7. Synthesis of imidazolium O,S-dimethylphosphorothioates via salt metathesis using sodium O,S-dimethylphosphorothioate.

82

5. Future business potentialAn economically successful manufacture of regenerated fibers in Europe is only possible if highest demands on product quality, product flexibility, environmental impact, process costs and social and so-cietal acceptance are accomplished. Re-generated cellulosic fibers have a clear future. This has been recently confirmed by renowned scientists from the Coper-nicus Institute, Utrecht University, Neth-erlands, as a result of a comprehensive life cycle assessment of man-made cellu-lose fibers (LB, Vol 88, 2010). The study revealed that a stand-alone Lyocell plant comprising the latest technology is ad-vantageous compared to the production of cotton, PET and PLA. It has the lowest impact on abiotic depletion, terrestrial ec-otoxicity and photochemical oxidant for-mation. The only way to further improve the eco-friendliness of regenerated cel-lulosic fiber production may be realized by its full integration with pulp produc-tion. Thus, the vision for the Finnish pulp industry could be the integrated produc-tion of regenerated cellulosic fibers and other products.

It is the aim of the FuBio 2 phase to contribute to the first step in the realiza-tion of this “vision”.


During the FuBio 1 the necessary infra-structure to conduct high level, industry-relevant research was installed and exist-ing knowledge acquired from all over Eu-rope. Preliminary experiments were per-formed at renowned textile institutes in Germany to bridge the gap of the long manufacturing time of the spinning ma-chine.

Successive research will consist of the production of (staple) fibers depicting the quality stated in the background informa-tion as well as the in-depth investigation of the structure formation process. This includes the attempt to shed light on the relationship between the spinning param-eter setup and fiber properties. Further, the influence of the pulp quality on the final fiber properties will be investigated in detail. Finally, there will also be efforts to reduce fiber fibrillation by a modified structure formation process and thereby make low fibrillation to an intrinsic prop-erty of the fiber.


Hummel, M; Froschauer, C; Laus, G.; Röder, T.; Kopacka, H.; Hauru, L.K.J.; Weber, H.K.; Sixta H.; Schottenberger H. 2011. Dimethyl phosphorothioate and phosphoroselenoate ionic liquids as sol-vent media for cellulosic materials. Green Chem. 13(9), 2507-2517.

Lehtonen, H.; Plog, J.P; Ingildeev, D.; Heinämäki, M.; Hummel, M.; Sixta, H. Extensional rheology of cellulose–ionic liquid solutions, American Chemical So-ciety (ACS) meeting, 21.25.3.2010, San Francisco, CA (oral presentation).

Hummel, M; Sixta H.; Froschauer, C; Weber, H.K.; Röder, T.; Kahlenber, V.; Laus, G.; Schottenberger H. 2011. Di-alkyl phosphate-related ionic liquids as advantageous solvation and fractionation media for (ligno)cellulosic materials, ACS meeting, 21-25.3.2010, San Francisco, CA (poster).

83

Hauru, L.K.J.; Hummel M.; Sixta H. Studies on the treatment of birch and spruce with the ionic liquid 1-ethyl-3-me-thylimidazolium acetate, EWLP 2010, 16.-19.8.2010 Hamburg, Germany (oral pre-sentation and proceedings pp. 129-132).

Hummel, M.; Lehtonen, H.; Heinämä-ki, M.; Sixta, H. Applicability of exten-sional rheology for the characterization of cellulose-ionic liquid solutions, EWLP 2010, 16.-19.8.2010 Hamburg, Germa-ny (poster and proceedings pp. 493-496).

Hummel, M.; Froschauer, C.; Hauru, L.; Laus, G.; Schottenberger, H.; Six-ta, H. Dialkylimidazolium phosphate de-rivatives as solvents for lignocellulosic materials and their subsequent regener-ation, Pacifichem 15.-20.12.2010, Hono-lulu, Hawaii, USA (oral presentation).

Hummel, M.; Michud, A.; Sixta, H. In-fluence of the molecular weight distribu-tion of cellulose on the extensional rhe-ology of cellulose-ionic liquid solutions, ACS meeting, 27-31.3.2011, Anaheim, CA (oral presentation).

Hauru, L.K.J.; Hummel M.; Sixta H. Fractionation of birch in 1-ethyl-3-methy-limidazolium acetate, ACS meeting, 27-31.3.2011, Anaheim, CA (oral presenta-tion).

Hummel, M.; Michud, A.; Sixta, H. Extensional rheology of cellulose-ionic liquid solutions, Nordic Rheology Confer-ence, 6-8.6.2011, Helsinki, Finland (post-er and proceedings).

Mikko Heinämäki (2010) The Hildeb-rand solubility parameters of ionic liquids and the rheological comparison of pulp-ionic liquid solutions by rotational and ex-tensional rheology.

Anne Michud (2011) Influence of molar mass distribution on the elongational be-havior of cellulose-IL solutions.

Joni Saastamoinen (2011) Influence of the molecular weight distribution on the shear rheological properties and “spin-nability” of cellulose-ionic liquid solutions.

8. References1. Lenziger Berichte (2010), Vol. 88,

p. 1-59 and 60-66.

84


Editors

Partners:


Pedro Fardim and Jani Trygg

Key researchers:

Pedro Fardim, Jani Trygg, Ari Ivaska, Johan Bobacka, Tingting Han

85

AbstractCellulose beads are porous spherical particles with a wide range of applications. They are suit-

able for stationary phase in chromatographic techniques, carriers for fertilizers and drugs, sub-

strates for immobilization of enzymes and bacteria and numerous other technological purpos-

es. Beads can be prepared from synthetic or biomaterials depending on the purpose and prop-

erties needed. Cellulose is a preferred option for biotechnological applications due its biodegra-

dation and abundance. Cellulose beads have usually been prepared by using cellulose xanthate

(CX) as a precursor. Additionally, different cellulose and cellulose derivative solutions are appro-

priate to be used as raw material depending on the bead preparation process. The method of

spinning disk atomization (SDA) has been successfully applied to produce beads with controlled

properties such as size, shape or porosity. However, the development of a sustainable technol-

ogy for bed production is dependent on the utilization of water-based cellulose solvents. In this

first phase of the FuBio we have been focusing on developing a suitable pre-treatment for pulp

fibres to enhance the dissolution of cellulose in NaOH-urea-water solvent system. We developed

a new method based on one step chemical pre-treatment using ethanol–hydrochloric acid prior

to the dissolution of cellulose in NaOH–urea–water. The dissolution mechanism of the pre-treat-

ed sample was initially examined in diluted cupri-ethylenediamine and 7% NaOH–12% urea–

water solvent using optical microscopy methods and field emission scanning electron micros-

copy. The apparent energy of activation for the viscous flow of ethanol–acid pre-treated pulp in

NaOH–urea–water was calculated using rheological methods. Our results showed that the dis-

solution of pre-treated pulp was achieved up to 4% cellulose concentration. We suggest that

the enhancement of dissolution was due to a combination of degradation of remnant primary fi-

bre wall layer and reduction of degree of polymerization of cellulose. Functional cellulose beads

were successfully prepared using preliminary experiments of coagulation after manual drop for-

mation. New alternatives of bead design and up-scaling using SDA, cellulose dissolved in NaOH-

urea-water and blends with cellulose derivatives will be investigated in the FuBio Products from

dissolved cellulose programme.

86

1. Background

Cellulose beads or porous cellulose micro-spheres can be prepared using different methods of shaping/coagulation based on two main techniques: microemulsion and drop formation. The micro-emul-sion method yields particles with dimen-sions between 10–300 µm and is based on the formation of micelles (stabilized by a surfactant) when a cellulose solution is mixed with an immiscible non-polar sol-vent followed by solidification of spheres due to temperature treatment or addition of solidification agent. The drop forma-tion method is based on the generation of droplets manually from a pipette, by fil-ament formation or by spinning drop at-omization (SDA) prior to coagulation in an anti-solvent bath. The range of particle diameter is 50–500 µm. The SDA meth-od is a very attractive alternative for up-scaling of bead production and also for enhanced control of bead properties such as size, shape, pore size distribution and

surface morphology. Figure 1 illustrates beads produced using viscose solution and SDA method.

The utilisation of beads in different technological applications is dependent on the type of the cellulose solution used for processing. Viscose or cellulose xan-thate dissolved in NaOH is the most com-mon cellulose solution used commercial-ly. However, this type of cellulose solvent has several drawbacks due to the utili-sation of carbon disulfide, a highly toxic hazardous chemical. The advances of ap-plications of beads require the utilization of an environmentally friendly and cost competitive cellulose solvent. The dissolu-tion of natural cellulosic fibre takes place in two steps; first the solvent penetrates into the fibre structure and then sepa-rates the polymer chains from each oth-er. Dissolution mechanism depends on fi-bre type and mostly on the solvent sys-tem. Derivatising solvents react chemi-cally with hydroxyl groups and reduce hy-drogen-bond network which makes it pos-

Figure 1. Electron microscopy images of functional cellulose beads. The morphology of the beads can be tailored during manufacture using SDA technique. Magnification: a) 100x, b) 500x, c) 1,000x, d) 10,000x.

87

sible to dissolve cellulose. Derivatisation prolongs the dissolution procedure and in the case of viscose, for example, re-generation of cellulose releases toxic CS2 gas. Non-derivatising solvents disrupt in-tra-molecular hydrogen-bonding and dis-solve cellulose directly without chemical reactions.

The solvent quality has a great ef-fect on the mechanism of how a fibre dis-solves. Good non-derivatising solvents break down the fibre structure by frag-menting whereas moderate and poor sol-vents first dissolve the inner parts of the fibre with less hemicelluloses. This causes swelling and ballooning of the outer lay-ers. They are also unable to dissolve cel-lulose in greater extent so the dissolution of the fibre is often partial. Many good solvents, e.g. ionic liquids are able to dis-solve even 25% of cellulose directly with-out any pre-treatment. However, they are still debatable because of their water sen-sitivity and difficult purification.

NaOH–water with or without addi-tives has got more attention as a water-based, cheap and environmentally friend-ly solvent. It has been shown, however, that cellulose molecules are not com-pletely dissolved in NaOH–water solvent but form aggregates. Also dependence of solubility on the degree of polymeriza-tion has been studied in NaOH–urea–wa-ter systems and concluded that only low DP cellulose can be dissolved. This means that native cellulose has to be pre-treat-ed in order to reduce the polymer chain length. Acid hydrolysis has been applied usually as a pre-treatment to make cel-lulose molecules shorter and more acces-sible for chemicals or complete hydroly-sis to produce glucose. This is benefi-cial when aim is to produce, e.g. biofu-els. Traditionally hydrolysis has been car-ried out at high temperature and low ac-id concentration or vice versa. Howev-er, this can be very energy demanding and causes degradation of sugar produc-ing unwanted by-products, such as furfu-

ral, (hydroxymethyl)-furfural and formic and acetic acid. Acid treatment on cellu-lose is performed in aqueous environment almost without exceptions. It has been shown that acid hydrolysis in alcohol im-proves the degradation of starch and cel-lulose. This might be due to higher activ-ity coefficient of hydrogen ions in alcohol. Differences in acid hydrolysis of cellulose in water and in ethanol have rarely been studied and it is concluded that degrada-tion rate of cellulose has a dependence on the media favouring ethanol environment. However, connection of the acid hydroly-sis in ethanol environment to the solubili-ty and dissolution mechanism of cellulose has not been studied before.

In the first phase of FuBio we fo-cused our research in developing a new pre-treatment to enhance the dissolution of cellulose in NaOH-urea-water system and allow the preparation of functional cellulose beads using an environmentally friendly process.

2. ObjectivesOur main objective was to study the ef-fect of the ethanol–acid pre-treatment of pulp on the solubility of cellulose in NaOH-urea-water using microscopical and rheological methods. The feasibility of this cellulose solution for preparation of func-tional cellulose beads was also assessed.

3. Research approach

3.1 MaterialsDissolving pulp (Cellulose plus) was pur-chased from Domsjö Fabriker, Sweden. Pulp is a mixture of spruce and pine (60%/40%) with 93%-cellulose content and 0.6% of lignin (Domsjö 2007). In-trinsic viscosity of the delivered pulp was reported to be 530 ± 30 cm3 g-1 accord-ing to SCAN-CM 15:99 standard (com-

88

plies with ISO 5351 standard) after two stage sodium based cooking. Pulp is To-tal Chlorine Free (TCF)-bleached. All re-agents were obtained from commercial sources, analytical grade and used with-out further purification unless otherwise mentioned.

3.2 MethodsEthanol–acid pre-treatment of pulpEthanol (100 cm3, 92.5 w%) was pre-heated in water bath to the reaction tem-perature (25–75°C) and 4 cm3 of hydro-chloric acid (37 w%) was added. After temperature had stabilized, 4.0 g of dry-weighed fibrillated dissolving pulp was added. Stirring was applied in the begin-ning to make sure that pulp was evenly distributed in the reaction vessel. Treat-ment time was varied from 15 min to 5 h. Reaction was stopped by adding cold distilled water (900 cm3) and the mixture of ethanol, acid and water was filtered off instantly on a glass filter (porosity 1). Pulp was washed with distilled water until the pH of the filtrate was neutral and left in 1,000 cm3 of distilled water overnight to ensure solvent exchange from ethanol–acid to water. On the next day pulp was filtered, suction dried and kept in an ov-en at 60°C overnight.

Optical microscopyDissolution mechanisms and visual trans-parency of cellulose solutions were stud-ied with a Wild M20 optical microscope attached to a Nikon Coolpix 990 digital camera. Dissolution steps were recorded when ethanol–acid treated (5 h at 75°C, see above) and untreated pulp was intro-duced to a 0.2 M cupriethylene- diamine (CED) solution. Undissolved fragments were examined after dissolving 0.2% of pre-treated cellulose from total weight in 7% NaOH–12% urea–water solution (see Dissolution of cellulose). Cellulose was treated in ethanol–acid solution for 2 h at 25, 45, 55 and 65°C like described above.

Surface morphology using FE-SEMThe morphology of the untreated pulp and samples treated in ethanol–acid so-lution for 2 h at 25 and 75°C was exam-ined by a Leo Gemini 1530 field emission scanning electron microscope with an In- Lens detector. Samples were coated with carbon in Temcarb TB500 sputter coater (Emscope Laboratories, Ashford, UK). An optimum accelerating voltage was 2.70 kV and magnifications were 5,000 and 50,000.

Residual hemicelluloses in pulp by acid methanolysis-GC8–10 mg of pulp treated in ethanol–acid for 2 h at 25, 45, 75°C and one untreated sample were accurately weighed in pear-shaped flasks. Samples were subjected to acid methanolysis (Sundberg et al. 1996) by adding 2 cm3 of 2 M HCl in dry metha-nol and kept in an oven at 100°C for 5 h. After methanolysis, the resulting solution was neutralized with pyridine and 4 cm3 of sorbitol standard in methanol (0.1 mg cm-3) was added. 1 cm3 of clear solution from each sample was transferred into another pear-shaped flask and dried first under nitrogen flow and then in a vacu-um oven. Samples were dissolved in pyr-idine and silylated with 150 mm3 of hexa-methyldisilazane and 70 mm3 of trimeth-ylchlorosilane overnight. About 1 mm3 of silylated sample was injected via a split injector (250°C, split ratio 25:1) into col-umn coated with dimethyl polysiloxane (HP-1, Hewlett Packard), the film thick-ness being 0.17 lm.

Column temperature was first stabi-lized 1 min at 100°C, then heated 4°C min-1 to 170°C and followed by 12°C min-

1 to 300°C where temperature was kept stable for 7 min. The flame ionization de-tector (FID) temperature was 310°C and carrier gas was hydrogen. Peak posi-tions were recognized comparing peaks to the prepared reference solution con-taining exact amounts of arabinose, xy-

89

lose, galactose, glucose, mannose, rham-nose, glucuronic acid and galacturonic ac-id. Concentrations were calculated from peak areas using calibration factors de-termined from internal sorbitol stan-dard. Two parallel measurements were analysed from each sample and an av-erage value was calculated. Changes in the amount of cellulose were studied by measuring the yield of samples treated for 2 h at 25 and 75°C.

Degree of polymerization of celluloseIntrinsic viscosities of ethanol–acid treat-ed samples were measured according to standard ISO/FDIS 5351:2009 and aver-age degrees of polymerization were cal-culated from intrinsic viscosity values (Immergut et al. 1953) to observe chang-es in cellulose chain length as a function of treatment time and temperature. One untreated sample and one treated for 5 h at 65°C in ethanol without hydrochloric acid were used as references. Oven-dry samples were freeze-dried before disso-lution into 1.0 M CED solution. Capillary temperature was 26.0 ± 0.1°C. Results were used in evaluating recyclability of ethanol–acid solution. Pre-treatment was carried out with 10 times higher volumes and masses than described earlier for 5 h at 65°C and then treating new pulp us-ing same solution without concentrating it between treatments. Intrinsic viscosity was measured and degree of polymeriza-tion was calculated like described above and values were compared with previous results.

Dissolution of celluloseCellulose solutions were prepared by us-ing dissolving pulp which was treated in ethanol–acid for 2 h at 75°C (see Etha-nol–acid pre-treatment of pulp). Humidity of the oven-dry pulp was measured and calculated amount of pulp was activated in 6% NaOH aqueous solution at room

temperature through mechanical stirring. After fibres swelled and mixture became homogeneous it was enriched with NaOH and urea. Final concentrations were 7% NaOH and 12% urea from the solvent weight and 0.2–5% cellulose from total weight of the final solution. Vigorous stir-ring was applied with magnetic stirrer to obtain a homogeneous mixture and then cooled down to -10°C. Mixing was contin-ued until clear transparent solution was obtained, usually less than 20 min.

Rheological measurementsRheological measurements of 0–5% cel-lulose in 7% NaOH–12% urea–water so-lutions were carried out using an Anton Paar Physica MCR 300 rotational rheom-eter with DG 26.7 double-gap cylinder. Temperature was controlled with a TEZ 150P thermostat with an external water cycle. Approximately 4 cm3 of sample was pipetted into the cylinder and tempera-ture was stabilized at 10, 15, 20 and 25°C with accuracy of 0.01°C. Apparent activa-tion energies Ea of the viscous flow were calculated according to Arrhenius law from extrapolated zero-shear rate (Roy et al. 2003) and viscosity values from con-stant shear rates 10, 100 and 1,000 s-1. Cellulose solutions were stored at -5 to 0°C between measurements.

Preparation of functional beadsBead–making procedure involved coag-ulation and regeneration in an acidic liq-uid environment. Small drops of cellulose solutions were dropped in anti-solvent to obtain beads. Factors like acid used, con-centration of acid, drop height, contact time and stirring effect were varied to ob-tain different sets of beads. After coagu-lation/regeneration, each set of bead was washed with distilled water (at room tem-perature). After rinsing the beads were left to stand in distilled water overnight (at room temperature). Subsequently, they were washed with distilled water to

90

remove the residue of acid by monitor-ing the pH of the filtrate. Washing was repeated several times till a neutral pH was obtained. Prior to analysis, the beads were conditioned at 20±2°C and 50±2% RH for 72 hours.

4. ResultsDissolution mechanism was studied with CED solution since it is known to dissolve cellulosic fibres directly at high concentra-tions (ISO/FDIS 5351:2009). When sol-vent is diluted, the dissolution of fibres proceeds via ballooning phase and not by fragmenting. Swelling and ballooning were observed when untreated pulp was added into a 0.2 M CED solution (Figure 2A). Ballooning was rapid and usually be-gan from the tip of the fibre but balloons were seen also in the middle of the fi-bres. Balloons continued to swell un-til they dissolved completely, but leav-ing rings (knots between balloons) undis-solved. Ethanol–acid treated pulp did not show any ballooning (Fig. 2B).

Fibres were dissolved directly through fragmenting and only arbitrary undis-solved fragments were left behind. Etha-

nol–acid treated pulp was dissolved in 7% NaOH–12% urea–water and solution was studied for undissolved fragments. Un-dissolved fibres and ballooning was ob-served after dissolution procedure when pulp was treated in ethanol–acid for 2 h at 25 and 45°C (Figure 3A, B). Swollen fi-bres or balloons were not observed when treatment temperature was 55°C (Fig-ure 3C). However, in solution, rings and fragments from ballooning could be seen, which indicated that dissolution proceed-ed through ballooning. When treatment temperature was 65°C or higher, frag-ments or rings could not be observed at all in NaOH–urea–water solution (Figure 3D). This might be due to the weakened cell wall structure so that solvent can ac-cess throughout the fibre dissolving all the components without creating osmot-ic pressure inside and causing balloon-ing. This would lead to direct dissolution through fragmenting.

The effect of the ethanol–acid treat-ment on fibre wall was studied using FE-SEM. When pulp was treated for 2 h at 25°C fibre wall showed some minor changes compared to untreated pulp (Fig-ures 4A–D). Primary fibre wall or its rem-nants were present on both samples. In the sample treated at 75°C fibril bundles could be seen (Fig. 4E–F). Orientation of the fibrils in the bundles indicated that they are on the secondary fibre wall and primary wall was extensively removed.

Hemicellulose content of the sam-ples and reference pulp was analysed us-ing acid methanolysis and gas chroma-tography (data not presented). Glucose content in samples treated at 25, 45 and 75°C was 8, 5 and 11% lower than in ref-erence. Glucose formed in acid methano-lysis is mainly non-cellulosic. Amounts of galactose, mannose and xylose decreased the most during 75°C pre-treatment (14, 14 and 18%). Differences between ref-erence and samples treated at 25 and 45°C were too low to make any conclu-sions about the connection to the treat-

Figure 2. A Untreated and B ethanol–acid treated pulp (5 h at 75°C) in 0.2 M CED solution.

91

ment temperature. Arabinose was pres-ent in such a low quantities in all sam-ples that it was not possible to conclude any changes due to reaction temperature. Though relative changes in contents are notable, especially between reference and samples treated at 75°C, the overall con-centration of the hemicellulose constitu-ents decreased only from 4.8 to 4.3%. It can be concluded that ethanol–acid treat-ment does not have an impact on the hemicellulose content of the pulp. Yields of samples treated at 25 and 75°C were 98 and 97%.

Intrinsic viscosities of the samples were measured using capillary viscome-try. Viscosity value of the pulp treated in ethanol (5 h at 65°C) without hydrochlo-ric acid is only 1 cm3 g-1 lower than that of untreated reference. Both values are in good agreement with the viscosity val-ue of the starting material. This shows that ethanol treatment alone does not cause degradation of cellulose. When hy-drochloric acid was added, degradation of cellulose chains took place even at low temperatures (Figure 5). Previous study showed that a similar treatment in water–

Figure 3. Pulp treated in ethanol–acid for 2 h at A 25, B 45, C 55 and D 65°C. Cellulose content was 0.2% from total weight and solvent was 7% NaOH–12% urea–water. Scale bars are 100 µm.

Figure 4. SEM images of reference pulp A, B and pulp pre-treated in ethanol–acid for 2 h at C, D 25 and E, F 75°C. Magnifications are in top row 5,000 and in bottom 50,000×.

92

hydrochloric acid mixture needed nota-bly longer time and temperature to reach similar changes in intrinsic viscosity. Use of ethanol instead of water enhances the degradation kinetics drastically. This was due to higher mobility of the molecules close to the boiling point of ethanol and increased activity coefficients of hydro-chloric acid in ethanol.

The degree of polymerization de-scended more rapidly when reaction tem-perature increased (Figure 5). Degrada-tion of cellulose chains also reached the levelling-off degree of polymerization (LODP) faster at higher temperatures and degradation slowed down substantially af-ter the first hour. At lower temperatures degradation seemed to continue even af-ter five hours. Ethanol–hydrochloric acid treatment was repeated at 65°C for 5 h with 10 times higher volumes and mass-es. Ethanol–acid solution from the first treatment was re-used and treatment was repeated with new pulp without concen-trating the solution. Intrinsic viscosities were measured and degrees of polymer-

ization were calculated. DP values were only slightly higher (198 and 203) than calculated on the first time (189) with lower volume and mass (Figure 5). Simi-lar DP values indicated that ethanol–acid solution was possible to use at least twice and gain practically same DP values. This indicated that the pre-treatment process can be up-scaled and the solution can be recycled by small supplemental addi-tions of acid and ethanol (Figure 6). After pre-treatment ethanol–acid solution could be pressed out and enriched before re-use.

Ethanol–acid treated (2 h at 75°C) cellulose was dissolved successfully in 7% NaOH–12% urea–water in various amounts (0.2–5% from total weight) and clear solutions were obtained. At shear rates below 10 s-1 and at low cellulose concentrations, viscosity was fluctuating and hence left out from calculations. So-lutions behaved like Newtonian liquids on higher shear rates at low cellulose con-centrations (Figure 7). However, when cellulose concentration

Figure 5. Viscosity average degree of polymerization (DPv) of ethanol–acid treated dissolving pulp at various temperatures as a function of time.

93

Increased, shear-thinning behaviour was observed uniformly on the whole shear rate range. At each cellulose con-centration viscosity increased as tempera-ture decreased. Time to stabilize the tem-perature and measure the viscosity was approximately 10 min for all the samples. This is notably shorter than gelation time for the temperature and cellulose concen-tration used here.

Arrhenius plots could be fitted linearly at all cellulose concentrations and appar-ent activation energy of the viscous flow

was calculated. Zeroshear- rate viscosities were extrapolated from viscosity- shear-rate curves. On shear rates up to 100 s-1 activation energies were growing until cellulose concentration reached 3% (Fig-ure 8). When shear rate was 1,000 s-1, the activation energy increased up to a concentration level of 4%. The increase of the activation energy as a function of cellulose concentration indicates that cel-lulose was dissolved completely. At high-er concentrations and lower shear rates a decline was seen in the activation energy

Figure 6. A diagram of the pre-treatment prior to cellulose dissolution steps.

Figure 7. Viscosity of 0–5% cellulose solutions in 7% NaOH–12% urea–water solvent system at temperatures 10–25ºC as a function of shear rate.

Figure 8. Apparent activation energies of the viscous flow calculated on shear rates 0, 10, 100 and 1,000 s-1 as a function of cellulose concentration. Lines are given to guide the eye.

94

values. This indicates that addition of cel-lulose does not change the solution prop-erties but cellulose forms aggregates and inclusion complexes.

Cellulose beads were prepared used the resulting cellulose solution in NaOH-urea-water after ethanol-hydrochloric ac-id pre-treatments. The preparation meth-od was generation of manual drop using an automatic pipette and coagulation in acidic bath. A set of preliminary experi-ments were carried out aiming at under-standing the effects of anti-solvent condi-tions and cellulose concentration in solu-tion on bead shape as observed using op-tical microscopy. Effect of stirring was al-so assessed. The results are summarized in Figure 9. Cellulose solutions with con-centration of 5% yielded spherical beads with regular shape when combined to in-organic acid concentration in the coagula-tion bath of 10%. Stirring caused chang-es in the shape of the beads. Organic ac-ids could also be used but the coagula-tion process was slow and released vola-tiles. Pure water was not feasible to ob-

tain spherical beads. However, the addi-tion of table sugar to reduce the surface tension of water allowed the formation of spheres. Several new conditions of coag-ulation will be tested in the second phase of the programme using our spinning drop atomizer device.


Preparation of cellulosic beads with spin-ning disk atomization with further devel-oped nozzle atomizers was proven to be a procedure with good controllability. Due to its capability to large scale production it enables simple and economical meth-od for preparation of spherical cellulos-ic beads to versatile purposes in several technological applications. We foresee a business potential related to pharmaceu-tical technology, chromatographic appli-cations, water cleaning and fractionation of biopolymers in biorefinery. When inves-

Figure 9. Cellulose beads obtained using different concentrations of cellulose solution and different conditions of coagulation bath.

95

tigating materials that could replace mi-crocrystalline cellulose (MCC) in produc-tion of pharmaceutical tablets, there is a need of improvement of material proper-ties in comparison to MCC, such as flow ability and particle size and mass varia-tion. Our bead making technology makes possible to produce cellulosic bead with superior flow ability compared to MCC and uniform mass. The beads are also suit-able for tablet making. However, if the drug release profile of the material needs to be tailored, it is necessary to control the pore size, connectivity and geometry of the beads and we have several options to tune these properties. Cellulose beads can also be widely utilized in biotechno-logical applications, e.g. for immobiliza-tion support material in such processes as ethanol production, lactose hydrolysis or enzyme immobilization. Beads from cellu-lose solutions have been tested as an en-vironmentally friendly method for water purification such as Cu2+ ion adsorbents. Beads made of cellulose/chitin have also been tested for Cd2+, Cu2+, and Pb2+ removal in aqueous solution. Composites products made of cellulose beads loaded with solid and liquid hydrophilic solubilis-ers have been also reported and have a good potential to be used in several val-ue chains.


The breakthrough of cellulose beads in commercial applications is dependent in advances of process technology to up-scale the production and to tailor bead properties by combining chemical and physicochemical approaches. The tailor-ing of bead morphology, chemistry and surface chemistry and consistent produc-tion of large amounts will be achieved by cooperation with different research and innovation partners which will help us

in designing new beads for several val-ue chains.


Trygg, J., Fardim, P. 2011. Enhance-ment of cellulose dissolution in water-based solvent via ethanol–hydrochlo-ric acid pre-treatment, Cellulose (2011) 18:987–994.

8.References1. Domsjö (2007) Specification

Domsjö Cellulose. http://www.domsjoe.com - Produkter - Specialcellulosa - Produktinformation Domsjö Cellulose. Accessed 10 November 2010.

2. Immergut E, Schurz J, Mark H (1953) Viskositätszahl-Molekulargewichts-Beziehung für Cellulose und Untersuchungen von Nitrocellulose in verschiedenen Lösungsmitteln. Monatsh Chem 84:219-249, DOI 10.1007/BF00899186.

3. Roy C, Budtova T, Navard P (2003) Rheological properties and gelation of aqueous cellulose-NaOH solutions. Biomacromolecules 4:256–264.

4. Sundberg A, Sundberg K, Lillandt C, Holmbom B (1996) Determination of hemicelluloses and pectins in wood and pulp fibres by acid methanolysis and gas chromatography. Nordic Pulp and Paper Res J 11:216–219, DOI 10.3183/NPPRJ-1996-11-04-p216-219

96


Editors

Partners:

Aalto University

Jukka Seppälä, Sami Lipponen and Eve Saarikoski

Key researchers:

Jukka Seppälä, Sami Lipponen, Eve Saarikoski

97

AbstractIn an attempt to move from petroleum-based polymers to bio-based polymers, blending the

two is typically a step in the development path. In this specific work, we present a feasible way

of synthesising cellulose/polyethylene-co-acrylic acid (PE-co-AA) blends. Cellulose and PE-co-AA

were mixed in alkaline water solution and the analyses of the dried blend revealed a well-mixed

morphology over the whole composition range. The cellulose/PE-co-AA blends were found to be

thermo-formable, when the cellulose concentration remained below 50%. The melt rheology of

the blend was ‘normal’ at cellulose concentration of 0-15%, where after the ‘saturating’ point

was reached. The mouldability of the blend was harder after this point, but it remained mould-

able up to 50 % of cellulose. The stiffness of the blend at the 50 % cellulose concentration was

found to be comparable with the polyethylene-based composites. At higher cellulose concentra-

tions (>50-wt%) the thermo-formability of the blend was poor, but films could be prepared via

solution casting. These films showed excellent water vapour barrier properties, when compared

with the neat cellulose.

98

1. Background

The polymers from renewable resourc-es raise a great deal of interest today because of the continuous increase in oil prices. Continuous research and de-velopment will be made in the near fu-ture when looking at alternative materi-als for example for packaging plastics and for other plastic materials. The very sub-stantial, but so far little exploited cate-gory of cellulose based material, which has led to some of the very first indus-trial-products such as celluloid and cel-lophane, still offers numerous new pos-sibilities for polymeric materials. In addi-tion, cellulose, starch and chitin are the most abundant polymers in the earth and therefore highly accessible for the plastic manufacturers.[1] Cellulose is the most common natural polymer but its usage is strongly limit-ed by its poor solubility and inability to thermoform. The strong hydrogen bond-ing between cellulose fibres results in a tightly packed crystalline material which is insoluble to water. These strong inter-molecular interactions are the main rea-son why cellulose rather degrades than becomes deformable when heated. These disadvantages limits strongly its suit-ability e.g. in natural-filler/thermoplas-tic blends and composites. Still, cellulose plays in an important role as a raw ma-terial for future plastic materials. Basical-ly two main groups of cellulose-materials can be distinguished, regenerated cellu-lose and cellulose derivatives, which are used in industrial scale processes.[1]

Regenerated cellulose grades are dissolvable in special solvent systems. Steam exposed wood pulp, RC from cu-prammonium solution and microcrystal-line cellulose could be dissolved in alka-line aqueous solution (NaOH 8-10 wt-%). For the normal grades of cellulose the solvent power needs to be increased by e.g. urea or thiourea.[2] Normally re-generated celluloses are suitable for fi-

bre and film production from convention-al and new processes. In contrast to re-generated cellulose grades, cellulose de-rivatives can be used for extrusion and moulding processes. The thermoform ability (as well the solubility) of cellulose can be enhanced by derivatisation. There are a variety of thermoplastic cellulose grades where the free OH-groups are re-acted with a certain reagents resulting in such as cellulose acetate, cellulose aceto-butyrate, benzyl cellulose, ethyl cellulose etc. As a drawback, this means in prac-tice that most of the OH-groups needs to be reacted (DS~3) which in contrast re-duces e.g. biodegradability. This can be avoided by forming longer biodegradable chains, e.g. grafting with ε-caprolactone [3], and thermoplastic cellulose derivative is gained with a lower DS level.[1]

Blending cellulose with synthetic poly-mers is an important process to prepare functional polymeric materials. Conven-tionally all polymer blends are prepared in a melt extruder, and this is also the case if thermoplastic cellulose derivatives are blended e.g. with polyethylvinylace-tate [4]. Likewise, regenerated cellulose grades are blended with synthetic poly-mers in solutions using new kinds of non-aqueous solvent systems, like N-meth-yl-2-pyrrolidinone/Lithium chloride [5] or N,N-dimethylacetamide/Lithium chloride [6,7] for blending cellulose with polyvi-nylalcohol, and dimethyl sulfoxide/para-formaldehyde for blending cellulose with poly-vinylpyrrolidone [8], poly-4-vinylpyr-idine [9], or polyacrylonitrile [10]. To re-duce the pollution from organic solvents, new aqueous systems for synthetic poly-mers have also been developed, like N-methylmorpholine-N-oxide/H2O [11,12].

In parallel of cellulose, starch has been used more widely due to its ‘less problematic nature’. Especially thermo-plastic blends of starch and poly(ethylene-co acrylic acid) were studied for a number of years starting at the end of the seven-ties by Otey et al. [13], and a series of

99

articles and patents describes their prepa-ration and properties. In addition Fanta et al. [14] and many other research groups [15,16] studied starch-poly(ethylene-co-acrylic acid) blends and other starch con-taining plastics for agricultural purposes in early 90s.

In addition to natural polymer based blends, natural fibre reinforced compos-ites are finding much interest as a sub-stitute for glass and carbon fibre rein-forced polymer composites. Spun cellu-lose fibres from the viscose, lyocell and carbamate processes have been used to reinforce thermoplastic commodity poly-mers, such as polypropylene (PP), poly-ethylene (PE) and (high impact) poly-styrene (HIPS) as well as poly(lactic ac-id) (PLA) and a thermoplastic elastomer (TPE) for injection moulding applications. In addition to the eco-friendly character-istics, the natural fibres are favoured by lower price and lighter weight as well due to the reduced abrasion in the processing machines. The applications are found es-pecially in composites used in automotive industry.[17,18]

2. ObjectivesThe objective of this work was to study the possibility of prepare blends of cel-lulose and poly(ethylene-co-acrylic acid) (PE-co-AA) by mixing their alkaline water solutions. The mixing method was var-ied for finding the way for best morphol-ogy for the suspension. The morpholo-gy of the formed suspension was charac-terised by rotational rheometry and op-tical microscopy. Also some spinning tri-als were performed for checking the spin-nability of the formed cellulose/PE-co-AA suspension. Finally, the final dry blends were characterized by Differential Scan-ning Calorimeter (DSC), Scanning Elec-tron Microscope (SEM), Fourier Transform Infrared (FTIR), and Rotational Rheome-ter. In addition, the water vapour barrier

property of the blends was studied. Also the mouldability of the blends were stud-ied by preparing 3D-objectives from the blend using injection moulding

3. Research approachAlkaline cellulose water solution (3 wt-% cellulose, 6 wt-% NaOH, 1.3 wt-% ZnO) was received from the Tampere Universi-ty of Technology. The preparation of this solution is described in the patent ap-plication of Vehviläinen et al. [19]. Af-ter preparation the cellulose solution was kept frozen in freezer (-20 °C). Alkaline polyethylene-co-acrylic acid water solu-tion (20 wt-% PE-co-AA, pH ~10) was ob-tained from BIM Finland Oy and stored in refrigerator (+5°C). The both solutions were allowed to warm up just prior the use. All the preparation steps were made at 23°C unless otherwise mentioned.

Mixing of the abovementioned solu-tions was done in 100–500 ml glass ves-sel under vigorous magnetic stirring. Dosages of PE-co-AA solution were do-ne with a syringe and needle by feeding it slowly to the vortex of cellulose solution (cellulose batch ~20–100 ml, the over-all amount of the final dry blend ~4–6 g.). After mixing for 24 h, the mixed solu-tion was frozen up at -20°C where after it was slowly warmed back at RT and mixed for another 24h. The mixing step was fol-lowed by centrifugation of the solution for 90 min. during a thin layer of non-uni-form phase was separated at the surface of the solution. The non-uniform phase was removed and the rest of the solution was injected with a needle in 100–300 ml of acidic solution (20 wt-% of H2SO4) to regenerate the cellulose/PE-co-AA blend composite. Finally, the precipitation was filtered, washed several times with water (until neutral) and then dried in vacuum at 40°C overnight. The prepared polymer suspensions and their relevant polymer blends are gathered in table 1.

100

All the analyses were performed at 23°C unless otherwise mentioned. The rheological characterization of the solu-tions and suspensions were done using a Physica MCR 301 rotational rheome-ter (Anton Paar GmbH, Austria) equipped with DIN concentric cylinders geome-try with bob and metal cup diameters of 26.665 mm and 28.918 mm, respectively (CC 27). In order to estimate the changes in suspension structure during mixing, a series of measurements were done. First a frequency sweep measurement with frequency ranging from 0.1 to 50 rad/s at strain of 0.2% in linear viscoelastic re-gion was performed, with the exception of the neat cellulose solution that was mea-sured with 5% strain due to the low vis-cosity. Next, rotational measurement of steady shear flow (0.01–1000 1/s) was performed for all of the suspensions and solutions. In addition, a 10 min recovery time was used before the steady shear flow test.

To clarify the rheological results, changes in the suspension structure were also estimated with centrifugation (Here-us Multifuge 3s, 4000 rpm) and Olympus BH-2 optical microscopy equipped with a digital camera. A drop of the suspension was placed between microscope glass plates and the pictures were taken with 100x and 200x magnification.

The thermal behaviour of the dry blends was measured with a Mettler To-ledo DSC 821e under a nitrogen atmo-sphere. The thermal history of the blend was destroyed by heating the sample to 150°C at 20°C/min. The crystallization behaviour was then determined from the peak area and the peak tempera-ture of the crystallization exotherm (TC), which was obtained at a cooling rate of 10 ºC/min (from 150°C to -30°C). After the cooling step, the melting endotherms (∆H) and the peak melting temperatures (Tm) were measured by reheating the sample at 10°C/min. The size of the sam-

Table 1. The compositions of the suspensions/blends.

Suspension/Blend ID

Cell-100

Cell-90

Cell-75

Cell-50

Cell-40

Cell-30

Cell-15

Cell-5

Cell-0

Cell./PE-co-AA solutions volume ratio in susp.

-

100/1.5

100/5

100/15

100/22.5

100/35

100/85

100/285

-

Polymer concen. in susp.(wt-%)

3

3.3

3.8

5.2

6.1

7.4

10.8

15.6

20

Cellulose(wt-%)

100

90

75

50

40

30

15

5

0

PE-co-AA(wt-%)

0

10

25

50

60

70

85

95

50

101

ples varied between 10–20 mg depend-ing on the PE-co-AA concentration in the blend (PE-co-AA amount in the sample 5 mg).

The morphology of the blend was an-alysed with Hitachi TM-1000 field emis-sion SEM using an acceleration voltage of 15 kV. The SEM-micrographs were tak-en from the sample before and after the sample was leached with hot toluene (at 70° for 3 h to remove PE-co-AA phase) or with H2SO4 (at RT for 4 h to remove cellu-lose phase). Before analysis the samples were sputter-coated with gold.

4. Results

4.1 Blending of cellulose and PE-co-AA in alkaline water solutionThe basic idea of blending cellulose and PE-co-AA relies on the fact that both of the polymers are soluble in alkaline water phase, where they were expected to form a homogeneous blend, and that both of the polymers regenerate fast and simul-taneously via acidic treatment, where the homogeneous morphology of the blend should be maintained (i.e. diminished phase separation during regeneration).

In practice, however, the preparation of the cellulose/PE-co-AA blend via solution mixing was not so straightforward. The main drawback was that the PE-co-AA so-lution started to form gel particles (Fig-ure 1a) when it was added in cellulose solution. It became obvious that the PE-co-AA was soluble only in modestly alka-line solution (pH 10) whereas at higher pH’s (cellulose solution pH 14) the solu-bility was diminished. In order to clarify the structural changes in the cellulose/PE-co-AA suspensions during the mixing, mi-croscopy pictures from the solutions were taken after 15 min, 3 h and 24 h of mix-ing (Figure 1a, 1b, and 1c, respective-ly). As the pictures show, large particles of PE-co-AA were seen after 15min, but when mixing was continued, the particles and aggregates started to decay. When the mixing was continued to 24 h, most of the remaining aggregates were broken-down and spread evenly throughout the suspension.

To get more information of the solu-tion stage properties the produced sus-pensions were also subjected to dynam-ic oscillation frequency sweep tests with rheometry. These results clearly showed how the complex viscosity of the suspen-sion increased in parallel of mixing time (Figure 2). This evolution was the sum

Figure 1. Optical microscopy pictures of the cellulose/PE-co-AA suspension (Cell-50) after mixed for a) 15 min, b) 3 h, or c) 24 h. The size of the scale bar is 100μm.

102

of several different phenomena, i.e. the solubility levels (soluble vs. gel vs. ag-gregate) of cellulose and PE-co-AA as well the size and amount of these differ-ent phases during the mixing. Therefore it was difficult to find any direct correla-tion between the rheological result and the miscibility of the polymers.

In order to see the influence of cel-lulose gel formation on the rheologi-cal properties, a freeze-melting pro-cedure was performed after 24 h mix-ing. This freezing-melting was expected to dissolve any gelled cellulose phase. Fig. 3 shows the storage and loss mod-ulus of Cell-50 suspension as function of different mixing procedures. These re-sults clearly showed that if the suspen-sion was mixed for 48 h straight without freezing and melting, the gel like behav-ior was seen as a rise of G’. However, when the freezing-melting procedure was performed after 24 h mixing, the evo-lution of the suspension was stopped, i.e. the G’ and G’’ levels remained un-changed during the last 24 h mixing (i.e.

the curves of Cell-50-24h and Cell-50-fm overlaps in figure 3). 4.2 Thermal properties of the cellulose/PE-co-AA blendsThe thermal behaviour of the blends was analysed with DSC. Over the measured temperature range there was no indica-tion of thermal transition points of cellu-lose phase, but the thermal transitions of PE-co-AA phase were clearly observed. These transitions were not influenced, ex-cept the crystallization temperature, which was clearly higher in all the blend samples than in neat PE-co-AA (Figure 4B, 4C vs. 4A; TC=72°C vs. TC=65°C, respectively). The increase in crystallization temperature, as also observed in the C-5 sample where the concentration of cellulose was the low-est, clearly indicated that the rigid/solid cel-lulose acted as a strong nucleation agent for the molten PE-co-AA phase. One expla-nation for this strong nucleation effect in cellulose/PE-co-AA blends is that the phas-es were well mixed. This enables large in-terfacial area between cellulose and PE-co-

Figure 2. Complex viscosities of neat cellulose solution (º), Cell-50-15min (-), Cell-50-3h (+), Cell-50-24h ( ) and Cell-50-48h (◊).

Figure 3. Frequency sweeps of Cell-50-24h ( ), Cell-50-48h ( ), and Cell-50-fm(●), suspensions. Storage modulus G’ (filled symbols) and loss modulus G’’ (open symbols).

103

AA which would intensify any weak nucle-ation effects.

4.3 Morphology of the cellulose/PE-co-AA blendsThe morphology of the synthesized blend composites was studied more detailed us-ing scanning electron microscopy (SEM). The SEM pictures of the selected surfac-es are presented in Figure 5. The H2SO4 leached samples Cell-15, Cell-30, and Cell-50 contained small ~0.5-5 µm size small holes indicating the past location of the cellulose phase. In Cell-15 (Fig-ure 5A) the holes were clearly apart from each other while the PE-co-AA formed the continuous phase in the blend. The situ-ation was slightly different in the com-posite Cell-30 and Cell-50 (Figure 5B and 5C) where the surface was more robust as the cellulose phase was no more isolat-ed. It was obvious that both of the poly-mers in, cellulose and PE-co-AA, formed the continuous phase, i.e. co-continuous morphology. The toluene leached Cell-75 sample (Figure 5D) differentiated partly

from the other samples as only small frac-tion of the PE-co-AA phase was removed from the sample. It seemed quite obvi-ous that most of the PE-co-AA phase was bounded tightly within the cellulose and therefore out of reach of toluene leaching. Still, this result confirmed those observa-tions which indicated that the polymers were dispersed homogenously, and only a small fraction of the PE-co-AA phase re-mained unbound/free and within reach of the toluene leaching.

4.4 The rheological properties of the cellulose/PE-co-AA blendsThe blends with lower cellulose concen-tration (from Cell-50 to Cell-0) were ther-mo-formable and therefore also their melt rheology was able to study. As expect-ed, the viscosity increased in parallel of the cellulose concentration (Figure 6A). Noteworthy should be mention that the zero shear viscosity started to approach infinity at rather low cellulose concentra-tion (~15-20 wt-%). At this point also the storage modulus started to be the domi-

Figure 4. Crystallization endotherms of A) Cell-0, B) Cell-5, and C) Cell-50.

104

nating component in the melt (Figure 6B). This kind of behaviour is usually seen in composites where the filler concentration start to reach the saturation/percholation point [20-22], which, in this case, would in-dicated that the cellulose phase started to form a continuous phase in the melt. This observation correlated well with the SEM results which indicated that the perchlo-ration point (co-continuous morphology) located somewhere between the cellulose concentrations of 15-30 wt-% in the blend.

4.5 Mechanical properties of the blend compositesThe DMA was used to determine the stiff-ness of the blend composites. As could be expected, the modulus of the compos-ite increased foreseeable within the cellu-lose concentration (Figure 7). When com-paring with the more traditional polyolefin composites, the stiffness of cellulose/PE-co-AA blend composites was high enough to compensate e.g. particulate filled poly-ethylene composites.

Figure 5. The SEM-micrographs of H2SO4 leached A) Cell-15, B) Cell-30, C) Cell-50, and toluene leached D) Cell-75 samples.

105

Figure 6. Frequency sweep study of the Cell-0 ( ), Cell-5 (●), Cell-10 ( ), Cell-15 ( ), Cell-20 ( ), and Cell-30 (+). A) Complex viscosity, and B) Storage modulus G’ (filled symbols) and loss modulus G’’ (open symbols) as a function frequency f (Hz). Blends Cell-5 and Cell-30 were removed from Figure B for clarification.

Figure 7. Storage modulus measured with DMA. cellulose/PE-co-AA blend composites (solid lines) and reference composites PE/ATH(40 wt-%) (), PP/μSi (30 wt-% (o).

106

4.6 The barrier properties of the cellulose/PE-co-AA blendsDue to the hydrophobic nature of the PE-co-AA the barrier properties of the mod-ified celluloses were of interest. Howev-er, mostly due to requirement to devel-op new sample preparation method, on-ly preliminary results were obtained. Still, the cellulose/PE-co-AA blend showed promising results as barrier for water va-pour. The preliminary results are present-ed in figure 8. It was obvious that blend-ing a small amount of polyolefin based polymer brought some hydrophobicity in-to cellulose which decreased markedly the water vapour transition rate. 4.7 Spinning trials of the cellulose/PE-co-AA suspensionsThe preliminary tests of characterizing spinnability of the cellulose/PE-co-AA suspension were performed by preparing wet-spun fibres from Cell-90 suspension. In addition, the results were compared with the parallel spinnability test of the neat cellulose solution. As a results it can be said, that it was possible to spin fibres from the cellulose/PE-co-AA suspension.

Still, the cellulose/PE-co-AA fibres were weaker than the fibres made of pure cel-lulose. Also the quality of the neat cellu-lose fibres was better. The spinning meth-od with long coagulation length was ben-eficial for the cellulose/PE-co-AA suspen-sions and improved the quality and me-chanical strength of these fibres.


Blending cellulose with other polymers is an important procedure to prepare functional polymeric composites. Con-ventionally all the polymer blends are prepared in melt extruder, and that is al-so the case if thermoplastic cellulose de-rivatives are used. In these cases, how-ever, the cellulose derivatives lack of free OH-groups (if DS 3) which in contrast re-duces e.g. biodegradability. To remain the true nature of the cellulose, regenerat-ed cellulose grades are used and these blends are prepared in solution. The solu-tion mixing enables also better morpholo-

Figure 8. The water vapour transition rates of Cell-90 vs. Cell-100 and some homopolymers.

107

gy in the blend which is crucial when the properties of the two different kinds of polymer are to be combined.

In this study the cellulose/polyolefin blends were prepared in alkaline aque-ous solution systems, which enabled good polymer/polymer morphology. At cellulose concentrations up to 50 % the final blend was thermo-formable which enabled the preparation of rigid 3D-shaped speci-mens. In general, natural fibre reinforced composites are finding much interest as a substitute for glass and carbon fibre re-inforced polymer composites, and there-fore these new blends/composites could found applications in the field of polyole-fin composites.

At higher cellulose concentrations (>50-wt%) the thermo-formability of the blend was poor but the good polymer/polymer morphology remained. The prop-erties of polyolefin and cellulose (with OH-groups) are effectively combined and the films can be prepared via solution cast-ing. These films showed excellent barri-er properties and the applications would be e.g. on the packing where good bar-rier properties are required. Still, these barrier results were only preliminary, and needs to be studied more detailed.


Lipponen, S., Saarikoski, E., Seppälä, J. 10.03.2010. Synthesis and character-ization of cellulose/polyethylene copoly-mer blends. Poster presentation in FuBio Workshop.

Lipponen, S., Saarikoski, E., Seppälä, J. 06.05.2010. Natural polymer based blends and composites. Literature survey.

Lipponen, S., Seppälä, J., Saarikoski, E., Salminen, A. 07.09.2010. Cellulose/Polyethylene-co-acrylic acid blend via al-kaline solution mixing. Invention disclo-sure.

Lipponen, S., Saarikoski, E., Seppälä, J. 11.10.2010. Cellulose/PE-co-AA blend; synthesis, characterization, and applica-tion. Poster presentation in FuBio T2 Post-er Session.

Saarikoski, E., Seppälä, J. 10.6.2011. Gelation Behaviors of Cellulose/PE-co-AA Alkaline Solutions, Nordic Rheology Con-ference, Helsinki. Oral Presentation.

Saarikoski, E., Lipponen S., Seppälä, J. 2011. Gelation Behaviors of Cellulose/PE-co-AA Alkaline Solutions, Annual Trans-actions of The Nordic Rheology Society. 19: 157-162

Saarikoski, E., Lipponen, S., Ris-sanen, M., Seppälä, J. 2011. Blending Cellulose with Polyethylene-co-acrylic ac-id in Alkaline Water SuspensionRheologi-cal Properties of Cellulose/PE-co-AA Alka-line Solutions, Manuscript send to Cellu-lose, Revisions required.

Lipponen, S., Saarikoski, E., Ris-sanen, M., Seppälä, J. 2011. Prepara-tion and properties of Cellulose/PE-co-AA blend composites, Manuscript.

Rissanen M., 19.5.2011. Wet spinning of Cellulose/PE-co-AA, Report.

108

7. References

1. Simon J., Müller H.P., Koch R., Müller V. 1998. Thermoplastic and biodegradable polymers of cellulose, Polym. Deg. Stab. 59: 107-115.

2. Zhou J., Zhang L. 2001. Structure and properties of Blend Mebranes Prepared from Cellulose and Alginate in NaOH/Urea Aqueous Solution, J. Polym. Sci. Pol. Phys. 39: 451-458.

3. Yan C., Zhang J., Lv Y., Yu J., Wu J., Zhang J., He J. 2009 Thermoplastic Cellulose-graft-poly(L-lactide) Copolymers Homogeneously Synthesised in an Ionic Liquid with 4-Dimethylaminopyridine Catalyst, Biomacromolecules 10: 2013-2018.

4. Girija BG, Sailaja RRN, Sharmistha B, Deepti MV. 2010. Mechanical and Thermal Properties of EVA Blended with Biodegradable Ethyl Cellulose, J Appl Polym Phys. 116: 1044-1056.

5. Schartel B, Wendling J, Wendorff JH. 1996. Cellulose/Poly(vinylalcohol) Blends. 1. Influence of Miscibility and Water content on Relaxations, Macromolecules 29: 1521-1527.

6. Nishio Y, Manley R.St.J. 1988. Cellulose/Poly(vinyl alcohol) Blends Prepared from Solutions in N,N-Dimethylacetamide-Lithium Chloride, Macromolecules 21: 1270-1277.

7. Nishio Y, Haratani T, Takahashi T, Manley J.St.R. 1989. Cellulose/poly(vinylalcohol) Blends: An Estimation of Thermodynamic Polymer-Polymer Interaction by Melting Point Depression Analysis, Macromolecules 22: 2547-2549

8. Masson JF, Manley R.St.J. 1991. Miscible Blends of Cellulose and Poly(vinylpyrrolidone), Macromolecules 24: 6670-6679.

9. Masson JF, Manley R.St.J. 1991. Cellulose/Poly(4-vinylpyridine) Blends, Macromolecules 24: 5914-5921.

10. He C, Pang F, Wang Q. 2002. Properties of Cellulose/PAN Blend Membrane, J Appl Polym Sci 83: 3105-3111.

11. Biganska O, Navard P, Bédué O. 2002. Crystallization of cellulose/N-methylmorpholine-N-oxide hydrate solutions, Polymer 43: 6139-6145.

12. Liu X, Chen Q, Pan H. 2007. Rheological behavior of chitosan derivative/cellulose polyblends from N-methylmorpholine N-oxide/H2O solution, J Mat Sci 42: 6510-6514.

13. Otey F.H., Westhoff R.P., Russell C.R. 1977. Biodegradable Films from Starch and Ethylene-Acrylic Acid Copolymer, Ind. Eng. Chem. Prod. Res. Dev. 16: 305-308.

109

14. Fanta G.F., Swanson C.L., Shogren R.L. 1992. Starch-poly(ethylene-co-acrylic acid) composite films. Effect of processing conditions on morphology and properties, J. Appl. Polym. Sci. 44: 2037-2042.

15. Swanson C.L., Shogren R.L., Fanta G.F., Imam S.H. 1993. Starch-Plastic Materials--Preparation, Physical Properties, and Biodegradability (A Review of Recent USDA Research), J. Env. Polym. Deg. 1: 155-166.

16. Gould M.J., Gordon S.H., Dexter L.B., Swanson C.L., Chapter 7 Biodegradation of Starch-Containing Plastics, Agricultural and Synthetic Polymers: Biodegradability and Utilization, edited by J.E. Glass and G. Swift, American Chemical Society 1990.

17. Taha I., Ziegmann G. 2006. A comparison of Mechanical Properties of natural Fiber Filled Biodegradable and Polyolefin Polymers, J. Comp. Mat. 40: 1933-1946.

18. Gauthier R. ,Joly C., Coupas A.C., Gauthier H., Escoubes M. 1998. Interfaces in polyolefin/cellulosic fiber composites: chemical coupling, morphology, correlation with adhesion and aging in moisture, Polym. Comp. 19: 287-300.

19. Vehviläinen M, Nousiainen P., Kamppuri T., Järventausta M., A method of dissolving cellulose and a cellulosic product obtained from a solution comprising dissolved cellulose, WO 2009/135875.

20. Supaphol P., Harnsiri W. 2006. Rheological and isothermal crystallization characteristics of neat and calcium carbonate-filled syndiotactic polypropylene, J. App. Polym. Sci. 100: 4515-4525.

21. Samsudin M.S.F., Mohd Ishak Z.A., Jikan S.S., Ariff Z.M., Ariffin A. 2006. Effect of filler treatments on rheological behavior of calcium carbonate and talc-filled polypropylene hybrid composites, J. App. Polym. Sci. 102: 5421-5426.

22. Wang K., Jingshen W., Hanmin Z. 2006. Effect of interfacial modification on the rheological properties and nucleation behaviour of PP-BaSO4 composites, Polym. Polym. Comp. 14: 473-483.

110


Editors

Partners:


Aalto University


Elina Laatikainen and Elias Retulainen

Key researchers:

Harri Heikkinen, Elina Laatikainen, Jaakko Pere, Elias Retulainen, Sauli Vuoti, Xiling Zeng

Leena-Sisko Johansson

Ilkka Kilpeläinen, Jorma Matikainen, Pirkko Karhunen

111

AbstractPaper and board must have improved formability, when used as future packaging materials.

Extensibility is a crucial component of formability. In this work, the aim was to increase form-

ability by mechanical treatment of the fibres, by affecting the chemical composition of fibres,

and by chemical treatments. Several physical factors were found to affect the extensibility of

paper: Certain fibre deformations induced mechanically, sheet shrinkage, amount of moisture

and chemical modification of fibres. Several chemical modifications were tested. For example

the hydroxypropylated highly-refined fibres showed promising results and produced sheets with

high transparency, low porosity, increased hydrophobicity and increased elongation (8.5%),

when rewetted. When the sheet was allowed to dry freely, the elongation was increased up to

16%.

112

1. Background

Packaging materials require formability especially in converting operations, where different forms are imposed on the paper and board. The production of both rigid and flexible packages includes normally several stages where the formability and extensibility is required. When three-di-mensional forms, such as plates or trays are produced this is commonly accom-plished by applying a series of creasing lines into the material. With higher ex-tensibility and formability the number of creasing lines can be reduced, or more demanding structures formed, and even deep-draw type of forming could be in-troduced.

Conventionally the research of me-chanical properties of paper has concen-trated in increasing the strength and stiff-ness properties. The stretch at break or elongational properties often have not even been reported in the scientific lit-erature. Relatively little effort has been put in increasing the formability of paper.

2. ObjectivesIn this sub-project the objective was to increase the formability of paper and board by combining simultaneously sev-eral measures: mechanical, enzymatic and chemical treatments. • Mechanical treatment: The objective

was find the effective methods to create axial compression zones (‘locally activated areas’) and other deformations that would increase the formability and extensibility of individual fibres and further converted to increased extensibility and formability of paper and board.

• Enzymatic treatment: This part of the work aimed at clarifying the role of hemicellulose in plasticisation. Target was a controlled removal of

hemicellulose from fibre cell wall using purified enzymes and alkaline extraction.

• Chemical treatment: The aim was to introduce plasticity and formability into the fibre by chemical treatment, for example by hydrophobisation of the cellulose.

3. Research approachWhen studying the effect of mechanical treatments on formability, the hypothe-sis was that fibre deformations would in-crease the elongation potential of the pa-per sheet and better formability of paper would be reached. The deformation po-tential of conifer fibres was mechanically increased by creping and in-plane com-pression of the pulp sheets. These treat-ments introduced kinks and compression zones to fibres increasing their deforma-tion potential. Several other methods cre-ating fibre deformations were also com-pared. Pulp were refined at high consis-tency (at 30%, wing refiner), at low con-sistency (Valley beating) and processed with E-compactor (at 40%). The effect of free shrinkage during drying of paper sheet was tested as well as the effect of rewetting.

In order to improve plasticity and mouldability of fibres measures to pre-vent or block hornification are needed. A working hypothesis was that the method of extraction (chemical vs. enzymatic) ef-fects on resulting cell wall structure and properties. Hemicellulose content of birch pulp was reduced by using chemical and enzymatic means and the effects of ex-tractions on intrinsic fibre properties were analysed. Controlled removal of hemicel-lulose from fibre cell wall was carried out using purified enzymes (xylanase) and al-kaline extraction. A never dried, bleached birch pulp was used as the raw materi-al. Removal of xylan was carried out both

113

with purified xylanases and by chemical means (NaOH extraction). After extrac-tion the pulps were analysed by chemical, physical and microscopical methods in or-der to differentiate the effects of extrac-tion method on cell wall structure. Labo-ratory sheets from pulps of variable xy-lan content were prepared and analysed for mouldability in pressing tests.

Ionic liquids (ILs) have been used to modify wood. The work performed so far has shown that wood chips (spruce and pine) can be fibrillated into individual fi-bres in a very mild conditions using ILs. In the treatment no delignification oc-curred and only pectin was dissolved. In-terestingly, after treatment fibres showed twisted, ´cork screw` appearance, which might give rise to interesting properties for paper (elasticity, mouldability). In this work ILs treated wood samples were dis-integrated to fibres, hand sheets were prepared and elongation properties were tested.

The plasticity of the fibre was also im-proved by using chemical treatments for example by increasing its hydrophobici-ty. External plasticizers were also tested. The chosen chemical modification meth-ods were esterification and etherifica-tion, which are traditional and cost effi-cient methods. Refined (SR70) cellulose was hydrophobised by preparing its es-ter derivatives (palmitoyl, stearyl, butu-ryl, naphtoyl, diphenyl acetyl, benzyl) in toluene. After the refining process, the fi-bres resided in water. The fibre materi-al was solvent exchanged to toluene be-fore the reactions. Hydrophobisation was also introduced to the fibre by preparing cellulose ether derivatives. The targeted compounds were hydroxypropylcellulose, benzyl glycidyl ether cellulose and cellu-lose silyl ether. External softeners were used for the external plasticization of cel-lulose. Three amine salts somewhat relat-ed to known cellulose solvents were used in testing.

Figure 1. Effect the type of treatment (high consistency vs. low consistency) (Wing refiner vs. Valley) on the elongation potential of wet and dry paper (first-thinning pine kraft pulp).

114

4. Results

4.1 Mechanical treatmentMethods increasing the fibre deformations (dislocations, curl, kinks, etc.) have been found to increase the formability of wet paper. Fibre treatment at high consistency yields higher elongation potential. The in-creased treatment time gave higher elon-gations. The elongation was also clearly higher with wet paper.

However, also low consistency beating that straightens fibres and reduces kinks can increase formability of paper, if free shrinkage is allowed to take place dur-ing drying of the sheet (Figure 1). This suggests that extensibility of freely dried sheets can be increased with methods that increase the shrinkage potential of paper.

By creping the pulp sheet (Figure 2), or by using a ’E-compactor’ device the

number of deformations in the re-slushed fibres were increased. For wet paper the stretch is known to increase considerably with fibre deformations, however, trans-ferring the deformation potential of fi-bres into essentially greater deformabil-ity of paper was challenging. It was no-ticed that although the fibre deformations increased the elongation potential of the sheet they do not necessary result in in-creased elongation of paper. In dry paper the interfibre bonding restricts the paper elongation. This suggests that the inter-fibre bonds should be dissolved/ plasti-cised during the forming/moulding oper-ation of paper.

Dry fibre cell wall openness and po-rosity can be controlled over a large range by mechanical pre-treatments and dry-ing. The amount of shrinkage during dry-ing increases the strain at break. Reduced restraint or higher shrinkage tendency of fibre material is beneficial.

Figure 2. Inducing fibre deformations by creping the pulp sheet. Creping takes place in one dimension. From left to right: 1) Reference: uncreped sheet, 2) Creped and dried sheet, 3) Creped, autoclave treated and dried sheet, 4) Creped and oven-dried sheet. The sheets were photographed after first drying.

115

Moistening of the paper before strain-ing generally improves the strain at break. However, with free dried samples the added moisture was found to reduce the elongation. This is an indication of the sensitive balance between the stress lev-el needed to straighten the fibres and the strength of (softened) fibre joints trans-mitting the stresses.

4.2 Enzymatic treatment and chemical composition

The hemicellulose removal affected fi-bre morphology and dimensions: average fibre length and shape factor was the low-est after alkaline treatment as compared with the xylanase treated sample and the untreated reference. Number of kinks was also the highest in the case of the alka-line treated pulp.

By the alkaline and xylanase treat-ments about 58%, 42% (Xyl 2) and 31% (Xyl 1) of the original xylan was removed, respectively. The xylanase treatment re-duced more efficiently acidic residues from the xylan, which can also be de-duced from the charge results.

The method used for xylan removal affected the swelling of the pulps; the al-kaline treatment reduced swelling (mea-sured as WRV and FSP) whereas gentle removal of xylan with xylanase retained mostly the swelling properties of pulp.

Molar mass distributions of the pulps were also determined by GPC. The

amount of xylan was decreasing in the or-der: reference > xylanase treated > alka-line treated. According to the results av-erage molar mass of xylan was decreased in the xylanase treated sample, but re-mained unchanged in the alkaline extract-ed sample. Some degradation of cellulose was also detected in the alkaline treat-ed sample.

The pulps were further characterised by high resolution SEM (Figure 3). Out-look of fibres was greatly affected by the alkaline treatment: fibres were rugged, heavily fibrillated, soft and “fluffy”, and cell wall structure was opened and loos-ened. In the case of the enzyme treat-ed samples fibre walls appeared compact, but fibril structure of the cell wall (P lay-er?) was clearly visible and no external fibrillation was observed. The untreated reference fibres were also compact and smooth and in some parts of the fibre surface fibril structure was poorly visible, perhaps due to a thin layer of re-precipi-tated xylan. One might assess that the al-kaline and enzymatic treatments result-ed to more porous fibre structure as com-pared with the initial pulp, but that needs further analysis.

Regardless of the method used for xylan removal extensibility of the hand sheets was reduced.

The correlation between strain at break and xylose content of the pulps are shown in Figure 4. There was no differ-

Sample Amount of xylan, %

Carboxylic acids, mmol/kg

WRV, g/g

FSP, g/g

FU1 Initial pulp 100 46 1.59 1.28 FU2 NaOH extracted 42 20 1.29 1.07 FU3 Xylanase treated 1 69 24.5 1.70 1.18 FU4 Xylanase treated 2 58 21 1.71 1.18

Table 1. Charge, WRV and FSP of the modified pulps.

116

ence whether the removal was carried out with NaOH or the xylanase. This effect was clearly related to xylan as a plasti-cizer between cellulose fibrils; when xy-lan amount was decreased hornification was increased due to interfibrillar hydro-gen bonding.

4.3 Wood defibrated with ionic liquids (ILs)Ionic liquids may show interesting results also when wood samples are defibrated. Three softwood wood pulp samples defi-brated at University of Helsinki were ex-amined. The samples are listed in Table 2.

Figure 3. High resolution images of the initial pulp and the alkaline and xylanase 2 treated pulp fibres.

Figure 4. The effect of xylan removal on the strain at break of hand sheets made of birch.

Sample IL 1. Spruce EMIM diethyl phosphate 2. Spruce EMIM methyl phosphonate 3. Pine EMIM dimethyl phosphate

Table 2. Wood samples treated with ILs.

117

Wood chips treated with ionic liquids rendered defibrated pulp, that showed unusual combination of strength and elongation (Figure 5).

These results confirmed the earli-er finding that the strain at break was much higher for the EMIM phosphonate treated fibres than for spruce TMP fibres. When dried freely or under restraint the strain at break was 5.5% and 4.9%, re-spectively. This positive effect on strain

can, at least partly, be due to twisted or cork screw like structure of fibres, which enables increased extension of the sheet under strain. Twisted outlook of fibres are shown in SEM images taken from the re-jects (Figure 6). Rather low fibre length, 1.38mm together with high swelling (WRV 6.5 g/g) of the EMIM phosphonate treated fibres in wet state further sup-port the idea of twisted and contracted fi-bre structure.

Figure 5. The effect of defibrating the wood with ILs on the tensile strength and elongation of hand sheets compared with TMP having different fractional composition.

Figure 6. SEM images on the ILs treated spruce (left) and pine fibres (right).

118

Spruce and pine wood chips treat-ed with ILs were disintegrated and af-ter coarse fractionation the slurries were used for preparation of hand sheets. Sep-arated spruce and pine fibres had a twist-ed appearance in the light microscope. Testing of the hand sheets gave very in-teresting results: The bulk was high, ten-sile strength was good and strain at break was doubled when compared with TMP. Increased strain of sheets might indicate

improved mouldability upon pressing test-ing.

4.4 Chemical treatmentThe hydroxypropylation treatment of fi-bres was found to be one promising chemical modification method that in-creases elongation but also hydrophobic-ity. Hydroxypropylation was carried out using a new, dry method, where the fi-bre is activated with a KOH/ethanol mix-

Scheme 1. Hydroxypropylation of cellulose.

Figure 7. Effect of hydroxypropylation on the elongation potential of re-wetted and dry paper.

119

ture after which the ethanol is removed, and then the fibre is hydroxypropylated with propylene oxide under high pres-sure. The reaction does not require a sol-vent. The reactions yielded hydroxypropyl ethers of cellulose with degree of substi-tution of 0.45, 0.23, 0.15 and 0.1 (deter-mined by XPS).

The first batch of hydroxypropylcel-lulose showed a degree of substitution of 0.45 and improved properties fibre net-works. The sheets of hydroxypropylat-ed fibres with DS 0.1 and 0.15 were also quite transparent and had increased elon-gation potential as well, whereas the less refined sheets with DS 0.23 were more paper like.

Hydroxypropylation was found to in-crease the elongation potential of the pa-per especially with freely dried sheets (Figure 7). Sheets made from hydroxy-propylated cellulose were transpar-ent (Figure 8), nearly poreless, with in-creased hydrophobicity and showed si-multaneously good elongation (8.5 %) when the paper was rewetted. Using free shrinkage drying the highest elongation of paper was 15.7 %. The hydroxypropylat-ed sample with the degree of substitution 0.45 (XPS) showed highest transparency and also good formability

4.5 ConclusionsThe research on the basic aspects of formability gave a very good picture of the fibre-level factors that contribute to the formability of fibre network. The re-sults also suggest that formability and transparency can be combined.

Several physical factors were found to affect the extensibility of paper: fibre de-formations, sheet shrinkage, and moist-ening, and chemical modification. Sev-eral chemical modifications were tested. For example the hydroxypropylated high-ly refined fibres showed promising results and produced sheets with high transpar-ency, low porosity, increased hydropho-bicity and increased elongation (8.5%) when rewetted. When the sheet was al-lowed to dry freely, the elongation was in-creased up to 15%.

The hemicellulose content and meth-od of hemicellulose removal was found to affect sheet properties and also the fibre structure and surface chemical properties of fibres. The method used for xylan re-moval affected the swelling of the pulps; the alkaline treatment reduced swelling (measured as WRV and FSP) whereas gentle removal of xylan with xylanase re-tained mostly the swelling properties of

Figure 8. Sheets made of hydroxypropylated fibres with originally different degree of refining and substitution.

120

pulp. Regardless of the method used for xylan removal extensibility of the hand sheets was reduced. This suggest that high hemicellulose content is beneficial for formability.

The results of chemical modifications show that the fibres can be modified with various hydrophobising groups, but high-er reagent to product –ratios and degrees of substitution demand the removal of the water from the fibre. Depending on the chemical treatment applied and the final degree of substitution, fibres with a con-siderable range of properties (from very hydrophobic, and poorly bonding, to fi-bres with considerable high WRV) can be produced.

These first results obtained indicate that by combining suitable mechanical, chemical and enzymatic methods the properties of fibres and fibre network can be modified, and the extensibility of fibre network can be increased by mechanical, enzymatic and chemical means. However, more research is needed in determining the best individual treatments and how the treatments should be combined.

5. Future business potentialSeveral converting and end-use appli-cations of paper-based products require suitable amount of deformability, rup-ture elongation, and energy absorp-tion capability. Especially when forming three dimensional packaging structures, the formability and extensibility is one of the basic requirements. Obtaining higher formability means better packages, more freedom in designing the packages and possibly new ways of forming the pack-age. If the formability of wood fibre based material can be increased, we are mov-ing toward mouldable webs, which could in certain applications replace oil-based, non-renewable polymer materials.

121

6. Key development needs and future plansThis research showed several, mainly ba-sic fibre-level methods that affect the ex-tensibility and formability of fibre net-works. They are however the first steps towards mouldable webs. More research is needed to identify the simplest ap-proaches and their combinations that are also applicable in industry. As these were fibre level methods, approaches and methods that can be applied to fibre networks need also to be evaluated and combined with fibre level methods.

7. Publications and reportsVuoti, S., Laatikainen, E., Sahari-nen, E., Heikkinen, H., Johansson L.-S., Retulainen, E. 2011. Chemical modifi-cation of cellulosic fibres for better con-vertability in packaging applications. Man-uscript submitted to Carbohydrate Poly-mers.

Pere, J., Retulainen, E., Laatikai-nen, E., 2010. Enzymatic and mechan-ical modification of fibres. Poster presen-tation at the Fubio T2 Workshop, May 10, 2010, Otaniemi, Espoo. (poster)

Laatikainen, E., Pere, J., Retulainen, E., Vuoti, S., Zeng, X., 2010. Improving the formability and extensibility of paper and board. Poster at the Annual Fubio Sem-inar, June 11, 2010, Hanasaari, Espoo. (poster)

122


Editors

Partners:

University of Oulu


Henrikki Liimatainen

Key researchers:

Henrikki Liimatainen, Juho Sirviö

Heikki Pajari

123

AbstractThe primary aim of this investigation was to examine a novel method for the production of dial-

dehyde cellulose (DAC) microfibers based on the concept of reactive milling. This concept com-

bines two processes, namely production of cellulose microfibers and the derivatisation of cel-

lulose by periodate oxidation. In addition, further derivation of DAC to a cationic water-solu-

ble product and its feasibility as flocculation agent was studied. The principle of operation and

unit operations of the novel process was proven and demonstrated in lab-scale. As an outcome

highly oxidized DAC fibres, microfibers and submicron particles were produced. The efficiency

of periodate oxidation reaction was successfully improved by using metal salts and high tem-

peratures as cellulose activators during oxidation. DAC was successfully further modified to wa-

ter treatment purposes. A cationic, water-soluble derivative (CDAC) was synthesized from DAC.

This derivative was shown to possess a high potential as a flocculation agent.

124

1. Background

Functionalized cellulose fibres and mi-cro- and nanoparticles are considered to be amongst the most potential can-didates for several future high-end ap-plications. Oxidation offers one effective route to obtain aldehyde and carboxyl-ic functionalities on cellulose. Many oxi-dants are capable of oxidizing cellulose, but the most of them possess a lack of selectivity. Amongst the few specific oxi-dation agents are 2,2,6,6-tetramethylpi-peridine-1-oxyl (TEMPO) and periodic ac-id and its salts which show high selectivity towards specific hydroxyl groups of cellu-lose. Furthermore, these oxidation treat-ments have also been found to promote the disintegration of cellulose fibres into its constituent particles.

Periodate is able to oxidize vicinal hy-droxyl groups of cellulose at positions 2 and 3 to aldehyde groups, simultaneous-ly breaking the corresponding carbon-carbon bond of the glucopyranose ring in order to obtain 2,3-dialdehyde cellu-lose (DAC). The aldehyde groups of DAC in turn have high reactivity towards fur-ther modification such as Schiff base re-action, cationisation and further oxida-tion to 2,3-dicarboxylic acid cellulose. Both DAC and its derivatives possess a great potential e.g. in medical materials and biodegradable composites.

The low reactivity of cellulose is one of the key-problems in periodate oxida-tion reaction. Cellulose has a high-or-dered intra- and inter-molecular hydrogen bond network which restricts the avail-ability of reactive free hydroxyl groups and also results in poor solubility of the native cellulose.

Both intensive mechanical and chem-ical treatments can be used to loosen the rigid hydrogen bonded network, reduce high crystallinity of cellulose and promote chemical reactions. In this study simulta-neous wet stirred media milling process

and periodate oxidation of cellulose fibres (reactive milling) was examined. In addi-tion, the effect of metal salts and use of elevated temperatures (>55°C) on oxida-tion reaction was clarified.

2. ObjectivesThe aims of the investigation were: 1. To study a novel process for the

production of dialdehyde cellulose microparticles. This is based on the concept of reactive milling, i.e. cellulose material is milled with a stirred media mill and further oxidized to DAC during milling. Sub-tasks of this subject were:

• To study how wet stirred media milling affects comminution and material properties of cellulose, and optimization operating

conditions of milling • To improve the efficiency of DAC reaction2. To study DAC applications and

further derivations of DAC. • Cationic polyelectrolyte based on DAC and its flocculation performance

3. Research approachThe concept of DAC production by reactive milling combines the idea of simultaneous cellulose comminution and its derivatisation by periodate oxidation. The main unit operations of the studied process are simultaneous milling and oxidation of cellulose raw material by periodate in the presence of chloride salts. In addition of these main operations, separation and washing operations are needed to separate the reagent solution from DAC material after reactive milling.

125

4. Results4.1 Preparation of DAC microfibers by reactive millingReactive milling, i.e. simultaneous peri-odate oxidation (Figure 1) and microni-sation of wood pulp fibres (hardwood kraft pulp), was conducted with a hori-

zontal agitated laboratory pearl mill (Ho-sokawa Alpine AHM 90). The experiments were performed in batch mode by milling 4 grams of oven-dry disintegrated pulp at a constant stirrer speed of 2000 rpm at room temperature (20°C) using yttrium oxide (Y2O3), stabilized zirconia (ZrO2) pearls and filling volumes of 0.70 dm3. The oxidation agent (NaIO4) was dosed directly into the milling chamber at the beginning of each experiment.

Milling significantly enhanced the cel-lulose reactivity towards the periodate ox-idation by reducing crystallinity and in-creasing the specific surface area of cel-lulose (Table 1). DAC microfibers with a high aspect ratio and aldehyde content of 0.26 mmol/g were obtained already after the first 15 minutes of milling. Advanced milling completely destroyed the cell wall of fibres producing smaller fragments in-cluding microfibers and even smaller par-ticles depending on conditions applied. Figure 2 shows a typical FESEM image of the DAC microfibers from reactive mill-ing process. Visually the samples can be characterised as white, gel-like high-ly viscous fluids. This new way to simul-taneously modify cellulose material me-chanically and chemically offers an effec-tive route to produce highly functional-ized cellulose microfibers within short re-action times.

Figure 1. Oxidation of cellulose by periodate.

Table 1. Aldehyde content of DAC produced by reactive milling and non-milled reference conditions.

Aldehyde content (mmol/g) Temperature

(¡ C) Reaction time (h) Non-

milled Milled

RT 0.25 0.09 0.26

0.5 0.11 0.43

1 0.26 0.65

3 0.43 0.88

55 0.25 0.36 0.51

0.5 0.61 0.71

1 0.95 1.16

2 1.31 1.86

3 1.68 1.90

65 0.25 0.43 0.80

0.5 0.71 0.78

1 1.26 1.05

2 1.86 1.80

3 2.20 1.87

126

4.2 Optimisation of DAC reaction: periodate oxidation of cellulose at elevated temperatures using metal salts as cellulose activatorsThe periodate oxidation is traditionally do-ne in water solution, but a great amount of periodate has to be used to achieve high aldehyde contents due to low reac-tivity of cellulose. This reduces the effec-tiveness of the oxidation, leads to large amounts of iodine contain waste and pro-longs the oxidation time.

Here the role of metal chlorides and elevated temperatures (>50 °C) in re-action efficacy was clarified. The results show that the oxidation efficiency can be significantly improved by means of heat-ing and by use of metal salts as cellulose activators (Table 2). Even though peri-odate decomposes at 55 °C and above with time, higher temperatures can be used if the oxidation time is sufficiently short. Even temperatures up to 85 °C can be used with short reaction times. Use of metal chlorides further promoted the periodate oxidation reaction. This effect was the most significant with CaCl2 (Ta-ble 2).

Figure 2. FESEM images of the original pulp (left), pulp after 10 min of milling (middle) and pulp after 45 min of milling (right).

Table 2. The promotion of periodate oxidation reaction by metal salts (reaction time 1 h and temperature 55 °C).

Metal salt Amount of LiCl (mmol)

LiCl/AGU ratio

Aldehyde content

(mmol/g) ZnCl2 3.5 1 1.058

7 2 1.208 14 5 1.164 21 7 1.208 28 9 1.136

CaCl2 3.5 1 1.186 7 2 1.322 14 5 1.25 21 7 1.386 28 9 1.258

MgCl2 3.5 1 1.022 7 2 1.072 14 5 1.158 21 7 1.272 28 9 1.272

NaCl 3.5 1 1.028 7 2 1.158 14 5 1.114 21 7 1.136 28 9 1.172

Metal salt Amount of LiCl (mmol)

LiCl/AGU ratio

Aldehyde content

(mmol/g) ZnCl2 3.5 1 1.058

7 2 1.208 14 5 1.164 21 7 1.208 28 9 1.136

CaCl2 3.5 1 1.186 7 2 1.322 14 5 1.25 21 7 1.386 28 9 1.258

MgCl2 3.5 1 1.022 7 2 1.072 14 5 1.158 21 7 1.272 28 9 1.272

NaCl 3.5 1 1.028 7 2 1.158 14 5 1.114 21 7 1.136 28 9 1.172

127

4.3 Cationic water-soluble flocculation aid based on DAC (CDAC)DAC has a high reactivity toward further chemical derivatization such as the Schiff base reaction. The formation of stable im-ine functionalities, such as hydrazones, is an attractive synthesis route to introduce cationic charge on DAC as it can be con-ducted in an environmentally friendly way in mild reaction conditions and without to use of any hazardous solvents. These materials have in turn high potential to be used as water treatment chemicals. The aim here was to synthesise a cat-ionic, water-soluble derivative from DAC to be used as a flocculation agent in wa-ter treatment applications. Synthesis was performed in water phase at pH 4.5 using carboxymethyltrimethylammonium chlo-ride hydrazide (Girard’s reagent T) as a reagent (Figure 3). Water soluble cation-ic derivatives were successfully obtained from DAC fibres (hardwood kraft pulp) having initial aldehyde content of 11.35 mmol/g (DS= 1.56). The derivative pos-sessed high degree of substitution (DS)

as shown in Table 3 which presents the amount of cationic groups in DAC as a function of reaction time.

The flocculation performance of cat-ionic DAC derivative was demonstrated in GCC filler suspension. In Figure 4 origi-nal GCC filler sample and samples floccu-lated with cationic DAC and commercial flocculation aid (CPAM with a molecular weight of 1.5*106 g/mol and charge den-sity of 4 meq/g) are presented. It can be seen that water-soluble CDACs resulted in clear flocculation. More detailed floc-culation study conducted with an analyti-

Figure 3. The synthesis of CDAC from DAC.

Table 3. Cationicity of CDAC produced from highly oxidized cellulose.

Reaction time (h)

Cationicity (mmol/g)

1 2.85 2 2.95a

3 3.29a

12 3.70a

24 4.07a

48 4.03a

72 4.27a

96 3.89a

aSoluble at room temperature

128

cal centrifuge showed that the maximum flocculation efficiency was about 85% of the value obtained with the reference polymer (CPAM). Also the optimum dos-age was found to be in similar range that with the reference (Figure 5.).


Reactive milling proved to be an efficient concept to produce dialdehyde cellulose fibres and microfibers. In this study DAC was further modified to cationic, water-soluble polyelectrolyte, which possessed to be a promising material for flocculation purposes. It is also possible to produce other water-soluble cellulose derivatives using environmentally benign mild water solutions. Key factors determining the vi-ability of DAC materials in the commer-

cial solutions are the energy efficiency of milling process and the regeneration pro-cess for periodate solution after oxidation. As periodate is expensive and toxic sub-stance it should be recycled and regener-ated after oxidation. A detailed study to develop this aspect of this process is cur-rently on-going in FuBio.


The critical issues to be addressed in fu-ture are the function of the process as a unity taking into account the techni-cal and economical questions as well as characteristics of products derived from the DAC. In particular, the feasibility of DAC microfibers in potential applications should be clarified.

Figure 4. Image of original GCC suspension and suspensions flocculated with CDACs and a commercial reference flocculant (CPAM). Flocculant dosage: 0.5 mg/g (CDAC II and CPAM) and 1.0 mg/g (CDAC 1).

Unflocculated CDAC I CDAC II Reference

Figure 5. Flocculation efficiency of CDAC and reference flocculation aid in terms of residual transmission.

129

7. Publications and reportsHyväkkö, U. 2010. Tehostettu selluloo-san perjodaattihapetus, M.Sc thesis, Uni-versity of Oulu.

Liimatainen, H., Sirviö, J., Sundman, O., Visanko, M., Hormi, O., Niinimäki, J. (2011) Flocculation performance of a cat-ionic biopolymer derived from a cellulos-ic source in mild aqueous solution. Biore-source Technology 102: 9626-9636.

Liimatainen, H., Sirviö, J., Haapa-la, A., Hormi, O., Niinimäki, J. (2011) Characterization of highly accessible mi-crofibers generated by wet stirred media milling, Carbohydrate Polymers 83(4): 2005-2010 .

Liimatainen, H., Haapala, A., Hormi, O., Niinimäki, J. Production of cellulose mi-croparticles by wet stirred media mill-ing. 64th Appita Annual Conference, Mel-bourne, April 18-22, 2010, 57-62.

Sirviö, J., Liimatainen, H., Niinimäki, J., Hormi, O. (2011) Dialdehyde cellulose microfibers generated from wood pulp by milling-induced periodate oxidation, Car-bohydrate polymers 86(1): 260-265.

Sirviö, J., Honka, A., Liimatainen, H., Niinimäki, J., Hormi, O. (2011) Synthe-sis of highly cationic water-soluble cellu-lose derivative and its potential as novel biopolymeric flocculation agent, Carbohy-drate polymers 86(1): 266-270.

Sirviö, J., Hyväkkö, U., Liimatainen, H., Niinimäki, J., Hormi, O. (2011) Peri-odate oxidation of cellulose at elevat-ed temperatures using metal salts as cellulose activators, Carbohydrate Poly-mers 83(3): 1293-1297.

Sirviö, J. 2009. Selluloosan katalyytti-nen hapetus natriumperjodaatilla, M.Sc. thesis, University of Oulu.

130


Editor

Partners:



Lappeenranta University of Technology

Harri Setälä

Key researchers:

Jorma Ihalainen, Kari Kammiovirta, Hanna Kyllönen, Katri Mäkinen, Pentti Pirkonen, Pekka Ruuskanen, Juha Sarlin, Harri Setälä

Heikki Tenhu, Kati Salo

Mika Mänttäri, Marianne Nyström, Mehrdad Hesampour, Tiina Huuhilo

131

AbstractStimuli-responsive materials are able to change their physic-chemical properties in response

to changes in their environment and present one of the most interesting and unexplored com-

mercial applications. Stimuli-responsive materials have been designed to respond in a control-

lable and often reversible way to changes in their environment such as pH, temperature, ionic

strength, light, electric and magnetic fields, solvent composition, or other chemical agents. This

article will give a short introduction to these materials and some novel examples of interesting

applications and demonstrations especially concerning so-called thermo-responsive materials

that can be used in several different applications, such as in drug release systems, adsorption

materials, and filtration systems. Most of these materials and examples presented in this arti-

cle are based on so-called activated polysaccharide derivatives prepared to be used as macro-

molecular cross-linkers and coupling components in preparation of stimuli-responsive materials.

132

1. Background

Stimuli-responsive materials are a large group of different kinds of (polymer-ic) materials that are generally able to change their properties such as viscosi-ty, volume, colour, conductivity, or shape in a smart and usually in a reversible way as a function of an external stimuli such as temperature (thermo-sensitive), pH, ionic strength, solvent content and type, light, electric current, or magnetic field. The stimuli-responsive polymers reviewed in this article are able to change their vol-ume by shrinking or swelling, and hydro-phobicity/hydrophilicity balance as a func-tion of temperature and/or pH and/or ion-ic strength. The so-called Lower Critical Solution Temperature (LCST) (or pH/ion-ic strength), where these thermo-respon-sive transitions happen, may be adjust-ed, for instance, by using different kinds of monomers, for example, N-isopro-pylacrylamide (NIPAM) yielding polyNI-PAM with LCST 32–34°C, or N-vinylcap-rolactam (VCL), or by copolymerizing dif-ferent kinds of monomers together, such as NIPAM with acrylic acid (AA) yielding PNIPAM-co-PAA. Crosslinking agents such as N,N’-methylene bisacrylamide (MBA) are often used to adjust their structural

and mechanical properties. Cellulose and hemicelluloses has already been used as starting compounds for stimuli-respon-sive materials such as cellulose-g-PNI-PAM with LCST at about 35°C [Hao et al. 2009], or cellulose based membranes grafted with copolymer of PAA-b-PNIPAM [Pan et al. 2010]. They could be expect-ed to have a potential application in drug release system, adsorption materials and functional membrane materials. This ex-tremely interesting and growing research topic of stimuli-responsive materials is al-so very well reported in some review arti-cles, for example, published by Wandera et al.2010 or Liu et al. 2010.

2. ObjectivesThe aim of research has been to develop and perform demonstrations in some se-lected application areas such as 1) in hy-drogels and (super)adsorbents, 2) con-struction materials for a moisture control in buildings, 3) ion-exchange materials, and 4) filtration materials, and 5) also to develop some novel coating methods, for example, for cellulosic or synthetic fibrous materials.

Figure 1. Reversible behaviour of thermo-responsive polymers in filtration materials.

133

3. Research approachThe most important starting point and idea was to develop novel cellulose or hemicellulose based derivatives which could be used as reactive and crosslink-ing or coupling additives in many different kind of construction and coating process-es and at the same time to use them also as reactive starting materials when stim-uli-responsive or other polymers were at-tached to those products using grafting techniques. Cellulose in fibrous or crys-talline forms, cellulose derivatives such as hydroxypropylcellulose, and hemicel-luloses such as birch xylan were used as starting materials and derivatised to allyl-ic substituents (double bonds, A) as re-active substituents for polymerization and grafting chemistry, and/or to epoxy func-tionalities (E) for “click” type of coupling chemistry, see Figure 2. Also other type of so-called non-reactive substituents

with alkyl or aromatic side-chains were attached for adjusting, for example, hy-drophobicity-hydrophilicity balance, ther-moplasticity, or solubility of polysaccha-ride derivatives. Allylic double bonds are able to react with vinyl, acryl, or acryl-amide type of monomers in grafting reac-tions, and at the same time they are able to crosslink among themselves forming a 3D network. Epoxy functionalities are able to react with suitable nucleophilic re-agents or substituents such as hydroxyl groups of polysaccharides or polyvinylal-cohol, phenolic hydroxyl groups, or amino functionalities also utilized commonly in so-called epoxy resins. These functional-ities and one example of a grafting reac-tion are presented in Figure 2. So-called thermoresponsive and/or pH-responsive polymers were studied in more detail and grafted to reactive polysaccharide deriva-tives. The utilization of allylated polysac-charide derivatives was studied and dem-

Figure 2. A schematic presentation of derivatisation (i) of cellulose to derivatives with allyl (A) and/or epoxy (E) groups which can be utilized (ii) for crosslinking (CL) (A) and/or for grafting with stimuli-responsive polymer such as PNIPAM forming a 3D network/structure onto a matrix material or in a hydrogel. Epoxy groups can react (iii) with suitable nucleophiles e.g. amino compounds (H2N-R).

134

hydrogels is presented in Figure 3. Xylan hydrogel was prepared either by cross-linking and grafting with PNIPAM or with-out any grafting but only using crosslink-ing reaction of allylic double bonds among themselves. In this case, the xylan based hydrogel without PNIPAM seemed to ad-sorb much more water than the grafted one. The volume of xylan-graft-PNIPAM hydrogels was increased approx. 6 times in a swollen state, which means 570% adsorption of water, whereas the cross-linked xylan was able to adsorb 1000 % of water (10 times volume change). This result is very similar with the results pub-lished by Lindblad et al. 2005. The main difference between these hydrogels is that the xylan-graft-PNIPAM has a tempera-ture-dependent and reversible absorption behaviour at 32–34°C, and therefore it can be used as a “smart” thermo-sensi-tive absorption material. This “smart” be-haviour was also demonstrated to be use-ful, for example, in thermally reversible flocculant / dispersant systems by Nichi-for et al. 2003.

4.2 pH sensitive hydrogels and ion-exchange materialsThe pH-hydrogel is a hydrophilic porous mixture, constructed by three-dimension-al cross-linked polymeric network with in-terstitial water or biological fluid. The hy-drogels are increasingly used for biomed-ical and biotechnology applications since they are relatively inexpensive and can perform excellently in biocompatibility with tailorable characteristics. The swell-ing modulation of the hydrogels by re-sponding to environmental pH stimuli is able to dynamically control the conver-sion of electrochemical energy into me-chanical one and then to provide effec-tive diffusivity or permeability of the sol-utes. The capabilities of the smart hydro-gels in sensing and actuating applications make the pH-stimulus-responsive hydro-gels gain popularity in vast number of BioMEMS applications, including the con-

onstrated in coating methods onto sever-al fibrous or membranes materials, and in selected applications.

4. Results

4.1 Hydrogels and adsorbentsHydrogels are chemically or physical-ly cross-linked networks that are water-insoluble but capable of absorbing large amounts of water. They can be made of synthetic or natural starting materials but commercial hydrogels have traditionally been prepared mainly from toxic acrylates and acrylamides. Hydrogels based on nat-urally occurring products are of interest not only for their renewable character and nontoxic nature but because they may of-fer biocompatibility and biodegradabili-ty. Hydrogels possess a degree of flexi-bility due to their significant water con-tent and they are potential material can-didates, e.g., in tissue engineering, con-trolled drug release, agriculture and hy-giene products. Superabsorbent hydro-gels are three-dimensional cross-linked hydrophilic, linear or branched polymers with the ability to absorb large quanti-ties of water, saline or physiological so-lutions compared with general absorbing materials. Cellulose and also hemicellu-loses having abundant hydroxyl groups can be used to prepare hydrogels easi-ly with fascinating structures and prop-erties. There is a need to study polysac-charide-based hydrogels in both funda-mental research and industrial applica-tion. [Chang et al. 2010]

In this study both cellulose fibres and hemicelluloses such as birch xylan were used as starting materials for hydro-gels and adsorbents and they were also grafted with stimuli-responsive and oth-er polymers. Both starting materials were first activated and derivatised with allylic groups for crosslinking and grafting pur-poses. One demonstration of xylan based

135

ing pH and temperature as presented in Figure 4. Phenyl red was used as a pH in-dicator (pH 5.2-6.8) being red in alka-line and yellow in acidic conditions. Water phase was first adjusted slightly to alka-line (pH > 7) and a grafted cellulose sheet was put into the solution at 60°C. A co-lour change was very slow (over 15 min) at 60°C but very fast at 25°C. This means that this kind of a stimuli-responsive ma-terial in hydrophobised and collapsed stage at 60 ºC decreases an ion-exchange phenomenon remarkable because of pre-vented diffusion of ions through a hy-drophibised cellulose sheet. The change of colour from red to yellow one occupied

trolled drug release, microscale actuators or sensors, microfluidic flow control and filtration as well as separation processes. [Li et al. 2009] The responsive materials can be also so-called dual-stimuli-sensi-tive carrying at the same time, for exam-ple, pH- and thermo-responsive proper-ties. [Ma et al. 2010]

Allylated cellulose fibres were grafted with N-isopropylacrylamide (NIPAM) and acrylic acid (AA) yielding cellulose-graft-PNIPAM-co-PAA yielding a pH and ther-moresponsive material. The thermo-re-sponsive/pH-dependent behaviour of PNI-PAM-co-PAA grafted filter paper sheets were characterized tentatively by chang-

Figure 3. Xylan based hydrogels grafted with PNIPAM and without it, in their swollen (left) and dry (right) states.

Figure 4. Thermo-responsive pH-dependent ion-exchange behaviour of P(NIPAM-co-PAA) grafted cellulose sheets.

136

approximately in 90 minutes. When the temperature was below the so-called low-er critical solution temperature (LCST) of PNIPAM (32–34°C), in this case at 25°C, the ion-exchange was much faster and the colour of solution changed immedi-ately to bright yellow when the PNIPAM-co-PAA grafted cellulose sheet was set in-to that solution.

4.3 Membranes and filtration materialsPerhaps the most interesting application area may be the use of stimuli-respon-sive polymers in filtration and other sepa-ration processes and especially in the pu-rification processes of industrial and oth-er waste waters, and also in desalina-tion of sea and brackish waters. The re-view published by Wandera et al. 2010 gives a very good overview of that very important research and application area. The temperature-dependent behaviour is illustrated in filtration type of applica-tions in Figures 1, 5–7. The unpublished results presented in Figure 5 have been performed in the Smart Filter project fi-nanced by Tekes during the years 2003-

2007. However, some main results of the Smart Filter project have also been pub-lished [Ihalainen et al. 2006, Hesampour et al. 2008, Pirkonen et al. 2010] as well as reported in project reports. [Ihalainen et al. 2005; Sarlin et al. 2007]

The pore size of filter material mod-ified with a thermo-sensitive polymer is increased above the LCST (lowest critical solution temperature) of a stimuli-respon-sive polymer where it is in a collapsed stage, and the flux through a filter lay-er is highest (see Figure 1). A stimuli-re-sponsive polymer is there rather hydro-phobic and it has lost a main part of wa-ter. A volume change can be 10 or even 100 fold. The stimuli-responsive polymer is again in a swollen state below the LC-ST, and the flux is highly decreased or even stopped if the amount of a stimu-li-responsive polymer is high enough on a filter material, see Figures 1 and 5. At this state, polymer is again hydrophilic and it may contain water up to 95–98 wt-%. The shift or temperature range of vol-ume transition may be very narrow such as 2-4°C of PNIPAM or rather broad (10–20°C) depending mainly on a composi-

Figure 5. An industrial filter paper T 750 modified with PNIPAM shows clearly the temperature-dependent behaviour compared to the unmodified filter paper sample (top line).

T 750 FILTER PAPER

14000 15000 16000 17000 18000 1900020000 21000 22000

20 25 30 35 40 45 50 55 60 65 70 Temperature, o C

Flux

, kg/

(m 2 hb

ar)

T 750-reference T 1. down T 2. up

T 2. down

T 3. up

T 3. down

137

tion of a stimuli-responsive (co)polymer and/or matrix material (membrane, filter fabric etc.). For example, when PNIPAM - having normally a very narrow temper-ature range, LCST 32-34 ºC - is grafted on a filter paper, the temperature range was increased from 2 to 15 ºC units and the volume change took place between 20 and 35 ºC. This is illustrated in Fig-ure 5. The industrial filter paper is pre-treated first using water and/or water –organic solvent soluble allylated hydroxy-propylcellulose derivatives as primers and macromolecular cross-linkers which can then be grafted with PNIPAM [Setälä et al. 2011].

The same technique has also been used for a treatment of other kind of fil-ter fabric materials based on polypro-pene, polyacrylated, polyvinylalcohol, or polyethyleneterephtalate (PET). For ex-ample, an industrial PET type of filter fab-ric (S2209 Tamfelt) was first primerised using an allylated hydroxypropylcellulose (A-HPC) and then grafted with PNIPAM. The amount of A-HPC-g-PNIPAM was 4.1 wt-% on the filter fabric. The stimuli-re-sponsive and other properties such as stability of the polymer layer at differ-ent pHs and temperatures were studied and published [Pirkonen et al. 2010]. The original flow of untreated filter fabric was between 4000–4600 kg/(m2 h bar) de-pending on a temperature (see Figure 7). After modification with PNIPAM the flow was decreased down to 750 kg/(m2 h bar) at 20°C and to 1200 kg/(m2 h bar) at 40°C, see Figure 6. The flux of filter fab-ric with stimuli-responsive polymer lay-er was not changed when the tempera-ture cycle (from 20 to 70°C) was repeat-ed even more than 50 times which in-dicates that the stimuli-responsive poly-mer layer is very well attached onto the filter fabric. The polymer layer was also very stable against acidic (pH 2) and al-kaline (pH 12) conditions. The regener-ation or defouling of PNIPAM filter fab-ric after fouling or clogging states with a

real process water such as a white water of a paper mill was also tested by using a backwashing procedure at 20 and 40°C without any washing or defouling chemi-cals. The defouling process was especially effective at 40°C for the filter fabric mod-ified with A-HPC-g-PNIPAM and only few washing steps were needed for returning the original flux. This was not possible to perform only using backwashing process at 20°C and the flux was not returned to-tally to the initial level. This is present-ed in Figure 6. When filter fabric without any PNIPAM treatment was tested, it was not at all possible to normalize the origi-nal flux, see Figure 7.

The advantages of stimuli-respon-sive filtration materials are, for exam-ple, according by Wandera et al. [2010] that “stimuli-responsive membranes have strong potential for future applications in tissue engineering, bioseparations, an-tifouling surfaces, and drug delivery among others. Reversible changes to sur-face composition, surface energy, adhe-sion and wettability of stimuli-responsive membranes will provide ways of fabricat-ing membranes with new functions, such as self-cleaning and self-refreshing abili-ties. Switchable membrane surface prop-erties will improve the efficiency of many technological processes.” The self-clean-ing or fast defouling process in filtration applications has also been demonstrated to be very efficient with stimuli-respon-sive filter fabrics by Pirkonen et al. 2010, see Figures 6 and 7.


The main potential of stimuli-responsive materials seems to be mainly in applica-tions such as:• Separation of impurities and

components from municipal or industrial waste- or side-streams

138

Figure 6. Filtration tests using real white water and filter fabric coated with a PNIPAM polymer layer. Start = original values, 5thCl = flux after 5 clogging cycles, 1th … 7th are fluxes after (back)washing with 20ºC water (yellow column) or 40°C water (red column).

Figure 7. Filtration tests using real white water and filter fabric without PNIPAM. Start = original values, 5thCl = flux after 5 clogging cycles, 1th … 7th are fluxes after (back)washing with 20°C water (yellow column) or 40°C water (red column).

139

using stimuli-responsive materials as adsorbents [Akiyama et al. 2008] or flocculants [Nichifor et al. 2003]

• Filtration of municipal or industrial waste- or side-streams using stimuli- responsive filter materials [Wandera et al. 2010; Pirkonen et al. 2010]

• Stimuli-responsive hydrogels and materials in medical or biotechnological applications such as drug realeasing systems, self-healing systems, tissue engineering etc. [Urban 2009; Brun et al. 2010, Motornov et al. 2010]

• Stimuli-responsive coatings in textiles [Kulharni et al. 2010]

The business potential concerning the fil-tration and separation techniques was evaluated and reported, for example, in the reports of Sirius Consulting [2006] or TheBerkelyMBA [2005].


Although “smart” polymers have been known for more than four decades, on-ly recently design and synthesis of stim-uli-responsive systems with controlla-ble properties have been tackled and will continue to be of great challenges to sci-entists and engineers. [Liu et al. 2010] Stimuli-responsive membranes have strong potential for future applications in tissue engineering, bioseparations, antifouling surfaces, and drug delivery among others. Reversible changes to sur-face composition, surface energy, adhe-sion and wettability of stimuli-responsive membranes will provide ways of fabricat-ing membranes with new functions, such as self-cleaning and self-refreshing abili-ties. Switchable membrane surface prop-erties will improve the efficiency of ma-ny technological processes. [Wandera et al. 2010]

The research in the field of stimuli-re-sponsive materials has increased greatly in Finland from the year 2003, when the first VTT-related project “Smart filtration” begun. Currently the R&D is mainly fo-cused on applications in separation and filtration combined with the development of novel membrane and filtration mate-rials, especially based on cellulosic raw materials. Additionally, some projects are related to the development of intelligent drug releasing systems based on stimuli-responsive polymers [Ropponen et al.], or to a smart moisture control in some oth-er applications. [Mäkinen et al. 2005] At the moment, the research in the field of stimuli-responsive materials is going on in the several projects or projects, such as FuBio and Naseva. It will also be a fo-cus of the EU-project Nanoselect, which will start in the beginning 2012. The gen-eral aim of all these projects is to devel-op new hydrogels, adsorbents, filtration or membrane materials based on cellu-lose combined often with stimuli-respon-sive polymers.


Akiyama, Y., Shinohara, Y. Hasega-wa, Y., Kikuchi, A., Okano, T., Prepa-ration of novel acrylamide-based thermo-responsive polymer analogues and their application as thermoresponsive chro-matographic matrices. J. Polym. Sci: Part A: Polym. Chem. 46 (2008) 5471–5482.

Brun-Graeppi, A.K.A.S., Richard, C., Bessodes, M., Scherman, D., Merten, O.-W., Thermoresponsive surfaces for cell culture and enzyme-free cell detach-ment. Prog. Polym. Sci. 35 (2010) 1311–1324.

140

Chang, C., Zhang, L., Cellulose-based hydrogels: Present status and application prospects. Carbohydr. Polym. 84 (2011) 40–53.

Denbo, D., Kannan, H., Tran, L., Yan, L. Smart Filter. Observations and Recom-mendations for VTT Processes Team. The-BerkeleyMBA. 2005.

Hao, Y., Peng., J., Li, J., Zhai, M., Wei, G., An ionic liquid as reaction media for radi-ation-induced grafting of thermosensitive poly (N-isopropylacrylamide) onto micro-crystalline cellulose. Carbohydr. Polym. 77 (2009) 779–784.

Hesampour, M., Huuhilo, T., Mäkinen, K., Mänttäri, M., Nyström, M., Grafting of temperature sensitive PNIPAAm on hy-drophilised polysulfone UF membranes. J. Membrane Sci. 310 (2008) 85–92.

Ihalainen, J., Mäkinen, K., Pirkonen, P., Ruuskanen, P., Sekki, H., Setälä, H., Tenhu, H., Salo, K., Suodatin, mene-telmä sen valmistamiseksi ja sen käyttö / Filter, förfarande för framställing därav och dess använding. Patent application: FI 117272 B, 2006.

Ihalainen, J., Mäkinen, K., Pirkonen, P., Ruuskanen, P., Salo, K., Setälä, H., Tenhu, H. Älykkäät suodattimet, TEKES report 2005.

Kulharni, A., Tourrette, A., War-moeskerken, M.C.G., Jocic, D., Micro-gel-based surface modifying system for stimuli-responsive functional finishing of cotton. Carbohydr. Polym. 82 (2010) 1306–1314.

Li, H., Yew, Y.K., Simulation of soft smart hydrogels responsive to pH stimu-lus: Ionic strength effect and case stud-ies. Materials Sci. Eng. C 29 (2009) 2261–2269.Lindbland, M.S., Albertsson, A.-C., Ranucci, E., Laus, M., Giani, E., Biode-gradable polymers from renewable sourc-es: Rheological characterization of hemi-cellulose-based hydrogels. Biomacromol. 6 (2005) 684-690.

Liu, F., Urban, M.W., Recent advances and challenges in designing stimuli-re-sponsive polymers. Prog. Polym. Sci. 35 (2010) 3–23.

Liu, B., Hu, J., The application of tem-perature-sensitive hydrogels to textiles: a review of Chinese and Japanese inves-tigations. Fibres Text East Eur 13 (2005) 45–9.

Ma, L., Liu, R., Tan, J., Wang, D., Jin, X., Kang, H., Wu, M., Huang, Y., Self-as-sembly and dual-Stimuli sensitivities of hydroxypropylcellulose-graft-poly(N,N-di-methyl aminoethyl methacrylate) copo-lymers in aqueous solution. Langmuir 26 (2010), 8697–8703.

Motornov, M., Roiter, Y., Tokarev, I., Minko, S., Stimuli-responsive nanoparti-cles, nanogels and capsules for integrated multifunctional intelligent systems. Prog. Polym. Sci. 35 (2010) 174–211.

Mäkinen, K., Kosteuden hallinta huo-koisuuden keinoin – älykkäiden materi-aalien mahdollisuudet kosteuden hallin-nassa. TEKES-raportti, 2005.

141

Nichifor, M., Zhu, X.X., The use of N-alkylacrylamide-styrene copolymers as thermally reversible dispersants / floccu-lants for emulsions and suspensions Col-loid Polym. Sci. 281 (2003) 1034-1039.

Pan, K., Zhang, X., Ren, R., Cao, B., Double stimuli-responsive membranes grafted with block copolymer by ATRP method. J. Membr. Sci. 356 (2010) 133–137.

Pirkonen, P., Setälä, H., Kyllönen, H., Sarlin, J., Salo, K., Tenhu, H., Ruus-kanen, P., Thermal Stimuli Controlled Functional Filter Cloth For Liquid Filtra-tion, Filtration 10 (2010) 144-152. Ropponen, J., Nikkola, J., Mahiout, A., Novel Responsive Surfaces based on Ac-tive Hybrid Coatings Utilizing Encapsu-lation Technologies (RESCOAT). TEKES Functional Materials program. 2011-2014.

Sarlin, J., Pirkonen, P., Kyllönen, K., Setälä, H., Immonen, K., Vuori, T., Kati Salo, Niskanen, J., Tenhu, T., Hesampour, M., Mänttäri, M., Nyström, M., Hovila, P., Smart Filter II. Thermo responsive functional filters. TEKES report, 2007.

Setälä, H., Harlin, A., Talja, H., Mene-telmä spesifisten tuotteiden valmista-miseksi polysakkaridimolekyylista. Pat-ent application BP203541/JJA/SSO, 2011.

Sirius Consulting. Älykäs suodatin, metsäteollisuuden sovellutukset. Inno-TULI selvitys. 2006.

Urban, M.W., Trends in polymer science. Stratification, stimuli-responsiveness, self-healing, and signaling in polymer net-works. Prog. Polym. Sci. 34 (2009) 679–687.

Wandera, D., Wickramasinghe, S.R., Husson, S.M., Review. Stimuli-respon-sive membranes. J. Membr. Sci. 357 (2010) 6–35.

Zhang, J., Xu, X.-D., Wu, D.-Q., Zhang, X.-Z., Zhuo, R.-X., Synthesis of thermo-sensitive P(NIPAAm-co-HEMA)/cellulose hydrogels via ‘‘click” chemistry. Carbohy-dr. Polym. 77 (2009) 583–589.

Zhao, S.P., Cao, M.J., Li, L.Y., Xu, W.L., Synthesis and properties of biodegrad-able thermo- and pH-sensitive poly[(N-isopropylacrylamide)-co-(methacrylic ac-id)] hydrogels. Polym. Degrad. Stabil. 95 (2010) 719-724.

142


Editors

Partners:

Lappeenranta University of Technology

Esa Saukkonen and Katja Lyytikäinen

Key researchers:

Esa Saukkonen, Katja Lyytikäinen, Isko Kajanto

143

AbstractTwo pulps with low hemicellulose (xylan) content were prepared and used to make paper. One

of the pulps was prepared by treating bleached birch kraft pulp with sodium hydroxide, while

the other was a softwood pre-hydrolysis pulp. The effect of xylan removal on pulp/fibre and pa-

permaking properties were thoroughly determined in order to clarify whether hemi-lean pulps

could replace kraft pulp as a raw material in papermaking. The effects of decreased hemicel-

lulose content on the wet-end behaviour were mainly evaluated in terms of filler retention and

performance of cationic additives. Results show, although partly surprisingly, that hemi-lean

pulp would be an interesting raw material for certain paper grades. The removal of xylan does

not only change the fibre properties, but might also have remarkable effects on the interactions

in the wet-end. This study gave new insights on the possibilities to utilise hemi-lean pulps in pa-

permaking.

144

1. BackgroundThe finite nature of fossil fuels and their contribution to carbon dioxide emissions is currently impelling the development of green technologies and harvesting for al-ternative resources for both fuels and the production of chemicals. The more com-plete use of naturally abundant lignocel-lulosic, non-food biomass as a feedstock offers a logical alternative and a possibil-ity for the development of a lignocellulos-ic-based economy. Consequently, a grow-ing interest in the utilization of lignocellu-losic biomass has led to the development of various biorefinery concepts.

The integrated forest biorefinery (IF-BR) concept presents a promising oppor-tunity for an enterprise transformation of the pulp and paper industry by offering new sources of revenue and significantly improved industry profitability. One pro-posed next generation technology for an IFBR is partial extraction of hemicellu-loses. In addition to the utilization of the hemicelluloses stream for biofuels/chem-icals production, this process would al-low the production of special-grade pulps in which the absence of hemicelluloses are preferred. Despite of the current re-search focus on biorefineries and past ex-perience on dissolving pulp, there is still a lack in understanding concerning the use of hemi-lean fibres in papers. Although it is generally accepted that hemi-lean fi-bres have negative effect on paper prop-erties, the fundamental mechanisms have not been extensively determined.

This study has focused on the paper-making behaviour and paper properties of hemi-lean fibres in order to increase the knowledge on the possibilities to uti-lize these kinds of fibres in the papermak-ing purposes. This work addresses impor-tant aspects related to the integration of a hemicelluloses extraction step, the effect on fibre properties, and ability to scale up the concept. Considerable value poten-tial of utilizing the hemicellulosic material

stream should be realized to maintain an economically feasible process.

2. ObjectivesThe main objectives of the project were to find out:• Effect of alkaline hemicellulose

extraction conditions on xylan yield from bleached birch kraft pulp

• Effect of alkaline hemicellulose extraction of bleached birch kraft pulp on pulp and paper properties

• Effect of pre-hydrolysis of pine wood chips prior to the kraft pulping on pulp and paper properties

• Effect of alkaline hemicellulose extraction of bleached birch kraft pulp on wet end chemistry

• How the papermaking process behaves in the presence of hemi-lean fibres

• The possibilities to utilize hemi-lean fibres in the papermaking purposes

3. Research approachA full understanding on the chemical and physical properties of hemi-lean fibres is essential in order to make it possible to use these kinds of fibres in paper and/or paperboard products. Therefore, the goal of our research work was to deter-mine which kind of effects can be expect-ed when paper is made of pulps with sub-stantially reduced hemicellulose content.

The research was carried out in two stages. The papermaking properties of the hemi-lean pulps were first evaluat-ed. Alkaline extracted birch kraft pulp and pre-hydrolysed pine kraft pulp were se-lected to demonstrate the distinct effects of hemicellulose extraction method on the chemical and physical properties of the fi-bres. Secondly, the performance of he-mi-lean birch fibres in papermaking pro-

145

cess was investigated using a Moving Belt Former (MBF), which is a dynamic sheet-forming device that enables the shear and drainage conditions of paper and board machines to be simulated in laboratory environment.

In order to make the hemicellose ex-traction process economically feasible, the extraction conditions should be op-timized. Therefore, efforts were made to study the extraction conditions of alka-line extraction of birch kraft pulp more in detail.

4. Results

4.1 Fibre and paper properties of hemi-lean pulpsThe extraction yields determined for both pulps indicated that about half of the hemicelluloses were removed by the treatments. It is shown in Table I that the alkaline extraction reduced the amount of xylan from 23.6% to approximately 13%, whereas the amount of glucomannan re-mained at the initial level. The chemi-cal compositions of alkali-extracted birch pulp and the unextracted birch pulp are shown in Table 1. The chemical composi-tion of the softwood pulps was not anal-ysed in detail.

The main findings related to the paper properties of the alkali-extracted birch pulp are reported in Table 2. For compar-ison, some properties of the pre-hydroly-sis softwood pulp are also given.

In case of the post-bleaching alkaline extracted birch pulp, some interesting findings related to fibre and paper prop-erties were observed.

Firstly, the strain to failure increased although hemicelluloses were removed. This increase is probably due to the high kink content of the fibres after alkaline extraction. Due to the increased strain to failure alkaline extracted birch kraft pulp fibres could possibly be utilized in packag-ing papers, for example in the top ply of paperboard, where high strain improves the convertability. In comparison, with the pre-hydrolysed softwood pulp, strain to failure decreases, as was expected. Secondly, water retention value (WRV) of alkaline extracted birch kraft pulp was higher than that of reference pulp. Alka-line extraction seems to change the fibre wall morphology in such way that the wa-ter up-take is increased despite the re-duction in fibre charge. A possible reason for this is the increased fibre wall porosi-ty following xylan dissolution. More fines were also found in hemi-lean pulp which, however, does not explain the differenc-

Chemical component Unextracted birch pulp Alkaline extracted birch pulp

Cellulose * 74.8 85.8 Xylan * 23.6 13.0 Glucomannan * 1.1 1.1 Cellulose/hemicellulose ratio 3.04 6.09

Table 1. Chemical composition of the unextracted and alkaline extracted birch kraft pulp. About 50 % of the hemicelluloses were removed by extraction.

*The polysaccharide composition was calculated based on the monosaccharide content of the pulp samples according to Janson, J. (1974) Analytik der Polysaccharide in Holz und Zellstoff, Faserforsc. Textiltech., 25(9): 375-382.

146

es in WRV; the WRV was higher for hemi-lean pulp also when fines-free pulps were compared. In case of softwood pulps, an opposite behaviour was observed i.e. the WRV was lower for hemi-lean pulp in comparison to reference pulp. This indi-cates that the mode of hemicellulose re-moval might play an important role.

Thirdly, investigations on the recy-clability of hemi-lean birch kraft pulp in-dicated that alkaline extracted birch kraft pulp did not differ substantially from the reference pulp when incorporating several recycling rounds. It is known that hemi-celluloses tend to have deterrent effect against hornification. Apparently, alkaline extraction of pulp inhibits hornification at some extent overruling the effect of par-tial hemicelluloses removal. This finding may be interlinked with the non-expect-ed increase in WRV of alkaline extracted birch kraft pulp.

As a summary, for both hemi-lean pulps the main changes in fibre and pa-per properties were that the demand for beating energy increases, tear strength of paper increases and tensile strength of paper decreases. However, our find-ings indicate that there are also some differences when the post-bleaching al-kaline extraction process is compared to

pre-hydrolysis of wood chips prior to the kraft pulping. Properties of fibres and the properties of paper are more or less the same despite the extraction method, but it is notable that the hemicellulose extrac-tion method induces some distinct char-acteristics.

4.2 Paper properties of a mixture of post-bleaching alkaline extracted and unextracted birch kraft pulpAs the papermaking properties of post-bleaching alkaline extracted and unex-tracted birch kraft pulp are markedly different, mixtures of low hemicellulose content and unextracted birch kraft pulp might possess interesting properties com-pared to properties that these pulps have alone. Mixtures of alkaline extracted (he-mi-lean) and unextracted birch kraft pulp have following features:• Optical properties follow quite linear

pattern in mixtures• The dry strength properties of paper

are not significantly affected when adding low amounts (up to 10-15%) of alkaline extracted fibres to unextracted birch kraft pulp furnish

o The tear-tensile performance of alkaline extracted birch kraft

Alkaline extracted birch pulp Prehydrolyzed pine pulp

WRV increases WRV decreases More flexible fibres than reference Lots of fines generated in refining →

stiffer fibres than reference The strain to failure increases probably because the extraction process develops

lots of kinks

Reinforcement index slightly higher than for reference

Recyclability does not change compared to reference

Table 2. The main findings related to the partial removal (~50 %) of hemicelluloses from pulp (birch) or chips (pine). Novel findings denoted in italics.

147

pulp benefits more from addition of unextracted birch kraft pulp than unextracted birch kraft pulp suffers from addition of alkaline extracted birch kraft pulp (see Figure 1)

Results shown in Figure 2 suggest that slight additions of hemi-lean pulp in-to unextracted birch kraft pulp based fur-nish could even have a positive effect on the tear-tensile behaviour of paper.

4.3 Performance of hemi-lean fibres in papermaking processThe behavior of hemi-lean fibres in pa-permaking process, especially in the wet end of a paper machine, was mainly stud-ied from the retention point of view. In this context, the electrostatic charge in-teractions are very important and there-fore the effects of post-bleaching alkaline extraction on fibre charge properties were determined.

Figure 1. Tensile (left) and tear indices (right) for mixtures of alkaline extracted and unextracted birch kraft pulp.

Figure 2. Tear-tensile curve of pulp mixtures made from alkaline extracted birch kraft and unextracted birch kraft pulp.

148

The fibre charge measurements were based on the polyelectrolyte adsorption technique. Poly-DADMAC with molecular weight of 200,000–300,000 g/mol was used for determination of surface charge while Polybrene with molecular weight of 8,000 g/mol was used for determination of total charge. Before the charge mea-surements, the pulps were converted in-to their Na-form and washed. The charge measurements were made in fixed elec-trolyte concentration (0.01 M) at pH 8. The main conclusions related to these charge measurements were that:• The surface charge of the pulp was

slightly reduced due to alkaline extraction

• The reduced surface charge of hemi-lean pulp was mainly due to decreased charge density of fines

• No differences were observed in surface charge density of fines-free pulps

• The total charge of the fines fraction separated from hemi-lean pulp has a similar total charge to that of corresponding fibre fraction. For comparison, the fines in unextracted pulp are clearly more charged than the long fibres.

The results suggest that the alkaline extraction may have removed charged groups more effectively from fines than from fibres.

The altered charge properties of the alkali extracted pulp reflect to the cat-ionic demand behaviour of this particular pulp in wet end of a paper machine. The retention of ground calcium carbonate (GCC) in presence of cationic polyacryl-amide (c-PAM) was investigated in a lab-oratory using a MBF device. The results showed that the filler retention was nota-bly higher for the hemi-lean pulp in pres-ence of fines. In absence of fines, how-ever, there was no differences in the filler retention of the alkali extracted and unex-tracted pulps (Figure 3). The results sug-gest that the less charged fines in hemi-lean pulp consume less cationic retention aid which can then more effectively inter-act with fibres and filler.

Besides filler and retention polymer, other papermaking additives normally used in the production of fine paper were evaluated. From these, alkyl ketene dimer (AKD) and cationic starch were selected for further examination. The behaviour of these chemicals with alkali extracted pulp was evaluated using a MBF device. The

Figure 3. The effect of hemicellulose extraction and presence of fines on filler retention with varying C-PAM dosages. 30 % ground calcium carbonate was included in furnish. The injection of C-PAM took place 10 seconds before drainage. Sheets were prepared using the Moving Belt Former.

0

10

20

30

40

50

60

70

0 200 400

Filler reten)

on [%

]

C-‐PAM dosage [g/t]

Reference

Hemi-‐poor

Reference (fines-‐free)

Hemi-‐poor (fines-‐free)

149

main conclusions related to these experi-ments can be listed as follows:• In absence of filler, the sizing

response (ISO 535:1991) of AKD was better or similar for alkali extracted pulp in comparison to unextracted pulp

• In presence of GCC filler, the sizing response was notably lower for alkali extracted pulp than for unextracted pulp

• The alkali extracted pulp becomes more easily recharged by the addition of cationic starch in comparison to unextracted pulp

• The efficiency of cationic starch in improving the total retention of the stock at wet end seems to be impaired by the alkaline extraction of birch pulp

During the MBF trials, the water re-moval properties of the alkali extracted

and unextracted pulps were compared by continuously monitoring the water remov-al time and sheet dry content after blot-ting. The results showed no significant differences in the water removal proper-ties between the two pulps.

4.4 Effects of extraction conditions in alkaline extraction of birch pulpThe effects of the two most important ex-traction parameters, time and alkali con-centration, on extraction yield were inves-tigated. Of these parameters, the alka-li concentration was found to be the most important parameter that determines the amount of extracted xylan whereas the long extraction time was found to be of less significance.

Note, that by using an alkali concen-tration of 2.5 wt-% during extraction, same amount of xylan can be extracted in four minutes that normally is obtained

Figure 4. Effect of time on the efficiency of xylan extraction. 5% consistency of pulp slurry, 2.5% NaOH concentration, room temperature (22±1°C) and mixing with laboratory stirrer 120 r/min.

150

during two hours extraction time (see Fig-ure 4). These results thus reveal that at a certain alkali concentration, a short ex-traction time can be achieved to remove the extractable xylan. This indeed dem-onstrates that the process is feasible for scale up and is interesting solution for an integrated biorefinery process.


The papermaking, paper properties and the processability of hemicellulose ex-tracted and unextracted kraft pulp are markedly different. Therefore, mixtures of low hemicellulose content and unex-tracted kraft pulp might possess inter-esting properties compared to proper-ties that these pulps have alone. Extract-ed hemicellulose product could be sold for further processing and product upgrading or utilized in-house for increased profit-ableness of the process. If done careful-ly and tailored to individual mill to fulfil adequate market demand, the extraction of hemicelluloses might provide improved profitability for pulp mill producing pine or birch kraft pulp.

The time in alkaline extraction of xy-lan from bleached birch kraft pulp seemed to have negligible effect on the amount of xylan to be extracted; in fact, the alka-li concentration was the only critical pa-rameter. Therefore, the extraction of xy-lan could be considered to be relatively easy to implement in an industrial scale if only short extraction times are required. The shorter extraction time simplifies the process extraction arrangements signif-icantly and makes it economically more attractive. However, a major downside of the alkaline extraction is the high need of sodium hydroxide (1 M NaOH corresponds to 40 g/l NaOH solution) and the poor re-covery rate of reusable NaOH. In the lit-

erature it is stated that the efficiency of the recovery is only 82 %. Consequent-ly, the separation and utilization of xylans from the strongly alkaline extracts could be economically and technically challeng-ing. This problem related to the NaOH re-covery must be solved in order to reach a feasible process if the xylan in the ex-tract is separated to be utilized for prod-uct upgrading.

A possible option to avoid the cost-ly separation of xylan from the alka-line extracts would be to utilize the xy-lan-rich material stream as such in-situ. The strongly alkaline xylan-rich material stream from the extraction process could be possibly be used as an alkali source for e.g. oxygen delignification. It can be hypothesised that in these process con-ditions the xylan would precipitate on the surface of the oxygen delignified pulp.


The hemicellulose extraction seems to increase the energy required in refin-ing no matter how the hemicelluloses are removed. Especially, the swelling of the fibres is limited due to the removal of hemicelluloses from the fibre cell wall that facilitate fibre swelling. Therefore, the pulps with low content of hemicellu-loses could benefit from mechanical treat-ments that increase fibre conformability and their swelling tendency in water-fibre interaction prior to conventional refining. Providing that the problem of increased beating energy demand can be solved, some technological advantages could be achieved by replacing a portion of kraft pulp with hemi-lean pulp in production of conventional paper and paperboard.

The post-bleaching alkaline extract-ed and unextracted birch kraft pulps have different papermaking and paper physic-

151

chemical properties. A part of the chang-es entailed by the alkaline extraction are undesirable. Nevertheless, it might be possible to utilize the hemi-lean birch fi-bres by mixing them with unextracted birch kraft pulp, thus introducing some specific properties for the selected end-products. Further research on mixtures of hemicellulose extracted and convention-al kraft pulps could give an indication if it makes sense to integrate the alkaline hemicellulose extraction step of bleached birch pulp to an existing fibre line. Thus, the hemi-lean fibres would be utilized among the main product that is the un-extracted birch kraft pulp. The extracted xylan product could be utilized either in-house or sold for further processing, thus entailing in improved profitability.


Lyytikäinen, K., Saukkonen, E., Ka-janto, I., Käyhkö, J. (2011) The ef-fect of hemicellulose extraction on fiber charge properties and retention behavior of kraft pulp fibers, BioResources 6(1): 219-231

Saukkonen, E., Kajanto, I. (2009) Ef-fect of hemicellulose extraction on pa-permaking properties of kraft pulp fibers, 2nd Nordic wood biorefinery conference, 2-4 Sept. 2009, Poster Proceedings II, p. 40-44, Helsinki, Finland

Saukkonen, E. (2009) Hemicellulose extracted kraft pulp as raw material of paper, MSc Thesis, LUT, Department of chemical technology, Lappeenranta, Fin-land

Saukkonen, E. (2010) Post-bleaching al-kaline extracted kraft pulp for paper pro-duction, PAPSAT yearbook 2010: Inter-national Doctoral Programme in Pulp and Paper Science and Technology in Finland, Järvelä, H. (Ed.) p. 89-93

Saukkonen, E., Lyytikäinen, K., Käyh-kö, J., Kajanto, I. (2011) Papermaking performance of a mixture of low hemi-celluloses content and normal birch kraft pulp, 3rd Nordic Wood Biorefinery Con-ference (NWBC), 22-24 March 2011, Pro-ceedings, p. 300-301, Stockholm, Swe-den

152


Editor

Partners:

Metla

University of Eastern Finland


Pöyry Management Consulting Oy

Pekka Saranpää

Key researchers:

Pekka Saranpää

Riitta Julkunen-Tiitto

Jari Yli-Kauhaluoma

Katja Bergroth

153

AbstractTrees are biological organisms and wood, pulp and paper are natural biological materials. Wood

and paper, as such, require biochemical protection in many applications. Products packed in pa-

per or paperboard, on the other hand, also require protection, for example, food and cosmet-

ic products against oxidation. Many small molecules present in wood can be extracted and po-

tentially used, in terms of wood and paper products, to generate protection against, e.g. oxida-

tion and microbiological activity. They can also have an impact on human health, even as medi-

cines (see separate chapters in this report). The work presented in this chapter focuses on tan-

nins and stilbenes, the chemical up-grading of extracted molecules, as well as a desk-top study

on market opportunities.

154

1. BackgroundTrees comprise, in addition to the well-known polymers cellulose, hemicellulose and lignin, a large variety of small mole-cules. The total concentration of pheno-lics can in leaves and stem bark represent more than 10 and 20% of the total dry weight, respectively. 30 to 80% of pheno-lics are tannins. About half of the tannins are insoluble. They are bound to the cell wall cellulose-matrix. The soluble half is, on the other hand, found in the cell vacu-oles. Most of the pinosylvins are found in the heartwood and knots of trees.

The functional roles of tannins and stilbenes are not well known, but there are infers that they are structural wood components just as much as cellulose, hemicelluloses and lignin. Considerable data show that both compounds possess strong antimicrobial activity towards bac-teria, fungi and viruses and thus, may serve as a natural defence against a wide range of microbial infections in wood ma-terial.

Conifer roots may have a very dif-ferent composition due to the symbiot-ic associations that form between the roots and fungi (mycorrhiza) as well as between roots and endophytic micro-or-ganisms. These symbioses are character-ized by bi-directional movement of nu-trients where carbon flows to the fun-gus and inorganic nutrients move to the

plant, thereby providing a critical linkage between the plant root and soil. In infer-tile soils, nutrients taken up by the my-corrhizal fungi can lead to improved plant growth and reproduction. As a result, my-corrhizal plants are often more compet-itive and better able to tolerate environ-mental stresses than are nonmycorrhizal plants.

The terpene resin acids from pines (Pinus sp.), spruce (Picea sp.) and firs (Abies sp.) constitute an affordable and ubiquitous byproduct of the Finnish forest industry. The Kraft chemical pulping pro-cesses release tall oil soap, which consist mainly of sodium salt of fatty acids and resin acids, and unsaponifiable neutral components such as sterols. After acid-ification, free fatty and resin acids are formed and crude tall oil is obtained. The fatty and resin acids can then be separat-ed by vacuum distillation in special plants.

In nature, resin acids protect wood against insects and several microorgan-isms, from bacteria to fungi. Natural res-in acids are hydrophobic mixtures of com-pounds and can be found in many iso-meric forms, such as abietic acid, dehy-droabietic acid, pimaric acid, isopimar-ic acid neoabietic acid, levopimaric acid and palustric acid (Figure 1). Their anti-microbial activities, especially antifungal and antibacterial activities, have gained interest in recent years.

Figure 1. Representative chemical structures of common resin acids.

155

2. ObjectivesThe objectives of the work can be defined as follows:1. To identify, separate and test

selected extracts and compounds from trees or forest industry waste streams for their biological properties, especially antioxidative, antimicrobial, stability as well as binding properties.

2. Utilize the new knowledge on bioactive extracts and compounds to design products for antioxidative, antimicrobial and abiotics protection of various products, especially wood products.

3. Describe the chemical composition of spruce and pine stumps and thick roots.

3. Research approachCondensed tannin fractions of bark, stems and cones of spruce (Picea abies), pine (Pinus silvestris) and willow (Salix pyro-lifolia) were extracted and characterized (purity, structure, average chain length)

using HPLC, HPLC/MS, NMR and phloro-glucinol degradation methods. The sta-bility of tannins during UVB-exposure at the dose adjusted to equate the UVB level measured in June in Joensuu was tested in laboratory conditions. Condensed tan-nins and stilbenes were added to the pa-per and wood chips, exposed to the UVB radiation and the residual tannins and stilbenes were quantified. Studies on microbial tannins and stilbene interac-tion have been conducted using bacteria and specified fungal tests. For nanoparti-cle association of tannin extracts in order to use tannins in a controlled release sys-tem to preserve wood products indoors and outdoors Sol-gel method (silica) and monolithic silica was preliminarily tested as binding matrices for tannins on wood pieces in soil contact tests.

The accelerated soil contact test in which the impregnated sapwood samples with tannins, tannins in silica matrix and stilbenes were exposed to soil-inhabiting microorganisms has proceeded also.

Samples of Norway spruce and Scots pine were taken from the stump and un-derground positions of the roots at three

Figure 2. Root sections of Norway spruce (left) and Scots pine (right).

156

0

0,1

0,2

0,3

0,4

0,5

0,6

Bark, crude

Bark, Concentrated 1

Bark, concentrated 2

Cone, crude 1

Cone, crude 2

Cone, crude 3

Cone, concentrated

Sarja1

different distances from the root collar (Figure 2). The trees chosen from both tree species differentiated between their ages and size. The samples were ex-tracted with different methods including Soxhlet, sonication and ASE. Analyses of the extracts were performed with chro-matographic methods (GC-MS, HPLC-UV and -MS). Isolation of pure compounds was accomplished with preparative-HPLC.

In the subproject that aimed at a chemical modification of abietic acid and dehydroabietic acid methods of modern organic synthesis were used. In addition, state-of-the-art methods of structural characterization of the synthesized new compounds (NMR, FTIR, MS, UV) com-bined with methods for determining the purity of the prepared compounds (HPLC-MS, GC-MS, GPC, preparative chromatog-raphy) were applied.

4. Results

4.1. Tannins and stilbenes from conifers and willowCondensed tannin fractions of bark, stems and cone of spruce, pine and willow were

based on the HPLC- analyses almost free of small-molecular-mass impurities, and NMR-analyses confirmed the profiles of condensed tannins. The average size of tannins obtained by the phloroglucin-ol degradation method (average chain length) indicated 4–12 catechin units in the tannin molecules. The tannin con-tent is highly variable depending on the conifer part extracted (Figure 3). More-over, the exceptionally high content of to-tal tannin (7 to 9 % on dry weight basis) was detected in the willow bark samples over different ages.

The stability of condensed tannins and stilbenes on paper chips was low, they disappeared quite rapidly and linear-ly under UVB radiation. However, the pilot tests of the stability of condensed tannins impregnated into the wood chips indicat-ed high tannin stability under UVB over 6 months treatment.

Antifungal properties of several tan-nin fractions were evaluated using a high throughput method in 24-well plates in liquid malt broth in total with 15 decay fungus species (8 brown rot, 3 white rot and 4 soft rot species). Inhibiting ef-fects of the tannin fractions on the fungal growth was measured by weighting the

Figure 3. The content of condensed tannins in conifer fractions.

Spruce tannin fractions: relative amount of tannins

Abs

orba

nce

550 n

m(a

cid

buta

nol

sam

ples

)

157

mass of fresh fungal hyphae in the be-ginning and after one week of cultivation. Four tannin fractions out of eight showed high antifungal activity at rather low con-centrations against brown rot fungi.

Nanoparticle association of tan-nin extracts, in order to use tannins in a controlled release system to preserve wood products indoors and outdoors, showed that Sol-gel method (silica) gave nanoparticles and the tannins did not in-hibit formation of nanoparticles. Howev-er, catalyst interaction with tannins was a problem. Next monolithic silica was used as a binding matrix for tannins on wood pieces in soil contact tests. The accelerat-ed soil contact test with tannins, tannins in silica matrix and stilbenes measured as mass losses over several months showed minor differences between treatments.

4.2. Stilbenes from root barkThe total amounts of acetone extracts in Norway spruce and Scots pine roots and stumps are represented in Figure 4. The

Tannin

fractions

Minimal inhibitory concentration

(mg/ml)

within all tested decay fungi

1. 0,5

2. 0,5

3. 0,75

4. 0,75

5. 1,5

6. 1,5

7. 2

8. >2

Table 1. Concentration of tannin fraction for inhibition of fungal growth.

root and stump extracts of spruce and pine wood comprised mainly mono- and oligosaccharides and resin and fatty ac-ids. Pine stump wood contained also two stilbenes: pinosylvin and its monometh-yl ether. Lignans were present in some spruce wood samples. Catechin and β-sitosterol was also identified in the bark of both species. The amount of extrac-tives varied between the samples because of different age and size of felled trees.

We found the bark of Norway spruce roots to be a rich in the stilbene gluco-sides: astringin and isorhapontin (Figure 5). Also resveratrol glucoside, piceid was identified. The bark in the root neck con-tains more stilbenes than the root tip. The results suggest that the bark of the roots close to the stem of Norway spruce may be a potential source of these bioactive compounds.

Polyphenolic stilbene compounds in Norway spruce root bark can be success-fully purified from the extracts (Figure 6). Enzymatic hydrolysis offers a method to

Figure 4. Total amounts of acetone extracts from different zones of Norway spruce (left) and Scots pine (right) roots and stumps.

158

Figure 5. The amount (mg/g, dry mass) of major compounds in Norway spruce root bark sample from a peatland site (sample zones S, A, B, C according to Figure 4).

get stilbene aglycones which may act as antioxidants or antimicrobial agents.

4.3. Abietic acid and dehydroabietic acid derivativesWe have found out that chemically pure isolated abietic acid is a relatively unsta-ble compound due to the easily isomer-izable and conjugated double bond sys-tem in its tricyclic diterpenoidA structure. Thus, our synthesis work has been fo-cused mainly on modification of chemi-cally more stable dehydroabietic acid and its derivatives. During the first phase of the FuBio reseach programme we synthe-sized thirty phenyl urea derivatives of de-hydoabietic acid. In addition, we devel-oped a method for preparing previously unknown norditerpenoids, such as norde-

hydroabietylamines, which showed very interesting properties in various bioactiv-ity assays.

The antileishmanial properties of the abietic acid and dehydroabietic acid deriv-atives were assayed in Professor Charles Jaffe’s laboratory at the Hebrew Univer-sity of Jerusalem in Israel. Leishmania-sis is a parasitic disease that affects mil-lions of people in developing countries and has been designated as a neglected tropical disease by the World Health Or-ganization2). It is caused by the proto-zoan parasite Leishmania and transmit-ted by sand flies belonging to the genus Phlebotomus and Lutzomyia in the Old and New World, respectively. Due to the severe side effects of current antileish-manial drugs, there is an urgent need for

0

5

10

15

20

25

30

35

40

45

amou

nt (m

g/g, dry weight)

Norway spruce root bark, Peatland (April 2009) n=5

159

Figure 6. Chemical structures of stilbene compounds in spruce bark.

(R1=R2=OH; trans-piceatannol (astringenin), R1=OH, R2=H; trans-resveratrol, R1=OH, R2=OCH3; trans-isorhapontigenin, R1=OGlc, R2=OH; trans-astringin, R1=OGlc, R2=H; trans-piceid, R1=OGlc, R2=OCH3; trans-isorhapontin, *OGlc=3-O-β-D-glucoside).

the development of new compounds for the treatment of all clinical forms of leish-maniasis. The preliminary screening as-say of dehydroabietic acid-derived phe-nyl ureas gave five inhibitors of Leishma-nia donovani parasite at 12.5 µM concen-tration3). Promising and distinct struc-ture-activity relationships were found, and these compounds can be regarded as significant new hit compounds for fur-ther improvement and optimization.

Furthermore, professor Pia Vuorela’s group at the Åbo Akademi University in Finland identified that some amino deriv-atives of abietic acid and dehydroabiet-ic acid are capable of destroying mature Staphylococcus aureus biofilms. Staphy-lococcus aureus is a very versatile oppor-tunistic pathogen due to its ability to form

biofilms in medical devices as well as di-verse resistance mechanisms4). Identi-fied hit compounds were able to prevent biofilm formation and more importantly destruct mature biofilms with good po-tency values (IC50 of 15-120 µM for pre-vention and 70-290 µM for destruction). One of the hit compounds inhibited bio-film formation by a mechanism that does not involve killing of planktonic bacteria.

Finally, Sami Alakurtti at VTT Techni-cal Research Centre of Finland carried out UV-activity tests. In these assays abietic acid and dehydroabietic acid derivatives possessed good UV-A absorbance.

4.4. Regulative issues: Lessons learntModern research is highly regulated through various processes. This is a con-sequence of the need to steer research and development towards products with as minimally harmful an effect on hu-mans and the environment as possible. Likewise, while not in any way minimis-ing the need for basic research, a link to market perspectives is often missing yet potentially highly useful. The challenge of this has, however, been the high cost and long regulatory time span which is needed to fulfil all the requirements be-fore a sales permit for a new product can be granted. Moreover, the regulatory field remains complex which makes it hard for researchers and SME’s to gather critical knowledge and information. In line with these considerations the regulatory ques-tions were also addressed in FuBio.

One of the targets of this research project was to study how regulatory is-sues and market perspectives could be integrated into a research project already at an early stage. The aim was to draw product roadmaps and evaluate market possibilities and address regulatory issues at the same time. The study focused on wood-derived polyphenols and on prod-ucts where antioxidant properties are es-

160

sential. The market studies were comple-mented by laboratory research. Potential application areas include natural antioxi-dants and biocides for protection of food and other products, e.g. cosmetics as well as specialty dietary products. Further-more, if polyphenols are deemed accept-able for human food use, without harmful effects, they could be used also as com-ponents in food packaging materials. Al-so, the suitability of polyphenols as func-tional components in cosmetics was eval-uated. Evaluations were made collecting existing and project-generated knowledge on market, cost and business concept.

The regulatory issues were mapped by considering the effects of REACH, Nov-el Foods and health claims on the process of reaching markets. These combined ef-forts were deemed to make it possible to identify and develop novel products from polyphenols which could be essential for protection of various products or human health, and which could create profit for the forest cluster.

The REACH-legislation was linked in

at the very beginning of the research pro-cess in order to fully assess all aspects having an impact on final market oppor-tunities. The REACH-directive regulates the manufacturing of substances that are brought to the market in larger quantities than 1 ton/year. In addition, the research and development activities of substanc-es meant to be produced or produced for R&D purposes in aforementioned quan-tities are under the requirements of the directive.

During the study it became clear to what a great degree the regulative pro-cesses vary for different end-use product markets. The regulatory scheme is very different between e.g. food and pharma-ceuticals. Also, the REACH process has its own test set requirements which have to be followed. Moreover, the cost of tak-ing new products through one regulato-ry scheme is very high. In practice this means that the researcher has to, already at a quite early stage; determine how to position the research, meaning which end-use the products are best suited for.

Figure 7 illustrates the process scheme applied during the project.

161

The reason is that it might be impossi-ble from the financial point of view to ini-tiate regulatory tests in all industry cat-egories. In research which aims at im-proving known products this might not be a problem, as the target market and end-use is already specified. However, in research which targets finding com-pletely new products from nature derived raw materials such as the polyphenols in this case, the product positioning is often hard. Here, the research case is the op-posite, the raw material is known but the end-product is not.

In this study it became ever more ap-parent that in order to speed up research, a platform for the regulatory processes should be developed. This would not on-ly save costs, research efforts but also increase speed and hopefully improve throughput of research. Modern research and development practices should be de-veloped towards shorter development cy-cles and enabling greater flexibility. Reg-ulatory processes should not work in the way that they slow down and make it hard to initiate research and development possibilities already at the beginning.


In the world today, there is a need and also a desire for new and more environ-mentally friendly products and agents in different industries. The roots and stumps might offer a base for sustainable value chain for producing novel biochemicals without consuming raw materials of tra-ditional wood working and paper industry. Possible applications from the extracts and their polyphenolic and terpene-based compounds for the industry are products for improving health (i.e. anti-cancer, an-ti-inflammatory) or antifungal and anti-bacterial products for example for protec-

tion of wooden products against mould, rot and insects.


Extraction, hydrolysis and purification methods for producing bioactive stilbenes from Norway spruce root bark need to be optimised and integrated into a whole process. Also suitability of compounds for utilization needs to be evaluated with pre-cise testing methods (antioxidative and antimicrobial properties). The binding methods to increase the stability and con-trolled release of tannins and stilbenes on the products will need the further stud-ies and developments. Next steps and on-going studies include the development of purification methods for stilbene frac-tions with preparative HPLC to test and enhance their stabilities. The optimisation of extraction methods, stability, decom-position activity and nanoparticle binding tests of the fractions, and large scale iso-lation of compounds such as tannins and pinosylvin will be subject for further stud-ies. Finally, further studies aiming at com-prehensive structure-activity relationship analyses and resolving the biochemical mechanisms of leishmanicidal and anti-biofilm activities by the semisynthetic de-rivatives of abietic acid and dehydroabiet-ic acid are underway.

162


Editor

Partners:

University of Tampere

Eeva Moilanen

Key researchers:

Mirka Laavola, Tiina Leppänen, Riina Nieminen, Eeva Moilanen

163

AbstractNatural products and their derivatives have been invaluable as a source of therapeutic agents

throughout the history. Natural plants still continue to have a significant role in drug develop-

ment and there are a variety of novel chemical entities from natural sources undergoing clinical

trials. The bioactive compounds in Finnish natural plans, including forest trees, remain poorly in-

vestigated and hold a great potential for health promoting and other applications. The aim of the

present sub-project was to screen knot and bark extracts of Finnish forest trees and compounds

isolated from the extracts for their immunomodulatory and anti-inflammatory activities. Some

of the tested preparations and compounds proved to have immunomodulatory properties and

also previously unknown effects were discovered. The results encourage to continue investiga-

tions with selected compounds to obtain further evidence to support development of wood ma-

terial based health-promoting applications

164

1. BackgroundNature and natural plants are a rich source of bioactive compounds. For thou-sands of years, natural products have played an important role all over the world in the treatment and prevention of human diseases. Over 50 % of drug molecules in clinical use stem from nat-ural sources, being either natural prod-ucts or analogues inspired by them. Good examples of health-related successes o iginating from the nature are acetylsal-isylic acid, cyclosporin A and docetaxel (Figure 1).

Throughout human history salicylate containing willow extracts have been used in reducing pain, inflammation and fever. In 1899 acetylated salicylic acid, Asper-in®, was created and it is still one of the world’s most used therapeutic agents.

Cyclosporin A was initially isolated from a Norwegian soil sample and is pro-duced by the fungus Beauveria nivea. Cyclosporin A is an immunosuppressant, which revolutionally improved the prog-nosis of e.g. severe kidney diseases. To-day it is widely used to prevent organ re-jection after transplantation, and also as

Figure 1. Acetylsalicylic acid, Cyclosporin A and Docetaxel.

an immunosuppressive drug in the treat-ment of diseases like rheumatoid arthri-tis.

Docetaxel is an esterified product of 10-deacetyl baccatin III, which is extract-ed from the leaves / needles of the Taxus baccata (European yew tree). It is used as an antimitotic chemotherapy medica-tion in the treatment of breast, ovarian and non-small cell lung cancer.

Natural products continue to have a significant role in drug development. The structures of plant derived compounds have characteristics of high chemical di-versity, biochemical specificity and other molecular properties that make them fa-vourable as lead structures for drug dis-covery, and differentiate them from li-braries of synthetic and combinatorial compounds. The natural compounds are unique as compared to libraries of the ar-tificially designed molecules because they have undergone nature’s own evolution-ary selection process for the optimization of biologically active compounds. Natu-ral compounds tend to possess well-de-fined three-dimensional structures, which enhances the likelihood of specific bind-ing and focused effects and reduces the

165

risk of side effects related to unspecif-ic binding characteristics typical for ma-ny synthetic small molecule compounds. The bioactive compounds in Finnish nat-ural plans, including forest trees, remain poorly investigated and hold a great po-tential for health promoting and other ap-plications.

The current research project aims to characterize immunomodulatory/anti-in-flammatory activities and compounds in by-products of forest industry. In general, inflammation is a protective response in-tended to eliminate microbes and other of-fending agents and to repair tissue dam-age. The response is usually beneficial but both inflammation and repair processes have also considerable potential to cause harm. Inappropriately focused, dysregulat-ed or prolonged inflammation may result in a chronic inflammatory disease such as asthma and arthritis (Figure 2).

Osteoarthritis is one of the most prev-alent and disabling chronic diseases af-fecting the elderly. In Finland nearly 0.5 million people suffer from symptomatic osteoarthritis. The most prominent fea-ture of this disease is the progressive de-struction of articular cartilage, which re-sults in impaired joint motion, severe pain, and ultimately, disability. Rheuma-

toid arthritis is an autoimmune disease affecting about 1% of population. The lymphocyte and macrophage drived in-flammation in the joints causes pain and stiffness and gradually the process leads to joint destruction and physical disabil-ity. Examples of other inflammatory dis-eases are listed in Table 1.

Today, various anti-inflammatory agents are available for the treatment of inflammatory diseases but still a remark-able number of patients do not respond to the treatment sufficiently. Therefore, there is a critical need to develop more effective and less toxic agents to treat the signs, symptoms and long-term conse-quences of inflammatory diseases. More importantly, compounds and function-al food products which could prevent in-flammatory and allergic diseases are ur-gently anticipated.

2. ObjectivesThe purpose of the project was to identi-fy bioactivities or compounds in domes-tic wood extracts which could have ben-eficial immunomodulatory or anti-inflam-matory properties suitable for health-re-lated applications such as functional food products, medicines or natural drugs. The

Figure 2. Typical features of rheumatoid arthritis (A) and osteoarthritis (B) in the hand.

Table 1. Examples of inflammatory diseases.

• Arthritis (rheumatoid arthritis and osteoarthritis)

• Alzheimer’s disease

• Asthma and allergy

• Atherosclerosis

• Celiac disease

• Diabetes

• Inflammatory bowel disease

• Multiple sclerosis

• Obesity and metabolic syndrome

• Parkinson’s disease

• Septic shock

• Stroke

166

project aimed to develop and validate in vitro and preliminary in vivo methods to model the responses typical for arthritis and allergic inflammation and exclude di-rect cytotoxicity, and apply those meth-ods in the studies aiming to identify novel immunomodulatory properties and com-pounds.

3. Research approach and results

Eleven different polyphenol-rich wood ex-tracts and a series of compounds isolat-ed from them were selected for biotest-ing of immunomodulatory activities. The extracts and compounds were first test-ed for their direct cytotoxic properties by using standard assays to detect mito-chondrial dehydrogenase activity and cell death. Just a few compounds showed de-tectable toxicity at tested drug concentra-tions, and those were excluded from the further testing.

Anti-inflammatory properties of the extracts and compounds were investi-gated on expression of an array of in-flammatory genes in macrophages, lym-phocytes and chondrocytes activated by toll-like receptor ligands or by other in-flammatory stimuli, and the effects were compared to standard anti-inflammato-ry drugs. Many of the tested extracts and compounds showed immunomodulatory effects by down-regulating inflammatory genes including interleukin-6, tumor ne-crosis factor alpha and inducible nitric ox-ide synthase in a dose-dependent man-ner at reasonable drug concentrations. All of those genes are aberrantly activated in inflammatory arthritis and responsible for joint inflammation leading to cartilage degradation.

Two extracts and some compounds isolated from them showed significantly improved efficacy and were selected for further studies to investigate the mecha-

nisms of action in further detail. Some of the active extracts and compounds were found to inhibit the activation of inflam-matory transcription factor NF-kB while some other compounds seemed to cause their action by another previously un-known mechanism.

Based on the results from the in vi-tro anti-inflammatory tests we select-ed the most potent extracts and isolated compounds to be tested in vivo. The an-ti-inflammatory properties of the select-ed extracts/compounds were investigat-ed in carrageenan-induced inflammato-ry paw oedema in the mouse which is a widely used in vivo model for initial an-ti-inflammatory testing of pharmaceuti-cal compounds. Four of the tested com-pounds showed a significant anti-inflam-matory efficacy when using dexameth-asone as a positive control compounds. The results also suggest that the four compounds have a reasonable bioavail-ability, and no toxicity was recorded.

4. Future plans and business potential

In the first phase of FuBio a large se-ries of different wood extracts and com-pounds derived from them were investi-gated. Some of the tested extracts and compounds were shown to have previous-ly unknown anti-inflammatory properties and might therefore be potential ingredi-ents and/or lead molecules in health ben-eficial products targeted for immunomod-ulation to prevent or treat inflammato-ry diseases including arthritis and allergy. Based on the results of these screening type experiments, more focused investi-gations with selected compounds are en-couraged to obtain further evidence of the discovered novel biological activities in the by-products of forest industry to sup-port development of added-value prod-ucts with health-promoting properties.

167

Figure 3. Examples of inflammatory genes and mechanisms related to the pathogenesis of inflammatory diseases which were addressed in the research project.

5. Publications and reports related to the projectHämäläinen, M., Nieminen, R., As-mawi, M.Z., Vuorela, P., Vapaatalo, H. Moilanen, E., 2011. Effects of flavonoids on prostaglandin E2 production, and on COX-2 and mPGES-1 expressions in acti-vated macrophages. Planta Med 77:1504-1511.

Leppänen, T., Laavola, M., Nieminen R., Korhonen R., Tuominen, R.K., Moi-lanen, E., Downregulation of protein ki-nase Cd inhibits inducible nitric oxide syn-thase expression through IRF1. Submit-ted for publication.

Laavola, M., Nieminen R., Yam, A., Sa-dikun, M.F., Asmawi, M.Z., Basir, R., Welling, J., Vapaatalo, H., Korhonen R., Moilanen, E., In Vivo and in Vitro anti-inflammatory activity of extracts from Or-thosiphon stamineus leaves. Submitted for publication.

168


Editor

Partners:

University of Turku


Sari Mäkelä

Key researchers:

Sari Mäkelä, Lauri Polari, Niina Saarinen-Aaltonen, Emrah Yatkin

Christer Eckerman, Bjarne Holmbom, Annika Smeds

169

AbstractWood and bark are abundant raw materials for different organic compounds. Many natural

chemicals present in wood are postulated to have positive, health-promoting effects, based on

the information available on similar compounds present in edible plants. Especially heartwood,

knots and bark are rich in bioactive compounds that serve as defense substances for living trees.

Researchers have recognized the value of trees as a useful source of natural biochemicals, and

several pharmaceuticals and food components based on wood biochemicals have been devel-

oped during the last decades. However, a vast reserve of wood biochemicals still remains unex-

plored and unexploited.

Regarding health-promoting properties, several interesting groups of compounds are pres-

ent in wood. Here, the focus is on phenolic compounds. High concentrations of various pheno-

lic compounds such as flavonoids, stilbenoids and lignans commonly occur in species of trees

processed by forest industry. These compounds are of special interest, with respect to develop-

ment of novel applications of wood-derived biochemicals. Epidemiological studies suggest that

high consumption of similar phenolic compounds as natural components of edible plants is asso-

ciated with low risk of Western lifestyle diseases, such as breast, prostate, lung and colon can-

cer, metabolic syndrome and cardiovascular disease. In this project phenolic extracts were bio-

logically characterized with different in vitro and in vivo models, in order to identify extracts and

compounds for further development as health-promoting ingredients.

Main objective in this project was to build the scientific knowledge base for novel polyphenol-

based health promoting products by using various biomedical test systems, consisting of specific

and human-relevant bioassays related to possible prevention of cancer, metabolic disorders and

endocrine disturbance. Ten wood-derived extracts were first screened in in vitro bioassays, and

one extract was selected for further studies, using different in vivo models. Our data indicate that

this extract and its components may, indeed, target a number of pathways that are critical for

obesity and obesity-related diseases, such as cancer. In conclusion, results obtained in FuBio1

strongly suggest that certain wood extracts and their components possess properties that can

be utilized in the development of pharmaceuticals and novel health-promoting food ingredients.

170

1. BackgroundPrevalence of obesity-associated diseas-es (e.g. cardiovascular diseases, diabe-tes and cancer) has increased significant-ly during the last decades, affecting ma-jority of Western populations. In partic-ular, there is a great need and market potential for effective preventive means, such as functional foods that aid in weight management, or attenuate the biological processes associated with obesity, con-tributing to the development of diseas-es, such as cancer. The complex interre-lationships of various lifestyle factors, bio-logical targets, and diseases are present-ed in Figure 1.

Phenolic compounds are the main de-fence compounds in plants and trees to abate attacks by other organisms. Here the focus is on lignans, stilbenoids and flavonoids. Lignans are widely distribut-ed in the plant kingdom, and present in high amounts in certain parts of trees, in particular knots. Their chemical structure is two C6C3 moieties, coupled at their β carbons. C6 is a benzene ring usual-

ly substituted with m-methoxy and p-hydroxy groups. Chemical and physical properties of lignans are well character-ized and they are known to modulate a number of important biological processes in mammalian cells. Several plant lignans are known as dietary precursors of en-terolignans (enterolactone and enterodi-ol), lignan metabolites that are produced by bacterial fermentation in mammali-an gastrointestinal tract. High serum en-terolactone concentrations have been as-sociated with reduced risk of e.g. breast cancer and cardiovascular diseases. In addition to lignans, many other pheno-lic compounds are present in wood, such as stilbenoids, flavonoids, hydrolysable tannins and lignin-derived phenolics, al-so suggested to possess health promoting properties. However, when working with wood derived extracts, it must be kept in mind that they are highly complex mix-tures and, therefore, may include other potentially bioactive non-phenolic com-pounds such as resin acids, betulinol or sucrose.

2. ObjectivesMain objective in this project was to build the scientific knowledge base for novel polyphenol-based health promoting prod-ucts by using various biomedical test sys-tems, consisting of specific and human-relevant bioassays related to possible prevention of diseases such as cancer, metabolic disorders and endocrine dis-turbances.

3. Research approachWe applied a tiered screening strategy composing of: • Stage I bioassays: cell survival,

proliferation and cell death

Figure 1. Interrelationships of lifestyle factors, biological targets, and Western diseases.

171

• Stage II bioassays: target specific cell-based assays (cell types, signalling pathways, indication specific assays)

• In vivo assays: metabolism of wood-derived molecules and anti-carcinogenic and metabolic effects in selected models

3.1 Stage I and Stage II bioassays – wood-derived extracts and pure compoundsThe effects of ten different wood knot and bark extracts and their main con-stituents were first screened in vitro in three different human cancer cell lines; androgen-independent PC3 prostate can-cer cells, metabolically active HepG2 liv-er cancer cells, and U-2 OS osteosarco-ma cells. Based on results obtained here, extracts were selected for further in vitro and in vivo studies.

3.2 In vivo assays – metabolism of wood-derived extractsMale and female mice were fed with a pu-rified high-fat diet containing 0.03% of the selected extract; control groups were fed with the same diet without the ex-tract. After seven weeks of exposure, 24-h urine samples were collected and phe-nolics and their metabolites were iden-tified and quantified using high-perfor-mance liquid chromatography-mass spec-trometry.

3.3 In vivo assays – anti-carcinogenic effectsAnti-carcinogenic effect of the selected wood extract was studied in vivo by using an orthotopic human PC3 prostate cancer xenograft model in immune-compromised mice. We used a transgenic PC3 cell line which expresses luciferase reporter allow-ing assessment of prostate tumor size by in vivo optical imaging during the study period. Wood extract was administered to mice for 5 weeks by daily oral gavage

(i.e. intragastrically) in two doses trans-lating into daily dietary polyphenol expo-sure levels in humans. In addition to tu-mour growth profiles, several growth-re-lated biomarkers were analysed from tu-mour samples.

3.4 In vivo assays – metabolic effectsThe metabolic effects of the same select-ed wood extract (in comparable dietary exposure levels) were tested in a mouse model that allows investigation of endo-crine-modulatory and metabolic effects simultaneously. This is a diet induced obesity model (DIO) in transgenic hARO-Luc reporter mice, expressing luciferase reporter gene under the control of full-length human aromatase gene promoter. In DIO/hARO-Luc mouse model, tissue-selective expression and regulation of hu-man aromatase gene can be determined by measuring luciferase activity. Aroma-tase is the key enzyme in estrogen bio-synthesis, thus having a widespread im-pact on a number of important physio-logical and pathological processes (e.g. breast cancer). Previous studies suggest that aromatase expression and/or activi-ty is modulated by various dietary factors (e.g. plant phenolics), as well as obesity-related metabolic disorders.

DIO/hARO-Luc mice were fed with a “Western” type of high-fat diet to induce weight gain, adiposity, hyperglycemia, hyperinsulinemia, and changes in serum adipokine profile (e.g. leptin and adipo-nectin). First, a pilot study was performed with intact adult males and females to confirm that hARO-Luc mice respond ad-equately to high-fat feeding. Next, two full-size studies were performed with adult males and with ovariectomised fe-males, modelling postmenopausal ovari-an hormone milieu in women. In addition to reporter gene activity, several biomark-ers of metabolic disorders were analysed from serum and tissue samples.

172

4. Results

4.1 Effects on cell proliferation in vitroIn the project we tested ten wood ex-tracts and their main constituents in three different cancer cells lines. Based on the obtained results on cell proliferation (Fig-ure 2) one extract was selected for in vivo testing in a prostate cancer model.

4.2 Metabolism of wood-derived extracts in vivoAs expected, several of the components (stilbenes and lignans) of the selected wood extract were detected in the urine of mice fed with the test diet. Further-more, significant concentrations of res-veratrol, which, per se, is not present in the wood extract fed to the animals, were present in the urine. This is of special in-terest, as resveratrol is known to possess health-promoting properties. We have, thus, identified a novel source of resve-ratrol and other potentially beneficial bio-chemicals.

4.3 Anti-carcinogenic effectsIn prostate cancer xenograft model, we were able to show that the higher dose

of selected wood extract reduced the growth of orthotopic human prostate can-cer growth in vivo. Attenuation of tumour growth was associated with reduced mi-crovessel density, suggesting that the an-ti-tumorigenic effect of the extract is, at least partly, due to inhibition of angiogen-esis.

4.4 Metabolic and endocrine effectsThe metabolic effects of the selected wood extract were studied in DIO/hA-RO-Luc mouse model. Feeding of high-fat “Western” diet to these mice increased their body weight, adiposity, blood glu-cose, and altered serum adipokine lev-els, indicating adequate response to high-fat feeding and applicability of the model to study targets associated with diet in-duced obesity. Moreover, feeding of se-lected wood extract to adult male and ovariectomised female DIO/hARO-Luc mice altered the human aromatase gene expression in selected tissues. This dem-onstrates that the wood-derived polyphe-nol can, indeed, modulate expression of key steroidogenic enzymes in tissue-se-lective manner and may have significance in development of hormone-driven dis-eases such as breast and prostate cancer.

Figure 2. Proliferation of different cancer cell lines in the presence of a selected wood extract.

173

5. Future business potentialWood and bark are abundant raw materi-als for different organic compounds. Ma-ny natural chemicals present in wood are postulated to have positive, health pro-moting effects, based on the information available on similar compounds present in edible plants. The beneficial proper-ties of plant polyphenols have previously been studied in both epidemiological and experimental settings with promising re-sults, and the positive effects, associat-ed with foods rich in plant phenolics, are also recognized by consumers. Further-more, there are several well-known ex-amples of wood derived compounds from different chemical groups, used in drug or food development, resulting in well-es-tablished products, such as salicylic acid, stanol esters, xylitol and quinine. For ex-ample, technology for separation of knots has been developed and is now used in industrial scale to provide the raw mate-rial for the production of the HMRLignan, which is marketed since 2006 as a dietary supplement.

Obviously, there is a great need and market potential for novel therapies and effective preventive means to combat the Western lifestyle diseases, such as met-abolic disorders and cancer. Results ob-tained in FuBio 1 strongly suggest that certain wood extracts and their compo-nents possess properties that can be ex-ploited in the development of pharmaceu-ticals and novel health-promoting food in-gredients.


The interest on the use of polyphenols in reducing the risk of developing cancer or metabolic disorders is increasing. Wood materials as an exceptionally rich source

of broad spectrum of bioactive compo-nents offer a good source of selected polyphenols. New knowledge on bioac-tive extracts and compounds, obtained in FuBio1, will aid for development of new health promoting products. Both pharma-ceutical and food-applications may be rel-evant development concepts, and will be considered in the on-going FuBio 2 proj-ect. Now the biotesting, focusing on the selected wood extract, is expanded to in-clude more specific and human-relevant disease models, related to possible pre-vention of cancer, metabolic disorders and endocrine disturbances. The main aim of the future work is to obtain proof-of-con-cept, the key milestone in the develop-ment chain of novel health-promoting products (see figure below).


Smeds A, Yatkin E, Polari L, Ecker-man C, Willför S, Holmbom B, Mäkelä S: Metabolism of scots pine knotwood phenolics in mice. 5th International Con-ference on Polyphenols and Health, 17-20.10. 2011, Barcelona, Spain.

Figure 3. Proof-of-concept in development of novel health-promoting products.

174

www.forestcluster.fi

Date post:	06-Sep-2014
Category:	Education
Upload:	finnish-bioeconomy-cluster-fibic-oy
View:	1,514 times
Download:	20 times

Forestcluster Fubio Biorefinery programme report

Education