Ageing and stabilisation of paper
edited by
Matija Strlič University of Ljubljana, Ljubljana, Slovenia
Jana Kolar National and University Library, Ljubljana, Slovenia CIP - Kataložni zapis o publikaciji Narodna in univerzitetna knjižnica, Ljubljana
676.017
AGEING and stabilisation of paper / edited by Matija Strlič, Jana Kolar. - Ljubljana : National and University Library, 2005
ISBN 961-6551-03-5 1. Strlič, Matija 219327232
© The authors, 2004. Apart from any fair dealing for the purposes of research, education or private study, this publication may not be reproduced, stored or transmitted, in any form or by any means, without the prior permission by the authors.
The authors gratefully acknowledge the support of the European Community. The work is the sole responsibility of the authors and does not represent the opinion of the Community. The Community is not responsible for any use that might be made of the data appearing herein.
Distributed by the National and University Library, Turjaška 1, 1000 Ljubljana, Slovenia. Preface
Paper is a material, which is often taken for granted – hardly any user is aware of how complex and variable material it actually is. Even fewer are informed of problems with its durability. Much of the information on paper is perhaps not meant to be long-lived in this day and age. Yet, among all information carriers it remains the most durable and printed or written information is still the most generally understandable. It is the obligation of many libraries, archives and museums worldwide to continuously provide access to the information and its carrier. Since the quality of paper varied considerably through centuries, problems with its preservation are equally varied. Acidic paper, produced roughly between 1850 and 1990 is among the most pressing topics. Its remaining lifetime possibly reduced to a mere century (as demonstrated in this book), the real danger exists that much of the written knowledge generated in this era of social, artistic, scientific and political turmoil will become obliterated. Will the 20th century become a new dark age in five hundred year’s time? Perhaps not, if appropriate actions are taken. The mission of this book is to provide the foundation on which the conservation chemist will be able to design and test conservation treatments. Descriptions of the necessary research methodology, general knowledge of paper degradation pathways and possibilities for stabilisation are summarized. Cellulose, the main structural component of paper, is a macromolecule, so it can be regarded from a wider perspective – chapters on degradation of polymers in general are also provided. It is wished that the book will thus aid not only the conservation chemist, but also students of polymer and paper chemistry and of material science on the whole. Especially the concluding chapters should be of interest also to the conservation practitioners and the responsible officers. This book is the result of much enthusiasm and scientific excellence of all the contributors, many of which took part in the innovative project Papylum – "Chemiluminescence – a novel tool in paper conservation studies" which ran from 2001 to 2004. We are deeply indebted to all. With particular gratitude, however, we welcomed the participation of A. Barański, J. B. G. A. Havermans, J. M. Łagan, T. Łojewski and T. A. G. Steemers, who made it possible for the book to grow far beyond what was initially planned. Through the 5th Framework programme, the European Commission provided vital support to research on many important subjects covered in this volume. It is not possible to overestimate its crucial role in bringing together relevant experts working together on the subject of research into preservation of cultural heritage. Therefore, we express our belief and expectation that the support will continue in future research programmes. M. Strlič and J. Kolar
Ageing and stabilisation of paper iii List of contributors Andrzej Barański Jasna Malešič Jagiellonian University National and University Library of Slovenia Laboratory for Physicochemical Analyses National Centre for Preservation of Library and Structural Research Materials & Ljubljana, Slovenia Department of Chemistry Kraków, Poland José Luiz Pedersoli Jr. Netherlands Institute for Cultural Heritage Dominique Fromageot Amsterdam, The Netherlands Centre National d’Evaluation de Photoprotection Ensemble Universitaire des Cézeaux Boris Pihlar Aubière, France University of Ljubljana Faculty of Chemistry and Chemical Technology Olivier Haillant Ljubljana, Slovenia Netherlands Institute for Cultural Heritage Amsterdam, The Netherlands Lyda Rychlá Polymer Institute of the Slovak Academy of John B. G. A. Havermans Sciences TNO Environment and Geosciences Bratislava, Slovakia Delft, The Netherlands Jozef Rychlý Drago Kočar Polymer Institute of the Slovak Academy of University of Ljubljana Sciences Faculty of Chemistry and Chemical Technology Bratislava, Slovakia Ljubljana, Slovenia Steph Scholten Jana Kolar Netherlands Institute for Cultural Heritage National and University Library of Slovenia Amsterdam, The Netherlands National Centre for Preservation of Library Materials Ted A. G. Steemers Ljubljana, Slovenia National Archives The Hague, The Netherlands Janusz Marek Łagan Jagiellonian University Matija Strlič Laboratory for Physicochemical Analyses and University of Ljubljana Structural Research Faculty of Chemistry and Chemical Technology Kraków, Poland Ljubljana, Slovenia Jacques Lemaire Vid Simon Šelih Centre National d’Evaluation de Photoprotection University of Ljubljana Ensemble Universitaire des Cézeaux Faculty of Chemistry and Chemical Technology Aubière, France Ljubljana, Slovenia Tomasz Łojewski Jagiellonian University Department of Chemistry Kraków, Poland
iv Ageing and stabilisation of paper Contents List of contributors iv Contents v Abbreviations viii Introduction Paper and durability 3 1.1 Introduction 3 1.2 Paper and its production 4 1.3 Paper degradation 5 1.4 References 7 Degradation and ageing of polymers 9 2.1 Introduction 9 2.2 Modes of degradation 10 2.3 Thermolysis 10 2.4 Oxidation 13 2.5 Light-induced degradation 14 2.6 Propagation and advanced degradation phases 16 2.7 General stabilisation strategies 18 2.8 Conclusions 21 2.9 References 22 Methodology Methodology and analytical techniques in paper stability studies 25 3.1 Introduction 25 3.2 Mechanical properties 25 3.3 Paper pH 26 3.4 Determination of alkaline reserve 27 3.5 Viscometry 28 3.6 Size exclusion chromatography 29 3.7 IR spectroscopy 30 3.8 Diffuse reflectance spectrophotometry, colourimetry and spectrofluorimetry 31 3.9 Degradation of linear polymers 32 3.9 The Arrhenius equation 34 3.10 Accelerated degradation studies 38 3.11 Conclusions 40 3.12 References 40 Experimental techniques in studies of photo-stability 45 4.1 Introduction: Accelerated photoageing – a century-long story 45 4.2 Review of outdoor weathering sites and accelerated photoageing units 46 4.3 Acceleration of photooxidation reactions 52 4.4 Conclusions 58 4.5 References 59
Ageing and stabilisation of paper v Contents
Chemiluminescence from polymers 65 5.1 Introduction 65 5.2 General classification of chemiluminescence from polymers 65 5.3 The effect of temperature and atmosphere 68 5.4 The effect of molar mass 69 5.5 The effect of antioxidants and polymer stabilizers 70 5.6 Mechanisms leading to chemiluminescence during polymer oxidation 71 5.7 A simple kinetic model of non-isothermal chemiluminescence 74 5.8 Fit of isothermal experimental data 77 5.9 Instrumentation 79 5.10 Conclusions 81 5.11 References 81 Degradation Acid-catalysed degradation 85 6.1 Introduction 85 6.2 Paper acidity in library collections. The example of Poland 85 6.3 Acid-catalysed degradation – background 87 6.4 Studies at Jagiellonian University 92 6.5 Diffusion phenomena inside books 97 6.6 The concept of mixed-control kinetics – acid hydrolysis and oxidation 97 6.7 Perspectives 99 6.8 References 99 Thermo-oxidative degradation 101 7.1 Introduction 101 7.2 Autoxidation 102 7.3 Initiation of oxidation 103 7.4 Peroxides and hydroxyl radicals 107 7.5 Role of water 109 7.6 Thermo-oxidative degradation: Ca or Mg? 112 7.7 Antioxidants 115 7.8 Conclusions 117 7.9 References 118 Chemiluminescence of cellulose and paper 121 8.1 Introduction 121 8.2 Sample morphology and size 122 8.3 Chemiluminescence in inert atmosphere 122 8.4 Chemiluminescence in oxidative atmosphere 124 8.5 Chemiluminescence following irradiation with light 132 8.6 Conclusions 134 8.7 References 134 Photooxidative degradation 137 9.1 Introduction 137 9.2 Photooxidative degradation of polymers 137 9.3 Free radicals during irradiation of cellulose 140 9.4 Depolymerisation of cellulose due to irradiation 141 vi Ageing and stabilisation of paper Contents
9.5 Natural and accelerated photoageing 142 9.6 Natural light and accelerated photoageing 144 9.7 Conclusions 147 9.8 References 148 Stabilisation Air pollution and its prevention 153 10.1 Introduction 153 10.2 Materials used in the study 154 10.3 Archival storage conditions and air purification 156 10.4 Study of pollutants 157 10.5 Evaluation of pollution 157 10.6 Study of stored materials 159 10.7 Conclusions 164 10.8 References 165 Stabilisation strategies 167 11.1 Introduction 167 11.2 Aqueous deacidification 167 11.3 Thermo-oxidative stabilisation study using model paper 172 11.4 Thermo-oxidative stabilisation studies using papers for common use 175 11.5 Stability of real-life paper samples during photodegradation 180 11.6 Conclusions 182 11.8 References 183 Outlook 185 12.1 Introduction 185 12.2 Analytical methodology 185 12.3 Degradation mechanisms 186 12.4 Experimental approach to ageing 186 12.5 Standards on paper ageing 187 12.6 Storage 187 12.7 Development of applications 188 12.8 References 188 Index 191
Ageing and stabilisation of paper vii Abbreviations
A absorbance AU arbitrary unit BHT 2,6-di-t-butyl-p-cresol (butylated hydroxytoluene) CL chemiluminescence, chemiluminometry DP degree of polymerisation DRIFT diffuse reflectance infrared Fourier-transform DSC differential scanning calorimetry DTA differential thermal analysis DTPA diethylenetriaminepentaacetatic acid Ea activation energy EDTA ethylenediaminetetraacetic acid eq equation viscosity intrinsic viscosity FTIR Fourier-transform infrared Mn number-average molecular weight Mw mass-average molecular weight n number of experiments PL50% time to 50% property loss SD standard deviation SEC size exclusion chromatography SEM-EDX scanning electron microscopy with energy-dispersive x-ray analysis Tg glass transition temperature UV-VIS ultraviolet-visible
Frequently used samples C cotton pulp sheets, Radeče, Slovenia, 2002, 60 g m-2, DP 6080 SA bleached sulphate pulp sheets, Pöls, Austria, 2002, 60 g m-2, DP 3090 WH Whatman filter paper No. 1, Maidstone, UK, 2002, DP 2900 S1 factual paper sample from "Vestnik slovenskega kemijskega društva", Slovensko kemijsko društvo, 1984, Ljubljana, 50% bleached sulphate hardwood pulp, 50% bleached sulphite softwood pulp, pH 5.7, DP 1600 S2 factual paper sample from "Bauffs Märchen", Leopold Bebhardt Verlag, 1870, Leipzig, 70% cotton, 30% wheat straw, pH 5.0, DP 550 S3 factual paper sample from "Lesen und Reden", Huber und Co. Verlag, 1922, Leipzig, 70% groundwood, 30% bleached sulphite or sulphate pulp, pH 4.3 S4 factual paper sample from “Slovstvena zgodovina Slovencev, Hrvatov in Srbov”, Jugoslovanska knjigarna, 1938, Ljubljana, 100% sulphite softwood pulp, pH 5.3, DP 520 S5 factual paper sample: office paper (Radeče, Slovenia, 2002), 70% bleached sulphate softwood, 30% bleached sulphate hardwood pulp, pH 9.9, CaCO3 18% m/m, DP 1750
viii Ageing and stabilisation of paper Introduction
Chapter 1 Paper and durability
Matija Strlič, Jana Kolar, Steph Scholten
1.1 Introduction The degradation of paper is an inevitable process. Like all materials, paper will eventually decompose, however slowly that process may seem to take place. It is a cause for great concern: up to the late 20th century, paper was -and often still is- the most important carrier of information about our cultures, sciences, business, politics and history. Many documents are so important that they have to be available to the posterity and without the relevant information contained in them, all achievements are forgotten. The information on paper that we may want to preserve should be regarded broadly: not only the written or printed information on paper in books and archives, but also the wealth of information in the form of music scores, drawings and graphic arts, photographs, also the type of paper, bookbinding, materials used etc. The information needs to be preserved in the original form for a core of valid reasons, national, political, legal, historical, economic, scientific, emotional. An example: the exhibition of Dürer’s famous works in the Albertina collection in Vienna, Austria, was visited by 427,000 visitors in three months in 2003.2 This is why we need to understand paper, develop ways of slowing down the degradation process and prolong its lifetime. At the same time, it would appear that paper’s apparent longevity prevents people from thinking far ahead. Indeed, how can we be sure of the rate of the process of paper degradation? If a medieval illuminated book has lasted for so many centuries, why should we be concerned at all? But people do not know, or tend to forget, how much historical material has already been lost and much of the existing material is threatened by decay. Some argue that in these modern times, we shouldn’t be concerned. And up to a point they are right: as a consumer material, most of the paper produced today is not meant for long- term use. Yet, two millennia after its invention, paper is still one of the most durable information carriers. In our era of electronic data processing, this is still true: the lifetime of electronic data on most contemporary devices for storage is several decades at most; shorter still is the lifetime of software needed to process it and read it. In the era of electronic data processing, the consumption of graphic paper has increased by more than 35% in the last decade.1
Ageing and stabilisation of paper 3 Chapter 1: Paper and durability
Fortunately, enormous economic efforts are occasionally made. Even if only its 15 stars are made of cellulose, the Star-Spangled Banner Preservation Project serves well to illustrate the point. The Smithsonian’s National Museum of American History’s project of conservation of the US flag, including research, education and a new display, was budgeted at US$ 18 million.3 Unfortunately, these initiatives are exceptions to the rule: there are but a few institutions which are able to afford sufficient funds for preservation. To be able to tackle the preservation of paper in the best and most economical way, research is of vital importance. First of all to understand the complex degradation processes of paper itself and secondly to develop and evaluate effective and efficient preservation methodologies. Optimisation of storage appears to be the least costly first step, of interest to practically every museum, library or archive, yet even in this area, few actions are taken globally. The first piece of knowledge that we need is about the nature of paper and the way paper was produced. When we speak about paper, it is important to understand that the term describes a material with strong similarities, but also very important differences which developed over time. The second piece of information is about the understanding of the degradation processes and factors that influence the lifetime of paper.
1.2 Paper and its production The process of paper production changed during the centuries in response to market needs, and so did the quality of paper.4 In medieval times, the production in mills depended almost entirely on linen, hemp and cotton rags so that the raw cellulosic material was of high quality. With the invention of printing with movable type in the 16th century, the demand for paper had increased tremendously. More than 10 million books, nowadays called incunabula, were produced in the first half of the century following the great invention. The effect this had on our civilisation was such that J. Gutenberg was voted the greatest inventor of the millennium.5 More efficient and productive pulp beating was ensured with the invention of "hollander" beaters made of iron, which introduced an increased amount of corrosive metal ions into the material. Soon there was a need to introduce other fibres besides cotton. Coloured textiles were used, but they had to be bleached first, and straw was also used as a low- quality substitute. Not until the second half of the 19th century did lower quality wood fibres (either as groundwood or chemical pulp) decisively enter the process. By that time the process of paper production was also fully mechanized. The paper of this era is generally of very poor quality. Since a sheet of cellulosic fibres is water absorbent, its surface is usually sized to facilitate writing. The early common types of sizing were flour and starch, which were added to the pulp, while gelatine sizing was introduced in the 13th century and was done on already formed sheets. Up to 20% of alum (potassium aluminium sulphate), was added to gelatine to reduce biodeterioration and to control its viscosity. It was occasionally replaced by zinc sulphate (white vitriol). In the presence of water, both give rise to acids, the pH of the sized surface thus decreased. Gelatine sizing did not change much until well into the 19th century, when rosin sizing was added to the pulp and precipitated with papermaker’s alum (aluminium sulphate), also an acidic substance. A low pH is generally an indication of poor(er) durability. The first half of the 20th century saw changes in the machinery, leading to an increased production. Later, new pulping processes were introduced, and the use of
4 Ageing and stabilisation of paper Chapter 1: Paper and durability deinked recovered paper also became widespread. New fillers, dyes, coatings, sizing appeared. With an increased concern for the environment, the use of neutral sizing became a norm only in the 1990s.
1.3 Paper degradation Any librarian or archivist is aware that most of the material, produced between 1850 and 1950 is now fragile. In fact, surveys of library collections have shown that an alarming proportion of items in libraries, 70-85%, are prone to rapid degradation.6 As expressed recently, we may be facing a catastrophic situation,7 as these objects are likely not to survive and remain readable in the 22nd century. But why would paper, produced in 1500, be more stable than the one produced four centuries later (Figure 1.1)?
Figure 1.1: Left: a well preserved page from Hartmann Schedel: “Liber chronicarum”, Nürnberg, 1493; right: a degraded leaflet from the 1900s. ©National and University Library, Ljubljana, Slovenia. Of all the causes for paper degradation, biological, physical and chemical, the latter are probably the most difficult to control. The type and quality of the raw material is crucial for the longevity of the product: already degraded recycled fibres will lead to a less durable paper than high-quality cotton fibres. The intermediate quality of wood fibres will depend on the type of the pulping process. Even if made of best quality fibres, the durability of cellulose will crucially depend on its acidity/alkalinity. Already in 1910, T. Edlund questioned the harmful effect of acidity in paper, then introduced through sizing.8 Due to the detrimental chemical process of acid hydrolysis, lifetime expectancy of acid papers can be compared with that of a human being. The rate of this degradation process depends crucially on paper pH, a measure of its acidity/alkalinity.
Ageing and stabilisation of paper 5 Chapter 1: Paper and durability
The high quality papers produced in the 15th century will easily last for the next millennium if properly cared for. If paper is neutral to moderately alkaline, its lifetime increases greatly. It will still degrade, but the process is slower: for these papers, oxidation is the predominant cause for concern.9 Proper care implies that we know the proper conditions of storage. We don’t or, to be more precise, we don’t know exactly. We know that higher temperature accelerates the process, but research indicates that it is not sure what relative humidity is optimal for long- term storage. The effect of light on cellulosic materials is also a controversial topic. Although there was much research dedicated to the effect of light on lignin-containing paper, the data on cellulosics are scarce, again for the apparent reason that cellulose was long thought to be relatively photo-stable. Fortunately, there are ways of reducing the acidity (increasing the pH) in paper. Already in 1936, O. Schierholz published a patent10 on “deacidification” with barium, strontium or calcium bicarbonate or hydroxide. The term deacidification implies that not only neutralisation is achieved, in fact, the pH of paper after deacidification is frequently higher than 7. The deposited carbonate in paper leads to an alkaline reserve, effectively preventing damage that future build-up or absorption of acids from the environment might exert upon paper. Nowadays, deacidification is performed as a basic paper conservation technique, either manually or as a mass treatment. In manual techniques, aqueous solutions of calcium or magnesium bicarbonate or of calcium hydroxide are most often used. Mass deacidification is carried out in batches with organic solvents and several procedures are available.11 In this book, we will focus on the aqueous procedures, although many findings are relevant irrespective of the way the alkalis enter the cellulosic network. While pH is undoubtedly crucial, there are a number of other parameters influencing paper durability. We may roughly divide them in endogenous and exogenous (Figure 1.2), i.e., those introduced during production or formed in the material during ageing, and those which affect the process either as environmental parameters or by deposition or absorption.
Figure 1.2: A simplified scheme of factors affecting paper stability. It may be an oversimplification, but in an experiment, scientists tend to isolate a certain parameter in order to learn more about its influence on the studied system. Therefore, we frequently speak about models and model samples. By the use of models we also like to
6 Ageing and stabilisation of paper Chapter 1: Paper and durability promote the idea of experimental repeatability so that the findings can be confirmed by others. While the selection of models is crucial, the findings have to be tested in practice, meaning on real historic papers. This is frequently difficult, given the stability of our material at room conditions. Fortunately, there are ways we can extrapolate the behaviour of some material at more favourable experimental conditions to conditions during use. Much of the scientific effort is dedicated to this aspect, a topic of great concern.12
OH OH H H H H H H OH OH O OH O OH O OH OH H H H H H H H OH O OH O H O O OH OH H H H H H H OH n OH Figure 1.3: The cellulose macromolecule. Cellulose (Figure 1.3) is the major structural component of paper. Its degradation will also lead to the degradation of mechanical paper properties, relevant to the user. Other components of paper also influence its stability, the major one being another natural polymer, lignin. Although of extreme interest, the influence of lignin is deliberately not a subject of this book. The reasons are of practical nature: the complex subject simply deserves a complex treatise of its own. Similarly, it is impossible to cover in one volume the compound effects of other important paper components, e.g. sizing or inks. Since cellulose is a macromolecule, many aspects of its degradation are common to degradation of other polymers. The methods of study are identical, and so are some of the degradation mechanisms. Here, we would like to stress the similarities and differences, but above all, underline the need for new approaches, new techniques of study, shedding new light on old problems. Fortunately, measuring the emission of light can be used as an experimental technique, i.e. chemiluminometry. As a technique, it started developing in the 1960s,13 and reached industrial labs in the past decade. It is a technique offering supreme sensitivity. However, complex data interpretation and the fact that many concerted chemical phenomena lead to light emission, obstruct its general implementation. Nevertheless, it offers complementary information to the more traditional techniques, especially in studies of oxidative degradation phenomena. Research on paper durability is typical applied research, and following a few intensive periods in the past, we seem again to be witnessing more exciting developments lately.14 The need for this line of research is demonstrated through a number of standards for permanent paper and through a number of standardized procedures for estimation of its stability. Furthermore, research on optimal storage and conservation is called for by the users and collection keepers in museums, libraries and archives. Much of the knowledge can be used in other areas of research and implemented in other industries of natural (e.g. textiles) and synthetic materials (e.g. polymers), but also food industry or biology. After all, degradation is a ubiquitous process.
1.4 References 1. http://www.paperonline.org, accessed 10/10/2004.
Ageing and stabilisation of paper 7 Chapter 1: Paper and durability
2. http://www.theartnewspaper.com, accessed 10/10/2004. 3. http://americanhistory.si.edu, accessed 10/10/2004. 4. D. Hunter, Papermaking: History and Technique of an Ancient Craft, Dover Publ., New York, 1987. 5. http://news.bbc.co.uk/hi/english/static/events/millennium/jan/winner.stm, 10/02/2005. 6. S. Buchanan, S. Coleman, Deterioration survey on the Stanford University Libraries Green Library stack collection, College and Research Libraries, 1987, 48, 102-147. 7. A. Barański, A. Konieczna-Molenda, J.M. Łagan, L.M. Proniewicz, Catastrophic room temperature degradation of cotton cellulose, Restaurator, 2003, 24, 36-45. 8. T. Edlund, Schwefelsäure im Papier, Papier Fabrik. 1910, 8, 765. 9. J. Kolar, Mechanism of Autoxidative Degradation of Cellulose, Restaurator, 1997, 18, 163-176. 10. U.S. Patent 2,033,452, Process for the Chemical Stabilization of Paper and Product, 1936. 11. A. Blüher, B. Vogelsanger, Mass deacidification of paper, Chimia, 2001, 55, 981-989. 12. H. Porck, Rate of paper degradation. The predictive value of artificial ageing tests, ECPA, Amsterdam, 2000. 13. M.P. Schard, C.A. Russell, Oxyluminescence of Polymers. I: General Behavior of Polymers, J. Appl. Polym. Sci., 1964, 8, 985-995. 14. M. Strlič, J. Kolar, Evaluating and enhancing paper stability - the needs and recent trends, Proc. 5th EC Conf. Cultural Herit. Res., 16-18 May 2002, Cracow, Poland, 79-86, http://www.heritage.xtd.pl/pdf/full_strlic.pdf.
8 Ageing and stabilisation of paper Chapter 2 Degradation and ageing of polymers
Jozef Rychlý, Matija Strlič
2.1 Introduction The term degradation of macromolecules denotes all processes which lead to a decline of polymer properties. It may eventually involve physical processes, such as polymer recrystallization, or denaturation of protein structures. Chemical processes related to degradation may lead to a reduction of average molar mass due to macromolecular chain bond scission or to an increase of molar mass due to cross-linking rendering the polymer insoluble. A particular group of degradation processes involves reactions where the average molar mass remains practically unchanged while dissociation of side groups or their modification takes place. The term ageing of polymers is usually associated with long-term changes of polymer properties under the conditions of weathering and may involve any of the above processes. In relation to degradation, the term corrosion is also encountered, which has been borrowed from metal chemistry. In certain cases, an advanced stage of polymer degradation may be seen by the naked eye, as it is accompanied by formation of cracks and a catastrophic reduction of mechanical properties (Figure 2.1). Most users understand polymer degradation in terms of reduction of average molar mass and we shall use it in that sense throughout the following text. This type of degradation represents a part of the global cycle of inorganic and organic processes and leads to regeneration of original natural resources. The large scale production of synthetic polymers that we are witnesses of and the increasing amounts of waste demand of us a proper understanding of the degradation processes so that they can be controlled and so that a positive impact on the environment can be achieved. This chapter will review some of the basic chemical aspects of polymer degradation, common to many polymeric materials, also to cellulose and paper. Due to the abundance of available general literature,1-5 we only attempt to provide a suitable basis for a more easy understanding of the chapters to follow.
Ageing and stabilisation of paper 9 Chapter 2: Degradation and ageing of polymers
Figure 2.1: A foil of low density polyethylene at an advanced stage of degradation.
2.2 Modes of degradation The mode of degradation is determined by the character of its initiation. Accordingly, we thus differentiate photo-degradation (photolysis), thermal degradation (thermolysis), photooxidation and thermo-oxidative degradation (which involve the former two degradation processes with the assistance of oxygen), radiolysis, mechano-chemical degradation etc. Polymers may further degrade under the effect of ozone, peroxides, acids and alkalis, halogens or other aggressive compounds, under the effect of electric field, plasma and corona discharge, ultrasound, laser irradiation. A particular field of polymer degradation studies is high temperature degradation, which occurs along with polymer burning and carbonization and leads to destruction of the material.
2.3 Thermolysis Chemical species which initiate the degradation process may have the character of ions, free radicals or excited states of molecules. Their formation in the polymer depends on the number of weak sites in the macromolecular structure, which are integrated into the polymer during its synthesis and production. They may be related to impurities or residues of catalysts which remain in the polymer after its synthesis. The inherently present and potential weak sites in a polymer structure are head-to-head or tail-to-tail linkages of monomer units in a macromolecular chain, which are thermodynamically less stable than head-to-tail linkages; terminal groups which are usually more reactive than other units in the chain; and, finally, O-O bonds which may be introduced into the polymer during polymerisation of monomers not sufficiently deprived of oxygen, or during polymer storage. The head-to-head structural irregularity, however, need not necessarily be the source of lower stability of polymer product. E.g., for polystyrene, poly(vinyl cyclohexane) and poly(methyl methacrylate) both types of structural units, i.e. head-to-head and head-to-tail show comparable thermal stability.4 The strength of bonds which may form the macromolecular backbone decreases in the following sequence: C=N > C=C > P=N > N=N > Si-O > C-O > C-C > C-N > C-Si > N-O > O-O, as in Table 2.1. The scission of macromolecular backbone may take place as a secondary
10 Ageing and stabilisation of paper Chapter 2: Degradation and ageing of polymers process following the rupture of some bond in the side group; the order of thermal stability of such bonds is as follows: Si-F > O-H > C-F > C-H > Si-Cl > C-Cl > Si-Br > C-Br > C-S > O-O. The dissociation energy of bonds forming the structure of a macromolecule is thus a first criterion in estimation of thermal stability of a given polymer. The fraction of bonds with energy equal to dissociation energy D is determined by the Boltzmann’s factor, exp(-D/RT), where T is absolute temperature and R is the universal gas constant. We can thus easily calculate that the temperature, at which in one mole of C-C bonds at least one is dissociated into radicals, is 486 °C while in one mole of O-O bonds this temperature equals 30 °C, only. This explains why in inert atmosphere at temperatures above 350 °C, most hydrocarbon polymers decompose into volatile products more or less rapidly. At lower temperatures, defects in the regular structure of macromolecular chains are obviously the initiating sites of decomposition.
Table 2.1: Dissociation energies of selected bonds in kJ mol-1 potentially forming a polymer backbone structure.
A-BH C NOF Cl Br I S Si H (436) 413 391 463 563 432 366 299 339 C 348 292 352 441 329 276 240 259 290 N 160 222 270 200 O 139 185 203 369 F (153) 254 Cl (243) 250 Br (219) 193 178 I (151) S 213 Si 541 359 289 A=B C 615 N 615 418 O 716 607 498 A≡B C 812 N 891 946
The thermal stability of a given polymer is usually determined by following the rate of degradation using thermogravimetry, or by following the formation of unsaturations and carbonaceous residue in the case of polymers which degrade predominantly by cross- linking. From the temperature dependence of rate constants, the activation energy Ea and the pre-exponential factor of the degradation process are calculated in order to perform
Ageing and stabilisation of paper 11 Chapter 2: Degradation and ageing of polymers extrapolation of rates to another temperature region. A relation quite useful for estimation of the activation energy Ea of degradation is based on the comparison of dissociation energies D of the newly formed (j) and disappearing (i) bonds so that:
DDE jia , (eq 2.1) where is an empirical coefficient with the value 0.5-1, depending on the mechanism of the degradation process. Qualitatively, the more stable the degradation products are, the lower is the activation energy of the process and the higher is its rate.
Scheme 2.1: An exemplary degradation pathway of a linear polymer. As an example of a possible degradation pathway, we can consider the degradation of a linear macromolecular chain in which a reaction centre () is formed through some initiating event (heat, light, oxygen, shear stress etc.). The complementary reaction site is denoted by (-). E.g., when () is a free radical site, (-) is also free radical site, if () is cation, (-) is anion etc. (Scheme 2.1). The dissociation energies of bonds situated in a -position to the reaction site () are considerably lower than in the corresponding saturated compounds and the subsequent fragmentation occurs predominantly there. If () is a free radical site, the following dissociation energies have been found for the C-C bond in butane and 2-butyl radical:
-1 CH3CH2―CH2CH3; D = 348 kJ mol (saturated compound),
-1 CH3C H―CH2CH3; D = 105 kJ mol (free radical). Such -fragmentation propagates the reaction site either through scission of the main chain or through splitting off of side groups. The resulting degradation pattern is an unzipping process (peeling off) with formation of either an unsaturated low molar mass compound (e.g. degradation of poly(methyl methacrylate) to methyl methacrylate) or polyconjugated unsaturation on the main chain (degradation poly(vinyl chloride) to HCl and polyconjugated system of double bonds). Another related degradation process is that involved in degradation of cellulose acetate, in which the splitting off of the side groups and production of acetic acid is often referred to as the vinegar syndrome (Figure 2.2).
12 Ageing and stabilisation of paper Chapter 2: Degradation and ageing of polymers
Figure 2.2: The vinegar syndrome – degradation of cellulose acetate accompanied by production of acetic acid.
2.4 Oxidation Regardless of the mode of initiation, a very efficient co-factor in the reduction of polymer stability is the presence of oxygen. It is related to a quite fast introduction of the labile O-O bond into the polymer structure and development of a free radical degradation mechanism according to Scheme 2.2 which was proposed by Bolland and Gee for autoxidation of rubber compounds. Free radicals (P) generated in the initiation process by the effect of heat, light, oxygen or by other initiating routes are converted to peroxyl radicals (PO2 ) in the presence of oxygen, and subsequently to hydroperoxides – intermediate compounds which lead to a chain reaction unless stabilizers (InH or D) are used to interrupt it. Thermal decomposition of dialkyl peroxides, diacyl peroxides, hydroperoxides and peracids usually occurs at a measurable rate above 60 °C, depending on the structure. Diacyl peroxides and peracids are considerably less stable than dialkyl peroxides and hydroperoxides. Some traces of transition metals and metal ions may catalyse the decomposition of hydroperoxides even at room temperature. Traces of metal ions are present in virtually each polymer and may affect polymer oxidation and subsequent degradation considerably. The sequence of efficiency of metal ions crucially depends on the oxidation state and ligand type. In an aqueous reaction system at pH 7, the following order of catalytic activity was found: Cu(II) > Cr(III) > Co(II) > Fe(III) > Mn(II) > Ni(II), while Cd(II) and Zn(II) did not exhibit any catalytic activity.6 The activity is related to the ability of the transition metal to undergo the pseudo-Haber-Weiss cycle of hydroperoxide decomposition: ROOH + Men+→→ RO + Me(n+1)+ + HO–,
(n+1)+ n+ + Me + ROOH → RO2 + Me + H , in which R can either be a hydrogen, a low molecular weight compound or a polymer. In
Ageing and stabilisation of paper 13 Chapter 2: Degradation and ageing of polymers the second reaction, the transition metal can be reduced back into its catalytically active state by some other reducing compound, e.g. superoxide anion or an appropriate organic radical. The above scheme is simplified and the mechanism for a particular metal ion is more complex and may involve e.g. reaction of the lower oxidation state of metal ion with peroxyl radicals. Ions of Al, Ti, Zn and V usually reduce the rate of oxidation.
wi productsPPH (1) initiation
k3 (2) oxygen absorption POOP 22 k 4 PPOOHPHPO (3) DSC, DTA, determination of 2 hydroperoxides kmono HOPOPOOH (4)
kbi (5) DSC, DTA, chemiluminometry 22 OHPOPO2POOH k 4a PPOHPHPO (6) determination of hydroxyl groups k fragm productsPPO (7) embrittlement due to molar mass reduction
k61 (8) 22 productsO2PO
k62 (9) chemiluminometry O2PO2PO 22
k7 (10) 2 POOPPPO k8 P-P2P (11) cross-linking kInH1 productsInHP (12) inhibitor consumption
kInH 2 (13) inhibitor consumption 2 productsInHPO kInH3 productsDPOOH (14) inhibitor consumption
Scheme 2.2: The general autoxidation scheme for oxidative degradation of hydrocarbons (PH); including inhibitors (InH) and peroxide decomposers (D). Some possibilities for following the consequences of an individual step are given. An increase in electronegativity of heteroatoms in the polymer main chain or in its side groups also leads to an increase of polarity of the respective bonds. In such cases, ionic degradation mechanisms start to predominate, e.g. in polymers with ester, ether and acetal bonds, C-O and P-O bonds in biopolymers, or Si-O bonds in polysiloxanes. An ionic course of degradation is less probable for amide or imide bonds, C-N and C-halogen bonds.7
2.5 Light-induced degradation Absorption of light by a macromolecule involves a specific interaction of some functional group (chromophore) with a photon. Bonds such as C-C, C-H, O-H and C-Cl absorb wavelengths shorter than 200 nm and polymers composed of such bonds are in principle photolytically stable. In practical terms, however, almost all polymers are sensitive to radiation at longer wavelengths, including sunlight ( > 300 nm). This is due to
14 Ageing and stabilisation of paper Chapter 2: Degradation and ageing of polymers unsaturated groups, e.g. C=O (carbonyl groups), and C=C, which may be present in the polymer structure as a result of synthesis, processing and thermal degradation of the material. These functional groups have absorption maxima between 200-400 nm. Photolysis is also enhanced by additives and impurities which absorb light more efficiently than the polymer itself. The effect of respective chromophores depends on the amount of energy absorbed which is determined by the concentration of chromophores, by its absorption spectrum and by the yield and reactivity of radicals which are subsequently formed. As the concentration and the nature of chromophores may change with time, polymer photoreactivity is determined by a rather broad range of quantum yields, i.e. number of chemical reactions brought about by one absorbed photon (Table 2.2).
Table 2.2: Quantum yields, of chain scissions at irradiation by light of = 254 nm. Polymer Poly(methyl isopropyl ketone) 0.22 Poly(methyl acrylate) 0.013 Poly(methyl methacrylate) 1.7·10-2 – 3·10-2 Poly(methyl vinyl ketone) 2·10-2 – 2.5·10-2 Cellulose nitrate 1·10-2 – 2·10-2 Cellulose 0.7·10-3 – 1·10-3 Polyamide-6 5·10-3 Poly(-methyl styrene) 1·10-3 Cellulose acetate 2·10-4 Polystyrene 9·10-5
The quantum yield for bond scission depends also on segmental mobility of the macromolecule, chromophores in side groups being more reactive than those in the main chain. Rather significant changes of quantum yield occur at the glass transition temperature
(Tg). E.g., for copolymer of methyl methacrylate and methyl vinyl ketone, the quantum yield below Tg increases with temperature only slightly, from 0.04 to 0.08, at Tg (100 °C) it 4 changes to 0.28 and above Tg, it reaches 0.30 which is typical for copolymer solutions.
h MM *11 MM *21 eh 1*1 heatMM *3*1 heatMM 1*1 MM h 1*3 MM h Scheme 2.3: Possible reactions of chromophore M following absorption of a photon. Superscripts 1, 2 and 3 denote singlet, doublet or triplet states of the chromophore which are characterized by 2 electrons with antiparallel spins, by one electron, and by 2 electrons with parallel spins, respectively; the asterisk denotes an electronically excited state.
Ageing and stabilisation of paper 15 Chapter 2: Degradation and ageing of polymers
The effect of absorption of a photon in a chromophore (M) is demonstrated in Scheme 2.3. Initially, the ground singlet state is either converted to an excited singlet state or undergoes photoionization. An excited singlet state may convert to an excited triplet state by non-radiative inter-system crossing or a ground singlet state. The conversion to a ground singlet state may alternatively proceed through emission of light as fluorescence which occurs from a singlet state or phosphorescence which occurs from excited triplet state. The distribution of populated excited states has a statistical character. The presence of a suitable sensitizer (F) may increase the ratio of radiative to non-radiative conversion while a quencher (Q) will decrease it. Sensitizers or quenchers may prolong or shorten the lifetime of excited molecules in the system and thus in enhancement or suppression of the effect of photolysis. Following light absorption, chemical transformation of macromolecules usually occurs via molecular triplet states with relatively long lifetimes. In polyolefins, carbonyl groups are the most frequent chromophores and their excited triplet states decompose by -scission into two radicals, following the Norrish type I mechanism:
or by -scission to methyl alkyl ketone and alkene which is preceded by 1,4 abstraction of hydrogen of alkoxyl radical moiety from excited triplet state of the carbonyl group, i.e. Norrish type II mechanism:
In the presence of oxygen, alkyl or carbonyl radicals lead to formation of hydroperoxides (POOH) or peroxyacids, which on irradiation at ~ = 300 nm decompose either directly to free radicals:
* ,HOPOPOOH or, after energy transfer from an excited carbonyl group, M*:
** .POOHMPOOHM The latter energy transfer in photooxidation is important only if the content of hydroperoxides exceeds 10-2 mol kg-1 and may occur e.g. in localised micro-domains with an increased content of oxidation products.
2.6 Propagation and advanced degradation phases According to the propagation step, degradation reactions may be divided into statistical degradation, depolymerisation and unzipping processes. Statistical degradation is inherent to polymers which have no bulky substituents on the main chain, e.g. polyethylene, polypropylene, etc. Poly(methyl methacrylate) degrades predominantly by depolyme- risation reaction and the original monomer is formed. Such mechanism usually occurs in polymers having either relatively bulky or polar substituents on the main chain, also in poly(tetrafluorethylene), i.e. Teflon®. A particular case of degradation is unzipping where small molecules are released from the polymer backbone in a way which resembles the opening of a zipper. Poly(vinyl chloride),
16 Ageing and stabilisation of paper Chapter 2: Degradation and ageing of polymers
PVC, or poly(vinyl acetate) are typical representatives of such polymers, in which conjugated double bonds spread along the main macromolecular chain, the polymers turn yellow, then brown and finally black. The future of already degraded figurines in Figure 2.3 is apparently not very bright without an immediate conservation treatment. PVC degradation and stabilisation has been a well-discussed topic in the literature for the past thirty years. Its relatively low thermal stability is the consequence of a cooperative action of several factors such as the presence of some metal cations (Fe, Cu, Co, Zn, Al), – 2– certain anions (Cl ,B4O7 ), residual polymerisation initiators, alkaline compounds and double C=C bonds in the main polymer chain. Depending on the way of activation of the C-Cl bond, the above factors may induce either a free radical or an ionic pathway of dehydrochlorination of PVC. The initially formed double bonds promote subsequent propagation of dehydrochlorination with HCl having an autocatalytic effect on the further course of degradation. The longest polyene sequences which are formed by such unzipping mechanism involve 25-30 double bonds, with the average value being around 3 to 15 double bonds. The coloration of the degraded PVC is not only due to polyene sequences, but also due to the presence of coloured carbonium ions. This may be evidenced by the addition of NH3 which discolourates degraded PVC. The sequence of above reactions which start from some initiating sites and propagate throughout the material more or less slowly, occurs at a micro-level and changes in material properties may be controlled on the basis of changes of mechanical properties such as elongation and tensile strength, number of double folds in the case of paper, etc. Very useful analytical tools providing information on whether a material is reaching its limiting stability are spectroscopic techniques, with which we can analyze the content of carbonyl and hydroxyl groups. A decrease or increase of the average molar mass (degree of polymerisation) is also important; however, its relation to mechanical properties is usually not a continuous function.
Figure 2.3: Aged PVC. A particularly disturbing phenomenon during ageing of polymeric materials is formation of cracks (Figure 2.4). The mechanism explaining how macrostructural deformations appear from microstructural changes in the material is very complex and involves both elementary degradation processes connected with a release of mechanical tensions due to chain disruptions and with a heterogeneous character of the degradation process which is
Ageing and stabilisation of paper 17 Chapter 2: Degradation and ageing of polymers focused on micro-regions. These phenomena may be enhanced by crystallization of low molar mass products, structural changes of macromolecules from a coiled structure to more rigid structures.
Figure 2.4: The surface of an aged rubber product with well visible cracks. However, it is generally agreed that the best way of avoiding such a combination of degradation chemistry and spontaneously spreading stress in a material, is to inhibit the responsible elementary reactions. 2.7 General stabilisation strategies The prolongation of residual lifetime of a given polymer is of significant practical importance. This goal can be achieved by modifying the material in such a way that the weakest structural sites are eliminated; or by providing additives, which protect it through suppression of a degradation pathway in which active sites are generated. Such stabilizers absorb the energy of an initiation process more efficiently than the polymer itself. Once formed, the primary sites of degradation may be deactivated either physically (through quenching of excited states) or chemically (through inhibition of a chemical reaction which leads to propagation of oxidation). Additionally, the species which are formed from stabilizers should be less reactive than the species formed during degradation of the non-protected polymer itself. According to the predominating function of stabilizing additives we distinguish thermo- and photooxidation stabilizers, antioxidants, antiozonants, flame retardants, antirads (radiation protectants) etc. In a mixture of stabilizers, the result may be either synergistic, if the positive effect of the mixture is higher than a simple arithmetic sum of individual contributions; or antagonistic if the mixture is less efficient than the sum. When estimating the remaining service life of a polymer material for a particular application, the limiting value of some material property such as tensile strength, elongation at break, electrical conductivity, permeability to low-molar mass compounds, average degree of polymerisation or similar, should be established, at which the polymer is still usable. Stabilizers, which degrade through a free radical mechanism, usually scavenge free radicals or hydroperoxides from the system. Such additives take part in reactions (12)-(14) of
18 Ageing and stabilisation of paper Chapter 2: Degradation and ageing of polymers
Scheme 2.2 and by doing so, they suppress propagation of free radicals through reactions (3)-(5). If reactions (13) or (14) are predominant, we distinguish inhibition antioxidants or chain breaking antioxidants. They intervene directly with the autoxidation cycle by rapidly reacting with alkyl or peroxyl radicals. On the other hand, preventive antioxidants decrease the rate of formation of free radicals either by non-radical decomposition of hydroperoxides (reaction 15) or by deactivation of metals or other types of catalysts. The most widely used antioxidants are sterically hindered phenols and bis-phenols; other additives are mostly used in combination with phenols in synergistic mixtures. In Table 2.3, typical compounds for stabilisation of hydrocarbon polymers against thermal oxidation are represented. Hindered phenols are remarkably efficient in protecting saturated hydrocarbons against thermal oxidation. If present in a molar fraction in the range of 100-1000 ppm, we can increase the induction time by one order of magnitude. The mechanisms of stabilisation and structure-property relationships are generally well understood although quantitative (kinetic) evaluations of efficiency are rare and often based on questionable simplifications and false hypotheses, such as predominance of radical scavenging in termination reactions, stationary state assumption, or constancy of initiation rate (which is either explicitly assumed or can be derived from the assumption on steady state).
Table 2.3: Some typical antioxidants.
Several mechanisms of synergism for two different antioxidants were assumed until now, namely: 1. An interaction of both additives and formation of a new component which acts as more efficient stabilizer. Example: aromatic amines and mercaptobenzimidazol. 2. A redox mechanism. Oxidized form of a more efficient component of synergic mixture is reduced by a less reactive component. Example: aromatic amines and phenols. 3. Conjugated effect of chain breaking antioxidants and chelating compounds reducing the rate of initiation by inactivation of transition metal ions.
Ageing and stabilisation of paper 19 Chapter 2: Degradation and ageing of polymers
4. Interaction of scavengers of reactive free radicals and compounds which inhibit the transfer reaction of radicals with substrate. 5. Synergism between chain breaking antioxidants (e.g. sterically hindered phenols) and hydroperoxide decomposers (e.g. organic sulphides or phosphites). The above mechanisms may occur in parallel in a given system, e.g. 1 and 5 etc. Currently, the search for synergistic mixtures of stabilizers represents the most promising route for increasing the stability of a given polymer system. An example is shown in Figure 2.5 where the above described mechanism 5 was systematically investigated.8 A several times longer induction time in the presence of a mixture of a chain breaking antioxidant and peroxide decomposer in a molar ratio 1:1 illustrates the potential this approach. At the same time, one can see that in order to achieve a significant synergistic effect, purity of the original polymer (expressed by a low value of defect sites responsible for primary) is very important. Line 2 in Figure 2.5 shows that the effect of a synergistic mixture of two antioxidants on thermo-oxidative stability is negligible if the extent of initiation reactions independent of hydroperoxides is significant.
[D] (mol L-1) 0 0.010 0.008 0.006 0.004 0.002 0.000
20000
16000 ) s (
e 1 12000 m i t
n o i t c
u 8000 d
n 2 i
4000
0 0.000 0.002 0.004 0.006 0.008 0.010 [InH] (mol L-1) 0
Figure 2.5: Theoretical plot (curve 1) of induction time determined for wi = 0, (zero rate of initiation according to reaction (1) in Scheme 2.2) on the composition of mixture of inhibitors InH (chain breaking antioxidant) and D (peroxide decomposer) in a summary concentration of 0.01 mol L-1. Curve 2 represents induction times as above, but assuming wi = 5·10-8 mol L-1 s-1. Additives which act as light filters, light absorbers and quenchers of excited states are photo-stabilizers. An efficient stabilisation against the effect of light requires a prompt elimination of free radicals which appear during polymer photolysis. Light filters form a barrier between the light source and a polymer molecule. They may be used as surface coatings or as pigments in the polymer bulk. Their presence may also affect the rate of diffusion of oxygen through the polymer. Light absorbers absorb and dissipate the energy of ultra-violet light; their effect resembles that of specific filters for short range wavelengths. Quenchers of excited states deactivate excited states formed in the polymer system and thus inhibit scission reactions of macromolecules. The choice of a suitable photostabilizer for a given polymer is not simple at all. The most
20 Ageing and stabilisation of paper Chapter 2: Degradation and ageing of polymers frequently used are derivatives of 2-hydroxy-4-alkoxybenzophenone, benzotriazoles, salicylates and nickel(II) chelates, carbon black, etc. The stabilizer should exhibit high resistance towards light; such is the case of aromatic compounds having a hydroxyl group in position 2 with respect to a chromophore which forms an intramolecular hydrogen bond:
. During photoexcitation of 2-hydroxybenzophenone, a proton from the hydroxyl group to the oxygen atom of the carbonyl group is transferred and a zwitterion is formed and after relaxation, the original structure is restored. In the sequence of reactions, the absorbed quantum of light is transformed to heat which is harmless under given conditions. In non- substituted benzophenone, where such a sequence of reactions cannot occur, excited singlet and subsequently triplet states of the carbonyl group are formed, which promote oxidative reaction. Benzophenone is thus a photosensitizer while 2-hydroxybenzophenone acts as a photostabilizer. Carbon black has a combined antioxidant and light-filtering effect and its efficiency depends significantly on particle size and dispersion in the polymer medium. The photo- stabilizing efficiency of other pigments is usually considerably lower. Again, mixtures of components exhibiting a synergistic effect are particularly effective. The photostabilizing effect of sterically hindered amines (HALS) exhibited in polyalkenes and polydienes is still a puzzle. It has been shown that the efficiency of HALS cannot be interpreted in terms of quenching or absorption but by some chemical process of scavenging peroxyl or alkyl radicals. In this overall reaction, semi-stable nitroxyl radicals are regenerated periodically, e.g. through reaction with peroxyl radicals:
It was demonstrated that HALS associate with hydroperoxides and are present at the site of oxidative attack. The complementary stabilizing effect may be a result of the reaction of nitroxyl radicals with metal ions in a lower oxidation state: 32 ,FeNOHFeNO which inhibits decomposition of hydroperoxides in the following reaction: 32 ,HOPOFePOOHFe in which the extremely reactive alkoxyl radicals are formed.
2.8 Conclusions Research into degradation and ageing of polymers is extremely intensive and new materials are being synthesized with a pre-programmed lifetime. New stabilizers are becoming
Ageing and stabilisation of paper 21 Chapter 2: Degradation and ageing of polymers commercially available although their modes of action are sometimes not thoroughly elucidated. They target the many possible ways of polymer degradation: thermolysis, thermo-oxidation, photolysis, photooxidation, radiolysis etc. With the goal to increase lifetime of a particular polymeric material, two aspects of degradation are of particular importance: ― Storage conditions, and ― Addition of appropriate stabilizers. A profound knowledge of degradation mechanisms is needed to achieve the goal. 2.9 References 1. L. Reich, S.S. Stivala, Autoxidation of Hydrocarbons and Polyolefins; Kinetics and Mechanisms, M. Dekker, New York, 1969. 2. B. Rånby, J.F. Rabek, Photodegradation, Photo-Oxidation and Photostabilization of Polymers, Wiley, Chichester, 1975. 3. H.H.G. Jellinek, Aspects of Degradation and Stabilization of Polymers, Elsevier, Amsterdam, 1978. 4. M. Lazár, T. Bleha, J. Rychlý, Chemical Reactions of Natural and Synthetic Polymers, Ellis Horwood, Chichester, 1989. 5. G. Scott, Atmospheric Oxidation and Antioxidants, Vol. I, Elsevier, Amsterdam, 1993. 6. M. Strlič, J. Kolar, V.-S. Šelih, D. Kočar, B. Pihlar, A Comparative Study of Several Transition Metals in Fenton-Like Reaction Systems at Circum-Neutral pH, Acta Chim. Slov., 2003, 50, 619-632. 7. G.E. Zaikov, Russ. Chem. Rev., 1975, 44, 833-847. 8. J. Verdu, J. Rychlý, L. Audouin, Synergism between polymer antioxidants - kinetic modelling, Polym. Degrad. Stab., 2003, 79, 503-509.
22 Ageing and stabilisation of paper Methodology
Chapter 3 Methodology and analytical techniques in paper stability studies
Matija Strlič, Jana Kolar, Boris Pihlar
3.1 Introduction In line with the complexity of the material, the techniques used for paper characterization are numerous. Rather than to provide a review of literature we would like to focus on the variety of possibilities so that the choice of a relevant analytical technique in studies of paper degradation is easier. There are many general and readily available works where the interested reader can obtain more detailed information on a particular procedure. Especially the classical methods of analysis, such as determination of functional groups, have been used almost unchanged for decades.1,2
3.2 Mechanical properties The usual use of paper includes its handling, and retention of mechanical properties should be the principal focus of all stability studies. The popular folding test, used during collection surveys,3 is, after all, a useful simplification of a similar automated method. The review of mechanical permanence by Gurnagul et al. is an authoritative reference work.4 The mechanical properties of most papers are governed by its major structural component, cellulose (Figure 1.3).5 Cellulose macromolecules are arranged into elementary fibrils, and these are arranged into micro and macrofibrils.6 Macrofibrils form cell walls, from which pulp fibres are consisted. Although the structure of a fibre is complex, its mechanical properties, and thus the mechanical properties of a formed sheet, are determined by both cellulose macromolecular chain length and intermolecular and interfibre bonding.6 In a study by Zou et al.,7 it is however indicated, that the cause of paper strength loss is predominantly fibre strength and not bond strength loss (Figure 3.1). This is convenient, as for determination of mechanical properties, large samples are frequently needed, and while the measurements exhibit a pronounced data scatter, many repetitions are required. Basing
Ageing and stabilisation of paper 25 Chapter 3: Methodology and analytical techniques in paper stability studies on the correlation between mechanical properties and degree of polymerisation (number of monomers in the macromolecular chain), degradation of material mechanical properties may be followed through determination of cellulose molecular weight. Thus, the amount of sample needed may be minute allowing us to take samples even on originals,8 and the repeatability of the measurements is improved. It should be noted that for determination of mechanical properties, standard atmospheric conditions are needed, in which the samples are properly conditioned.10 Most of the methods are also standardized: determination of tensile strength properties,11 tearing resistance,12 bursting strength,13 and folding endurance,14 as even if they are methodologically non-demanding, the results and repeatability depend on a variety of parameters, including the type of instrument.15 In certain cases, e.g. lignin-containing paper, where cellulose dissolution can be achieved using a pre-delignification procedure,16 determination of mechanical properties is probably the preferred method, e.g. for newspaper.17,18
Figure 3.1: Dependence of fold endurance on degree of polymerisation of cellulose for a bleached sulphite pulp sample. Reproduced from Zou et al.,9 with kind permission of Springer Science and Business Media.
3.3 Paper pH Determination of pH probably remains the most widespread chemical test in studies of paper stability. This is understandable as paper acidity/alkalinity is generally regarded as one of the crucial parameters regulating the mechanisms and rates of degradation. Although the measurement procedures are mostly simple, the evaluation of results is less straightforward and in most methods the analyst has to take into account a variety of systematic errors,19 in order that the measurements are comparable and not particular to a specific measurement system. Many methods are available and standardized, e.g. cold extraction,20,21 and determination of surface pH.22 Most extraction methods will give well reproducible results; however, a large amount of sample is usually needed, up to 1 g. While determination of surface pH is a popular, frequently used method, local wetting of the surface may lead to “tidelines”, along which degradation may proceed more rapidly.23 Thus, even determination of surface pH may not qualify as a non-destructive method. Besides, in most gelatine surface sized papers, the surface will be more acidic than the bulk
26 Ageing and stabilisation of paper Chapter 3: Methodology and analytical techniques in paper stability studies
(up to a few pH units), so that on the basis of a surface pH measurement, a wrong decision regarding the need for deacidification can be taken.19 Also frequently used, pH pens contain mixtures of acid/base colour indicators (cf. page 85). For acidic papers, measurement repeatability may be particularly disturbing, as evaluation of the indicator colour on a degraded yellowish paper is difficult. New pens are being developed, which contain one indicator only, for a more easy and straightforward decision-making.24 Micro-sampling in combination with micro-combined glass electrodes may be a good alternative, as samples of less than 1 mm diameter are sufficient,25 although sample inhomogeneities lead to low repeatability.19 Determination of pH of alkaline papers should take into account the role of atmospheric CO2, which is present in the atmosphere thus influencing paper pH during natural ageing. A new procedure was recently developed.19 In Table 3.1, the analytical parameters of the most frequently used procedures are summarized.
Table 3.1: A summary of the most frequently used methods of pH determination.
Method Sampling Remarks Literature Standard cold extraction Sample size: 1 g, Well repeatable but 20,21 destructive does not take CO2 into account Surface pH determination No sampling Well repeatable but 19,22 does not take CO2 into account, surface may have different pH than the bulk, systematic errors have to be evaluated before use Micro-pH determination Sample size: 10-50 g, Low repeatability 19,25 “micro-drilling” sampling procedure Determination using pH pens No sampling Low repeatability, 19 errors possible due to difficulties in estimation of colour hue Cold extraction with Sample size : 70 mg, Well repeatable, takes 19 CO2-equilibration destructive CO2 into account
3.4 Determination of alkaline reserve In standards for permanent paper,26 alkaline reserve plays an important role. It prevents acidification of paper due to absorption of acidic gasses from the atmosphere,27 and neutralises acids formed during material degradation. The procedure of determination is based on titrimetry and is also standardized,28 and standard laboratory equipment for titration is needed. The content is expressed as alkalinity in %CaCO3 (m/m) or as alkaline reserve in mol kg-1.
Ageing and stabilisation of paper 27 Chapter 3: Methodology and analytical techniques in paper stability studies
According to the standard, 1 g of paper is needed, yet this quantity may easily be downsized to 0.1 g. The repeatability is in the range of 5%. The titration procedure can also be automated. Spectroscopic methods are also used, e.g. diffuse reflectance FTIR spectroscopy,29 although with limitations in sample composition, especially regarding non-cellulosic components. SEM-EDX was used to check the homogeneity of alkaline deposit,30 where it should be taken into account that the content and distribution of Ca2+ do not necessarily correspond to that of CaCO3. Spectroscopic methods may provide valuable data on spatial distribution of the alkaline reserve.
3.5 Viscometry Determination of average molecular weight (or degree of polymerisation) is another frequently used analytical technique in paper degradation studies. Of the variety of available techniques for characterisation of polymer properties,31 two are frequently used, and in both cases the material needs to be sampled and dissolved. As it is a standardized procedure,32,33 viscometry is a robust, rapid, inexpensive and efficient technique. The sample is defibrillated in water and dissolved in a 0.5 mol L-1 2+ solution of [Cu(H2N-CH2-CH2-NH2)2] , i.e. cupriethylenediamine. Immediately after dissolution, which usually proceeds rapidly, measurements can be performed using a standard viscometer (Figure 3.2). Viscosity of solutions is critically dependent on temperature, so that a thermostated water bath at 25.0 ± 0.1 °C is also needed. The solvent is commercially available, although in- house preparation is easy and cost-effective.32 From the ratio t , (eq 3.1) t 00 where t is efflux time of a cellulose solution, t0 is the efflux time of the solvent, viscosity 34 ratio, i.e. /0, is calculated. It enters the Wetzel-Elliot-Martin’s equation, from which the intrinsic viscosity, [], is obtained. The relation between intrinsic viscosity of a polymer solution and average molecular weight of the polymer is described by the Mark-Houwink- Sakurada equation. If the average degree of polymerisation is required, the Evans and Wallis equation is equivalent:35
0.85 1.1DP . (eq 3.2) Viscometry only provides realistic data if the sample is reasonably alkali-resistant, otherwise, a pre-reduction treatment is needed.36 Furthermore, if cross-linking is one of the more prominent reaction pathways during degradation, as in laser cleaning,37 then viscometry will yield data burdened with a systematic error. In such cases, we may have to resort to size exclusion chromatography. Typical sample size for viscometric studies is 20 mg, time of sample preparation: 10 min, time of analysis: 5 min, typical repeatability better than 1%. Determination of moisture38 and dry matter in a sample may be necessary prior to analysis.
28 Ageing and stabilisation of paper Chapter 3: Methodology and analytical techniques in paper stability studies
Figure 3.2: The standard viscometer.32
3.6 Size exclusion chromatography Although chromatographic methods of analysis are extremely popular, size exclusion chromatography (SEC) is undoubtedly among the more complex. After dissolution, a sample is injected onto an appropriate chromatographic column, where it is separated according to hydrodynamic volume, which is roughly equivalent to the size of a molecule in solution. Bigger molecules will elute first, as they are not retained in the stationary phase pores. The columns usually need to be calibrated against a reliable set of macromolecular standards of known molecular mass, unless an absolute detector is available (e.g. laser light scattering detector39). Several solvents are available for this technique, while cupriethylenediamine cannot be used. The technique using LiCl/N,N-dimethylacetamide as a solvent was reviewed recently.40 Derivatisation and subsequent dissolution in tetrahydrofuran is another option, which was also downsized with sample requirement of a few micrograms only.8 Other components of paper, e.g. gelatine, can also be analysed using SEC.41 Size exclusion chromatography is considerably more instrumentally demanding than viscometry. Preparation of solutions is also time demanding and may last several days, depending on the solvent and the procedure employed. Dissolution also remains the main source of systematic errors, apart from phenomena taking place during separation. However, the technique also has many decisive advantages. The level of information obtained from an elugram is much more complex than in viscometry, where viscometric average molar mass is obtained only (from which DP can be calculated). From a size exclusion chromatogram, the distribution of molar masses in a sample is obtained (Figure
3.3), along with number-average molar mass ( M n ) and mass-average molar mass ( M w ). Mass distribution may namely reveal further information regarding the nature of degradation, an issue also discussed later in the case of photooxidation of cellulose (Chapter 9). Typical sample size for SEC studies is 5-10 mg, time of sample preparation up to a week,
Ageing and stabilisation of paper 29 Chapter 3: Methodology and analytical techniques in paper stability studies time of analysis: 30-60 min, typical repeatability in the range of 5 to 10%. Determination of moisture38 and dry matter in a sample may be necessary prior to analysis.
Figure 3.3: Number and weight distribution of molar masses of cellulose in a rag paper from ca. 1600.
3.7 IR spectroscopy Infra-red spectroscopy is practically indispensable in studies of polymer degradation. It offers real insight in the processes on the molecular level, while determination of DP only provides a measure of the overall effect of degradation. It is extensively used and a number of authoritative books have been published on the subject,42,43 also in conservation science,44 and a dedicated database is available for the users.45 A review on the possibilities and limitations of IR spectroscopy for characterization of paper was published recently by Workman,46 and older series by Fengel et al.47-50 The number of applications are overwhelming, with several focussing on degradation of cellulose, specifically.51-56 For transmission FTIR spectroscopy, the KBr pellet technique is most often used. Especially in photooxidation, processes are expected to take place primarily on the surface, so that surface-specific techniques are of special interest. The attenuated total reflectance spectroscopy (ATR) is a versatile surface technique for materials either too thick or too strongly absorbing to be analyzed by standard transmission FTIR. At the interface of two media of different refractive indices, i.e. the sample and the internal reflection element, an evanescent wave extends beyond the surface of the sample. The sample is placed in contact with the internal reflection element (Ge, ZnSe, Si, or diamond), the light is totally reflected (usually several times), and absorption spectra are obtained by measuring the interaction of the evanescent wave and the less dense medium (sample). For rough surfaces, diffuse reflectance FTIR (DRIFT) is a frequently used technique, in which scattered light is collected. It consists of light, reflected from uneven areas on the surface, and light, which partly enters the material and is scattered back by the interior. The samples can be used without pre-preparation, with dilution in KBr, or by rubbing off using commercially available abrasive sampling tools. Similar in performances is the photoacoustic (PA) detection technique. After absorption of IR light in a sample, which is kept in an air-tight sampling cell, the sample generates acoustic pressure waves due to thermal expansion. These photoacoustic waves can be
30 Ageing and stabilisation of paper Chapter 3: Methodology and analytical techniques in paper stability studies measured with, e.g., piezoelectric detectors. The spectra show the same features as absorption spectra, yet also depend on parameters such as particle size. The absence of special sample preparation techniques and the fact that the top 10-15 m layer is analysed, make the ATR, DRIFT and PA-FTIR particularly interesting non-invasive techniques. A lower sensitivity is compensated by the ability to provide spatially (or even depth-) resolved information, especially in micro-IR spectroscopic scanning applications. Quantification is difficult and much research focuses on qualitative or semi-quantitative studies.
3.8 Diffuse reflectance spectrophotometry, colourimetry and spectrofluorimetry Visual perception is an important aspect in conservation, especially of graphic art objects. Changes in colour, imposed by a stabilisation treatment, are usually an unwanted side- effect. However, it has to be appreciated that during natural ageing, both thermal and photo, colour changes are imminent. On irradiation of material, light is absorbed, transmitted, or reflected. Colour is a consequence of interactions of light with material, especially light absorption/reflection, and fluorescence, i.e. light emission following excitation with radiation of a shorter wavelength. During degradation of cellulose, both processes may occur simultaneously.57 In UV-VIS diffuse reflectance spectrophotometry, the sample is irradiated and the intensity of reflected light is measured. No sample preparation is needed, while for most commercially available UV-VIS spectrophotometers, diffuse reflectance accessories are available, in which the diffusely reflected light is collected by an integration sphere or a system of mirrors and transmitted to the detector. The result is a reflectance or absorbance spectrum, which is the most objective way of representing the light reflection properties of a sample (Figure 3.4). However, the two spectra in Figure 3.4 tell little about the differences in visual perception of the object’s colour.
88
86
84
82 ) 80 % (
t
n L* = 91.5 e
d 78 i c
n a* = 0.14 i I /
d 76
e b* = 3.6 t c e l f e r 74 I
72 L* = 91.8 70 a* = 0.45 b* = 5.6 68
66
400 450 500 550 600 650 700 (nm) Figure 3.4: VIS diffuse reflectance spectra of two paper samples and the corresponding L*a*b* values. The need for standardisation of colour measurement led to research into human colour
Ageing and stabilisation of paper 31 Chapter 3: Methodology and analytical techniques in paper stability studies vision and a frequently used system is the CIE L*a*b* system, in which the algorithms of calculation of L*, a* and b* values take into account the "standard human eye response".58 The maximum value for L* is 100, representing a perfectly reflecting (white) diffuser, and minimum is zero, which represents a black material. The a* and b* axes have no limits and represent red (positive a*), green (negative a*), yellow (positive b*) and blue (negative b*). While the reflection spectra in Figure 3.4 are different, the human eye will, looking at the two cellulose samples, probably only detect one of them being more yellow than the other (b* = 2), as other differences are small. As a consequence of light absorption in molecules, the absorbed energy may be released in several ways; one possible way is through fluorescence. The intensity of emitted light at different wavelengths, as a consequence of excitation, is measured by spectrofluorimeters. Emission of light in the yellow spectral region (570 nm) will thus make the item appear yellow. A comprehensive review on fluorescence of cellulose and paper was published.59 Light absorption and fluorescence are phenomena for which functional groups – chromophores and fluorophores are responsible, respectively. If such groups were perfectly isolated, the spectra would be well resolved, however, interactions in the material lead to band broadening, so that from absorption and fluorescence spectra little qualitative information can be extracted. 3.9 Degradation of linear polymers Cellulose is a linear homopolymer, meaning that each monomer is linked with two identical monomers, all bonds being the same, i.e. -1,4 glycosidic bonds (Figure 1.3). The length of the macromolecular chain is defined by the number of monomers, i.e. degree of polymerisation. In a typical cellulosic material, different macromolecules have different degrees of polymerisation, so that such material will be defined by both the average DP and by its distribution. The average DP is in turn defined by the number of monomers (N) and the number of bonds between monomers (x): N Nx . (eq 3.3) DP Instead of following the change in average DP, we can therefore follow the change in the number of bonds in the material. If the number is high (high DP), and if all bonds are equal, its change will follow the first-order rate law, as the degradation reaction A → B + B’ may be regarded as monomolecular: dx xk , (eq 3.4) dt
-1 -1 where t is time and k is the rate constant in molbonds molmonomers s . If instead of the number of bonds, which is difficult to determine, DP is introduced in (eq 3.4), and integration is performed, we obtain: N 1 DP 1 ln 00 tk 1 , (eq 3.5) N 1 DPtt where indices 0 and t indicate the values of variables at zero time and at time t. If the number of monomers does not change considerably during the experiment (meaning there
32 Ageing and stabilisation of paper Chapter 3: Methodology and analytical techniques in paper stability studies is no further degradation of monomers to simple products which could even be removed from the system), then 0 NN t . If the DP is large, then the simplification 1 1 1ln( ) (eq 3.6) DP DP leads us to 1 1 tk . (eq 3.7) t DPDP 0 This equation was worked out by A. Ekenstam,60 and has also been named after him. The above presumptions and simplification can be summarized in the conditions, under which the Ekenstam equation is valid: ― The sample must be of high molecular weight and monodisperse (narrow distribution of DP), and the degradation products are also macromolecules (the extent of “peeling- off” must be negligible); ― All bonds are equal; ― There is no loss of monomers during the experiment. The use of Ekenstam equation in studies of paper degradation was reviewed by Emsley and Stevens.61 In order to obtain the rate constant of degradation, one thus has to determine the DP in various points in time throughout an experiment. Such experiments are time- and work-intensive; however, this is the only way to provide comparable data on the rate of degradation of a cellulosic material. Experiments, in which only a single point in ageing time is used, give a much more simplified picture. This could be misleading, as we know that changes in mechanism are possible during cellulose degradation, which then result in inflection of the ageing curve, which should in principle be a straight line (cf. page 94). As prescribed by the Ekenstam equation (eq 3.7), if we plot 1 1 t DPDP 0 over time t, the line should go through the origin of the coordination system and the slope of the line should equal k, the rate constant (Figure 3.6).
In (eq 3.7), the rate constant is not dependent on the initial DP0. Sometimes, the product k·DP0 is therefore used to describe the rate constant; (eq 3.7) then transforms to:
DP0 1 DP0 tktk . (eq 3.8) DPt Rate constants obtained this way are of particular importance if we want to determine the temperature dependence of degradation processes.
Ageing and stabilisation of paper 33 Chapter 3: Methodology and analytical techniques in paper stability studies
y = (1.02 ± 0.03)·10-5·x + (9.1 ± 0.3)·10-5 0.00028 R = 0.9975 3000 0.00026
0.00024
0.00022
2500 0 0.00020 P D / 1 P - t
D 0.00018 P D / 1
0.00016 2000 0.00014
0.00012
0.00010 1500 0 2 4 6 8 10 12 14 16 18 20 time (day) Figure 3.6: Changes in DP of Whatman filter paper sample aged at 90 °C, 65% RH over time, and the same plot transformed according to the Ekenstam equation.
3.9 The Arrhenius equation The rate of chemical reaction depends on many parameters, among which the temperature is of particular importance for degradation studies. This is due to the fact that thermal degradation of paper is usually so slow that its consequences can usually not be measured at room temperature. At elevated temperatures the degradation process proceeds faster. The relationship between temperature, T, and rate constant, k, was described in 1889 by S. Arrhenius62 and has been named after him:
Ea eAk TR , (eq 3.9) where R is the universal gas constant (8.314 J mol-1 K-1), A is the pre-exponential factor -1 and Ea is the activation energy (J mol ). The latter two parameters can be calculated only for some simple gaseous reaction systems,63 and for reactions in condensed phase and heterogeneous reactions, the meaning of activation energy becomes more elusive. As a matter of fact, it can even be applied to many processes only distantly related to chemistry: cooking of meat apparently has an activation energy of 33 kJ mol-1.64 The concept of Arrhenius equation and its value was discussed by many authors, e.g. Priest,65 and comprehensively reviewed by Porck.66 The logarithmic form of the Arrhenius equation: E 1 Ak a .lnln , (eq 3.10) R T usually serves to construct the graph ln(k) vs. 1/T, as in Figure 3.7. In the Figure we demonstrate the validity of Arrhenius equation for a cellulosic sample in the temperature interval 60-220 °C. Knowing the rate constant k’ at the temperature T’, and the activation energy, the rate constant at any temperature T within the interval of study can be calculated:
k Ea 11 .ln . (eq 3.11) k TTR
34 Ageing and stabilisation of paper Chapter 3: Methodology and analytical techniques in paper stability studies
-14 220 oC -16
-18 140 oC -20 k -22 n l 90 oC -24
lnk = -(11200 ± 200)·1/T + (7.2 ± 0.6) 60 oC -26 R = -0.9986 E = 93±2 kJ mol-1 a -28
0.0020 0.0022 0.0024 0.0026 0.0028 0.0030 0.0032 1/T (K-1) Figure 3.7: Demonstration of validity of the Arrhenius equation for oxidation of a bleached sulphate pulp (pH 7.2) in dry air in a wide temperature interval. This is important, but for practical purposes, we would certainly like to know the rate of degradation at room temperature. In order to do this, we have to extrapolate the data to, e.g. 20 °C. This is linked with considerable uncertainty, as it is difficult to predict whether the continuity will apply throughout the desired extrapolation interval (if we refer to the previous example: meat will never be cooked at room temperature). However, the extrapolation interval in question is considerably small and judging from the relatively good fit over a wide temperature interval in Figure 3.7 we have grounds to believe that the extrapolated behaviour would be linear. For the first time, this assumption was experimentally confirmed by Zou et al. (Figure 3.8).67 For a number of pulp samples, the authors showed that the rates predicted from (eq 3.11) are in a comparatively good agreement with the rates of degradation observed during 22 years of natural ageing.
Figure 3.8: Comparison of predicted rates of degradation for 18 pulps with the rates, obtained during natural degradation for the same samples for 22 years (pH 3.2-5.7). Reproduced from Zou et al.,67 with kind permission of Springer Science and Business Media. The discussion of uncertainty in prediction is a topic that has to be addressed. The uncertainty is usually expressed in terms of a confidence interval in which the true value will appear with a certain degree of probability. For calculation of the extrapolation error, it would be most appropriate to take the exponential form of the Arrhenius equation (eq 3.9).
Ageing and stabilisation of paper 35 Chapter 3: Methodology and analytical techniques in paper stability studies
Using the exponential form, the confidence interval would also be symmetrical around the mean, i.e. predicted value. However, it has been shown that using the linearised form, i.e. (eq 3.10), is in principle not inferior to non-linear regression. If the relative error in the data is distributed normally, it can even be superior.68 In Figure 3.9, we replotted the values of ln k in Figure 3.7, in the temperature interval of 60-90 °C, in which most such studies are performed, together with confidence limits with 95% probability. The limits are narrowest in the centre of the studied interval, however, we can also see that they rapidly spread once extrapolation to lower temperatures is attempted: at 40 °C, the confidence interval for ln k will be (-27.6 ± 0.6) s-1, i.e. ~2%, whereas at 20 °C, it will be (-29.5 ± 0.9) s-1, i.e. ~3%. While using the linearised form of the Arrhenius equation, the confidence interval will be symmetrical around the mean for ln k, this will not be the case for k: the confidence interval will be (17.3·10-13 < 9.77·10-13 < 5.50·10-13) s-1 at 40 °C, whereas at 20 °C, it is (38.8·10-14 < 15.1·10-14 < 5.9·10-14) s-1.
-23 o 90 C ln k, determined -24 ln k, predicted Arrhenius model -25 confidence limits 60 oC -26
-27 k
n l o -28 40 C
-29
o -30 20 C
-31 0.0028 0.0029 0.0030 0.0031 0.0032 0.0033 0.0034 1/T (1/K) Figure 3.9: Extrapolation of ln k and the corresponding confidence limits for a typical experiment: ageing of a bleached sulphate pulp in dry air at four temperatures in the interval 60-90 °C. Each of the four rates of degradation was obtained experimentally by applying the Ekenstam equation to a number of points (DP) determined during an experiment, while the predicted two were calculated from the regression parameters. Calculation of the uncertainty and confidence interval for the rate constant at 20 °C is easiest from (eq 3.11):
Ea 1 1 kk 293 .lnln . (eq 3.12) 293 TR
Using regression analysis, the intercept ( ln k293 ) and the slope (Ea/R) can be calculated. In (eq 3.12), we now need to estimate the uncertainty in the intercept ( s ). The ln k293 corresponding equations are commonly available in statistics handbooks.69 For the ease of 1 1 notation, let us exchange x and ln ky . First, for n experiments, we 293 T calculate the value of s / xy :
36 Ageing and stabilisation of paper Chapter 3: Methodology and analytical techniques in paper stability studies
1 ( yy ˆ )2 2 ii i s xy , (eq 3.13) n 2 where values of yi are natural logarithms of rate constants at a particular temperature while
yˆi are the calculated (fitted) values of yi at the same temperature using (eq 3.10) with the regression data obtained for (eq 3.12). The value s / xy is also referred to as standard deviation of y-residuals. We can now easily calculate the uncertainty in the intercept: 1 x2 2 ss i , (eq 3.14) ln k293 / xy 2 i xxn i
1 1 where x is the average value of . The confidence limits for the intercept are 293 T now ln stk , where t is the Student factor for n-2 degrees of freedom, in our 293 ln k293 case its value is 4.3, as we performed four accelerated ageing experiments. In a similar manner, the uncertainty of the slope can be calculated:
s / xy s / REa 1 . (eq 3.15) 2 2 i xx i
From this value, the uncertainty in Ea can also be obtained. The quality of predictions (i.e. lower confidence limits of predicted values), can be improved by: ― Performing accelerated ageing experiments at temperatures as low as possible; ― Performing accelerated ageing experiments at many temperatures; ― Decreasing the experimental error in each data point. In a similar study, Bégin and Kaminska arrived at the same conclusions.70 Experiments at low temperatures are extremely time-demanding. Zou et al. compared natural ageing with accelerated ageing67 – the samples in his study were all acidic, and thus the rate of ageing high. However, for the least acidic sample with pH 5.72, no change in DP over 20 years was determined. It is obvious that for the more stable neutral or mildly alkaline samples, even longer periods of time would be necessary. The quality of extrapolation is also increased if the uncertainty in each data point is decreased. This can be done by repeated experiments, and by employing an analytical technique with less uncertainty. This is also the reason why determination of DP instead of a physical property might be more favourable.
Ageing and stabilisation of paper 37 Chapter 3: Methodology and analytical techniques in paper stability studies
3.10 Accelerated degradation studies Apart from increasing the temperature, acceleration of degradation can be achieved in many ways, by increasing the content of pollutants in the atmosphere surrounding the samples or by irradiation with light. Degradation in stacks or degradation of loose sheets can be performed. We can perform degradation at a variety of conditions, in oxygen, nitrogen, humid or dry, static or dynamic atmosphere, or by employing any variation of the parameters. The ways in which the degradation experiments can be performed are in fact so variable, that we would only like to outline the achievable goals. Besides, many of the ageing conditions frequently employed and their value were reviewed by Porck in his authoritative review.66 The goals of accelerated ageing studies can be identified as follows: ― Prediction of rates of degradation at room temperature (or similar); ― Comparison of stability of a variety of samples at some chosen conditions of ageing; ― Studies of degradation mechanisms. As outlined above, in order to be able to extrapolate the ageing behaviour of a sample to any temperature, several (in principle at least two) ageing experiments have to be performed. The changes in some property have to be followed (e.g. DP), and kinetics at the particular temperature evaluated. By obtaining the rate constants at several elevated temperatures, and by using the Arrhenius equation, the rate constant at room temperature, or similar, can be calculated. Using this approach, it is in principle possible to evaluate the stability of different papers. By comparison of the rate constants e.g. of a stabilised and a non-stabilised paper sample, we can also evaluate the beneficial effect of a stabilisation treatment. It is difficult, though, to predict the remaining lifetime of a paper. If we take that into account the Ekenstam equation (eq 3.7), we can calculate the time needed for DP to drop to some value we regard to be the lower limit of usability. Such a value is certainly highly disputable – a typical proposed value is 250-300.71 The time needed for the sulphate pulp (Figure 3.9) with starting DP0 of 3090 to degrade to final DP of 300, is then
1111 1 1 10 t 113 415s1031.1 years (eq 3.16) k t DPDP 0 s103.2 300 3090 in dry atmosphere at 20 °C. In this calculation, the biggest uncertainty contributing to the overall error will be the one in k. Bégin and Kaminska proposed to calculate the time to 50% 70 property loss (time to PL50%), which is another possibility, unfortunately, time to PL50% gives little idea of the actual lifetime. For a comparison of ageing behaviour of a variety of samples, they need to be tested at the same conditions. This is frequently done, although by ageing at one elevated temperature, we know fairly little about behaviour of the sample at any other temperature, e.g. at room conditions. Tests performed at one elevated temperature may only give data on stability at the particular temperature. Most usually, 80 °C, 65% relative humidity is used, although for faster results, 90 °C, 65% relative humidity is also an alternative.
38 Ageing and stabilisation of paper Chapter 3: Methodology and analytical techniques in paper stability studies
It is advisable, though, if the conditions during such experiments are standardised – in order for different studies to be comparable. There are several standard ageing procedures available; the International Standards Organisation prescribes four: ― 105 °C, dry atmosphere,72 ― 90 °C, 25% RH,73 ― 80 °C, 65% RH,74 ― 120 or 150 °C, dry atmosphere,75 and still other conditions are standardised by other organisations (TAPPI, ASTM etc.), the intention of a particular treatment being to test a paper sample for particular purpose. E.g., ageing at 150 °C is usually performed to test the stability of paper used in power transformers. Following their recent research, ASTM published three standards in 2002: for thermal,76 pollution77 and light78 ageing. The standards will be further considered in Chapter 12. Ageing in conditions of dynamic humidity is also often used – during ageing relative humidity is cycled, thus apparently simulating natural conditions. It was shown that degradation of loose sheets proceeds faster in such conditions, while the effect was less pronounced for books.79 Due to the effect such ageing has on migration of compounds in paper, this type of ageing is often used for evaluation of iron gall ink corrosion.80,81 Ageing of encapsulated material82 or ageing in closed vessels70 is also increasingly interesting and has been standardised.76 The experimental conditions during such experiments need some more control due to gradual oxygen depletion of the atmosphere and due to problems with leakage (cf. page 93). It is also evident that during ageing, volatile compounds, also acids are formed, which may easily migrate from the material if loose- sheet ageing is performed. Ageing in stacks has been shown to increase the degradation rate of acidic material considerably.83 Recently, migration of acids through a stack of papers was followed, accompanied by corresponding degradation.84,85 This important line of research was shortly reviewed by Baranski86 and is discussed in more detail later (page 97). Of similarly high importance are studies of the effect of pollutants, e.g. sulphur dioxide and nitrogen oxides,27,87 and acetic acid.88 It is important to appreciate that while some environmental gasses can be absorbed, the material itself may also be a source of volatile compounds, produced during degradation, thus accelerating the degradation of other objects in its vicinity. This will be discussed further in Chapter 11. For studies of degradation mechanisms, still a variety of techniques other than accelerated ageing at temperatures lower than 100 °C are used. It is difficult to demonstrate the validity of results obtained at high temperatures for the low temperature region and in the presence of humidity. Erhard et al. demonstrated that a variety of degradation products are produced at 150 °C, which are different to those at lower temperatures.89 Still, accelerated degradation experiments used in studies of ageing of paper in power transformers led to important knowledge on the mechanisms of oxidation.61 Another experimental condition not related to real situations, is inert atmosphere. The absence of oxygen during accelerated ageing experiments has been used for decades to discern between oxygen-dependent (oxidation) and oxygen-independent processes (thermolysis and acid-catalysed hydrolysis).90 Considering the important role of water during low-temperature degradation of cellulose and paper,67 and considering that above ~300 °C the degradation of cellulose proceeds at a
Ageing and stabilisation of paper 39 Chapter 3: Methodology and analytical techniques in paper stability studies similar rate in both inert and oxidative atmosphere, it might be of value to divide the regions of interest to: ― Low temperature region (T < 100 °C), ― Intermediate temperature region (100 °C < T < 300 °C), ― High temperature region (T > 300 °C). The three different intervals are evident in a dynamic thermogram in Figure 3.10.
1.01
1.00
0.99
0.98
0 0.97 m / m 0.96
0.95
0.94
0.93
0 50 100 150 200 250 300 350 T (oC) Figure 3.10: Dynamic thermogram of a sulphate bleached cellulose sample. Inert atmosphere, rate of heating 5 °C min-1.
3.11 Conclusions Regardless of the fact that there has been a lot of research invested in studies of degradation of cellulose and paper, a number of knowledge gaps still exist. There is a need for improvement of analytical methods, many of which have been unchallenged and used continuously for many decades. A number of contemporary analytical methods deserve to be employed more frequently, especially spectroscopic, e.g. X-ray photoelectron spectroscopy, Raman, but also IR, especially in combination with more advanced sampling techniques, and mass spectroscopy. Imaging techniques should also deserve our increased attention, particularly those with high spatial resolution. The research into accelerated ageing and degradation mechanisms also deserves more attention. Especially the time-demanding and work-intensive Arrhenius studies should be performed more often and at lower temperatures. The relationship between stack- and loose-sheet ageing should be elucidated. Finally, if experiments at a single temperature are performed, standard ageing conditions should be used to ensure the comparability of different studies.
3.12 References 1. B.L. Browning, Analysis of Paper, Marcel Dekker, New York, USA, 1977.
40 Ageing and stabilisation of paper Chapter 3: Methodology and analytical techniques in paper stability studies
2. T.P. Nevell, S.H. Zeronian, Cellulose Chemistry and Its Applications, Ellis Horwood, Chichester, 1985. 3. S. Buchanan, S. Coleman, Deterioration survey on the Stanford University Libraries Green Library stack collection, College and Research Libraries, 1987, 48, 102-147. 4. N. Gurnagul, R.C. Howard, X. Zou, T. Uesaka, D.H. Page, The Mechanical Permanence of Paper: A Literature Review, J. Pulp Pap. Sci., 1993, 19, J160-J166. 5. D.H. Page, The structure and strength of paper, Proc. Worksh. Effects Aging Print. Writ. Papers, ASTM Institute for Standards Research, 1994. 6. J.C. Roberts, The Chemistry of Paper, The Royal Society of Chemistry, Cambridge, 1996. 7. X. Zou, N. Gurnagul, T. Uesaka, J. Bouchard, Accelerated aging of papers of pure cellulose: mechanism of cellulose degradation and paper embrittlement, Polym. Degrad. Stab., 1994, 43, 393-402. 8. R. Stol, J.L. Pedersoli Jr., H. Poppe, W.Th. Kok, Application of Size Exclusion Electrochromatography to the Microanalytical Determination of the Molecular Mass from Objects of Cultural and Historical Value, Anal. Chem., 2002, 74, 2314-2320. 9. X. Zou, T. Uesaka, N. Gurnagul, Prediction of paper permanence by accelerated aging. Part I: Kinetic analysis of the aging process, Cellulose, 1996, 3, 243-267. 10. ISO 187:1990 - Standard Atmosphere for Conditioning and Testing and Procedure for Monitoring the Atmosphere and Conditioning of Samples. 11. ISO 1924-2:1994 - Paper and Board - Determination of Tensile Properties - Part 2: Constant Rate of Elongation Method. 12. ISO 1974:1990 - Paper - Determination of Tearing Resistance (Elmendorf Method). 13. ISO 2758:2001 - Paper - Determination of Bursting Strength. 14. ISO 5626:1993 - Paper - Determination of Folding Endurance. 15. W. Schneider, Problems in the Determination of the Folding Endurance in the Context of Material Fatigue, Papier, 2004, 58, 1-11. 16. P. Bégin, J. Iraci, D. Grattan, E. Kaminska, D. Woods, X. Zou, N. Gurnagul, S. Deschatelets, The impact of lignin on paper permanence. Part I: A comprehensive study of the aging behaviour of handsheets and commercial paper samples, Actes Trois. Journ. Intern. ARSAG, 1997. 17. V. Bukovsky, The influence of light on ageing of newsprint paper, Restaurator, 2000, 21, 55-76. 18. J. Hanus, M. Komorniková, J. Minariková, Influence of boxing materials on the properties of different paper items stored inside, Restaurator, 1995, 16, 194-208. 19. M. Strlič, J. Kolar, D. Kočar, T. Drnovšek, V.-S. Šelih, R. Susič, B. Pihlar. What Is the pH of Alkaline Paper?, e-PS, 2004, 1, 35-47. 20. TAPPI T 509 Om-02. Hydrogen Ion Concentration (pH) of Paper Extracts (Cold Extraction Method). 21. ASTM D778-97(2002) Standard Test Methods for Hydrogen Ion Concentration (pH) of Paper Extracts (Hot- Extraction and Cold-Extraction Procedures). 22. TAPPI T529 Om-88: Surface pH Measurement of Paper. 23. E. Eusman, Tideline Formation in Paper Objects: Cellulose Degradation at the Wet-Dry Boundary, Studies in the History of Art 51, Monograph Series II, National Gallery of Art, Washington, USA, 1995, 11-27. 24. A. Barański, K. Frankowicz, Z. Harnicki, Z. Koziński, T. Łojewski, Acidic books in libraries. How to count them?, Proc. 5th European Conf. Cultural Heritage Research: a Pan-European Challenge, R. Kozłowski, Ed., Cracow, Poland, May 16–18th 2002, Polish Academy of Sciences, Cracow, 2003, 283-285. 25. S. Saverwyns, V. Sizaire, J. Wouters, The acidity of paper. Evaluation of methods to measure the pH of paper samples, Preprints of the 13th Triennial ICOM-CC Meeting, ICOM committee for conservation, 2002, 2, 628-634. 26. ANSI/NISO Z39.48-1992(R1997): Permanence of Paper for Publications and Documents in Libraries and Archives. 27. A. Johansson, Air pollution and paper deterioration. Causes and remedies, PhD Thesis, Göteborg University, Göteborg, Sweden, 2000. 28. TAPPI T 553 Om-00 Alkalinity of Paper As Calcium Carbonate (Alkaline Reserve of Paper).
Ageing and stabilisation of paper 41 Chapter 3: Methodology and analytical techniques in paper stability studies
29. J.M.G. Vives, J.M.G. Monmany, R.A. Guerra, Non-destructive method for alkaline reserve determination in paper - Diffuse reflectance infrared Fourier transform spectroscopy, Restaurator, 2004, 25, 47-67. 30. A.L. Dupont, J. Barthez, H. Jerosch, B. Lavedrine, Testing CSC Book Saver®, a commercial deacidification spray, Restaurator, 2002, 23, 39-47. 31. H.G. Barth, J.W. Mays, Modern Methods of Polymer Characterization, J. Wiley & Sons, 1991. 32. ISO 5351/1: Cellulose in Dilute Solutions - Determination of Limiting Viscosity Number – Part 1: Method in Cupri-Ethylene-Diamine (CED) Solution, 1981. 33. SCAN-CM 15:88: Viscosity in Cupri-Ethylenediamine Solution, Scandinavian pulp, paper and board testing committee, 1988. 34. F.H. Wetzel, J.H. Elliot, A.F. Martin, Variable shear viscometers for cellulose intrinsic viscosity determinations, TAPPI, 1953, 36, 564-571. 35. R. Evans, A.F.A. Wallis, Comparison of Cellulose Molecular Weights Determined by High Performance Size Exclusion Chromatography and Viscometry, 4th Int. Symp. Wood Chem. 1987, 201-205. 36. M. Strlič, J. Kolar, M. Žigon, B. Pihlar, Evaluation of size-exclusion chromatography and viscometry for the determination of molecular masses of oxidised cellulose, J. Chromatogr. A, 1998, 805, 93-99. 37. J. Kolar, M. Strlič, M. Marinček, IR pulsed laser light interaction with soiled cellulose and paper, Appl. Phys. A, 2002, 75, 1-4. 38. ASTM Standard Test D 1348 - 89: Standard Test Methods for Moisture in Cellulose, 1989. 39. A.-L. Dupont, G. Mortha, Comparative evaluation of size-exclusion chromatography and viscometry for the characterisation of cellulose, J. Chromatogr. A, 2004, 1026, 129-141. 40. M. Strlič, J. Kolar, Size Exclusion Chromatography of cellulose in LiCl/DMAc. A Review, J. Biochem. Biophys. Meth., 2003, 56, 265-279. 41. A.-L. Dupont, Study of the degradation of gelatin in paper upon ageing using aqueous size-exclusion chromatography, J. Chromatogr. A, 2002, 950, 113-124. 42. H. Günzler, H.-U. Gremlich, IR Spectroscopy. An Introduction, Wiley-VCH, Weinheim, 2002. 43. J.L. Koenig, Spectroscopy of Polymers. 2nd Edition, Elsevier, Amsterdam, 1999. 44. M.R. Derrick, D. Stulik, J.M. Landry, Infrared Spectroscopy in Conservation Science. Scientific Tools for Conservation, The Getty Conservation Institute, Los Angeles, 1999. 45. Infrared and Raman users group, IRUG Spectral Database Edition 2000, http://www.irug.org/, accessed 27/09/2004. 46. J.J. Workman Jr, Infrared and raman spectroscopy in paper and pulp analysis, Appl. Spectrosc., 2001, 36, 139-168. 47. D. Fengel, M. Ludwig, Possibilities and Limits of the FTIR Spectroscopy for the Characterization of Cellulose. Part 1. Comparison of Various Cellulose Fibres and Bacteria Cellulose, Papier, 1991, 45, 45-51. 48. D. Fengel, Possibilities and Limits of FTIR Spectroscopy for the Characterization of Cellulose. Part 2. Comparison of Various Pulps, Papier, 1991, 45, 97-102. 49. D. Fengel, Possibilities and limits of FTIR spectroscopy for the characterization of cellulose. Part 3. Influence of attendant components on the IR-spectrum of cellulose, Papier, 1992, 46, 7-11. 50. D. Fengel, M. Ludwig, M. Przyklenk, Possibilities and limits of FTIR spectroscopy for the characterization of cellulose. Part 4. Studies on cellulose ethers, Papier, 1992, 46, 323-328. 51. K.L. Kato, R.E. Cameron, Structure-Property Relationships in Thermally Aged Cellulose Fibers and Paper, J. Appl. Polym. Sci., 1999, 74, 1465-1477. 52. C. Sistach, N. Ferrer, M.T. Romera, Fourier Transform Infrared Spectroscopy applied to the Analysis of ancient manuscripts, Restaurator, 1998, 19, 173-186. 53. A. Johansson, H. Lennholm, Influences of SO2 and O3 on the ageing of paper investigated by in situ diffuse reflectance FTIR and time-resolved trace gas analysis, Appl. Surf. Sci., 2000, 161, 163-169. 54. L.M. Proniewicz, C. Paluszkiewicz, A. Weselucha-Birczynska, A. Barański, D. Dutka, FTIR and FT-Raman study of hydrothermally degraded groundwood containing paper, J. Molec. Struct., 2002, 614, 345-353. 55. P. Calvini, A. Gorassini, FTIR - Deconvolution spectra of paper documents, Restaurator, 2002, 23, 48-66. 56. M. Ali, A.M. Emsley, H. Herman, R.J. Heywood, Spectroscopic studies of the ageing of cellulosic paper, Polymer, 2001, 42, 2893-2900.
42 Ageing and stabilisation of paper Chapter 3: Methodology and analytical techniques in paper stability studies
57. H. Tylli, I. Forsskahl, C. Olkkonen, A spectroscopic study of photoirradiated cellulose, J. Photochem. Photobiol. A-Chem., 1993, 76, 143-149. 58. K. McLaren, XIII-The Development of the CIE (L*a*b*) Uniform Colour Space and Colour-difference Formula, JSDC, 1976, 338-341. 59. J.A. Olmstead, D.G. Gray, Fluorescence spectroscopy of cellulose, lignin and mechanical pulps: A review, J. Pulp Pap. Sci., 1997, 23, J571-J581. 60. A. af Ekenstam, Behaviour of Cellulose in Solutions of Mineral Acids. Part II: A Kinetic Study of Cellulose Degradation in Acid Solutions, Ber., 1936, 69, 553-559. 61. A.M. Emsley, G.C. Stevens, Kinetics and mechanisms of the low-temperature degradation of cellulose, Cellulose, 1994, 1, 26-56. 62. S. Arrhenius, Über die Reaktionsgeschwindigkeit bei der Inversion von Rohzucker durch Säuren, Z. Phys. Chem., 1889, 4, 226-248. 63. K.J. Laidler, Chemical Kinetics, 3rd Ed., Harper Collins, New York, 1987. 64. A.L. Petrou, M. Roulia, K. Tampouris, The use of the Arrhenius equation in the study of deterioration and of cooking of foods. Some scientific and pedagogic aspects, Chem. Educ. Res. Pract. Eur., 2002, 3, 87-97. 65. D.J. Priest, Artificial aging of paper: correlation with natural aging, Proc. Worksh. Effects Aging Print. Writ. Papers, ASTM Institute for Standards Research, 1994. 66. H.J. Porck, Rate of Paper Degradation. The Predictive Value of Artificial Ageing Tests, ECPA, Amsterdam, 2000. 67. X. Zou, T. Uesaka, N. Gurnagul, Prediction of paper permanence by accelerated aging. Part II: Comparison of the predictions with natural aging results, Cellulose, 1996, 3, 269-279. 68. N. Brauner, M. Shacham, Statistical analysis of linear and nonlinear correlation of the Arrhenius equation constants, Chem. Eng. Proc., 1997, 36, 243-249. 69. J.C. Miller, J.N. Miller, Statistics for Analytical Chemistry. 3rd Ed., Ellis Horwood, New York, 1993. 70. P.L. Bégin, E. Kaminska, Thermal Accelerated Ageing Test Method Development, Restaurator, 2002, 23, 89-105. 71. D.H. Shroff, A.W. Stannett, A review of paper aging in power transformers, IEE Proc., 1985, 132, 312- 318. 72. ISO 5630-1:1991 Paper and Board -- Accelerated Ageing -- Part 1: Dry Heat Treatment at 105 °C. 73. ISO 5630-2:1985 Paper and Board -- Accelerated Ageing -- Part 2: Moist Heat Treatment at 90 °C and 25% RH. 74. ISO 5630-3:1996 Paper and Board -- Accelerated Ageing -- Part 3: Moist Heat Treatment at 80 °C and 65% RH. 75. ISO 5630-4:1986 Paper and Board -- Accelerated Ageing -- Part 4: Dry Heat Treatment at 120 or 150 °C. 76. ASTM D 6819-02, Standard Test Method for Accelerated Temperature Aging of Printing and Writing Paper by Dry Oven Exposure Apparatus. 77. ASTM D 6833-02, Standard Test Method for Accelerated Pollutant Aging of Printing and Writing Paper by Pollution Chamber Exposure Apparatus. 78. ASTM D 6789-02, Standard Test Method for Accelerated Light Aging of Printing and Writing Paper by Xenon-Arc Exposure Apparatus. 79. C.J. Shahani, F.H. Hengemihle, N. Weberg, The effect of variations in relative humidity on the accelerated aging of paper, ACS Symp. Ser., 1989, 410, 63-80. 80. J.G. Neevel, Phytate: A Potential Conservation Agent for the Treatment of Ink Corrosion Caused by Irongall Inks, Restaurator, 1995, 16, 143-160. 81. J. Kolar, M. Strlič, M. Budnar, J. Malešič, V.S. Šelih, J. Simčič, Stabilisation of corrosive iron gall inks, Acta Chim. Slov., 2003, 50, 763-770. 82. J.B.G.A. Havermans, Ageing behaviour of encapsulated paper, Restaurator, 1999, 20, 108-115. 83. C.J. Shahani, Accelerated aging of paper: can it really foretell the permanence of paper?, Proc. Worksh. Effects Aging Print. Writ. Papers, ASTM Institute for Standards Research, 1994. 84. H. Carter, P. Bégin, D. Grattan, Migration of Volatile Compounds through Stacked Sheets of Paper during Accelerated Ageing. Part 1: Acid Migration at 90 °C, Restaurator, 2000, 21, 77-84.
Ageing and stabilisation of paper 43 Chapter 3: Methodology and analytical techniques in paper stability studies
85. A. Bulow, P. Bégin, H. Carter, T. Burns, Migration of volatile compounds through stacked sheets of paper during accelerated ageing Part II: Variable temperature studies, Restaurator, 2000, 21, 187-203. 86. A. Barański, Ageing kinetics of cellulose and paper, Restaurator, 2002, 23, 77-88. 87. P. Bégin, S. Deschatelets, D. Grattan, N. Gurnagul, J. Iraci, E. Kaminska, D. Woods, X. Zou, The effect of air pollutants on paper stability, Restaurator, 1999, 20, 1-21. 88. A.L. Dupont, J. Tetreault, Cellulose degradation in an acetic acid environment, Stud. Conserv., 2000, 45, 201-210. 89. D. Erhardt, D. von Endt, W. Hopwood, The comparison of accelerated aging conditions trough the analysis of extracts of artificially aged paper, Preprints of Paper Presented at the Fifteenth Annual Meeting, The American Institute for Conservation of Historic and Artistic Works, Washington, D. C., 43-55. 90. J.S. Arney, C.L. Novak, Accelerated aging of paper. The influence of acidity on the relative contribution of oxygen-independen and oxygen-dependent processes, TAPPI, 1982, 65, 113-115.
44 Ageing and stabilisation of paper Chapter 4 Experimental techniques in studies of photo-stability
Olivier Haillant, Dominique Fromageot, Jacques Lemaire
4.1 Introduction: Accelerated photoageing – a century-long story Among the environmental parameters influencing degradation of cellulose and other components of paper (ligneous residues, fillers, sizing, dyes etc.), it is daylight, combined with the effect of heat and atmospheric oxygen, that acts as the main parameter of stress in outdoor conditions.1-5 This is also the case with indoor conditions, when a reasonable amount of daylight still passes through windows.6-10 Even fluorescent light usually used to reproduce daylight was revealed to be potentially damaging.11 Strictly speaking, only complete darkness would prevent the occurrence of photochemical reactions. The actual chemistry of photoageing of paper will be discussed in Chapter 9; here we will stress the instrumentation and some general principles concerning the study of the photooxidation of organic materials. Almost a century ago, when Hollywood did not exist and Chicago was the motion picture capital of the world, carbon arc light sources were used as stage lighting. Soon, it became apparent that certain textiles and garments worn by the actors faded after extended periods of time due to the lighting.12,13 In 1915, the Atlas Electric Devices Company developed the Solar Determinator instrument as a spin-off of the stage lighting business and in 1919, the first Atlas Color Fade-Ometer was introduced, which was a redesign of the earlier Solar Determinator unit. The German synthetic dye chemists, who faced the problem of textile fading, were the first to use the Atlas Color Fade-Ometer to evaluate the effect of light on the colour of dyed textiles.14 In 1922, results from exposure in a UV chamber using a mercury lamp were published.15 This was the first time an artificial light source other than carbon arc lighting was tested. Later on, in 1934, another generation of instruments using the carbon arc light source, the Sunshine Carbon Arc Weather-Ometer, was introduced. The carbon arc light has been proven to simulate natural daylight poorly, yet this unit remained the standard laboratory- accelerated testing device for over 30 years.
Ageing and stabilisation of paper 45 Chapter 4: Experimental techniques in studies of photo-stability
Although daylight is generally acknowledged as the key source of stress involved in weathering of objects exposed to outdoor conditions, other parameters like temperature, atmospheric moisture and direct application of water droplets were added in the following generations of accelerating ageing devices. During the same period, other light sources were researched. In 1954, Heraeus developed the first xenon arc-weathering instrument, the Xenotest 150, which is basically the ancestor of most existing artificial weathering instruments. In the 1960s, the first generation of accelerated weathering devices, Fresnel solar concentrators, was introduced. Within the past 30 years, three other major artificial sources of light were introduced and were shown to more or less replicate natural daylight: medium pressure vapour mercury arc lamps, fluorescent UV-lamps and metal halide lamps.
4.2 Review of outdoor weathering sites and accelerated photoageing units Outdoor or natural weathering Several locations worldwide have been identified as ideal for natural exposure tests, during which samples are exposed to direct daylight. This type of light ageing provides the best correlation with weathering during outdoor use (in subtropical, desert, temperate, marine, cold, polluted conditions etc.), but is also the most time-consuming. Among the companies offering natural exposure, the Atlas MTS Company provides a total network of 23 locations.16 Two of them, Southern Florida and Arizona Desert stations have long been considered as benchmark climates for natural outdoor weathering, due to the presence of natural climatic extremes: The Southern Florida sites (in Miami and the Everglades) are able to effectively represent any climate with high relative humidity, and the Arizona Desert site is used for dry climates. Although the first test site for natural outdoor exposures was established in North Dakota in 1906 for comparison of coatings,17 the Miami site is of historical importance as it is the oldest standardized outdoor weathering site with data reported since the beginning. Consequently, a large amount of data regarding ageing of many different materials has been collected over the past decades. Both locations provide mild acceleration of photooxidation reactions, as the angle of the target support can be optimized in order to maximize irradiation with daylight, and also because the two climates are recognized as some of the most severe for outdoor exposure of materials. Moreover, a wide range of exposure possibilities is provided, such as direct exposure on aluminium racks with either fixed or variable angles, as well as plywood-backed or open- backed racks, indirect exposure (under glass) for interior materials, and black box exposures for paints and coatings.18 The average irradiance (direct, diffuse and global in the UV, visible and IR ranges), temperature, relative humidity, rainfall, and wind speed are the main parameters continuously measured and numerous standards are devoted to static exposures undertaken on those three major sites.19
Units for accelerated photoageing A broad variety of artificial light sources were used over the last decades, aimed at reproducing the effects of daylight. Some, like the filtered xenon arc or the metal halide lamps were selected empirically due to the similarity of their spectral distribution to the
46 Ageing and stabilisation of paper Chapter 4: Experimental techniques in studies of photo-stability
UV, visible and IR range of daylight while others, e.g. the filtered mercury arc were chosen upon scientific considerations. Acceleration can generally be achieved in two ways: firstly, laboratory accelerated photoageing devices can be run continuously without being limited by the natural day/night cycles, the seasonal variations and weather conditions, and secondly, it is possible to increase the influences of irradiance and temperature.20 The effect of such changes on the degradation mechanism of the tested material must be controlled in order to ensure the relevance of the test. A description of the most frequently used accelerated photoageing units is given hereafter. They are based on the type of light, following the chronological order of their introduction.
Carbon arc light source The source consists of carbon rods, either enclosed in a Pyrex globe filled with an oxygen- deficient atmosphere (introduced in 1919) or operating in a free flow of air, filtered by flat filters (introduced in the 1930s). Both types exhibit a spectral distribution that is unsuitable for simulating daylight. Indeed, the former one emits light, which is much more intense than sunlight in the longer UV wavelength range and poor in the range below 310 nm. The latter one mimics the spectral distribution of daylight better, but emits wavelengths shorter than 295 nm, which is out of the daylight cut-on (lowest limit of the daylight spectral distribution).21 In addition to this, both carbon arc units require daily replacement of carbon rods and cleaning of filters or globes.22 Although photoageing units using carbon arc sources are not available anymore, standards referring to them are still in use in Japan and the USA in tests of material durability as such devices have been the only available accelerated photoageing devices for a long time.23
Xenon arc light source When properly filtered, this light replicates fairly well the spectral distribution of daylight in the UV-visible range.21 The use of a borosilicate glass filter meets this condition and is recommended in most standards using the source. However, some standards still recommend a combination of quartz/borosilicate or quartz/quartz filters, increasing the intensity at wavelengths below the daylight cut-on. For several reasons discussed in the following paragraph, such combinations of filters do not provide an appropriate source of light. Two categories of chambers are commercially available: a. Weather-Ometer and Xenotest devices (Figure 4.1) consist of a chamber equipped with a xenon arc lamp placed vertically at the centre around which a rack supporting the samples rotates. Irradiance, temperature and relative humidity can be adjusted. The chamber temperature is regulated by a cylindrical air flow, which ensures quite accurate control (1 °C). The brand names are Xenotest and Weather-Ometer (Atlas). b. In smaller chambers, samples are placed on a flat horizontal surface in the lower part of the chamber and the lamp is located above them (Figure 4.2). No rotation of either
Ageing and stabilisation of paper 47 Chapter 4: Experimental techniques in studies of photo-stability
the samples or the light is applied. Two features inherent to the geometry of such devices might limit the repeatable and reproducible characteristics of tests: ― The air-flow cooling system does not match the symmetry of exposure area, which means that the uniformity of temperature at the surface of the exposed samples is questionable. ― The large dimension of the flat exposure area with regard to the distance to the lamp guarantees no spatial uniformity of light over the whole exposure area. Although scientific publications report the use of such devices to effectively simulate natural weathering,24 their reproducibility and relevance to natural weathering have not been fully assessed. In a comparative study, the latest model of Q-Sun chamber (Q-Panel) was found not to meet the requirements of certain American automotive standards.25 Atlas manufactures the Suntest devices, geometrically similar to the Q-Sun and using the same light source. A general remark is relevant: the surface temperature of an object exposed to light depends not only on the temperature of the surrounding atmosphere but also on both its colour and its chemical composition.26 In both types of devices described above, the actual surface temperature of the exposed samples is unknown, though lower than the one indicated by a reference black specimen (referred to as black standard thermometer or black panel thermometer, placed in the same position as the samples), which by definition absorbs all radiation in the visible range. Conversely, the surface temperature of a white specimen, which scatters most of the incident visible light, can be up to 20 °C lower than that of a reference black specimen of which surface temperature is 70 °C.27 The use of xenon arc light sources is prescribed by several standards.28
Figure 4.1: Close-up of a Xenotest reaction chamber with the vertical lamp positioned in the middle of a rotor with sample holders (one in place). In Xenotest, the power setting refers to the wavelengths range 290-400 nm (or 320-400 nm if light is filtered to simulate daylight behind glass), the emission in this region is roughly 10% of the range 290-800 nm.
48 Ageing and stabilisation of paper Chapter 4: Experimental techniques in studies of photo-stability
Figure 4.2: The reaction chamber of a Suntest unit with a static sample in horizontal position under a xenon lamp. In Suntest, the power setting refers to the wavelength range 290-800 nm (or 320-800 nm with a glass-filtered light).
Mercury vapour arc light source This unit uses a medium pressure mercury vapour arc light filtered with borosilicate glass. It was developed in the mid 1970s by the Service d’Etude du Photovieillissement Acceléré des Polymères (SEPAP) at the Laboratoire de Photochimie Moléculaire et Macromoléculaire in Clermont-Ferrand, France, as a scientific laboratory tool aimed at studying the photooxidation mechanisms of a variety of synthetic polymers.29 The concept is based on a scientific approach to the processes involved during the weathering of organic materials at the molecular scale, described in detail later in this Chapter.
Figure 4.3: In the SEPAP chamber, four mercury vapour lamps are positioned vertically around a rotating sample-holder. The chamber consists of a central rotating sample holder accommodating up to 24 samples, with four lamps located in each corner of the chamber in geometrically equivalent positions (Figure 4.3). A thermosensitive resistor placed behind a film of the
Ageing and stabilisation of paper 49 Chapter 4: Experimental techniques in studies of photo-stability same colour and ideally the same chemical composition as the samples enables the determination of surface temperature of the exposed samples. In the latest model, the accuracy of a few tenths of degree centigrade is reached (e.g. 0.3 °C at 60 °C). Although the spectral distribution of a filtered mercury source is discrete (light is emitted as lines), the relevance of this type of accelerated photoageing to natural weathering has been demonstrated in numerous scientific publications.30 The method is covered by several standards.31 A variation of the instrument allows samples to be soaked in an aerated water bath during irradiation. An 8-lamp version enables us to conduct ultra-accelerated tests under controlled conditions. This version is used for fundamental photochemical studies or as a rapid way to test photo-resistant or heavily stabilized polymers.32 The brand names are SEPAP 12/24 (Atlas MTT BV), SEPAP 12/24 H (with water bath, Atlas MTT BV) and SEPAP 50/24 (ultra-accelerated version, not commercially available).
Fluorescent UV-lamps The characteristic spectral distribution of this source is Gaussian shaped, most of its energy being emitted in the UV range.33 Two types of fluorescent lamps are used. The UV-B type lamps emit a large percentage of light in the UV-B range, defined by the International Commission of Illumination (CIE) as the portion of the electromagnetic field lying between 280 and 315 nm,34 i.e. at wavelengths shorter than 295 nm (daylight cut-on). This spectral distribution makes this source irrelevant for tests of weathering.35,36 The American Society for Testing Materials (ASTM) even includes the following comment among the “non-mandatory information” of the Standard Method G53,37 describing UV-A and UV-B accelerated tests: “All UV-B lamps emit UV below the normal sunlight cut-on. This short wavelength UV can produce rapid polymer degradation and often causes degradation mechanisms that do not occur when materials are exposed to daylight. This may lead to anomalous results.”17 The UV-A type tubes do not emit light below 295 nm. However, the significant lack of emission in the visible range represents a drawback for the use of this source. As explained later in the Chapter, the visible range of the daylight spectrum accounts for fading of dyes and also for heat production, which is more or less important depending on colour of the object. Heat production increases the rate of photooxidation reactions. The quasi-absence of such wavelengths in the emission spectrum of fluorescent UV lamps may lead to results that do not correlate to what occurs during natural weathering. In natural weathering, photooxidation occurs at different rates for different colours. In contrast, UV lamps may cause photooxidation to occur at the same rate for different colours. Therefore, these lamps cannot be used to accurately simulate daylight. Both UV devices can be used in tests that vary day/night cycles, temperature (black panel), condensing humidity, water sprays and irradiance control. No motorized rotation of either the samples or the light is available. The brand names are QUV-A, QUV-B (Q-Panel), UV- CON (Atlas, discontinued), UV2000 (Atlas). The use of these light sources is also covered by standards.38
50 Ageing and stabilisation of paper Chapter 4: Experimental techniques in studies of photo-stability
Metal halide source The use of metal halides provides a variety of individual lines and a continuum of emission over a wide spectral range.21 Only stabilised, controlled power supplies and well-designed lamps offer appropriate operating conditions with a stable spectral power distribution. Consequently, only a few types of metal halide lamps, after their electrical and optical characteristics have been carefully checked and their power spectral distribution individually measured are suitable for solar simulation systems. Such lamps are referred as Metal Halide Global (MHG) lamps. When suitably filtered, the emission spectrum of MHG lamps is very similar to the daylight spectrum over the entire wavelength range. Because of their high efficiency, MHG lamps are ideally suited for use in large-scale chambers using several strategically placed radiation units to provide a uniform exposure. Large objects made from several components or materials like entire vehicles and even full railway coaches can be tested. Some smaller units have also been produced for testing medium-sized components and allow irradiance, relative humidity, and chamber and black panel (or black standard) temperature control. The main brand names are SolarConstant (largest units), SC 2000 and SC 600 Solar Simulation Chamber (Atlas). The use of these lamps is prescribed by several standards.39
Integrating-sphere-based UV chamber In 1998, NIST (National Institute of Standards and Technology, USA) scientists developed an accelerated artificial photoageing chamber based on the use of integrating spheres.26,40 According to the authors, this new device “appears to be capable of mitigating known sources of systematic errors”, by comparison with other accelerated photoageing devices.41 The device is designed to run simultaneously and independently several accelerated tests with different levels of stress during each test. The nature of the polychromatic light source is not specified but the authors stress the fact that the spectral distribution of the light is not necessarily required to match that of the daylight, provided no radiation lower than the daylight cut-on (295 5 nm) is emitted. Yet, removal of most of the visible and infrared part of the light entering the sphere can be criticized for the same reasons than those stated for the use of fluorescent UV lamps. Presently, this device, still in a development and testing phase,42 is not accredited by any standard.
Concentrated daylight This source of light is used in Fresnel Solar Concentrators that were originally developed in the 1960s, at the DSET Laboratories, in Arizona. The accelerated weathering devices, built to operate in the Arizona desert exclusively, employ ten highly reflective flat mirrors to concentrate direct sunlight onto samples mounted on a target board. The concentration factor was reported to be around 5.6 in the 315-400 nm range, showing a tendency to increase with increasing wavelengths.43 As the apparatus is tracking the sun, the amount of direct sunlight received is maximised. The temperature is monitored by a black panel positioned next to the samples on the target board and irradiance is measured via a radiometer mounted on the device.
Ageing and stabilisation of paper 51 Chapter 4: Experimental techniques in studies of photo-stability
Several studies refer to the use of this type of apparatus35,44 which has been acknowledged by some as the most relevant to sub-tropical and desert climate in terms of induced chemical changes on automotive coatings.45,46 However, although the 142 cm long and 15 cm wide target board on which samples are mounted lies under an air tunnel, along which an air deflector directs laminar air, the concentration of visible and infrared wavelengths brings into question the uniformity and control of the temperature occurring at the surface of samples. This is especially true of coloured samples or samples known to discolour during ageing, e.g. PVC or polyamide.47,48 Furthermore, no data is given by the manufacturer about the concentration factor in the most harmful part of the UV daylight range between 295 and 315 nm. The main brand names are EMMA (Equatorial Mount with Mirrors for Acceleration), EMMAQUA (EMMA with water spray; Atlas), Q-Trac (Q-Panel). The technology is covered by several standards.49 In 1998, scientists from the National Renewable Energy Laboratory set up an ultra- accelerated UV concentrator device, consisting of an array of faceted mirrors concentrating up to 100-fold the UV range of daylight while reflecting much less of the visible and near infrared radiations.50 Its developers claim that 10 years of outdoor weathering can be simulated accurately in just two months with this device.51 As explained in the next section, such intense UV radiation may bring about two major problems with reproduction of natural effects of daylight: firstly, the occurrence of biphotonic processes, possibly causing unrealistic photochemical reactions to occur. Secondly, the problem of oxygen depletion in the inner layers of the exposed material, encountered in all accelerated tests, is likely to be highly exaggerated. This can lead to a depth of photooxidation that is too shallow and thus does not mimic the degree of loss of mechanical properties if compared with natural weathering.
4.3 Acceleration of photooxidation reactions Introduction Although 95% of polymer weathering and accelerated ageing research is still based on an empirical approach developed in the early 1950s, a more scientific approach based on chemical changes during ageing is gaining increasing support. The empirical, or conventional, approach is associated not only with the application of physical and chemical stresses at a level as close as possible to the one occurring in real conditions, but also with the definition of various degradation criteria (degradation of mechanical, physical and visual properties, e.g. tensile strength, colour, electrical conductivity or permeability).3,4,20,52,53 Therefore, in this approach, the polymeric material is considered as a macroscopic system presenting functional properties (mechanical, electrical, optical etc.) and degradation due to ageing can be seen as the deterioration of those functional properties.54 However, since the early seventies, it has been recognized that most polymeric materials, subjected to environmental stress under normal conditions of use, should, in fact, be seen as photochemical reactors. Chemical ageing is the controlling route of deterioration and “mechanisms” of the complex chemical phenomena accounting for physical deterioration
52 Ageing and stabilisation of paper Chapter 4: Experimental techniques in studies of photo-stability should be identified at the molecular or supramolecular scale.20,55 In the scientific approach (also called the mechanistic approach) the degradation mechanism of a polymeric material is examined in its final formulation and after every processing operation.56,57 The studied samples should therefore be taken from the surface of the ready-to-use systems.
Chemical photoageing at the molecular scale Chemical ageing, i.e. ageing during which chemical modifications take place due to environmental stress, can occur via the following routes: ― Photochemical oxidation of the matrix, i.e. oxidation due to conjugated effects of light, heat and atmospheric oxygen; ― Non-oxidative photolysis of the matrix, i.e. transformation due to the conjugated effects of light and heat; ― Hydrolysis of oxidation products, i.e. transformation due to the conjugated effect of heat, atmospheric oxygen and water (in the gaseous or liquid state); ― Transformation of additives (stabilizers, dyes and pigments, anti-fog and anti-static agents etc.). The relative importance of each specific process should be established by analytic techniques, especially by means of in-situ spectrophotometric techniques carried out in the solid state.58 Each route has one or more macroscopic consequences: ― Photochemical oxidation of the matrix (and hydrolysis if it occurs) leads to chain scissions which directly affect the mechanical properties. Micro-cracks may occur from such chain scissions, resulting in loss of gloss or decrease of brightness, thereby affecting optical properties;5,59,60 ― Photochemical oxidation and/or non-oxidative photolysis of the matrix induces formation of compounds leading to yellowing (yellowing through absorption) and their subsequent bleaching, as well as formation of fluorescent substances (yellowing through emission) and secondary photooxidation of the fluorescent substances, therefore affecting the visual properties; 7,58,61 ― Transformation of additives affects the efficiency of the additives’ functions, and results in optical changes if additives are coloured compounds (fading) or precursors of substances leading to yellowing (discolouration).62,63 The relative importance of each specific process should be established using appropriate analytical techniques as each route has its own macroscopic consequences. However, several remarks need to be stressed: a. Combined effect of light and heat in the presence of atmospheric oxygen. It is well known that any oxidation mechanism involves photochemical and thermal elementary processes. Photochemical activation under polychromatic light and thermal activation (even at room temperature) cannot be examined separately. The oxidation rate, i.e. the rate of formation of any final inert oxidised group on the polymeric chain, is given by:
Ageing and stabilisation of paper 53 Chapter 4: Experimental techniques in studies of photo-stability
E 1 a RT ox a )(Rate eI , (eq 4.1) 2
where 1 and 2 are the limiting wavelengths of radiations inducing the elementary
process, Ia the absorbed light intensity at the wavelength , Ea the apparent activation energy (if existing), and T is the absolute temperature. The parameter is not
temperature independent and Ea is not independent on Ia. This leads to the conclusion that the two types of activation should be controlled simultaneously during artificial accelerated tests in order to ensure the reliability of results: the absorbed light intensity and the temperature of the exposed surface. b. The complexity of the effect of water. In environmental conditions, water may have various physical and chemical effects. Liquid water may lead to the removal of oxidized matter, swelling of matrices and extraction of additives soluble in water.64,65 Liquid or gaseous water may also induce hydrolysis of intermediates and/or final oxidation products.66 During artificial ageing in one single chamber, such perturbations can not be induced simultaneously without significant deviation from natural weathering conditions.67,68 Depending on the type of polymer, on the nature of stabilizers (and especially their solubility in water) and on the climatic parameters, the relative importance of different processes may be anticipated. The experimental conditions should be selected so that the most important effect of water is determined, i.e. the one which controls the overall process of chemical ageing.
Conditions for relevance of accelerated photoageing The chemical ageing observed in artificial accelerated ageing or in natural accelerated ageing should involve the same “mechanism” as in natural weathering. In the next section, we will stipulate what should be understood by the same mechanism. The physical and chemical stresses applied on a polymeric material in artificial or natural accelerated ageing should conform with the following conditions:20 a. Spatial uniformity of light exposure should be ensured by sample movement. If static, the stability of light intensity and of spectral distribution should be checked in different positions on the surface of the exposed object. Considering that the usual geometry of light sources is cylindrical or spherical, it is difficult to ensure the uniformity of the incident light over a large flat area.26 b. The incident light should not include radiation of wavelength shorter than 295 5 nm. Considering fundamental photochemistry laws, such radiation may induce photochemical reactions through excitation of electronic states otherwise not achievable with daylight. For instance, when aromatic polyurethanes are exposed to daylight, they are photooxidized into highly conjugated quinones, which lead to pronounced yellowing. Irradiation of such aromatic polymers with wavelengths shorter than 295 nm has been shown to inhibit the discoloration that occurs in natural conditions.69 Additives are also prone to undergo photochemical reactions which are not relevant for ageing of the material itself. The colour of any organic dye is a consequence of the fact that the wavelength of the corresponding complementary
54 Ageing and stabilisation of paper Chapter 4: Experimental techniques in studies of photo-stability
colour provides enough energy to excite its electrons. UV absorbers absorb radiation in the daylight UV range (295-400 nm) and dissipate it through a radiationless process.64 Excitation to more energetic electronic states through absorption of shorter wavelengths might potentially lead to photo-scission or photo-bleaching.26,70 c. Polychromatic light is generally used for practical reasons and its spectral distribution is not as important as it may seem. It is possible to use a source emitting a continuum (e.g. xenon lamp) or discrete lines (e.g. medium pressure or high pressure mercury arc lamps) if correct filtration is ensured. According to the fundamental concepts of photochemistry, especially the very fast relaxation of excited vibrational states after excitation occurring in excited electronic states in the condensed phase,71,72 should allow the use of medium-pressure mercury arcs. A study led on bis-phenol A polycarbonate has demonstrated that a mercury arc could induce chemical evolutions of PC more relevant to natural ageing than a xenon source.73 Excitation by light in the visible range is detrimental only in very particular cases in coloured materials (photooxidation due to singlet oxygen).74,75 d. The intensity of incident light should be high enough to maintain photolysis and photooxidation as controlling degradation routes over thermo-oxidation. If the intensity of incident light is low (e.g. when simulating daylight), temperature of the exposed surface should be kept low enough to prevent prevalence of thermo- oxidation as the degradation pathway. On the other hand, the use of very intense light sources such as a high-powered pulsed laser or a UV concentrator, concentrating the UV daylight up to 100 times50 must be avoided, as the nearly simultaneous absorption of two photons may occur.76 This so-called biphotonic process is equivalent to the absorption of one photon of twice the energy of the incident photons. For instance, irradiating an object with a intense laser source emitting at 500 nm may provoke the same photochemical processes as those that would be activated by radiations of 250 nm. This effect is used in the biology field as an imaging microscopy technique.77 e. Photooxidation of the surface is rarely affected by oxygen depletion in environmental conditions.78 Therefore, care should be taken to ensure that this effect does not limit the rate of photooxidation during accelerated artificial ageing.79,80 For instance, in accelerated ageing of “transparent” polymers which are not highly UV absorbing, if the permeability of the polymer is not high enough, oxidation occurs only in the surface layers whereas the core of the sample remains practically non-oxidised. In order to reduce this effect, the geometry of an exposed sample should be carefully designed (excessive thickness will lead to oxygen depletion, excessive thinness will lead to a fast loss of stabilizers). The sample geometric design should be suited to exposure requirements and not to the analytical methodology used as the criterion of degradation (e.g. mechanical changes). f. In environmental conditions, as indicated above, oxidation is activated both photochemically and thermally so that during accelerated tests the temperature of a light-exposed surface should be determined within 1 °C. For instance, (eq 4.1) enables us to calculate that hydroperoxides formed in polypropylene during an accelerated test decompose 1.6 times faster at 64 °C than they do at 60 °C, using the value of activation energy for their decomposition found in the literature.81 This illustrates the importance of temperature control for improving the repeatability of an
Ageing and stabilisation of paper 55 Chapter 4: Experimental techniques in studies of photo-stability
accelerated test. Temperature control using an external reference black specimen is only indicative and relates to the temperature at the surface of this specimen only, as the temperature of an illuminated body is partly determined by its colour.26 g. A correct evaluation of the effect of water (liquid or gas) is difficult due to the variety of physical and chemical processes in which water can be involved. Experimental design should be based on the best understanding of the prevalent effect of water: ― If the predominant effect of water is found to be erosion, sprinkling is relevant; ― If the predominant effect of water is found to be swelling and extraction of stabilizers, sprinkling should be limited to prevent exaggerated swelling or extraction; ― If the predominant effect is found to be hydrolysis of final photooxidation products, post-immersion of the exposed samples in water is enough to lead to hydrolysis and to induce changes in optical properties; ― If water is able to react with intermediate degradation products, a combined effect of light, heat, oxygen and water may be anticipated. The light-exposed sample should be in contact with a thin water film supplying oxygen (which is dissolved in a low concentration in water) while the film is re-saturated with oxygen outside the exposure chamber. Regulation of the sample temperature should be ensured by circulating thermo-regulated water (e.g. in the SEPAP 12/24 H chamber). However, in two common situations, it is possible to evaluate the effect without having to supply it additionally: Case 1: we rely on water being formed from hydroxyl radicals, which are the primary photooxidation products in many materials.4 Hydroperoxides, which are thermally and photochemically unstable, are dissociated and in the process, HO radicals are formed. Such radicals are extremely reactive and stabilize by abstracting any neighbouring hydrogen atom, thereby forming water. Water is thus supplied simultaneously with the hydroperoxides scission and in the same reaction zones as photooxidation intermediates. The intermediates are sometimes easily hydrolysed, e.g. imides in polyamide (as illustrated below) and this reaction proceeds with water which is formed internally in the material:66
Case 2: the effect can be evaluated analytically from kinetics of formation of an intermediate product (e.g. using appropriate spectrophotometric techniques), or from accumulation of compounds into which it is converted.82 In such a case, application of water would be required only if loss of a physical property had to be evaluated, which depends on a complete hydrolysis of the intermediate product.
56 Ageing and stabilisation of paper Chapter 4: Experimental techniques in studies of photo-stability
Control of relevance based on chemical analysis In the previous sections, we emphasized that the relevant character of data collected during accelerated ageing could be anticipated if the identified chemical reaction pathways are the same as during natural ageing. Yet, what is to be understood by reaction pathway? The most useful way of describing polymer photooxidation would be the following:4,82,83 ― The first sequence involves the formation of hydroperoxides. This formation proceeds as a chain reaction, the actual length of which is difficult to determine. Photo-initiation could occur via chromophores (species leading to formation of radicals through absorption of light) located on the macromolecular chain or via other absorbing species which are able to supply radicals in a photochemical way. ― The second sequence involves thermal or photochemical decomposition of hydroperoxides into intermediate products. This can be either a one-step reaction or a multi-step reaction (if unstable intermediate products decompose spontaneously into other intermediate products). ― The third sequence involves photochemical transformation of intermediate products and the accumulation of final inert products in the polymeric matrix. In the early 1970s, an approach to investigate the photooxidation mechanisms in several categories of polymers was developed. This approach is based on a useful description of the photooxidation reaction pathway rather than on the analysis of reactive species such as excited states, radicals or vibrationally excited products. It consists of the following steps:84,85 a. Identification of the most important intermediates, such as hydroperoxides, and the photochemically or thermally unstable decomposition products of hydroperoxides. b. Evaluation of transformation pathways of the main intermediate products (photochemical, thermal, hydrolytic etc.). c. Identification of the dominant final inert products which accumulate in the matrix, including the so-called critical photoproduct, which can be defined on the basis of the following criteria:86,87 ― It should be formed in the main degradation pathway and should therefore be formed as the consequence of a chain scission process; ― It should be the final and stable product which accumulates in the matrix. Therefore, it cannot be a product leading to yellowing, and should not be a low molecular weight product; ― It should accumulate in a linear fashion with exposure time until a complete physical destruction of the polymeric material. Since the physical deterioration of a polymeric material occurs after a very low extent of chemical evolution, the critical photoproduct is formed during the very early stages of photooxidation; ― It should be selected according to the best understanding of the degradation pathway. For practical reasons, our understanding of a degradation pathway can thus be based on identification of the final products during natural and artificial ageing. The approach is safe when the ageing mechanism is thoroughly understood (based on long-term analytical
Ageing and stabilisation of paper 57 Chapter 4: Experimental techniques in studies of photo-stability studies in laboratory conditions).87,88 In accelerated conditions, the material’s lifetime is then determined on the basis of an established correlation between the loss of a functional property of interest and accumulation of the critical photoproduct. Prediction of the lifetime in natural conditions is then derived using the so-called acceleration factor which is defined as the ratio between the accumulation rate of the critical photoproduct in accelerated ageing and the one in natural ageing.86,89 For obvious reasons, accelerated ageing should be performed in a time span much shorter than the material’s lifetime in natural conditions (e.g. during one summer). If a substance’s lifetime is defined on the basis of a physical property, which depends on oxidation, prediction is difficult, although not impossible.4 If the lifetime is based on a change of visual properties, the prediction is far more complex if it is controlled by more than one simple dynamic chemical process.90 If the visual properties are dominated by loss of gloss or discolouration (implying the occurrence of surface micro-cracks) then oxidation is the dominating mechanism and lifetime prediction is possible.91 However, if the change of visual properties is a consequence of several simultaneous processes, such as oxidation, formation and bleaching of products leading to discolouration (yellowing), formation and photooxidation of fluorophores, decomposition of pigments or dyes, then lifetime prediction is only possible if one of these processes is dominant.
4.4 Conclusions A recent review about service life of exterior wood coatings gives a fairly good picture of the current situation in the field of accelerated light ageing tests in the polymer industry in its wish to control the durability of products.92 Although it is widely recognized that more efforts should be put into understanding the underlying chemical processes occurring at the molecular scale, responsible for the macroscopic evolution of an object, durability studies still rely on correlating the evolution of physical properties between natural weathering and accelerated photoageing, like colour or gloss measurement.93 The photochemical degradation of macromolecules in general and of paper more specifically, appears to be fairly complex. Variations of physical properties and of chemical composition can easily be observed during photo-degradation in ambient conditions. However since long periods of time are involved, the variations in environmental parameters (temperature, humidity, light quality, light intensity) are not accurately known. At the laboratory level, a second difficulty is encountered if light-ageing of paper is performed in accelerated conditions. Each photoageing unit has its own specific characteristics depending on the light source used and the type of temperature and humidity control. The relevance of accelerated laboratory experiments to real ageing conditions cannot be reliably assessed on the variation of macroscopic properties like mechanical properties or aspect changes since the same variation can have several different origins. Reliability is gained when relevance is assessed by a more detailed survey of the various chemical evolutions i.e. oxidation, formation and bleaching of absorbing yellowing moieties, formation and photooxidation of fluorescent species. Due to our limited understanding of chemical changes at the molecular level, the control of relevance of accelerated light ageing tests is still difficult and more knowledge of paper photooxidation mechanisms is needed.
58 Ageing and stabilisation of paper Chapter 4: Experimental techniques in studies of photo-stability
4.5 References 1. J.F. McKellar, N.S. Allen, Photochemistry of Man-Made Polymers, Elsevier, London, 1979. 2. N. Grassie, Developments in Polymer Degradation, Vol. 1-7, Elsevier, London, 1978-1987. 3. R.B. Seymour, Long-Term Environmental Factors, in: C. Dostal (Ed.), Engineered Materials Handbook, Vol. 2: Engineering Plastics, ASM International, Materials Park, 1988, 423-432. 4. R.L. Feller, Accelerated Aging in Conservation Science, Getty Conservation Institute, Los Angeles, 1994. 5. G. Wypych (Ed.), Handbook of Material Weathering, 2nd edition, ChemTec Publishing, 1995. 6. S. Michalski, Damage to Museum Objects by Visible Radiation (light) and Ultraviolet Radiation (UV), Preprints of the Bristol Conference on Lighting in Museums, Galleries and Historic Houses, The Museums Association, 1987, 3-16. 7. A.L. Andrady, N.D. Searle, Photoyellowing of Mechanical Pulps, Part 2: Activation Spectra for Light- induced Yellowing of Newsprint Paper by Polychromatic Radiation, Tappi J., 1995, 78, 131-138. 8. B. Lavédrine, The Pink-Blue Scale: a New Light Dosimeter for the Exhibition of Photographs and Sensitive Artefacts, in: S. Clark (Ed.), Care of photographic, moving image and sound collections, Conference papers, July 20-24 1998, Institute of Paper Conservation, York, 1999, 124-128. 9. J. Ashley-Smith, A. Derbyshire, B. Pretzel, The Continuing Development of a Practical Lighting Policy for Works of Art on Paper and Other Object Types at the Victoria and Albert Museum, 13th Triennial Meeting Rio de Janeiro Preprints, ICOM Committee for Conservation, Vol. 1, 2002, 3-8. 10. K. Hallet, S. Bradley, Ultra-Violet Filtered Lighting and Cellulose Degradation: Evaluating the Effect of Light Exposure on Ethnographic Collections, The Conservator, 2003, 27, 3-12. 11. S.B. Lee, J. Bogaard, R.L. Feller, Damaging Effects of Visible and Near-Ultraviolet Radiation on Paper, in: S.H. Zeronian, H.L. Needles (Eds.), Historic Textile and Paper Materials II, Conservation and Characterization, ACS Symp. Ser. 410, Americal Chemical Society, Washington, 1989, 54-62. 12. P. Mole, The Evolution of Arc Broadside Lighting Equipment, Journal of the Society of Motion Picture Engineers, 1939, 32, 398-411. 13. C.W. Handley, History of Motion Picture Studio Lighting, Journal of the Society of Motion Picture and Television Engineers, 1954, 63, 129-133. 14. Atlas MTS Weathering Testing Guidebook, Ed. Atlas MTS, Chicago, 2001, 56. 15. H.A. Nelson, Proceedings of the ASTM, 1922, 33, 485. 16. http://www.bildanalytik.de/products/natural-weathering-testing-new/sites/, accessed 02/12/2004. 17. J.A. Chess, D.A. Cocuzzi, G.N. Van de Streek, Accelerated Weathering: Science, Pseudo-Science or Superstition?, in: D. R. Bauer, J.W. Martin (Eds.), Service Life Prediction of Organic Coatings: A Systems Approach, Oxford University Press, Washington, 1999, 130-147. 18. http://www.bildanalytik.de/products/natural-weathering-testing- new/sites/florida_static_weathering.shtml, accessed 02/12/2004. 19. For direct exposures: ISO 877 - Plastics - Methods of Exposure to Direct Weathering Using Glass-Filtered Daylight, and to Intensified Weathering by Daylight Using Fresnel Mirrors; ISO 2810 - Paints and Varnishes - Natural Weathering of Coatings: Exposure and Assessment; ISO 105-B03 - Textiles - Tests For Colourfastness - Colourfastness to Weathering: Outdoor Exposure; ASTM G7 - Recommended Practice for Environmental Exposure Testing on Nonmetallic Materials. For indirect exposures: ISO 877 - Plastics (full description above), ISO 2810 - Paints and Varnishes (full description above); ISO 105-B01 - Textiles - Tests For Colourfastness - Colourfastness to Light: Daylight; AATCC 111 - Weather Resistance: Exposure to Natural Light and Weather Through Glass; ASTM G24 - Standard Practice for Conducting Exposures to Daylight Filtered Through Glass. For black box exposures: ASTM Standard D4141 Method A - Standard Practice for Conducting Accelerated Outdoor Exposure Tests of Coatings. 20. J. Lemaire, Predicting Polymer Durability, Chemtech, 1996, 10, 42-47. 21. K.P. Scott, Shedding Some Numbers on Light, SunSpots, 2001, 64, 1-7. 22. Atlas MTS Weathering Testing Guidebook, Chicago, 2001, 56-57.
Ageing and stabilisation of paper 59 Chapter 4: Experimental techniques in studies of photo-stability
23. ISO 4892-4:2004 - Plastics - Methods of Exposure to Laboratory Light Sources - Part 4: Open- Flame Carbon-Arc Lamps; ASTM G152 - Standard Practice for Operating Open Flame Carbon Arc Light Apparatus for Exposure of Nonmetallic Materials; ASTM G153 - Standard Practice for Operating Enclosed Carbon Arc Light Apparatus for Exposure of Nonmetallic Materials; ASTM G23 - Practice for Operating Light-Exposure Apparatus (Carbon Arc Type) With and Without Water for Exposure of Nonmetallic Materials. 24. Klukowska A. et al., Photochromic Hybrid Sol-Gel Coatings: Preparation, Properties and Applications, Mater. Sci., 2002, 20, 95-104; A. Boudina, C. Emmelin, A. Baahiouamer, M.F. Grenier- Loustalot, J.M. Chovelon, Photochemical Behaviour of Carbendazim in Aqueous Solution, Chemosphere, 2003, 50, 649-655; L. Scrano, S.A. Bufo, T.R.I. Cataldi, T.A. Albanis, Surface Retention and Photochemical Reactivity of the Diphenylether Herbicide Oxyfluorfen, J. Environ. Qual., 2004, 33, 605-611. 25. K.P. Scott, A Comparison of Laboratory Instruments, SunSpots, 2004, 71, 1-15. 26. J.W. Martin, J.W. Chin, W.E. Byrd, E. Embree, K.M. Kraft, An Integrating Sphere-Based Ultraviolet Exposure Chamber Design for the Photo-degradation of Polymeric Materials, Polym. Deg. Stab., 1999, 63, 297-304. 27. R.M. Fischer, W.D. Ketola, Surface Temperatures of Materials in Exterior Exposures and Artificial Accelerated Tests, in: W.D. Ketola, D. Grossman (Eds.), Accelerated and Outdoor Durability Testing of Organic Materials, American Society for Testing and Materials, ASTM STP 1202, Vol. 120, Philadelphia, PA, 1994, 88-111. 28. ISO 4892-2:1994 - Plastics - Methods of Exposure to Laboratory Light Sources - Part 2: Xenon- Arc Sources; ISO 11341:2004 - Paints and Varnishes - Artificial Weathering and Exposure to Artificial Radiation - Exposure to Filtered Xenon-Arc Radiation; ISO 105-B02:1994 - Textiles - Tests for Colour Fastness - Part B02: Colour Fastness to Artificial Light: Xenon Arc Fading Lamp Test; ASTM G155-04 - Standard Practice for Operating Xenon Arc Light Apparatus for Exposure of Nonmetallic Materials. 29. P. Laurenson, J. Quemner, G. Roche, R. Arnaud, J. Lemaire, Photovieillissement et Environnement : Nouveau Dispositif de Photovieillissement Accéléré et Elaboration d’Isolants de Couleur Photo-Stables, Revue Générale d’Electricité, 1979, 88, 685-689. 30. J. Lemaire, The Photocatalyzed Oxidation of Polyamides and Polyolefins, Pure and Appl. Chem., 1982, 54, 1667-1682; A. Rivaton, D. Sallet, J. Lemaire, The Photochemistry of Bisphenol A Polycarbonate Reconsidered, Polym. Photochem., 1983, 3, 463-481; A. Gazel, P. Laurenson, G. Roche, J. Lemaire, Photooxidation of Silane Crosslinked Polyethylene, Makromol. Chem., Rapid Comm., 1985, 6, 81-85; E. Fanton, B. Athenor, H. Seinera, R. Arnaud, J. Lemaire, Vieillissement naturel et Photovieillissement Accéléré de Films Thermiques, Caoutchoucs et Plastiques, 1986, 7, 135-139; J.-L. Gardette, J. Lemaire, Prediction of Long-Term Outdoor Weathering of Poly(vinyl)chloride, J. Vinyl Technol., 1993, 15, 113-117; D. Fromageot, J. Lemaire, Long-Term Photo-Ageing of Soluble Polyamides Used in Conservation or in Modern Sculptures, Polym. Deg. Stab., 1994, 45, 39-45; J. Lemaire, J.-L. Gardette, J. Lacoste, P. Delprat, D. Vaillant, Mechanisms of Photooxidation of Polyolefins: Prediction of Lifetime in Weathering Conditions, Adv. Chem. Ser., 1995, 249, 577-598; P.N. Thanki, R.P. Singh, Photo-Oxidative Degradation of Nylon 66 Under Accelerated Weathering, Polymer, 1998, 39, 3363-3367; J. Mallegol, J. Lemaire, J.-L. Gardette, Durability of Oil-Based Varnishes and Paints: III. Photo- and Thermo-Oxidation of Cured Linseed Oils, J. Am. Oil Chem. Soc., 2000, 77, 257- 263; S. Bonhomme, A. Cuer, A.-M. Delort, J. Lemaire, M. Sancelme. G. Scott, Environmental Biodegradation of Polyethylene, Polym. Deg. Stab., 2003, 81, 441-452. 31. U51-159 - Drainage agricole - Tubes annelés en polychlorure de vinyle non plastifié - Essai de vieillissement accéléré, September 1990; XP T54-194 - Plastiques - Films pour ensilage et enrubannage - Spécifications et méthodes d’essai, December 1995; NF EN 13206 - Films thermoplastiques de couverture pour utilisation en agriculture et horticulture, January 2002; XP T-51 830 - Plastiques - Méthode d’exposition à des sources lumineuses de laboratoire - Lampe à vapeur de mercure moyenne pression, September 2002. 32. J. Lemaire, Relevancy of Accelerated and Ultra-Accelerated Photo-Ageing: The Example of Stabilised Polypropylene, Proceedings of the American Chemical Society Division of Polymeric Materials: Science and Engineering, Vol. 83, August 20-24 2000, Washington DC, 2000, 140. 33. Atlas MTS Weathering Testing Guidebook, Ed. Atlas MTS, Chicago, 2001, 50.
60 Ageing and stabilisation of paper Chapter 4: Experimental techniques in studies of photo-stability
34. A.F. McKinlay, B.L. Diffey, A Reference Action Spectrum for Ultraviolet Induced Erythema in Human Skin, CIE-Journal, 1987, 6, 17-22. 35. D.A. Cocuzzi, An Interim Report on the Subject of Accelerated Weathering: A Five-Year Update to ASTM’s D01.53 Ten-Year Exposure Study, Proceedings ot the National Coil Coating Association, Fall meeting, 2001. 36. D.R. Bauer, M.C. Peck, R.O. Carter, Evaluation of Accelerated Weathering Tests for a Polyester-Urethane Coating Using Photoacoustic Infrared Spectroscopy, Journal of Coatings Technology, 1987, 59, 103-109. 37. ASTM G53-96 - Practice for Operating Light- and Water-Exposure Apparatus (Fluorescent UV-Condensation Type) for Exposure of Nonmetallic Materials, withdrawn in 2000. 38. ISO 4892-3:1994 - Plastics - Methods of Exposure to Laboratory Light Sources - Part 3: Fluorescent UV Lamps; ISO 11507:1997 - Paints and Varnishes - Exposure of Coatings to Artificial Weathering - Exposure to Fluorescent UV and Water; ASTM G154-00ae1 - Standard Practice for Operating Fluorescent Light Apparatus for UV Exposure of Nonmetallic Materials. 39. DIN 75220 - Ageing of Automotive Components in Solar Simulation Units; IEC 60068-2-9 - Ed. 1.0 - Environmental Testing - Part 2: Tests. Guidance for Solar Radiation Testing; MIL- STD-810 - Environmental Test Methods and Engineering Guidelines. 40. J.W. Chin, J. Martin, E.J. Embree, W.E. Byrd, Use of Integrating Spheres as Uniform Sources for Accelerated UV Weathering of Advanced Materials, Proceedings of the Polymeric Materials: Science and Engineering (PMSE) Fall Meeting., Vol. 83, August 20-24 2000, American Chemical Society (ACS), Washington DC, 2000, 145-146. 41. J.W. Chin, E. Byrd, N. Embree, J.W. Martin, J.D. Tate, Ultraviolet Chambers Based on Integrating Spheres for Use in Artificial Weathering, Journal of Coatings Technology, 2002, 74, 39-44. 42. A.W. Signor, M.R. Van Landingham, J.W. Chin, Effects of Ultraviolet Radiation Exposure on Vinyl Ester Resins: Characterization of Chemical, Physical and Mechanical Damage, Polym. Deg. Stab., 2002, 79, 359-368. 43. Atlas MTS Weathering Testing Guidebook, Atlas MTS, Chicago, 2001, 34. 44. H.K. Hardcastle, Fractional Factorial Approaches to EMMAQUA Experiments, in: J.W. Martin, D.R. Bauer (Eds.), Service Life Prediction Methodology and Metrologies, ACS Symp. Ser. 805, 2001, 63-88; C. Capanescu, C. Cincu, UV Inhibitors in Polyester Gelcoats, Adv. Polym. Techn., 2003, 22, 365-372; L.S. Crump, Evaluating the Durability of Gel Coatings Using Outdoor and Accelerated Weathering Techniques: A Correlation Study, Proceedings of the Composites Institute’s 51st Annual Conference and Exposition, Session 22-B, 1996. 45. J.L. Gerlock, A.V. Kuchervo, C.A. Smith, Paint Weathering Research at Ford, SunSpots, 2001, 31, 1- 11. 46. J.L. Gerlock, C.A. Peters, A.V. Kucherov, T. Misovski, C.M. Seubert, R.O. Carter, M.E. Nichols, Testing Accelerated Weathering Tests for Appropriate Weathering Chemistry: Ozone Filtered Xenon Arc, J. Coat. Technol., 2003, 75, 34-45. 47. J.-L. Gardette, J. Lemaire, Photothermal and Thermal Oxidations of Rigid, Plasticized and Pigmented Poly(Vinyl Chloride), Polym. Deg. Stab., 1991, 34, 135-167. 48. D. Fromageot, A. Roger, J. Lemaire, Thermo-Oxidative Yellowing of Aliphatic Polyamides, Angew. Makromol. Chem., 1989, 170, 71-85. 49. ISO 877:1994 - Plastics - Methods of Exposure to Direct Weathering, to Weathering Using Glass-Filtered Daylight, and to Intensified Weathering by Daylight Using Fresnel Mirrors; ASTM G90-98 - Standard Practice for Performing Accelerated Outdoor Weathering of Nonmetallic Materials Using Concentrated Natural Sunlight; ASTM D4141-01 - Standard Practice for Conducting Black Box and Solar Concentrating Exposures of Coatings; ASTM D4364-02 - Standard Practice for Performing Outdoor Accelerated Weathering Tests of Plastics Using Concentrated Sunlight; SAE J1961 - Accelerated Outdoor Exposure Using a Solar Fresnel Reflective Apparatus; ASTM D5722 - Standard Practice for Performing Accelerated Outdoor Weathering of Factory-Coated Embossed Hardboard Using Concentrated Natural Sunlight and a Soak-Freeze-Thaw Procedure. 50. G. Jorgensen, C. Bingham, D. King, A. Lewandowski, J. Netter, K. Terwilliger, K. Adamsons, Use of Uniformly Distributed Concentrated Sunlight for Highly Accelerated Testing of Coatings, in: D.R. Bauer, J.W. Martin (Eds.), Service Life Prediction Methodology and Metrologies, ACS
Ageing and stabilisation of paper 61 Chapter 4: Experimental techniques in studies of photo-stability
Symposium Series 805, American Chemical Society, Oxford University Press, Washington, DC, 2002, 100-118. 51. http://www.nrel.gov/buildings_thermal/pdfs/concen-fs.pdf, accessed 02/12/2004. 52. A.M. Emsley, G.C. Stevens, A Reassessment of Low Temperature Degradation of Cellulose in Power Transformers, Proceedings of the 16th International Conference on Dielectric materials: Materials, Measurements and Applications, London, Institute of Electrical Engineers, 1992, 229-232. 53. A. Geburtig, U. Schulz, The Action of Weathering on the Scratch Behaviour of Automotive Clearcoats, Confederation of European Environmental Engineering Societies 1st European Weathering Symposium, September 25-26, 2003, Prague. 54. G.J. Jorgensen, A Phenomenological Approach to Obtaining Correlations between Accelerated and Outdoor Exposure test results for Organic Materials, in: W.C. Golton (Ed.), New Directions in Coatings Performance Technology, ASTM STP 1435, ASTM International, West Conshohocken, 2003. 55. J. Lemaire, R. Arnaud, J.-L. Gardette, J.-M. Ginhac, L. Tang, E. Fanton, Vieillissement des Polymères: Empirisme ou Science ?, Rev. Gén. Caoutch. Plast., 1979, 593, 147-152. 56. J. Lemaire, R. Arnaud, Evaluation of the Additive Effects on the Photo-Ageing of Solid Polymers, Polym. Mater. Sci. Eng., 1988, 58, 329-333. 57. R. Arnaud, P. De Monte, Evolution du Caoutchouc Naturel au Cours de sa Transformation, Caoutch. Plast., 1991, 708, 95-100. 58. X. Jouan, C. Adam, D. Fromageot, J.-L. Gardette, J. Lemaire, Microscopic determinations of photoproducts profiles in photooxidized matrices, Polym. Deg. Stab., 1989, 25, 247-265; J. Lemaire, J.-L. Gardette, J. Lacoste, Use of Micro-FTIR Spectrophotometry in the Determination of Polymer Photo-Ageing Mechanisms, Makromol. Chem., Macromol. Symp., 1993, 70-71, 419-431; P. Delprat, J.-L. Gardette, Analysis of Photooxidation of Polymer Materials by Photoacoustic Fourier Transform Infra-Red Spectroscopy, Polymer, 1993, 34, 933-937; Y. Israeli, J. Lacoste, J. Lemaire, R.P. Singh, S. Sivaram, Photo- and Thermal-Initiated Oxidation of High Impact Poly(Styrene). I – Characterization by FT-IR Spectroscopy, J. Polym. Sci. A-Polym. Chem., 1994, 32, 485-93; C. Wilhem, J.-L. Gardette, Infrared Analysis of the Photochemical Behaviour of Segmented Polyurethanes: 1. Aliphatic poly(ester-urethane), Polymer, 1997, 38, 4019-4031; L. Gonon, O. Vasseur, J.-L. Gardette, Depth-Profiling of Photooxidized Stryrene-Isoprene copolymers by Photoacoustic and Micro Fourier Transform Infrared Spectroscopy, Appl. Spectrosc., 1999, 53, 157-163; I. Forsskåhl, C. Olkkonen, H. Tylli, Depth Profiling of a Photochemically Yellowed Paper. Part1: UV-Visible Reflectance and Fluorescence Spectroscopy, Appl. Spectr., 1995, 49, 92-97; C.A. Smith, J.L. Gerlock, R.O. Carter, Determination of Ultraviolet Light Absorber Longevity and Distribution in Automotive Paint Systems Using Ultraviolet Micro-Spectroscopy, Polym. Deg. Stab., 2001, 72, 89-97. 59. J.Ph. Brandon, R. Cabala, A. Chambaudet, F. Jaffiol, Surface Photodegradation of a Polyolefin Copolymer, Nucl. Inst. Meth. B, 1988, 32, 155-159. 60. A. Vishwa Prasad, R.P. Singh, Photooxidative Degradation of Styrenic Polymers: 13C-NMR and Morphological Changes upon Irradiation, J. Appl. Polym. Sci., 1998, 70, 637-645. 61. A. Rivaton, The Bisphenol-A Polycarbonate Dual Photochemistry, Angew. Makromol. Chem., 1994, 216, 147-153. 62. S. Commereuc, P. Lajoie, V. Verney, J. Lacoste, A New ESR Study of Hindred Amine Stabilisers (HAS) and Their Oxidation Products, Polym. Int., 2003, 52, 576-580. 63. L. Motz-Schalk, J. Lemaire, Photochemical and Thermal Modification of Permanent Hair Dyes, J. Photochem. Photobiol. A, 2002, 147, 233-239. 64. J.E. Pickett, Permanence of UV Absorbers in Plastics and Coatings, in: S.H. Hamid (Ed.), Handbook of Polymer Degradation, 2nd Edition, Marcel Dekker, Inc., New York, 1998. 65. N.C. Billigham, Physical Phenomena in the Oxidation and Stabilization of Polymers, in: J. Pospisil, P.P. Klemchuk (Eds.), Oxidation Inhibition in Organic Materials, CRC Press, Vol. 1, 1990, 249-297. 66. A. Roger, D. Sallet, J. Lemaire, The Photo-Chemistry of Aliphatic Polyamides. 4- Mechanisms of Photooxidation of Polyamides 6, 11 and 12 at Long Wavelengths, Macromolecules, 1986, 19, 579-584. 67. N.C. Billigham, P.C. Calvert, The Physical Chemistry of Oxidation and Stabilisation of Polyolefins, in: G. Scott (Ed.), Developments in Polymer Stabilization, Applied Science Publishers, London, 1980, 139- 190.
62 Ageing and stabilisation of paper Chapter 4: Experimental techniques in studies of photo-stability
68. A. Tidjani, E. Fanton, R. Arnaud, The Oxidative Degradation of Stabilized LLPDE under Accelerated and Natural Conditions, Angew. Makromol. Chem., 1993, 212, 35-43. 69. J.-L. Gardette, J. Lemaire, Photo-Thermal Oxidation of Thermoplastic Polyurethane Elastomers: Part 3 - Influence of the Excitation Wavelengths on the Oxidative Evolution of Polyurethanes in the Solid State, Polym. Deg. Stab., 1984, 6, 135-148. 70. L. Song, C.A. Varma, J.W. Verhoeven, H.J. Tanke, Influence of the Triplet Excited State on the Photobleaching kinetics of Fluorescein in Microscopy, J. Biophys., 1996, 70, 2959-2968. 71. J. Guillet, Polymer Photophysics and Photochemistry, 1st ed., Cambridge University Press, Cambridge, 1985. 72. A. Rauk, Orbital Interaction Theory of Organic Chemistry, 2nd ed., J. Wiley, New York, 2001, 209-217. 73. A. Rivaton, D. Sallet, J. Lemaire, The Photochemistry of Bisphenol A Polycarbonate Reconsidered. Part 2: FTIR Analysis of the Solid State Photo-Chemistry in “Dry” Conditions, Polym. Deg. Stab., 1986, 14, 1- 22. 74. C.S. Foote, E.L. Clennan, Properties and Reactions of Singlet Dioxygen, in: C.S. Foote, J.S. Valentine, A. Greenberg, J.F. Liebman (Eds.), Active Oxygen in Chemistry, Chapman and Hall, London, 1995, 105-141. 75. R.L. Feller, Accelerated Ageing: Photochemical and Thermal Aspects, in: R.L. Feller, Accelerated Aging in Conservation Science, Getty Conservation Institute, Los Angeles, 1994, 51-52. 76. W. Denk, J. Strickler, W.W. Webb, Two Photons Laser Scanning Fluorescence Microscopy, Science, 1990, 248, 73-76. 77. D.W. Piston, B.D. Bennet, G. Ying, Imaging of Cellular Dynamics by Two-Photon Excitation Spectroscopy, J. Microsc. Soc. Am., 1995, 1, 24-35. 78. J.-L. Gardette, Heterogenous Photooxidation of Solid Polymers, Angew. Makromol. Chem., 1995, 232, 85-103. 79. A.V. Cunliffe, A. Davis, Photo-Oxidation of Thick Polymer Samples – Part II: The Influence of Oxygen Diffusion on the Natural and Artificial Weathering of Polyolefins, Polym. Deg. Stab., 1982, 4, 17-37. 80. R.L. Clough, K.T. Gillen, C.A. Quintana, Heterogeneous Oxidative Degradation in Irradiated Polymers, J. Polym. Sci.: Polym. Chem. Ed. 1985, 23, 359-377. 81. L. Matisová-Rychlá, Z. Fodor, J. Rychlý, M. Iring, Decomposition of Peroxides of Oxidised Polypropylene Studied by Chemiluminescence Method, Polym. Deg. Stab., 1981, 3, 371-382. 82. J.-L. Gardette, Fundamental and Technical Aspects of the Photooxidation of Polymers, in: S.H. Hamid (Ed.), Handbook of Polymer Science, Marcel Dekker, Inc., New-York, 2000, 671-698. 83. J.L. Bolland, F. Gee, Kinetic Studies in the Chemistry of Rubber and Related Materials: II. The Kinetics of Oxidation of Unconjugated Olefins, Trans. Farad. Soc., 1946, 42, 236-243. 84. J.-L. Gardette, J. Lemaire, A Methodology for the Study of Long-Term Photo-Ageing of Polymeric Materials, in: P. Hamelin, G. Verchery (Eds.), Textile Components in Building Construction, Vol 3: Mechanical Behaviour, Design and Applications; Pluralis Publ., Paris, 1990, 99-108. 85. J.-L. Gardette, Applications of Infrared Spectroscopy in the Study of the Weathering and Degradation of Polymers, in: J.M. Chalmers, P.R. Griffiths (Eds.), Handbook of Vibrational Spectroscopy Volume 4: Applications in Industry, Materials and the Physical Sciences, J. Wiley, New York, 2002, 2514-2522. 86. D. Fromageot, J. Lemaire, The Prediction of the Long-Term Photo-Ageing of Soluble Polyamides Used in Conservation, Stud. Conserv., 1991, 36, 1-8. 87. J. Lemaire, Prediction of the Long-Term Behaviour of Synthetic Polymeric Materials from Artificial Aging Experiments, in: D. Grattan (Ed.), Saving the 20th Century: the conservation of Modern Materials, Ottawa, Canada Conservation Institute, 1993, 123-134. 88. A. Tidjani, R. Arnaud, Photooxidation of Linear Low Density Polyethylene: A Comparison of Photoproducts Formation under Natural and Accelerated Exposure, Polym. Deg. Stab., 1993, 39, 285-292. 89. J. Lemaire, R. Arnaud, J. Lacoste, The Prediction of Long-Term Photo-Ageing of Solid Polymers, Acta Polym., 1988, 1-2, 27-32. 90. T. Bendaikha, C. Decker, Photodegradation of UV Cured Coatings I – Epoxy-Acrylate Networks, J. Radiat. Curing, 1984, 11, 6-13. 91. O. Guseva, S. Brunner, O. von Trzebiatowski, P. Richner, Service Life Prediction for Coatings Concerning Loss of Gloss, Confederation of European Environmental Engineering Societies 1st European Weathering Symposium, September 25-26, 2003, Prague.
Ageing and stabilisation of paper 63 Chapter 4: Experimental techniques in studies of photo-stability
92. J.A. Graystone, Service Life of Exterior Wood Coatings: A Review of Measurements and Performance Factors, http://www.vtt.fi/rte/bp/coste18/graystoneservicelifepaper.pdf, accessed 02/12/2004. 93. J. Custodio, I. Eusebio, L. Nunes, Durability of Varnishes – Are Subjective Assessments Enough?, http://www.vtt.fi/rte/bp/coste18/abstractdustodio.pdf, accessed 02/12/2004.
64 Ageing and stabilisation of paper Chapter 5 Chemiluminescence from polymers
Jozef Rychlý, Lyda Rychlá
5.1 Introduction Chemiluminometry as an experimental technique has only lately achieved a wider acceptance. The introductory experiments were done already in the early 1960s,1,2 but it took time for it to enter the industrial routine labs for various reasons, the frequently difficult interpretation of experiments undoubtedly being one of them. The early authors have indicated a large unexplored potential of the chemiluminometric method in the study of polymer oxidation and underlined the necessity of better understanding the phenomenon. Even more than 40 years after their pioneering papers much remains to be done. Chemiluminescence is light emission as a consequence of relaxation of electrons, which populate excited states in an elementary process of a chemical reaction. The process of population of excited states proceeds with some kinetics, which is related to the kinetics of the given chemical reaction. The emission of chemiluminescence in time should thus be related to the rate of the chemical reaction. A more general and comprehensive treatise of the topic was edited by Zlatkevich.3 It is our intention to provide an introduction to studies of cellulose and paper, and to review the recent developments in instrumentation.
5.2 General classification of chemiluminescence from polymers The reaction leading to excited states usually involves a transfer of one electron, such as free radical disproportionation (as in the case of two secondary peroxyl radicals) or transfer between an electron-donor and electron-acceptor, including intermediate formation of cation and anion radicals which are in a subsequent step converted to products, of which some may be formed in an excited state. The heat, released either simultaneously with light emission, or in a quasi-heterogeneous domain, where reactions of locally accumulated by-products of degradation may take place,
Ageing and stabilisation of paper 65 Chapter 5: Chemiluminescence from polymers
acts in favour of the chemiluminescence process. Therefore, it is not surprising that weak light emission accompanies oxidation processes involved in free radical chain reaction. One should be aware that the intensity of chemiluminescence signal from a heated polymer and kinetics of its change in time or with temperature is determined by:4,5 ― The quality of the polymer, the character of its terminal groups, the extent of previous oxidation and thermo and/or photooxidation history of the polymer sample, expressed in concentration of hydroperoxides, carbonyl groups or of other oxidised structures. The change of polymer mechanical properties as a consequence of oxidation may be related to the average molar mass and to its distribution (as shown in Figure 3.1) and to the ratio of amorphous/crystalline structures. ― Temperature and concentration of oxygen in atmosphere surrounding the oxidised sample. ― The extent and quality of polymer stabilisation. ― Occasionally, oxidation of a polymer additive may give a much stronger signal than oxidation of the polymer itself. This may lead to an erroneous correlation between the rate of polymer oxidation and chemiluminescence intensity. On one hand, the fact that one excited molecule out of one billion of other molecules can be detected and the fact that a possible relation between its formation and the chemical processes governing oxidation of the material may exist, testifies of the excellent sensitivity of the method. On the other hand, however, quenching and filtration of light may again lead to an erroneous interpretation of results. Qualitatively, two different patterns of chemiluminescence intensity in time may be observed for polymeric materials: Type 1: An autoaccelerating increase from very low values to a maximum followed by subsequent decay. Such curves, characterised by an oxidative induction time are typical for oxidation of polyolefins, polydienes and polyamides having relatively long oxidation chains (Figure 5.1). The initial level of chemiluminescence intensity at time t = 0 usually corresponds to the degree of polymer pre-oxidation. Type 2: Decay from some initial value of chemiluminescence intensity to a considerably lower steady-state level. This differs from curves of Type 1 in that the time to reach a maximum value of chemiluminescence intensity is extremely short and mostly only a steady decay is observed. Such runs are typical for oxidation of polymers with heteroatoms in the main chain, e.g. poly(2,6-dimethyl-1,4- phenylene oxide), shown in Figure 5.2, having a rather short length of oxidation chains. The decay of peroxide structures in inert atmosphere from pre-oxidised polymer samples can be classified in this category, as well. Some polymers show quite a complex pattern composed of both types 1 and 2 (oxidation of pre-oxidised polypropylene) or from two autoaccelerating curves of Type 1 (polyamides, polyethylene, as in Figure 5.1) which may be related to a change in oxygen diffusion in the system, the first wave corresponding to oxidation under excess of oxygen, the second to oxidation in an oxygen-depleted system. Thermo-oxidative stabilizers shift the increase of chemiluminescence in advanced stages of oxidation to longer times in Type 1 experimental curves and reduce the initial chemiluminescence intensity to lower values in experimental curves of Type 2. The relation of chemiluminometry and other experimental techniques of investigation relevant for the individual processes of the classical Bolland-Gee scheme of
66 Ageing and stabilisation of paper Chapter 5: Chemiluminescence from polymers hydrocarbon oxidation was already presented in the Scheme 2.2 (page 14), composed of initiation, propagation and termination. Chemiluminescence is assumed to be a consequence of step (8) and possibly of step (5) provided that it occurs in heterogeneous pre-oxidised domains of the polymer. Other possibilities cannot be excluded.
50000 polypropylene ) 1 -
g 40000 m
1 - s (
y
t 30000 i s n e t n i polyamide 66
L 20000 C
d
e polyethylene z i l
a 10000 m r o n 0
0 20000 40000 60000 80000 time (s) Figure 5.1: Chemiluminescence at 140 °C during oxidation of films of polypropylene, polyamide 66 and polyethylene in oxygen atmosphere (Type 1).
1600
1400 ) 1 - g 1200 m
1 - s ( 1000 y t i s n
e 800 t n i o o L 600 130 C 90 C C
d o
e 110 C z
i 400 l a m r
o 200 n
0
-500 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 time (s) Figure 5.2: Chemiluminescence at different temperatures during oxidation of films of poly(2,6- dimethyl-1,4-phenylene oxide) powder in oxygen atmosphere (Type 2).
As products of oxidation, ketones and O2 are formed as excited triplet state of ketone and singlet oxygen, respectively. They are converted to ground state either through emission of light: >C=O* >C=O + h
O2* O2 + h or as heat, dissipated in collisions. One should be aware of the fact that only a very small fraction of excited states is converted to the ground state through emission of light, while the prevailing part is quenched in a non-luminous way. Thus, a more or less complex relation of observed chemiluminescence intensity and rate of initiation is usually observed and deconvolution of the initiation mechanism from experimental chemiluminometric
Ageing and stabilisation of paper 67 Chapter 5: Chemiluminescence from polymers curves may turn out to be rather difficult. Following the Bolland-Gee scheme for non- inhibited oxidation occurring under stationary conditions, the following equation is frequently valid:
2 2 wi + kmono[POOH] + kbi[POOH] = k6[PO2 ] . (eq 5.1)
Provided that chemiluminescence intensity ICL is proportional to the rate of peroxyl 2 radicals termination, i.e. ICL ≈ k6[PO2 ] , which is often assumed in the literature, chemiluminescence intensity should achieve some quasi-stationary level when hydroperoxide concentration becomes stationary and its decay should correspond to consumption of oxidisable groups, PH, in a polymer. At the same time, the chemiluminometric curves of Type 1 are relevant for
2 kmono[POOH] + kbi[POOH] >> wi, (eq 5.2) while for Type 2 the following is valid:
2 kmono[POOH] + kbi[POOH] << wi, (eq 5.3) where wi represents the consumption of oxidisable (defect) sites in the polymer. In Type 1, the length of kinetic chains is sufficient to gradually develop a significant level of polymer hydroperoxides, while in Type 2, the concentration of hydroperoxides is rather low and the main oxidation events represent the initiation elementary step (1) in Scheme 2.2.
5.3 The effect of temperature and atmosphere An increasing temperature shortens the induction time related to advanced stages of oxidation and increases the maximum chemiluminescence intensity in the case of Type 1 chemiluminescence (Figure 5.3) whereas in Type 2, it increases the initial chemiluminescence intensity. This is not surprising as the rate of oxidation reaction increases with temperature as well.
18000
o 16000 110 C ) 1 -
s 14000 (
y t i 12000 s n e t
n 10000 i
e o c
n 8000 100 C e c
s 6000 e n i
m 4000 o u
l 90 C i
m 2000 o e 80 C h c 0
-2000 0 200000 400000 600000 800000 time (s) Figure 5.3: Chemiluminometric experiments for polypropylene powder at different temperatures in oxygen atmosphere. Molar mass of the polymer: 180,000 g mol-1. Oxygen increases chemiluminescence intensity significantly. This is not typical only of polyolefins as reported in many papers (see ref. no. 5 and references therein), but also of other polymers such as the polysaccharide pullulan (Figure 5.4). It is of significant interest
68 Ageing and stabilisation of paper Chapter 5: Chemiluminescence from polymers that chemiluminescence intensity is proportional to the square root of oxygen concentration,6 thus indicating a direct participation of oxygen in initiation of chemiluminescence reaction.
18000 220
16000 200
) 180 1
- 14000 s (
y 160 t i
s 12000 ) n oxygen e C
t 140 o ( n
i 10000
e r e c 120 u t n a e 8000 r c e
s 100 air p e m n
i 6000 e t
m 80 u l i 4000 m 60 e h c 2000 nitrogen 40 20 1000 2000 3000 4000 5000 6000 7000 time (s) Figure 5.4: Non-isothermal chemiluminometric experiment with pullulan samples of molar mass 24,000 g mol-1. Rate of heating 2.5 °C min-1. The same dependence may be observed for oxidised Whatman paper. From a single experiment with concentration of oxygen in the mixture with nitrogen decreased in a stepwise manner we see that the intensity of light emission is proportional to the square root of oxygen concentration (Figure 5.5). Comparing the levels of chemiluminescence intensities and thus the rates of oxidation of Whatman cellulose in oxygen, air and nitrogen, the following relation between respective rate constants may be estimated: koxygen = 2.1·kair = 27.6·knitrogen. (eq 5.4)
35000 100 40000
35000
) 83.3 1 - 30000 ) s 1 - (
s 30000 (
y 66.6
t i y s t 25000 i n 50 s 25000 e n t e n t i
n i e 20000 33.3 20000 c e n c e n c
e 15000 s
16.6 c e 15000 s n e i 13.3
n 10000 i m 10 u m l 6.6 i
10000 u l
i 5000 m 3.3 e m h e c h 0
5000 c 0 -5000 0 0 2500 5000 7500 -2 0 2 4 6 8 10 12 sqrt[O ] time (s) 2 Figure 5.5: Left: stepwise changes of the intensity of chemiluminescence signal following changes in concentration of oxygen in the surrounding atmosphere at 180 °C for Whatman paper. Numbers represent % of oxygen in the mixture with nitrogen (V/V), sample mass 5.6 mg. Right: Plot of the dependence of quasi-stationary levels of chemiluminescence intensity on the square root of oxygen concentration. 5.4 The effect of molar mass The effect of molar mass and stereoregularity on thermo-oxidation of polymers and on
Ageing and stabilisation of paper 69 Chapter 5: Chemiluminescence from polymers chemiluminescence was only rarely investigated, apart from polypropylene.7-9 Surprisingly, syndiotactic polypropylene was reported to be more stable than isotactic polymer. E.g., at 140 °C, the maximum chemiluminescence intensity was achieved after 2835 min for syndiotactic polypropylene while 45 min were necessary for isotactic polymer and similar was true also for atactic and isotactic polypropylene. The explanation was proposed that the structure of isotactic polypropylene is much more favourable for autoxidation proceeding via a “back-biting” mechanism where peroxyl radicals abstract adjacent tertiary hydrogens on the same polymer chain.
120000 40,000 g mol-1 45,000 g mol-1 ) 1 - s ( 80000 y t i s n e t n i
e c n
e 40000 c -1 s 80,000 g mol e n i 180,000 g mol-1 m u l i m e
h 0 c
-10000 0 10000 20000 30000 40000 50000 60000 70000 80000 time (s) Figure 5.6: Chemiluminescence experiments at 120 °C in oxygen atmosphere for polypropylene powder of different molar masses as indicated. On the other hand, Iring et al.10 found that the induction period for oxidation of solid isotactic polypropylene is longer than that of atactic polymer. Their result is in accordance with the predominantly crystalline character of isotactic polypropylene and with the resistance of crystalline zones to oxidation. Using well characterised non-stabilised polypropylene samples11 of molar mass 45,000-180,000 g mol-1 and of different tacticity and crystallinity it was demonstrated that an increasing molar mass (within some interval) leads to an increase of induction time and reduction of maximum chemiluminescence intensity (Figure 5.6). The polymer with higher average molar mass appears to be more stable than that with lower molar mass which may be ascribed to the effect of more reactive terminal groups in promotion of initiation of thermal oxidation.
5.5 The effect of antioxidants and polymer stabilizers In chemiluminometric curves of Type 1, antioxidants shift the autoaccelerating increase of chemiluminescence intensity to longer times. This is due to reactions (12)-(14) of Scheme 2.2, in which alkyl radicals, peroxyl radicals and hydroperoxides are scavenged until antioxidants InH and D are consumed. The typical example of such behaviour may be seen in the case of polypropylene containing 0.1% (m/m) of Irganox 1010 (sterically hindered phenolic antioxidant). The presence of antioxidants may even reduce the maximum chemiluminescence intensity (Figure 5.7).12 This may be explained by a quenching effect of the antioxidant on excited carbonyls but it may be related also to the mechanism of
70 Ageing and stabilisation of paper Chapter 5: Chemiluminescence from polymers oxidation of stabilised polypropylene, in which the stabiliser is distributed in a higher concentration in amorphous zones of the polymer of lower molar mass.
160000
140000 non-stabilised ) 1 - 120000 s (
y t i
s 100000 n e t n i 80000 e c n e
c 60000 s e n i
m 40000 u l i m
e 20000 h stabilised c
0
10 100 1000 10000 100000 time (s) Figure 5.7: Chemiluminescence during oxidation of a non-stabilised polypropylene film and a stabilised one with 0.1% (m/m) of the phenolic antioxidant Irganox 1010 at 140 °C, in oxygen atmosphere. The case of reduction of chemiluminescence intensity for stabilised polymers exhibiting the Type 2 chemiluminescence pattern has not been studied, yet. The parts of polymer system most vulnerable to oxidation are amorphous regions where macromolecules of higher molar mass form the skeleton responsible for mechanical properties. The mobility of chains on one hand, and easier access of low molar mass compounds (including oxygen) to reaction sites on the other hand, are the conditions leading to degradation processes starting predominantly in such regions. The domains of higher molar mass are, moreover, less protected by stabilizers which tend to dissolve better in lower molar mass regions. Oxidation of pure polymer starts, therefore, on lower molar mass molecules or on their terminal groups in the case of non-stabilised polymer. However, in the case of stabilised polymer these lower molar mass molecules are better protected by a higher local concentration of the dissolved stabilizer. In such a case, one may expect that degradation starts easier on higher molar mass domains which are less well protected.
5.6 Mechanisms leading to chemiluminescence during polymer oxidation Regardless of the fact that self-recombination of secondary peroxyl radicals defined by the Russell’s scheme:13 R1OO + R2R3HCOO R1OH + O2* + R2R3C=O* is quite accepted as the most probable reaction step providing excited triplet ketones and singlet oxygen thus leading to emission of light (chemiluminescence), the problem is not as simple as it appears. An argument in favour of the above scheme is its strong exothermicity which may lead to population of excited states after formation of an unstable tetraoxide and after its decomposition to products. However, an experimental proof of triplet atmospheric oxygen quenching light emission from triplet ketones (which should be a quite natural consequence) is still missing. The maximum of emission from excited triplet
Ageing and stabilisation of paper 71 Chapter 5: Chemiluminescence from polymers carbonyls is at 460 nm, while that from singlet oxygen is at 634 nm (dimol emission) and 1270 nm (monomol emission).14 Singlet oxygen is formed in the decomposition of hydrogen peroxide which appears during oxidation of polymers carrying hydroxyl groups (e.g. cellulose), following the stoichiometric equations:
–C(OH)H– + O2 –CO– + H2O2,
2H2O2 2H2O + O2*.
1 – The singlet oxygen ( g state) appears in a reaction of superoxide anion radicals (O2 ) which are reaction intermediates in the above reaction in an alkaline medium. In the presence of water, hydroperoxyl radicals HO2 can be formed:
– – O2 + H2O HO2 + HO and the energy required for excitation of oxygen is obtained from recombination of two hydroperoxyl radicals 2HO2 O2* + H2O2. According to Reshetnyak et al.,15 the sequence of subsequent reactions may lead either to ozone or to dimols of singlet oxygen:
H2O2 + O2* O3 + H2O,
H2O2 + O3 (O2)2* + H2O. As a consequence, the singlet dimol emission at 634 nm can be detected with photomultipliers, which are usually used for measurements of chemiluminescence from polymers. The above scheme, however, will require the presence of at least one molecule of hydrogen peroxide in a close vicinity of two recombining peroxyl radicals and assumes that the oxidation process is very heterogeneous. While the opinion of Quinga and Mendenhall16 is that luminescent processes in oxidation of polymers are due to disproportionation of two alkoxyl radicals, Audouin-Jirackova and Verdu17 expressed the idea that monomolecular fragmentation of alkoxyl radicals to carbonyl groups and alkyl radicals in which carbonyl groups are formed in the excited state, is responsible for chemiluminescence. The proposal by Osawa18 that light emission originates in polynuclear aromatics acting as photosensitizers is also worth noticing. They are supposedly sorbed to the oxidised polymer from the surrounding atmosphere during its storage. This opinion is based on a strong reduction of initial chemiluminescence intensity when polymer is extracted by low molecular solvents. Recently, Blakey and George19 have shown that chemiluminescence from oxidised polymer may also be due to the so called CIEEL (chemically initiated electron exchange luminescence) process, which involves electron exchange between a potential activator and peroxide which converts the activator to a radical cation and the peroxide to a radical anion. The next step is rearrangement of the radical anion of peroxide followed by charge recombination with formation of excited activators. It appears that either isolated carbonyls or carbonyls in the vicinity of C=C unsaturations may be the activators. The quantum yield
72 Ageing and stabilisation of paper Chapter 5: Chemiluminescence from polymers for the CIEEL mechanism appears to be by three to five orders of magnitude higher than that for a reaction of a pair of peroxyl radicals. Other speculative mechanisms may be proposed based on the presence of singlet oxygen and C=C unsaturations in the oxidised polymer. A reaction of both may lead to a transient formation of dioxetanes, the decomposition of which has an even higher quantum yield of luminescence than the CIEEL mechanism.20 In heterogeneous domains of oxidised polypropylene, dioxetanes may be formed from 1,3-hydroperoxylperoxyl radicals in a sequence of reactions:21
Here, the O-O bond of the hydroperoxyl group is split simultaneously with -scission of the alkoxyl radical formed, and ring closure of the alkyl peroxyl diradical may occur if the process has some extent of synchronisation. The process gives hydroxyl radical, methylcarbonyl end groups -CH2-CO-CH3 and dioxetane:
which is unstable and decomposes into an excited triplet state of formaldehyde and/or excited triplet state of methylcarbonyls. In the review by Ptschelintsev and Denisov,22 it was shown that in oxidised polyisoprene, one recombination of free radicals corresponds to approximately thirty C-C bond scissions. The authors expressed the idea that chain scission is a matter of propagation of peroxyl radicals and not of -scission of alkoxyl radicals. Recently, from non-isothermal chemiluminometric experiments with oxidised cellulose, comparable rate constants were obtained to those from experiments in which the DP was followed.23 This may be considered as an indication that scission of polymer chains is accompanied by light emission. A possible pathway, at the same time strongly exothermic, is also the scission of 1,3-hydroperoxyalkyl radicals which are formed by an intramolecular transfer of peroxyl radicals within a six-membered transition state:
The heat gain of the above reaction is surprisingly high (about 450 kJ mol-1) and in conditions of reduced mobility of a polymer chain such a mechanism may successfully compete with any other mechanism of light emission, described earlier. In Chapter 8, we will discuss other possible mechanisms with which we can interpret light emission
Ageing and stabilisation of paper 73 Chapter 5: Chemiluminescence from polymers phenomena accompanying the oxidation of cellulose. At the present level of knowledge one can not exclude that light from oxidised polymer originates from a number of elementary reactions and even from several different sources. However, the method has become very popular because of its extreme sensitivity. A persuading example of popularity of chemiluminometry is biochemistry and other fields of analytical chemistry,24 where mechanisms of light emission are also frequently unclear, but the method is still routinely used. This demonstrates a large and still unexplored potential of the chemiluminometric approach in other domains of chemistry. 5.7 A simple kinetic model of non-isothermal chemiluminescence The Ekenstam model of degradation of linear polymers, which was discussed in Chapter 3 is quite simple, however, it provides a satisfactory quantitative description of the process in time and/or with increasing temperature and thus became a necessary pre-requisite for any extrapolation or comparison of the effect of different additives. Similarly to (eq 3.3), we can define the degree of polymerisation (DP) as the ratio of the number of monomers (N) and polymer molecules (i): N DP . (eq 5.4) i The concentration of polymer molecules increases in time due to degradation and decays due to cross-linking. Provided that the process is taking place statistically, the kinetics of concentration of macromolecules increase (for the case of degradation) may be described by the equation: di mki n , (eq 5.5) dt where n stands for the reaction order of the process of chain scissions, and m = 1 in case of degradation, and m = -1 in case of cross-linking. The most frequent reaction orders encountered in the literature are n = 0 or n = 1.25 In Chapter 3, the Ekenstam equation was derived presuming n = 1. In this Chapter, the equation in the form of (eq 3.8) will be used as it provides the rate constant in s-1 and thus enables a direct comparison between samples of different initial DP. In order to develop a methodology of quantitative description of chemiluminescence data we have to have some idea about the reactions leading to light emission. We know that chemiluminescence intensity generally expresses the rate of sample oxidation w, so that: ΦwI , (eq 5.6) where is a proportionality constant involving the quantum yield of light emission, i.e. the fraction of excited molecules which are converted to the ground state through light emission; then sample geometry, the presence of quenchers and light filters, physical state of the polymer matrix etc. However, how to express the rate w? One has to be aware that oxidation may ultimately lead to: ― Degradation, i.e. reduction of molar mass, ― Cross-linking, i.e. increase of molar mass, or
74 Ageing and stabilisation of paper Chapter 5: Chemiluminescence from polymers
― No change in molar mass if oxidation is focused on side or terminal groups. Each of the above cases, which can be mutually inter-related, may occur in the case of oxidation of any polymer. This leads to the conclusion that the way we express w may be based on the shape of experimental curves CL intensity/time or CL intensity/temperature. Some possibilities are outlined below:
2 kw 26 ][PO , (eq 5.7) d[POOH] w 2 kk ][PH][PO[POOH]2 , (eq 5.8) dt bi 24
kw 24 ][PH][PO , and (eq 5.9) dDP w . (eq 5.10) dt The first case (eq 5.7) corresponds to the Russell’s scheme of disproportionation of peroxyl radicals. In this case the experimental curves of CL intensity/time attain a quasi-saturated level of light emission unless oxidisable material is not appreciably depleted. The second case (eq 5.8) describes well the isothermal experiments of oxidation of polyolefins and corresponds to the rate of hydroperoxide decomposition. The third case (eq 5.9) includes the possibility of exothermicity of propagation reactions which might contribute to the population of excited states. Formally, the simplest equation which enables an easy derivation of non-isothermal kinetics is the last case (eq 5.10), where chemiluminescence intensity is related to the rate of chain scissions, so that from (eq 5.6) we obtain: dDP I Φ[ ] . (eq 5.11) dt k dDP k By substitution k and derivation of (eq 3.8) we obtain DP 2 and in i 0 dt i DP00 non-isothermal conditions we have:
a/RTE dDP dT Ae 2 DP , (eq 5.12) dT dt i DP00 where T is temperature, A and Ea are the pre-exponential factor and the activation energy, dT respectively, and β is the linear rate of sample heating. After integration of (eq 5.12) dt and substitution into (eq 5.6) we finally obtain the equation
Ae a/RTE DP 0 I Φ T (eq 5.13) i 0 A a RTE dTe ]1[ 2/ i0 Troom for non-isothermal conditions. One can see that while the process of chain scission is assumed to be of zero or first order, the chemiluminescence runs formally correspond to the second-order scheme. Provided that we formally consider that the chemiluminescence
Ageing and stabilisation of paper 75 Chapter 5: Chemiluminescence from polymers process is governed by three independent mechanisms occurring in parallel, (eq 5.13) may be applied as follows:
3 P I i . (eq 5.14) T i1 A RTE 2/ i ai dTe ]1[ i0 Troom
Here, Pi is the proportionality constant including the corresponding terms from (eq 5.13). The chemiluminescence runs for Whatman paper (WH), sulphate bleached pulp (SA) and cotton pulp (C) oxidised under non-isothermal conditions are shown on Figure 5.8 and the corresponding parameters for two independent steps of degradation (i = 2) are summarised in Table 5.1. Agreement of the fit of experimental measurements with parameters of
(eq 5.14) is indeed very good. An average rate constant kav can also be defined:
P1 P2 P3 kav k11 k12 k13 , (eq 5.15) PPP 321 PPP 321 PPP 321
E a11 RT where 1111 eAk .
120000 ) 1
- 100000 g
m SA
1 - s (
80000 y t i s n e t 60000 n i
L C
d WH
e 40000 z i l a C m r
o 20000 n
0
20 40 60 80 100 120 140 160 180 200 220 240 temperature (oC) Figure 5.8: Experimental curves and theoretical fits for oxidation of Whatman paper (WH), sulphate bleached pulp (SA) and cotton pulp (C) in oxygen atmosphere. The rate of sample heating was 3 °C min-1. The points represent fitted values using (eq 5.14) for i = 2 and parameters in Table 5.1. The above kinetic analysis has its advantages and disadvantages. Using non-linear regression analysis we may obtain a highly satisfactory fit of the experimental data and get a series of rate constants characterised by the corresponding activation energies and pre- exponential factors. Without going into any further detail we may thus reveal whether the oxidation process under observation is typical autoxidation with an accumulation of hydroperoxides or not. If this is the case, activation energy obtained from the experimental curve is abnormally high due to contribution of autoacceleration to the increase of light emission. On the other hand, fitting the resulting curve by several first-order processes may lead to a certain compensation in parameters which becomes particularly evident if the experimental curves are not well developed or slightly distorted.
76 Ageing and stabilisation of paper Chapter 5: Chemiluminescence from polymers
Table 5.1: Parameters of (eq 5.14) obtained from non-isothermal runs in oxygen atmosphere for i = 2 for Whatman paper (WH), sulphate bleached pulp (SA) and cotton pulp (C). i = 1 (faster process) i = 2 (slower process) sample P1 A1/i0 Ea1 P2 A2/i0 Ea2 (counts mg-1) (s-1) (J mol-1) (counts mg-1) (s-1) (J mol-1) C 2.43·106 4.06·103 58141 1.3·108 1.34·1010 128409 WH 1.13·106 5.35·104 64327 1.11·108 9.85·109 126951 SA 6.31·107 1.19·103 60083 1.9·108 4.38·1012 149775
The question may arise why such an emphasis is put on non-isothermal kinetics. The reason is that if the “virgin” material is heated starting from a low temperature, then the oxidation proceed in small steps and there is a considerably lower probability of a reverse effect of the reaction products on the rate of the oxidation. The only assumption in the approach is that the proportionality constant is temperature independent.
5.8 Fit of isothermal experimental data Let us presume that oxidation of polymer PH takes place according to Scheme 2.2 from which we will only consider the reactions (2), (3), (5), (6), (8), (10), (11), which involve initiation, propagation and termination of intermediate free radicals. The production of primary radicals is soon governed by decomposition of hydroperoxides. From the shape of kinetic runs it seems that decomposition of hydroperoxides has the character of bimolecular reaction predominantly, even very shortly after the start of an experiment. Oxygen may be accounted for in the balance equation, in which the rates of production and consumption of alkyl and peroxyl are in equilibrium:
2 2 823 4 2 kkkk 6 2 ]PO[]PH][PO[]P[]O][P[ . (eq 5.16) Since the rates of initiation and termination should be equal, we have:
2 2 bi 6 2 kkk 27 ]P][PO[]PO[]POOH[ , (eq 5.17) provided that the rate of termination of two alkyl radicals is neglected. Formally, the same equations may be obtained by subtraction and addition of terms for d ]P[ d ]PO[ 0 and 2 0 , and the combination of (eq 5.16) and (eq 5.17) leads to: dt dt
k ]POOH[ 2 k ]PO[ 2 bi mk ]POOH[ 2 , (eq 5.18) 6 2 k k ]PH[ k ]PO[ bi 1 7 4 27 k6 k 23 ]O[ k 23 ]O[ provided that we neglect the third term in denominator, which is significant only at relatively low concentrations of oxygen in polymer. Thus we obtain m as a function of concentration of oxygen dissolved in polymer:
Ageing and stabilisation of paper 77 Chapter 5: Chemiluminescence from polymers
k ]O[ m 23 . (eq 5.19) k7 k 23 k4 ])PH[(]O[ k6 We have to stress that the above expression is simplified by presuming that: ― The initiation of the process is represented by a bimolecular decomposition of polymer hydroperoxides POOH; ― The transfer reaction of alkoxyl radicals PO to polymer PH is considerably faster than the corresponding reaction of peroxyl radicals; ― The concentration of alkoxyl radicals can thus be neglected if compared to that of peroxyl radicals; ― Differential terms for the rates of alkyl radicals and peroxyl radical concentration changes were assumed to be zero taking into account that these rates are negligible if compared with the rate of change of hydroperoxide concentration. Thus, for the time dependence of hydroperoxide concentration [POOH] we ultimately obtain the equation: X ]POOH[ , (eq 5.20) 2tkY )exp(1
]POOH[]POOH[ 0 where X ]POOH[ ,Y , while ]POOH[ 0
k PH mk ]POOH[ 4 and kk ]PH[ bi , (eq 5.21) 2 4 k 2 6 kk bi 6 where ]POOH[ and ]POOH[ 0 are steady-state and initial concentrations of hydroperoxides in the polymer, respectively. According to (eq 5.20) we thus obtain:
2tk d ]POOH[ 2 XYek 2 bi kmk 6 2 ]PH][PO[]POOH[ . (eq 5.22) dt Ye 2tk ]1[ 2 The experimental curves of chemiluminescence intensity vs. time show that the chemiluminescence intensity is directly proportional to the rate of hydroperoxide decomposition in which the rate of bimolecular decomposition of hydroperoxides and the rate of transfer of peroxyl radicals to the polymer chain are both included. The experimental curves go through a well developed maximum, followed by a decrease in intensity which is in accordance with the equation: d ]POOH[ I ( ) . (eq 5.23) dt The above equation is a first approximation of the chemiluminescence kinetics for homogeneous systems where oxidation takes place uniformly. However, as shown by several authors, different sections of polymer sample may oxidize with autonomous
78 Ageing and stabilisation of paper Chapter 5: Chemiluminescence from polymers kinetics, determined by different rates of primary initiation. The decrease of the resulting CL intensity/time curves is then not so sharp as predicted by (eq 5.12) and, as it was postulated by Rychlý et al.,11 an optimal fit of experimental data is obtained using (eq 5.12) corrected by a term describing oxidation spreading.
5.9 Instrumentation Currently, the only commercially available chemiluminometric instruments for polymers and non-volatiles are produced by Tohoku Electronic Industrial Co., Japan,26 and by the Polymer Institute of the Slovak Academy of Sciences, Slovakia.27 The company Atlas Material Testing Technology LLC, USA,28 also used to produce chemiluminometers. The instruments have found application in scientific and industrial polymer testing laboratories and on industrial scale for estimating the oxidizability of polymer materials, food components and crude oil products. They are of use for investigations of decomposition of free radical initiators, particularly of peroxides and explosives provided there are no corrosive decomposition products. If this is the case, as e.g. in poly(vinyl chloride), irreversible damage to mechanical parts of the instrument is possible. The examined samples may be in solid form (films, foils, powders etc.) or non-volatile liquids, in the amount usually not exceeding several milligrams. It is not recommended to use granules..
Figure 5.9: The Lumipol 2 chemiluminometer. All of the results presented in this book were obtained with the Lumipol instrument, so we will describe it in more detail. The instrument Lumipol 2 is capable of measuring luminescence within a large temperature interval (room temperature – 250 °C) and in an atmosphere of variable composition. The instrument is composed of two parts (Figure 5.9). In the base, electronic circuits necessary for the control of oven heating and motion, high
Ageing and stabilisation of paper 79 Chapter 5: Chemiluminescence from polymers voltage source for the photomultiplier and the corresponding connectors are situated. On the front panel, a regulator connected via a serial port to a computer is mounted. In this way, data on oven temperature and parameters of sample heating are stored. The regulator controls preset temperature parameters (operating temperature, heating programme etc.). Gas inlet and outlet are situated on the rear panel of the base part and lead to the oven compartment. The vertical part contains a mechanism controlling the motion of oven, which can be moved upwards or downwards. As the sample compartment on Figure 5.10 closes, the oven moves upwards, the photomultiplier shutter opens and the measurement can begin.
Figure 5.10: The open reaction compartment of the Lumipol 2 instrument, with a sample pan in its place on the oven.
Figure 5.11: A chemiluminescence measurement using an external sample compartment. There is no need for sample cutting. Humid atmosphere can be used. Chemiluminescence intensity is monitored by a photomultiplier and a photon counting system. The photomultiplier is separated from the reaction space of oven by an optical system and a shutter which protects it from intensive light shocks. It is situated in a thermally insulated casing protecting it from any radiation coming from the surroundings. A round sample of paper (diameter 1 cm or less) has to be cut from a sheet in order to be subjected to an experiment. However, a separate unit enables us to measure chemiluminescence without cutting the sample (Figure 5.11). Instead, the sample is
80 Ageing and stabilisation of paper Chapter 5: Chemiluminescence from polymers clamped between two gold-plated copper ovens so that the temperature gradient between the detector and the site of reaction is minimal. One oven is situated on a movable arm and serves only for sample heating, while the fixed one surrounds an optical cable through which the chemiluminescence signal is collected and transferred to the photomultiplier via an interface. An inlet and outlet of gas, preheated to the same temperature as that of both ovens, is also situated there. The gas may also contain a pre-set fraction of (gaseous) water. A typical experiment is performed in the following way. The arm with the heater approaches the part of a sample to be examined (e.g. a page of a book) and the sample is gently pushed towards the light-collecting end of the fibre optics, which is not covered by the sample, but is in fact a few millimetres away from it, thus forming a reaction compartment into which a humid atmosphere is introduced (Figure 5.11). The sample is locally heated to a desired temperature and is slowly equilibrated with the surrounding atmosphere, while the light can be collected. The sample position is easily adjustable.
5.10 Conclusions Measurements of chemiluminescence can provide useful data on kinetics of polymer oxidation. The correlation of rate constants obtained from non-isothermal chemilumi- nometric experiments and classical ageing experiments in ovens are particularly encouraging.23 Considering that micro-samples can be used (of diameter starting from 0.5 mm),29 the technique is promising also for studies of historical samples. Although much remains to be done to be able to interpret light emission of a given oxidised material properly, the potential of the method for analytical purposes is well recognised even without a detailed knowledge on the particular mechanism.
5.11 References 1. G.E. Ashby, Oxyluminescence from Polypropylene, J. Polym. Sci., 1961, 50, 99-106. 2. M.P. Schard, C.A. Russell, Oxyluminescence of Polymers. I: General Behavior of Polymers, J. Appl. Polym. Sci., 1964, 8, 985-995. 3. L. Zlatkevich, Luminescence Techniques in Solid-State Polymer Research, Marcel Dekker, New York, 1989. 4. L. Matisová-Rychlá, J. Rychlý, New Approach to Understanding Chemiluminescence from the Decomposition of Peroxidic Structures in Polypropylene, Polym. Degrad. Stab., 2000, 67, 515-525. 5. L. Matisová-Rychlá, J. Rychlý, Inherent Relations of Chemiluminescence and Thermooxidation of Polymers, in: R.L. Clough, N.C. Billingham, K.T. Gillen (Eds.), Polymer Durability: Degradation, Stabilization and Lifetime Prediction. Advances in Chemistry Series, American Chemical Society, Washington, 1996, 249, 175-193. 6. M. Strlič, J. Kolar, B. Pihlar, L. Matisová-Rychlá, J. Rychlý, Chemiluminescence during thermal and thermo-oxidative degradation of cellulose, Eur. Polym. J., 2000, 36, 2351-2358. 7. H. Mori, T. Hatanaka, M. Terano, Thermal stability of syndiotactic polypropene, Macromol. Rap. Comm., 1997, 18, 157-161. 8. Z. Osawa, M. Kato, M. Terano, Effect of stereoregularity on the thermo-oxidative degradation of poly(propylene)s estimated by chemiluminescence, Macromol. Rap. Comm., 1997, 18, 667-671. 9. M. Kato, Z. Osawa, Effect of stereoregularity on the thermo-oxidative degradation of polypropylenes, Polym. Degrad. Stab., 1999, 65, 457-461. 10. M. Iring, Z. Laszlo-Hedvig, F. Tudos, T. Kelen, Study of the thermal oxidation of polyolefins. 12. Thermal oxidation of isotactic and atactic polypropylene in the condensed phase and in solution, Polym. Degrad. Stab., 1983, 5, 497-480.
Ageing and stabilisation of paper 81 Chapter 5: Chemiluminescence from polymers
11. J. Rychlý, L. Matisová-Rychlá, P. Tiemblo, J. Gomez-Elvira, The effect of physical parameters of isotactic polypropylene on its oxidisability measured by chemiluminescence method. Contribution to the spreading phenomenon, Polym. Degrad. Stab., 2001, 71, 253-260. 12. G. George, M. Celina, Homogeneous and heterogeneous oxidation of polypropylene, in: S. Halim Hamid, (Ed.), Handbook of Polymer Degradation, M. Dekker, New York, 2000, 277-301. 13. G.A. Russell, Deuterium-isotope Effects in the Autoxidation of Aralkyl Hydrocarbons. Mechanism of the Interaction of Peroxy Radicals, J. Am. Chem. Soc. 1957, 79, 3871-3877. 14. E. Lengfelder, E. Cadenas, H. Sies, Effect of DABCO (1,4-diazabicyclo[2,2,2]-octane) on singlet oxygen monomol (1270 nm) and dimol (634 and 703 nm) emission, FEBS Lett., 1983, 164, 366-370. 15. O.V. Reshetnyak, E.P. Kovalchuk, P. Skurski, J. Rak, J. Blazejowski, The origin of luminescence accompanying electrochemical reduction or chemical decomposition of peroxydisulfates, J. Lumin., 2003, 105, 27-34. 16. E.M.Y. Quinga, G.D. Mendenhall, Chemiluminescence from hyponitrite esters - excited triplet states from dismutation of geminate alkoxyl radical pairs, J. Am. Chem. Soc., 1983, 105, 6520-6521. 17. L. Audouin-Jirackova, J. Verdu, Chemiluminescence of Hydrocarbon Polymers, J. Polym. Sci. A: Polym. Chem., 1987, 25, 1205-1217. 18. Z. Osawa, H. Kuroda, Y. Kobayashi, Luminescence emission of isotactic polypropylene, J. Appl. Polym. Sci., 1984, 29, 2843-2849. 19. I. Blakey, G.A. George, Simultaneous FTIR emission spectroscopy and chemiluminescence of oxidizing polypropylene: Evidence for alternate chemiluminescence mechanisms, Macromolecules, 2001, 34, 1873- 1880. 20. R.E. Kelogg, Mechanism of chemiluminescence from peroxy readicals, J. Am. Chem. Soc., 1969, 91, 5433- 5436. 21. R. Broska, J. Rychlý, L. Matisová-Rychlá, Chemiluminescence from tetramethyldioxetane in poly(methyl methacrylate) and polystyrene, Eur. Polym. J., 1996, 32, 1251-1256. 22. V.V. Ptschelintsev, E.T. Denisov, Mechanisms of oxidative degradation of diene rubbers, Vysokomol. Soed. (in Russian), 1985, 27A. 23. J. Rychlý, L. Matisová-Rychlá, M. Strlič, J. Kolar, Chemiluminescence from paper I. Kinetic analysis of thermal oxidation of cellulose, Polym. Degrad. Stab., 2002, 78, 357-367. 24. A.M. García-Campana, W.R.G. Baeyens, Chemiluminescence in Analytical Chemistry, M. Dekker, New York, 2001. 25. F. Shafizadeh, A.G.W. Bradbury, Thermal Degradation of Cellulose in Air and Nitrogen at Low Temperatures, J. Appl. Polym. Sci., 1979, 23, 1431-1442. 26. http://www.tohokueic.com/cl_analyzers_index.html, accessed 25/10/2004. 27. http://www.lumipol.com, accessed 25/10/2004. 28. http://www.atlas-mts.com, accessed 25/10/2004. 29. D. Kočar, J.L. Pedersoli Jr., M. Strlič, J. Kolar, J. Rychlý, L. Matisová-Rychlá, Chemiluminescence from paper II. The effect of sample crystallinity, morphology and size, Polym. Degrad. Stab., 2004, 86, 269- 274.
82 Ageing and stabilisation of paper Degradation
Chapter 6 Acid-catalysed degradation
Andrzej Barański, Janusz Marek Łagan, Tomasz Łojewski
6.1 Introduction Acid-catalysed hydrolysis is a relatively well studied degradation mode of cellulose. Therefore, the present research efforts are shifted somewhat towards applications and mutual interactions of acid hydrolysis and other degradation pathways. This is why the following topics will be emphasised in this Chapter: ― Determination of acidity of library collections as the prerequisite for mass-scale application of deacidification technologies; ― Phenomena caused by simultaneous degradation due to acid hydrolysis and oxidation. The second issue constitutes a background for introduction of the concept of mixed- control mechanism, as well as for better understanding of planning and interpretation of the accelerated ageing tests.
6.2 Paper acidity in library collections. The example of Poland In order to estimate the danger of acidic degradation for library collections, acidity of paper in several Polish libraries has been surveyed recently by means of a pH-pen.1,2 A comprehensive study was also carried out in Poland according to the Stanford method,3 in which not only pH, but also physical conditions of paper, text block and binding were evaluated. In this method, a representative sample of 384 volumes was selected randomly from the 19th and 20th century collection at a given library. In the case of the Jagiellonian Library, Kraków, pH of paper in the majority of selected books has been found to be well below 5.0 (Figure 6.1). As evident in Figure 6.2, books in the surveyed library can easily be divided into two broad categories: acidic and acid-free. Such a division is easy to obtain experimentally using a pH-pen filled with an aqueous solution of chlorophenol red. This indicator changes its colour from yellow (pH < 6.5) to lavender (pH > 7), and it is commonly used in the paper conservation practice.4
Ageing and stabilisation of paper 85 Chapter 6: Acid-catalysed degradation
The tests using this pH-pen were carried out in the Jagiellonian Library on 12,500 books and in 6 other Polish libraries in three cities (10,000 books in total). The results obtained are presented in Figure 6.3. The percentage of books printed on acidic paper equals nearly 100% for the years from 1938 to 1985, and drops dramatically in the mid-1990s to 3.5% in year 1998, and finally to only 0.7% in 2001. These facts correspond with changes in the Polish paper industry – in the 1995 and 1996 the major paper mills shifted the production process to acid-free technology.
9
8
7
H 6 p
r e p
a 5 p
4
3
2 1800 1820 1840 1860 1880 1900 1920 1940 1960 1980 2000 2020 year Figure 6.1: Paper surface pH in books from the Jagiellonian Library collection.5
50
40 s k
o 30 o b
f o
r
e 20 b m u n 10
0
2 3 4 5 6 7 8 9 pH Figure 6.2: Populations of acidic and acid-free books in the set tested by the Stanford method at the Jagiellonian Library.5 The Jagiellonian Library has a status of a national library and collects at least two copies of every book and periodical printed in Poland. Thus, collection of any other library in Poland could be considered as a sub-set (smaller or equal in the number of titles in possession) of the holdings of the Jagiellonian Library. As a result, similar profiles of acidity changes could be expected in all Polish collections. For all tested smaller libraries a very good agreement of obtained data was observed (Figure 6.3). Knowing the annual number of new books acquired by a library in a given period of time, the total number of acidic books can be calculated. Such calculation for the Jagiellonian Library is illustrated in Figure 6.4, where bars represent the number of new acquisitions in a
86 Ageing and stabilisation of paper Chapter 6: Acid-catalysed degradation given year, with the black part of a bar corresponding to the percentage of books with acid- free paper. The points in Figure 6.4 show cumulative numbers – the total number of books in the Jagiellonian Library grew steadily to 1.8 million in year 2001 (squares) but the number of acidic books stabilized in 1996 at 1.5 million volumes (circles), which makes for 80% of their 19th-20th century book collection.
100
) 80 % (
s k o
o 60 b
c i d i c
a 40
f o
n o i t
c 20 a r f all libraries Jagiellonian library 0
1930 1940 1950 1960 1970 1980 1990 2000 2010 year Figure 6.3: The percentage of acidic books in selected Polish libraries through time. Reproduced after Barański et al.1 ©Polish Academy of Sciences, with kind permission of the publisher.
60
) 1.8
s acid-free acquisitions d
n acidic new acquisitions a 50 1.6 s ) u
total number of acidic books s o n h total number of books o t
1.4 i l l i n i
( 40 m
s 1.2 n i n (
o i s t i k
s 1.0 i 30 o u o q b
c 0.8 f a o
r w e
e 20 b
n 0.6
f m o u
n r
0.4 l e a b 10 t o m t
u 0.2 n
0 0.0 1960 1970 1980 1990 2000 year Figure 6.4: The number of acidic and acid-free books at the Jagiellonian Library for years 1955-2001. Reproduced after Barański et al.1 ©Polish Academy of Sciences, with kind permission of the publisher. 6.3 Acid-catalysed degradation – background Methodology The topics of research discussed in the sections to follow can be divided in two categories: ― Taking into account the item size (single-sheets, books, and whole library collections); ― Considering the type of material (model papers prepared specially for research purposes, and commercially available printing papers).
Ageing and stabilisation of paper 87 Chapter 6: Acid-catalysed degradation
The importance of model systems in research is commonly recognized; therefore model papers are worth a more thorough consideration. It is reasonable to regard paper from pure cellulose without any additives as a model compound. For papers containing lignin, a cellulose sample with an added amount of lignin can be the basis for other model studies. Several model samples prepared for international research programmes were made available recently, and results obtained using such materials will be presented in the following text.
Mechanism In this section a molecular approach will be used for the description of acid hydrolysis of + cellulose. In the presence of an acid (i.e. proton-donating hydronium ions H3O , notation equivalent to H+) glycosidic bond is hydrolysed and the macromolecule splits into two shorter units, as shown by the scheme in Figure 6.5.
Figure 6.5: Schematic presentation of acid-catalysed hydrolysis of the glycosidic bond. As usual, the chemical reaction consists of some elementary steps. They are depicted in Figure 6.6. In the first step, the addition of a proton takes place, while later it is removed from the reaction product, thus acting as a catalyst.6 According to this scheme, all glycosidic bonds between anhydroglucose monomer units are equivalent in the ideal cellulose structure, and their splitting occurs randomly.
Figure 6.6: Some elementary steps of acid-catalysed hydrolysis of cellulose. The reaction rate is proportional to the number of available glycosidic bonds, easily expressed by the numerical value of cellulose DP (eq 3.3). This means that a first-order reaction takes place in the ideal cellulose structure. However, in reality some bonds are weaker than others, which is why frequently an initial fast reaction period is observed, followed by a slower one, especially in cellulose samples of high molecular mass. In order to explain this concept of weak links in terms of acid hydrolysis, various ideas have been put forward.7 The most convincing one seems to be: if some hydroxyl groups have been already oxidized to carbonyl and/or carboxyl groups in the real cellulose structure, then the glycosidic bonds in their vicinity become weaker. This turns our attention to the fact that degradation of cellulose may not be only a consequence of acid-
88 Ageing and stabilisation of paper Chapter 6: Acid-catalysed degradation catalysed hydrolysis, but also of other possible degradation pathways, such as oxidation involving atmospheric oxygen. Whitmore and Bogaard8 experimentally verified a method of discriminating the various chain scission routes. They determined the number of carbonyl and carboxyl groups in relation to the number of bond scissions (determined viscometrically), in paper samples subjected to various degradation conditions. In acid-hydrolyzed cellulose there was one carbonyl group per scission. Carboxyl production, or formation of more than one carbonyl per scission, was an indication of oxidation. This is a promising method, as it can help us determine the major scission chemistry of pure cellulose papers in various environments.
Paper acidity As elaborated by Carter,9 there are various reasons for paper acidity. However, it is commonly agreed that the addition of aluminium sulphate Al2(SO4)3·18H2O (alum) during final stages of the papermaking process is the primary cause of paper acidity. Other sources may be: ― Residual chemicals from pulping and bleaching processes; ― Oxidation of lignin to acidic products; ― Migration of acids from storage materials such as cardboard, and even from printing ink media that slowly oxidize over time to yield acidic materials; ― Major air pollutants, e.g. sulphur dioxide and nitrogen oxides (dealt with in more detail in Chapter 10). During the papermaking processes commonly employed from the mid-19th century until the final decades of the 20th century, alum was needed in the sizing step. It facilitates precipitation of the hydrophobic “rosin size” inside the paper structure. Alum is hydrolysed in the presence of water – hexaaquaaluminum(III) ions will hydrolyse to form an hydronium ion:
3+ 2+ + [Al(H2O)6] + H2O [Al(OH)(H2O)5] + H3O , with Ka similar to that of acetic acid.
The kinetic model by Zou, Uesaka and Gurnagul The group of authors published two highly relevant papers in 1996,10,11 in which the authors critically assessed kinetic data for cellulose and/or lignin-free paper degradation originating from the period of 1926-1995 and pointed at some inadequacies and inconsistencies. The main problems associated with the research were outlined in the following way:
― Arbitrary definition of paper degradation rate; ― The use of inappropriate variables for kinetic analysis (mechanical properties used for calculation of degradation rates);
Ageing and stabilisation of paper 89 Chapter 6: Acid-catalysed degradation
― No quantitative kinetic description on the effect of acidity and moisture content in paper; ― No critical examination of the Arrhenius equation; ― No quantitative comparison of the results of accelerated and natural ageing. The authors attempted to solve these problems. The degradation rate was rigorously defined in terms of decrease of cellulose chain length. It was also linked to the rate constant of Ekenstam equation, k (eq 3.7).12 They pointed out that the value of k is equal to the relative rate of scission of glycosidic bonds in cellulose, where the term relative reaction rate refers to the definition intervaltimeduringbrokenbondsofnumber t tk lim)( . (eq 6.1) t 0 tN where N denotes the total number of anhydroglucose units in the system (see eq. 7 in ref. 10). If k(t) is assumed to be constant, the Ekenstam equation is obtained. In further considerations, the degree of polymerisation (DP) was used as a true indicator of the degradation process, and multiple parallel reactions were considered. Acid hydrolysis, along with various oxidation processes, can be considered as the individual parallel reactions. The net reaction rate of the overall reaction is the sum of rates of the individual reactions. As a consequence, and considering the Arrhenius equation (eq 3.9), they obtained: E Akk exp i , (eq 6.2) ii RT where the subscript i denotes the contribution of an individual reaction. The essential result of Zou et al.10,11 is the proof that within a limited range – from reference temperature T0 to the considered temperature T – the net reaction rate constant k can be approximated by the following equation (cf. eq 3.9): E Ak exp a , (eq 6.3) a RT where Aa denotes the apparent pre-exponential factor, and Ea is an apparent activation energy. The smaller the temperature range (i.e. the ratio T0/T is closer to unity), the better is this approximation. The temperature of natural ageing, chosen arbitrarily as 293 K (i.e. 20 °C) can be considered as the reference temperature. These general considerations have been applied by authors to the case of degradation of cellulose pulps and papers, performed in a closed system, within the temperature range of 60-100 °C, RH range of 2-100%, and initial pH range of 4-6. The authors assumed that acid hydrolysis is the dominant degradation pathway, and that temperature, moisture content and pH of paper are the main variables affecting the degradation rate. These assumptions were also verified experimentally. The kinetic equation applied in the model is based on three dependences: ― Ekenstam equation (eq 3.7), ― Classical Arrhenius equation (eq 3.9),
90 Ageing and stabilisation of paper Chapter 6: Acid-catalysed degradation
― Dependence of the pre-exponential factor in the Arrhenius equation on the initial acidity of paper and on the moisture content of the studied sample. The final, condensed equation can be written as:
11 Ea Atkt a exp , (eq 6.4) DPDP 0 RT where
aaa 20 AAAA a5 22 OHHOH ]][[][ . (eq 6.5) The symbol [H+] denotes hydrogen ion concentration in water in paper (mol/L). As this concentration cannot be defined and measured, it must be replaced by (a) either the value calculated from pH measured by extraction of paper in a solution of NaCl,10 or (b) the value calculated from the initial pH value of paper, as measured by the surface method.13 The symbol [H2O] stands for moisture content in paper (m/m).
The parameters Aa2 and Aa5 are linked to the dependence of the reaction rate on water content in paper and on paper acidity, respectively. These dependences have been experimentally confirmed, as presented in Figures 6.7 and 6.8. The slopes and intercepts of lines fitted to the points in the graphs presented in the Figures lead to the following values: 9 -1 12 -1 16 -1 -1 Aa0 = 4.54·10 day , Aa2 = 2.83·10 day , and Aa5 = 9.85·10 L mol day .
The water-independent “free term” Aa0 in (eq 6.5), though numerically small, is of real consequence. It involves all degradation processes other than acidic hydrolysis of cellulose. Thus, a more general character of the model is noticeable.
Figure 6.7: Relationship between frequency factor and moisture content at constant acidity of paper. Reproduced after Zou et al.,10 with kind permission of Springer Science and Business Media. It is also important to point out that extrapolation of the results obtained at temperatures higher than the natural ageing temperature (e.g. 20 °C) should be done with caution. According to the authors, their model holds even when T0/T (in the absolute temperature scale) equals 0.786 (20 °C and 100 °C). In this case the relative error inherent in approximate calculations is about 10%. If temperatures 20 °C and 60 °C are taken, the T0/T ratio is 0.88, and the error in approximation would be about 6%. These theoretically calculated errors do not take into account the experimental errors, which make any
Ageing and stabilisation of paper 91 Chapter 6: Acid-catalysed degradation extrapolation of experimental results even less reliable.
Figure 6.8: Relationship between frequency factor and hydrogen ion concentration at constant RH. Reproduced after Zou et al.,10 with kind permission of Springer Science and Business Media.
Emsley’s model The Ekenstam equation was generalized by Emsley.14 He assumed that the rate constant k in the Ekenstam equation is a function of time, thus indicating that the molecular mass distribution and the mean chain length of cellulose change continuously with time. Additionally, a cellulosic fibre is inhomogeneous, and it is well known that its amorphous regions are more reactive than the crystalline ones. It was found by Emsley that the experimental data concerning degradation of paper at 120 °C (applicable to ageing of paper in power transformers) would be better described by an equation in which k decreased with time in a typical first-order way:
21 tkkk )exp( , (eq 6.6) where k1 and k2 are parameters of the equation. This is true especially for longer degradation times, i.e. when DP of cellulose decreases below 200.
6.4 Studies at Jagiellonian University Experimental information In order to further elaborate the kinetic models, a series of experiments was performed by the research group at the Jagiellonian University, Kraków.15,16 The studied samples were the well defined standard paper samples P1 and P2 developed for the European Research Project on the influence of air pollutants on paper degradation (cf. page 152),17 which have been stored in standard conditions (23 °C, 50% RH, darkness). The papers did not contain any filling agents, and were not sized. Paper P1 contained exclusively softwood cellulose, whereas P2 paper contained more than 95% of cellulose originating from cotton linters and some traces of softwood cellulose. A general description is given by Havermans elsewhere,18 and a detailed characterization can be found in the project report.17
92 Ageing and stabilisation of paper Chapter 6: Acid-catalysed degradation
In some experiments, the paper samples were previously impregnated with aluminium sulphate.16 The presence of aluminium compounds in samples modifies their acidity, e.g. pH of the non-impregnated model P1 paper is 5.70, and after impregnation, the pH value decreased to 4.0-3.8 for samples containing 0.8-4.0‰ of Al. According to Oye,19 the average amount of aluminium in 49 books published in the period of 1919-1957 was 0.2‰. An excess amount of sulphate ions was also found. During the accelerated ageing, closed polypropylene or Teflon® reaction vessels were used. Water or a saturated solution of LiCl, NaI, NaBr, or NaCl was added to the containers (V = 1 L) in order to attain a desired level of relative humidity during the experiment. Above the level of liquid (100 mL), sample strips (approx. 3.5 g) were placed, thus obtaining the ratio of 1 g sample per 260 cm3 of gas phase in the container. This experimental design is advantageous due to its simplicity, although it also has some disadvantages due to the following mass-transfer phenomena inside a container: ― Degradation products may migrate out of paper (a decrease of pH of water below the sample was noticed);15 ― Inorganic salts can migrate and precipitate on the paper strips.20 Also, the volatile degradation products could leak out of the closed reaction container (a decrease of water volume in the vessel was observed at 100 °C, equal to 1.5 mL day-1 for polypropylene and 0.2 mL day-1 for Teflon® container). The degree of polymerisation was determined viscometrically as a function of ageing time (2-21 days). The temperature during isothermal experiments was maintained at 40-100 °C, and relative humidity 10-100%, and the Al content was 0.8-4.0 mg of Al per 1 g of dry mass of paper, as determined by atomic absorption spectrometry.16 The results of this research were published in several papers.15,16,20–22 During further discussion, we will focus on some selected issues.
Limited applicability of Ekenstam equation In preliminary studies it was found15 that the Ekenstam equation described the data of acid degradation of paper P1 well. On the other hand, a linear decrease of DP with time, observed for paper P2, indicated a zero-order reaction. Thus, due to their typical behaviour, samples of paper P1 were used in further studies. A reaction mechanism cannot be proven by kinetic data; it can be, however, easily disproved. If, under certain conditions, the Ekenstam equation does not hold, then acid hydrolysis cannot be the only cellulose degradation pathway. This is why the following criteria of applicability of the Ekenstam equation have been chosen, admittedly somewhat arbitrarily:
― For a fitted kinetic curve [1/DP – 1/DP0] = f(t), the standard deviation of y-residuals (eq 3.13) should be less than 30 DP units;
― The plot [1/DP – 1/DP0] = f(t) should be a straight line with the intercept close to zero (where DP0 is the viscometrically determined degree of polymerisation of the non-aged paper sample);
Ageing and stabilisation of paper 93 Chapter 6: Acid-catalysed degradation
― Activation energies calculated for the rate constant of degradation should be typical for acid hydrolysis (in the arbitrarily chosen range of 100-135 kJ mol-1).
220 4.0‰ Al 200 ) s t i
n 180 u
P 160
D 0.8‰ Al (
n 140 o i t a i 120 v e d 100 d r a
d 80 n a t s
60 l a
u 40 d
i 0.0‰ Al s e
r 20
0 40 50 60 70 80 90 100 temperature (oC) Figure 6.9: Dependence of standard deviations of residuals for individual kinetic experimental curves on the temperature during ageing. Reproduced from Barański et al.,16 where time (days) was erroneously indicated instead of temperature (°C). ©Polish Journal of Chemical Technology, with kind permission of the publisher. These criteria have been applied to approximately a dozen experiments characterised by temperature and Al content in the model paper sample.16 In Figure 6.9, standard deviations of residuals for the linear Ekenstam model were calculated to express the quality of fit of the calculated curves to experimental data in the natural coordinates (DP, t). All three criteria for validity of the Ekenstam equation are fulfilled for the following temperature ranges for pure cellulose: 60–80 °C; for the sample with 0.8‰ Al: 40–50 °C; for the sample with 4‰ Al: 40 °C. These ranges are denoted with full points in Figure 6.9. One can see that these points form a borderline, and the area below this line in the Figure is shaded. At the conditions outside this region (at higher temperatures and higher Al contents) the Ekenstam equation does not hold, and consequentially, acid hydrolysis cannot be regarded as the dominating mode of degradation. It seems reasonable to postulate that only within the shaded area (i.e. at lower temperatures and lower Al contents) acid hydrolysis is the dominating mode of cellulose degradation.
Arrhenius plot curvature Different values of activation energies can be obtained if various temperature ranges are examined more closely.16 The Arrhenius plot for the sample not impregnated by Al is evidently curvilinear (Figure 6.10). This is in some disagreement with Zou et al.,10,11 who successfully applied the Ekenstam equation in the temperature range 60-100 °C. As indicated by the shaded area in Figure 6.9, we could apply it only in the range 60-80 °C. The obtained activation energies 103 18 kJ mol-1 for the paper sample P1,16 are in a good agreement with values 104-113 kJ mol-1, determined by Zou et al. On the other hand, the disagreement between Zou and our results can be ascribed to differences in the paper samples used and in the experimental setups.
94 Ageing and stabilisation of paper Chapter 6: Acid-catalysed degradation
-3.6 E = 133.0 kJ mol-1 (for 60-70 oC) a -3.8 E = 103 ± 18 kJ mol-1 (for 60-80 oC) -4.0 a
-4.2
-4.4 k
g -4.6 o l
-4.8 E = 78±9 kJ mol-1 (for 60-100 oC) -5.0 a
-5.2
-5.4
0.0027 0.0028 0.0029 0.0030 1/T (1/K) Figure 6.10: Arrhenius plots drawn for various temperature ranges for model softwood cellulose paper sample (P1) not impregnated with Al. Reproduced from Barański et al.16 ©Polish Journal of Chemical Technology, with kind permission of the publisher. Basing on our assumed range of activation energies, only the low-temperature range process may be attributable to acid hydrolysis. As the temperature interval considered in calculation of the activation energy is increased, the value of Ea becomes distinctly smaller (Figure 6.10). The value of 70-80 kJ/mol, obtained by applying the Arrhenius equation to the full temperature range, is in a reasonable agreement with the results obtained by other authors23 for chemiluminescence phenomena observed in the course of cellulose oxidation. This may indicate an increasing importance of oxidation as the temperature is increased.
Non-linear dependence of rate constant on aluminium content The effect of Al content on the degradation rate of cellulose is quantitatively expressed by its influence on the rate constant k in the Ekenstam equation – as seen in Figure 6.11. Although the plots in each case are based on three points only, a very distinct pattern of the data is observed. Linearity of the plots is more and more pronounced with lower temperature during the degradation experiments. It can be expected that in the region where acid hydrolysis dominates, at 40 °C, a linear dependence of the rate constant on Al-content exists.
Increase of acidity during advanced degradation phase Let us consider acidity changes of cellulose samples during ageing (Figure 6.12). The highest increase of acidity is observed at the highest temperature 90 °C and higher value of Al content, 4‰.16 At less severe degradation conditions, the increase of acidity is less pronounced. As seen in Figure 6.12 for the sample containing 0.8‰ Al at 40 and 50 °C (i.e. within the shaded area in Figure 6.9, where the Ekenstam equation describing acid hydrolysis is valid) the acidity value remains almost constant over time.
Ageing and stabilisation of paper 95 Chapter 6: Acid-catalysed degradation
o 0.00035 90 C
0.00030
80 oC ) 0.00025 1 - y a d (
t 0.00020 n a t s
n 0.00015
o o
c 70 C
e t a
r 0.00010 60 oC 0.00005
0.00000 0 1 2 3 4 Al content (‰) Figure 6.11: Dependence of rate constants, determined using the Ekenstam equation, on Al-content in model softwood cellulose paper samples (P1). Reproduced from Barański et al.16 ©Polish Journal of Chemical Technology, with kind permission of the publisher.
4.2
4.0
40 oC 3.8 o 50 C
3.6 o 70 C 60 oC H p 3.4
3.2 80 oC
3.0 90 oC
2.8 0 5 10 15 20 25 ageing time (days) Figure 6.12: Paper pH values as a function of ageing time for model softwood cellulose paper (P1) containing 0.8‰ of Al. Reproduced from Barański et al.16 ©Polish Journal of Chemical Technology, with kind permission of the publisher.
Degradation in inert atmosphere A preliminary experiment was also performed with the aim to compare degradation of a P1 paper sample consisting exclusively of softwood cellulose in air and in argon atmosphere. Accelerated ageing was done at 100 °C and 100% RH for 20 days in polypropylene containers. The results are given in Table 6.2.21 It can be clearly seen that due to the presence of oxygen, degradation of cellulose is more advanced.
Table 6.2: Degradation of paper in the presence of oxygen and in the oxygen-free environment in the reaction vessel. Gas DP [H+] (mol L-1)* pH of paper Argon 379 2.4510-4 5.2
96 Ageing and stabilisation of paper Chapter 6: Acid-catalysed degradation
Air 310 3.9810-4 5.0 *in water present in the reaction vessel.
A general conclusion arising from the experiments described above is that degradation of acidic cellulose (and paper) has to be regarded as a complex process, in which hydrolysis of cellulose is not the only possible mechanism of degradation.
6.5 Diffusion phenomena inside books The acidic products formed during accelerated ageing accumulate inside the paper and inter-sheet spaces of books or archival files, resulting in enhanced degradation. This means that a stack of sheets degrades faster than a single sheet. Hanus et al.24 have shown that the degradation rate inside a paper stack is higher than that in the top and bottom sheets. During the experiment done by Carter et al.25 the degradation products of accelerated ageing at 90 °C could only migrate in a stack in one direction and evaporate into the surrounding air from the top of the stack. Thus, the concentration of acids in the top pages was smaller. In a follow-up study26 it was found that this topochemical effect decreased with temperature; it may be supposed that at room temperature, i.e. during natural ageing, it is of minor importance. The effect of temperature was excluded in the study of Brandis and Lyall.27 They took into account 18 books published between 1862 and 1891 and found that the central pages of eleven books were weaker than the outer pages. The effect was not observed with the other seven books. They concluded that exposure to differing storage conditions remains a possible cause of the different ageing pattern of the two groups of books. These different storage conditions may have resulted in different ways of diffusion,21 which can be perpendicular or parallel to the sheets of a book. If a book is put loosely on a shelf, and the distances between adjacent sheets are large, the diffusion flux of the degradation products is parallel to the sheets. Then the products escape easily from the book without diffusing through the other, i.e. the inner sheets, which, therefore, are not attacked by them. On the other hand, if a book standing on a library shelf is tightly squeezed between other books, it is to be expected that the pressure exerted on the stack of sheets prevents swift diffusion of degradation products, thus resulting in a higher concentration of volatile acids in the paper and its enhanced degradation.
6.6 The concept of mixed-control kinetics – acid hydrolysis and oxidation The necessity and importance of analyzing the data obtained during accelerated ageing have been already presented in earlier sections of this Chapter. Porck28 listed 46 different methods of running the accelerated ageing tests. He pointed out that “the small number of research projects that have tried to verify the predictive value of artificial ageing analysis strongly contrasts with the widespread use of this analysis in practice”. This remark is essential for the discussion to follow. Extrapolation of the kinetic data is a crucial step in the procedure of analysis of accelerated
Ageing and stabilisation of paper 97 Chapter 6: Acid-catalysed degradation ageing. Application of this concept to the description of complex degradation phenomena will be discussed below. Three premises will be taken into account: 1. Complexity of degradation phenomena is indicated in the literature as well as from our own results. Degradation of whole books is affected by diffusion phenomena. The higher is the temperature and Al-content, the higher is the probability that acid hydrolysis, contributes to the degradation of paper, along with oxidation processes. This hypothesis is strongly supported by the observed difference between degradation rate of non-impregnated paper in air and argon atmosphere. Other arguments supporting this hypothesis have been outlined in ref. 16. 2. The process of cellulose oxidation is complex, and it can proceed through many reaction paths. For the sake of simplicity, let us assume that only one oxidation path is possible, and let us also ignore the diffusion phenomena mentioned above. 3. A fundamental methodological principle is the third premise applied here. If one wants to use Arrhenius equation for the description of temperature dependence of the rate constants, then one must be sure (or assume) that the considered kinetic equation is in agreement with the results of isothermal experiments in the whole range of temperatures studied. Applying these assumptions to the kinetic data presented above, we will try to draw some more general conclusions. The Arrhenius equation can be applied to the rate constant of Ekenstam equation inside the shaded area seen in Figure 6.9. The latter equation describes a single degradation route, this being acid hydrolysis. Two degradation routes exist outside this area – acid hydrolysis and oxidation – and, therefore, the Ekenstam equation does not hold. Another, at this time unknown, equation (let us call it a mixed-control equation) will properly describe the kinetic data. The postulated equation will contain two rate constants – kh and kox – for acid hydrolysis and oxidation respectively. There will be two Arrhenius plots: ln(kh) versus 1/T and ln(kox) versus 1/T, and, consequently, two activation energies can be calculated. It seems obvious that continuity of the system properties should be observed in such a case. One can expect that the mixed-control equation, if extended to the low-temperature region (inside the region defined by the borderline in Figure 6.9), will yield the Ekenstam equation (eq 3.7) as the limiting case.
Therefore the Arrhenius plot for kh should be linear (i.e. no deflections observed) in the whole range of experimental conditions – outside and inside the borderline in Figure 6.9. On the other hand, we may presume that the kinetic data providing kox values are of significance only outside the borderline. This is why the extrapolation of kox value to the temperatures below the borderline should be avoided as much as possible. Degradation of cellulose is an example of a more general scheme of kinetic studies, in which one needs to apply the Arrhenius equation to a complex reaction system. This can be summarized as follows. If during the studies of a complex process there are two ranges of some experimental variable (e.g. temperature), and: ― In the first range a single mechanism A exists, ― In the second range there are n reaction pathways, namely A, B, C, etc.,
98 Ageing and stabilisation of paper Chapter 6: Acid-catalysed degradation then mixed-control equation is needed to describe the kinetic phenomena in the second range, and, consequently, to extrapolate the data from the second range to the first one. No single Arrhenius plot, but n Arrhenius plots are necessary when describing the system in question. If this task can be achieved then the Arrhenius plot for mechanism A should be used for extrapolation purposes. 6.7 Perspectives A classification of research in the field of paper and cellulose degradation has been presented earlier in this Chapter, taking into account the size of the investigated items (single paper sheets, books, and library collections). It seems reasonable to consider perspectives of further research along the lines of this classification. A fundamental study can be performed most easily if we focus our attention on the simplest system – a sheet of model paper. The fundamentals of simultaneous description of both acid hydrolysis and oxidation, occurring during paper degradation, have been discussed in this Chapter. The most important research goal in this field is to design such a system of measurements in which it would be possible to determine – simultaneously and quantitatively – the advance of both hydrolysis and oxidation. As the oxidation process can proceed in many different ways, it is most desirable to find a way of further distinction between these oxidation pathways. Spectroscopic measurements in the infra-red region may be of pivotal significance in this respect. The degradation products of individual sheets, as well as acidic gaseous substances always present in the ambient atmosphere, are factors which accelerate the degradation of books. The former are endogenous (created inside books), whereas the latter are exogenous (penetrate the books from the environment). Books can be protected from sulphur dioxide and nitrogen oxides by wrapping, but this makes removal of acid degradation products impossible. The most advantageous solution of this problem would be to place suitable adsorbents in containers used for book storage. Such suggestions are already known from patents and literature,29 but the problem requires further research and optimization of costs.
6.8 References 1. A. Barański, K. Frankowicz, Z. Harnicki, Z. Koziński, T. Łojewski, Acidic books in libraries. How to count them?, Proc. 5th European Conf. Cultural Heritage Research: a Pan-European Challenge, R. Kozłowski, Ed., Cracow, Poland, May 16–18th 2002, Polish Academy of Sciences, Cracow, 2003, 283-285. 2. A. Barański, T. Łojewski, K. Zięba, Acidic Books in Poland, Conf. Chem. Technol. Wood, Pulp and Paper, G. Baudin, J. Fellegi, G. Gellerstedt, S. Katuscak, I. Pikulik, J. Paris, Eds., Slovak University of Technology, Bratislava, 2003, 398-401. 3. W. Sobucki, B. Drewniewska-Idziak, Survey of the Preservation Status of the 19th and 20th Century Collections at the National Library in Warsaw, Restaurator, 2003, 24, 189-201. 4. E. McCrady, pH Pens and Chlorophenol Red, 1995, http://palimpsest.stanford.edu/byorg/abbey/phpens.html, accessed 15/09/2004. 5. Data obtained from Mr. Z. Koziński, Jagiellonian Library, Kraków, Poland. 6. D. Klemm, B. Philipp, T. Heinze, U. Heinze, W. Wagenknecht, Comprehensive Cellulose Chemistry, WILEY-VCH, Weinheim, 2001, vol. 1. 7. A.-L. Dupont, Gelatine Sizing of Paper and its Impact on the Degradation of Cellulose During Aging, PhD Thesis, University of Amsterdam, 2003, 18-19. 8. P.M. Whitmore, J. Bogaard, Determination of the Cellulose Scission Route in the Hydrolytic and Oxydative
Ageing and stabilisation of paper 99 Chapter 6: Acid-catalysed degradation
Degradation of Paper, Restaurator, 1994, 15, 26-45. 9. H.A. Carter, Chemistry in the Comics. Part 3. The Acidity of Paper, J. Chem. Educ., 1989, 11, 883-886. 10. X. Zou, T. Uesaka, N. Gurnagul, Prediction of paper permanence by accelerated aging. I. Kinetic analysis of the aging Process, Cellulose, 1996, 3, 243-267. 11. X. Zou, T. Uesaka, N. Gurnagul, Prediction of paper permanence by accelerated aging. II. Comparison of the predictions with natural aging results, Cellulose, 1996, 3, 269-279. 12. A. Ekenstam, Über das Verhalten der Cellulose in Mineralsäure-Lösungen, II Mitteil: Kinetisches Studium des Abbaus der Cellulose in Säure-Lösungen, Ber., 1936, 69, 553-559. 13. Tappi Standard T529 om-99 Surface pH Measurement of Paper. 14. A.M. Emsley, R.J. Heywood, M. Ali, Ch.M. Eley, On the kinetics of degradation of cellulose, Cellulose, 1997, 4, 1-5. 15. A. Barański, R. Dziembaj, A. Konieczna, A. Kowalski, J.M. Łagan, L.M. Proniewicz, Methodology of kinetic investigation of cellulose degradation, Chemical Technology Between Centuries, Permanent Committee of Chemical Technology Congresses, Gliwice, Poland, 2000, 441-450. Available on-line: http://www.chemia.uj.edu.pl/~kp/deg_kinetics.pdf. 16. A. Barański, R. Dziembaj, A. Konieczna–Molenda, J.M. Łagan, S. Walas, On the applicability of Arrhenius equation to accelerated aging tests. The case of alum-impregnated cellulose, Pol. J. Chem. Technol., 2004, 6, 1-8. 17. Step Project CT 90–0100, EC–DG XII, The Effect of Air Pollutants on the Accelerated Ageing of Cellulose Containing Materials – Paper, Report BU3.94/1068/JH, TNO (Centre for Paper and Board Research), Delft, 1994. 18. J.B.G.A. Havermans, Effects of air pollutants on the accelerated ageing of cellulose-based materials, Restaurator, 1995, 16, 209-233. 19. R. Oye, T. Okayama, M. Akagi, T. Ohnishi, Degradation of Paper – Indication of degree of degradation; Appita, Proc. 6th Intern. Symp. Wood Pulping Chem., Vol 1, 1991, 397. 20. A. Barański, D. Dutka, R. Dziembaj, A. Konieczna-Molenda, J.M. Łagan, Effect of Relative Humidity on the Degradation Rate of Cellulose. The Methodology Studies; Restaurator, 2004, 25, 68-74. 21. A. Barański, Aging Kinetics of Cellulose and Paper, Restaurator, 2002, 23, 77-88. 22. A. Barański, A. Konieczna-Molenda, J.M. Łagan, L.M. Proniewicz, Catastrophic Room Temperature Degradation of Cotton Cellulose, Restaurator, 2003, 24, 36-45. 23. M. Strlič, J. Kolar, B. Pihlar, J. Rychly, L. Matisova-Rychla, Chemiluminescence during thermal and thermo-oxidative degradation of cellulose, Eur. Polym. J., 2000, 36, 2351-2358. 24. J. Hanus, M. Komorníková, J. Mináriková, Changes in some mechanical properties of paper during aging in an archival box, ICOM-CC 11th Triennial Meeting, Edinburgh, Sept. 1-6., Vol. 2, 1996, 510-516. 25. H. Carter, P. Bégin, D. Grattan, Migration of volatile compounds through stacked sheets of paper during accelerated aging. Part 1: Acid Migration at 90 °C, Restaurator, 2000, 21, 77-84. 26. A. Bülow, P. Bégin, H. Carter, T. Burns, Migration of volatile compounds through stacked sheets of paper during accelerated aging. Part II: Variable temperature studies, Restaurator, 2000, 21, 187-203. 27. L. Brandis, J. Lyall, Properties of paper in naturally aged books, Restaurator, 1997, 18, 115-130. 28. H.J. Porck, Rate of paper degradation. The predictive value of artificial aging tests. Amsterdam: European Commission on Preservation and Access, 2000. 29. F. Daniel, V. Hatzigeorgiou, S. Copi, F. Flieder: Etude de l’Efficacite d’un Nouveau Produit d’Archivage: le MicroChamber, Les Documents Graphiques et Photographiques: Analyse et Conservation. Travaux du CRCDG 1994-1998, Paris, 1999, 25-50.
100 Ageing and stabilisation of paper Chapter 7 Thermo-oxidative degradation
Matija Strlič, Jana Kolar, Drago Kočar, Jozef Rychlý
7.1 Introduction As shown in the previous Chapter, even in conditions in which acid hydrolysis is the predominant degradation pathway of cellulose, oxidation may take place. It is thought that as the pH of paper is increased, the relative importance of acid hydrolysis decreases. In a recent study,1 degradation of pure cellulose was shown to strongly depend on pH in the acidic region of the pH scale, which is a well-known2 and quantified fact,3 while there was no difference between the rate constants at pH 7.3 and 8 (Figure 7.1). In the pH region around neutral, the degradation thus slows down, and two phenomena are thought to play a major role: oxidation and alkaline degradation.
6
5 ) 1 - s
1 - l
o 4 m
l o m (
) 0 1 3 0 1 · (
k
2
1 3 4 5 6 7 8 9 pH Figure 7.1: Degradation rate constants for cellulose samples (WH) with different pH of aqueous extracts. Samples were immersed in phosphate buffer solutions with corresponding pH, dried in air and artificially aged at 90 °C, 65% RH. The rate constants were obtained using the Ekenstam equation (eq 3.7). The error bar represents SD (n = 3). Cellulose oxidation is a long-studied subject yet mostly in connection with pulping processes, where high temperature, pressure, pH and high concentrations of the oxidant
Ageing and stabilisation of paper 101 Chapter 7: Thermo-oxidative degradation regulate the kinetics.4 The processes taking place in conditions of atmospheric oxidation, i.e. with atmospheric oxygen, were reviewed recently5, and then further discussed by Kolar et al.6 While the oxygen-independent alkaline degradation7 proceeds in the vicinity of carbonyl functional groups, formed during oxidation, it is the latter process that has to be targeted in order to achieve a stabilisation effect. Judging from Figure 7.1, deacidification of paper, i.e. an increase of pH from 4 to 7.3, will lead to a three times lower rate of degradation at 90 °C, 65% RH. However, what is the chemistry of processes taking place after deacidification? We hope to answer this question, at least partly, in this and the following two Chapters.
7.2 Autoxidation As outlined in Chapter 2, the autoxidation scheme is thought to adequately describe the process of oxidation of organic polymers, including cellulose (Figure 7.2). The relative importance of certain reactions is different for certain polymer systems. In cellulose, the large content of functional groups has its influence. It is almost never studied as a chemically pure compound (in fact it is even difficult to obtain) so that it naturally contains a number of functional groups other than hydroxyl, e.g. carbonyl, aldehyde (other than end-groups) and carboxyl.
Figure 7.2: The Bolland-Gee autoxidation reaction scheme with the individual reactions outlined. The native cellulose polymer is denoted as PH. This is even more pronounced in bleached wood pulp, where the remains of lignin may also play an important role. Finally, the situation is even more complex in paper, where additives, fillers, coatings, and other admixtures further complicate the system. Even fully delignified pulps contain non-cellulosic components, e.g. hemicelluloses, whose chemical structure and composition has nothing in common with cellulose, apart from being a polysaccharide. The closest model for pure cellulose is probably filter paper (for qualitative purposes, i.e. acid-washed and without any additives), which is made of cotton and extensively purified in alkalis and acids. In order to study the contributions of individual reactions, we have to resort to such model systems, and later extend the knowledge to real paper systems, i.e. factual samples.
102 Ageing and stabilisation of paper Chapter 7: Thermo-oxidative degradation
However, the application of the autoxidation scheme, which was originally developed for autoxidation of simple hydrocarbons in solution, to heterogeneous systems, such as atmosphere/cellulose, is associated with a considerable risk. E.g., the addition of O2 to P is a diffusion-controlled reaction,8 with activation energy 0 kJ mol-1,9 and at ambient conditions, other reactions of P are negligible. However, if diffusion of oxygen to the reaction sites is impaired, e.g. due to slow diffusion in ordered (crystalline) regions, the relative importance of other reactions of P may increase. On the other hand, mobility of polymeric chains is lower in comparison with the mobility of low-molecular-weight compounds in solutions. Differences in mobility may easily lead to differences in rates of reactions; and instead of one rate of reaction we may well have to speak of a distribution of rates. A notable example of a sometimes sharp discontinuity in behaviour is the change in rate of degradation of polymers below or above the glass transition temperature. In cellulose, the glass transition temperature is poorly defined; however, the presence or absence of plasticizers, such as water in the case of cellulose, may greatly affect the mobility of polymeric molecules. The level of heterogeneity in polymer oxidation is a much discussed topic.10-12 The initiation reactions are of particular importance for the autoxidative process, both from the aspect of the rate of the overall degradation process and from the aspect of stabilisation (cf. page 18). Formation of radical species may occur both thermally, e.g. at elevated temperatures,13 and during irradiation.14 Neither pathway is of interest in this Chapter: instead, we would like to focus on reactions at low temperatures (room conditions) and in darkness. The fact that in air the rate of degradation is higher than in inert atmosphere,15 is itself a proof that oxidation of cellulose is an important degradation pathway. The individual reactions in Figure 7.2 are difficult to discern, however, with different experimental approaches we may observe the overall changes in the degree of polymerisation if one or the other reaction is promoted or suppressed.
7.3 Initiation of oxidation A direct reaction between ground state oxygen molecule and cellulose is unlikely, as it is a spin-forbidden process. The more common reactive oxygen species are superoxide anion – (O2 ) with its conjugated acid, i.e. hydroperoxyl radical (HOO ), hydrogen peroxide (H2O2) and hydroxyl radical (HO ). In aqueous solutions, at the partial pressure of oxygen above the solution 1 atm and pH 7, the following reduction potentials have been established:16
and even superoxide is supposed not to be able to abstract a hydrogen atom from D-glucose.17 The mobility of superoxide in a hydrated cellulose molecule must therefore be high, provided that the content of water is sufficient. On the other hand, hydroxyl radicals are unspecific due to their high reactivity and react at an almost diffusion-controlled rate with a variety of compounds.18
Ageing and stabilisation of paper 103 Chapter 7: Thermo-oxidative degradation
In a cellulose macromolecule (Figure 1.3), probably only the aldehyde end groups are capable of spontaneous reaction with oxygen, which is promoted by alkalinity. The mechanisms of carbohydrate oxidation in alkaline media were reviewed by Arts et al. and the reaction scheme as in Figure 7.3 may be put forward.19,20
Figure 7.3: Two possible reaction pathways for oxidation of carbohydrate end-groups by oxygen in a mildly alkaline environment. R denotes the rest of a monosaccharide. However, the reaction scheme was not applied to polysaccharides successfully until recently. A satisfactory model for such studies would be cellulose of different molecular weight, in which the only aldehyde group would be the end one. However, cellulose cannot be obtained in narrow distribution of molecular masses and of sufficiently high purity, so that other polysaccharides were studied instead. Pullulan differs from cellulose only in geometry of the glycosidic bonds, which is not expected to affect the mechanisms in question, considerably. In a recent study,21 pullulan samples of different molecular weights and thus different content of aldehyde end-groups were used.
3000
) 2500 U A (
a e r 2000 a
k a e p
e 1500 d i x o r e p 1000
500 10000 100000 1000000 peak molecular weight (g mol-1)
Figure 7.4: Peroxide peak areas for pre-oxidised (60 min, O2, 100 °C) pullulan samples of different initial peak molecular weights.21 As we will discuss in more detail in Chapter 8, during dynamic heating in inert atmosphere, a chemiluminescence phenomenon due to the presence of peroxides can be observed in pre-oxidised samples (cf. Figure 8.3).22 Such peaks were also observed in pullulan samples after pre-oxidation at 100 °C for 60 min.21 The duration of this pre-treatment is sufficient to allow for formation of a steady-state concentration of peroxides.23
104 Ageing and stabilisation of paper Chapter 7: Thermo-oxidative degradation
It seems that peroxides are indeed formed from aldehydes during oxidation of pullulan in oxygen atmosphere, possibly according to the pathway in Figure 7.3. Using pullulan samples of different molecular weights, it was shown that after the pre-oxidation treatment, the content of peroxides depends on the initial content of aldehyde groups (Figure 7.4).
1E-6 ) 1 - n i m
1 - r e m o n o m l o m
l o m (
t 1E-7 n a t s n o c
e t a r
n o i t a d a r g
e 1E-8 d 1E-5 1E-4 1E-3 0.01 0.1 content of aldehyde groups (mol mol -1) monomer Figure 7.5: First-order degradation rate constants for pullulan samples in a mildly alkaline environment in air, 80 °C, 65% RH. The line represents a linear fit of the experimental data in the log-log scale.21 The content of peroxides should undoubtedly influence the rate of the degradation process. Therefore, pullulan samples were also aged in a mildly alkaline environment ensured by surplus CaCO3. During a 13-day degradation period at 80 °C, 65% RH, the Ekenstam plots (page 33) were composed of two distinct parts: during an initial period the degradation was faster, while during the advanced period, the degradation of most samples proceeded at a similar rate indicating that a steady-state content of aldehydes had formed (cf. Figure 8.7). A comparison of the initial rates shows (Figure 7.5) that the content of aldehydes decisively influences the degradation rate. A linear fit of satisfactory quality (R = 0.9821) transforms into the curve in Figure 7.5 due to the logarithmic nature of the axes. In paper, aldehyde groups may originate in components other than cellulose. The addition of glucose to paper is known to accelerate the degradation considerably.6 Since aldehyde- group-containing compounds, glucose among them, are products of acid-catalysed hydrolysis, it is therefore extremely important that they are washed out of paper during stabilisation treatments (washing, deacidification). During delignification, wood pulps are subjected to various processes, during which a variety of oxidised groups on cellulose can be formed. Considering the above findings, bleached chemical pulps are thus especially prone to autoxidation. This was recently demonstrated in a study of a variety of differently deacidified pulps with different contents of carbonyl groups (Figure 7.6).24 Another feature is evident in Figure 7.6, i.e. the dependence on the type of deacidification. This may be due to two factors: the alkali-earth metal or pH. Both are known to influence the degradation mechanisms considerably. While we shall return to the question of alkali-earth metal on page 112, from reactions on
Ageing and stabilisation of paper 105 Chapter 7: Thermo-oxidative degradation
Figure 7.3 it is also evident that pH should be of prime importance for the rate of initiation, as dissociation of enediols is promoted by higher pH. However, since pH in real papers depends also on the wide variety of paper additives, the observed alkali sensitivity of bleached cellulose pulps, as it is sometimes referred to, may not be problematic anymore. After reduction of carbonyl groups this alkali sensitivity seems to be eliminated.24 Furthermore, the content of carbonyl groups in bleached chemical pulps is higher than in cotton-based pulps – bleached chemical pulps are thus appropriate models only for certain well-selected studies. Their increased susceptibility towards thermal-ageing-induced discolouration (yellowing) is well known.25
35
30 MgCO 3 R2 = 0.9538 25 ) % (
20 e s a e r
c 15 e
d CaCO 3 P R2 = 0.7694 D 10
5
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 content of carbonyls (mmol g-1) Figure 7.6: DP of various differently deacidified bleached chemical pulps during 7-day ageing at 80 °C, 65% RH vs. their carbonyl content prior to ageing.24
1.8
1.6 ) g
0 0 1 /
l 1.4 o m m (
1.2 t n e t n
o 1.0 c
p u o r
g 0.8
l y n o b
r 0.6 a c
0.4 0 200 400 600 800 1000 1200 time (min) Figure 7.7: Content of carbonyl groups in a notebook paper made of bleached pulp, during a reduction treatment with 0.01 mol L-1 aqueous solution of NaBH4.6 For stabilisation studies, it is important to realise that by reducing the carbonyls, we may thus greatly reduce the rate of degradation. A reduction treatment may consist of 6 immersion for several hours in a solution of NaBH4 (Figure 7.7). In this section, the importance of initiation was outlined, indicating that by removal of carbonyl groups, from which enediols may form, stabilisation of paper may be achieved. In
106 Ageing and stabilisation of paper Chapter 7: Thermo-oxidative degradation the following few sections we shall review other important factors influencing the rate of autoxidation.
7.4 Peroxides and hydroxyl radicals Among the more important oxidative species, hydroperoxides (HOOR) and certainly hydroxyl radicals (HO) play an important role. Kleinert and Marracini have already shown that peroxides are formed during cellulose oxidation.26 In fact, the reactions in Figure 7.3 lead to production of a number of radicals, including the possibly long-lived superoxide. Its 27 conjugated acid (pKa = 4.8) is easily formed:
– – O2 + H2O → HO2 + HO , and disproportionation of hydroperoxyl radicals 2HO2 →→ O2* + H2O2, leads to formation of both hydrogen peroxide and singlet oxygen. Using a chromatographic method, it was shown that the process of build-up of hydroperoxides during accelerated ageing of paper at 80 °C, 65% RH is fairly rapid and that the content is extremely low. Using chemiluminometric experiments, similar to the one shown in Figure 8.3, we can easily follow formation or decay of peroxide species. Even at room temperature, peroxides formed during oxidation for 60 min in oxygen atmosphere at 80 °C, degrade fast (Figure 7.9), with a steady state attained in 100 min, apparently following first-order kinetics.
50
45
) SA 1 - g k
40 l o m (
35 t C n e t
n 30 o c
e d i 25 x o r e p 20 WH
15
0 20 40 60 80 100 120 140 time (min) Figure 7.8: Kinetics of formation of hydroperoxides during oxidation of three cellulose pulps at 80 °C, 65% RH.23 The error bar represents SD (n = 3). Peroxides (hydrogen peroxide, hydroperoxyl and organic peroxyl radicals and hydroperoxides) have thus been shown to take part in the process of cellulose oxidation. As will be shown in the following Chapter, they are also thought to be the primary source of chemiluminescence during oxidation of cellulose and paper, especially in dry atmosphere. From hydroperoxides, even more reactive species may be formed, especially if transition metals are present in the material, at least in traces:
n+ (n+1)+ – Me + H2O2 → Me + HO + HO .
Ageing and stabilisation of paper 107 Chapter 7: Thermo-oxidative degradation
The reaction is named after H. J. H. Fenton, and more than a century after its discovery, there is intensive scientific debate about its exact mechanism in a variety of reaction systems.28-30 This is not surprising, considering its important role in living organisms and the environment, but also in degradation of materials.9 In paper,6,31 it is most often referred to in studies of metal tannate (iron gall) inks.32 The activity of transition metals in paper depends on many parameters, including complexation with various ligands,33 pH and type of metal;34 the effects of different metals may even be synergistic.34,35 The particular aspects of transition-metal mediated paper degradation and stabilisation strategies are dealt with in another volume in considerable detail.36
450
400 )
U 350 A (
a e r
a 300
k a e p
250 e d i x o r 200 e p y = 700·e-0.035·x + 130 150
100 0 50 100 150 200 250 300 350 400 450 time (min) Figure 7.9: Kinetics of decay of peroxides, followed chemiluminometrically, during storage of pre- oxidised (60 min, 80 °C, O2) sulphite bleached pulp in darkness at 22 °C. Formation of hydroxyl radicals in paper has not yet been indicated using appropriate instrumental methods, e.g. electron spin resonance.37 Use of organic compounds specially designed for the reaction,38 has lead to a positive proof of the role of HO during oxidation of paper.39 After introduction of the scavenger N,N’-(5 nitro-1,3-phenylene)bisglutaramide (NPG) into paper, it reacts with the formed HO, thereby turning into a coloured hydroxylated product. Using colourimetry, the rate of formation of HO can thus be conveniently followed during accelerated ageing (Figure 7.10). This result indicates the importance of formation of hydroxyl radicals for the chain-scission process during oxidation of cellulose. From the Figure, it also evident that in CaCO3- deacidified bleached pulp the formation of HO is much less pronounced than in MgCO3- deacidified bleached pulp, which is in line with the results discussed above (Figure 7.6). Due to the extreme reactivity of hydroxyl radicals and due to the postulated low number of autoxidation cycles in cellulose,40 the addition of substances scavenging the radicals has little effect on the rate of degradation.6
108 Ageing and stabilisation of paper Chapter 7: Thermo-oxidative degradation
14
12 M6 C20
10
) 8 C13 % (
e s
a 6 C2 e r c
e M2 d
4 P D 2 C2
0 C0
-2 0 5 10 15 20 25 diffuse reflectance decrease at 460 nm (%) Figure 7.10: Decrease of DP (of samples without N,N’-(5 nitro-1,3-phenylene)bisglutaramide – NPG) vs. decrease of diffuse reflectance (of samples with NPG) of Royal Kraft sulphate bleached pulp during accelerated ageing. C and M denote the deacidification type: CaCO3 and MgCO3, respectively, the numbers indicate the duration of accelerated ageing at 80 °C, 65% RH in days.39
7.5 Role of water In paper, water provides the reaction medium, in which ionic species are solvated at least to some extent. In amorphous parts of cellulose, the content of water may be up to 18%,4 whereas in crystalline regions the content is considerably lower. According to recent NMR studies, if the content of water in paper is <3%, the molecules are practically immobile. However, if the content is >9%, there is evidence of a condensed water phase, which means that it becomes available to all potential chemical reactions.41 There are several indications that some reactions in aqueous solutions and in paper proceed in a similar manner, although the medium is very different. Such examples are the catalytic behaviour of chelated iron,33 and the catalytic behaviour of a variety of other transition metals42: in both cases a correlation between their role in solution and in paper was obtained. At a constant temperature, the content of water in cellulosic materials depends on the humidity of the surrounding atmosphere.4 The so-called water absorption isotherms are not linear, instead, the content of water in cellulose increases sharply at the beginning and towards the end of the scale of relative atmospheric humidity (Figure 7.11). Up to 80% RH, moisture is taken up by the material in clusters in a uniform manner,43 whereas at higher RH, the process is thought to be non-uniform.44 This also indicates that, whatever the effect of water during degradation may be, it will most likely not be linearly dependent on relative humidity of the environment. Apart from providing the reaction medium, water has a plasticizing effect on cellulose.45 Thus, it will, by formation of hydrogen bonds with a macromolecule, shield its interactions with other macromolecules, decrease chain rigidity, facilitate segmental motion and decrease internal friction in the material. Due to its plasticizing effect, increased water content may lead to an increase of reaction rates of macromolecules.
Ageing and stabilisation of paper 109 Chapter 7: Thermo-oxidative degradation
14
12 ) % (
r
e 10 p a p
n i
t 8 n e t n o c
r 6 e t a w 4
2 20 30 40 50 60 70 80 90 100 RH (%) Figure 7.11: Water absorption isotherm (22 °C) for a Whatman filter paper sample. The error bar represents SD (n = 3). As a reactant, water participates in acid catalysed hydrolysis and it has been illustrated many times, that the rate of degradation of acidic papers depends on the relative humidity of environment.3,46-49 In Figure 7.12, the difference in behaviour of two pulps of slightly different pH during degradation at 90 °C is shown. The behaviour of Whatman paper is typical for an acidic sample, while the sulphate bleached pulp seemingly shows almost no dependence on relative humidity of the environment.1 In order to decrease the degradation rate of acidic papers during long-term storage, a cost effective strategy would thus seem to be to lower the relative humidity: the rate of degradation at 90 °C is decreased by 2.2 times if the relative humidity is decreased from 65% to 20%.
6 WH (pH 6.6) SA (pH 7.2) 5 ) 1
- 4 s
1 - l o m
l 3 o m (
) 0 1
0 2 1 · (
k
1
0 0 20 40 60 80 100 RH (%) Figure 7.12: Rate constants for degradation of two papers with different pH at 90 °C and different relative humidities in air, as indicated.1 Although the behaviour of sulphate pulp in Figure 7.12 apparently shows no interesting features, they are revealed if the scale is expanded. Indeed, the dependence of rates of degradation of the three deacidified pulps under study in Figure 7.13 on relative humidity shows some distinct features.
110 Ageing and stabilisation of paper Chapter 7: Thermo-oxidative degradation
1.4 CaCO 1.4 WH 3 MgCO SA 3 1.2 C 1.2 ) ) 1 1 - 1.0 - 1.0 s s
1 1 - - l l o o
m 0.8
m 0.8
l l o o m m ( (
0.6
) 0.6 ) 0 0 1 1 0 0 1 1 · · ( 0.4 (
0.4 k k
0.2 0.2
0.0 0.0 0 20 40 60 80 100 0 20 40 60 80 100 RH (%) RH (%) Figure 7.13: Degradation rates for samples containing CaCO3 (pH 7.6-8.4) and MgCO3 (pH 9.4-9.9), at 90 °C and different relative humidities in air.1 The error bars represent SD (n = 3). The experimental uncertainty is apparently high, however, it seems obvious that the rate of chain scission during oxidation of the alkaline samples at 90 °C increases as the relative humidity increases from 20-65%, and decreases again at very high levels of relative humidity. Kleinert and Marracini investigated the role of relative humidity during oxidation studies at temperatures lower than 100 °C,25 and they found out that a higher RH leads to a higher content of peroxides. The extremely low rate of degradation at 95% RH is of low practical value, but it is certainly of high scientific interest. Considering the various possible ways in which an increased content of water may lead to an increased rate of cellulose degradation, outlined at the beginning of this section, it seems curious that the minimum rate of degradation is observed at 95% RH. While higher humidity in the surrounding atmosphere decreases the partial pressure of oxygen from 21 vol.% (0% RH) to 7.4 vol.% (95% RH), this effect could not wholly account for the observed decrease in the rate of degradation.1 Considering that a sharp increase in water content in cellulose is not observed until high RH in the environment is reached (Figure 7.11), a possible explanation may be that a completely hydrated cellulose fibre is less susceptible to oxidation simply due to the fact that solubility of oxygen in water is low. Water may thus represent an efficient barrier. The decrease of rate of oxidative degradation as relative humidity is decreased can easily be explained with all the effects water apparently exhibits during degradation of paper: plasticizing, reaction medium, and reactant. Again, it should be of practical value to note that for sulphate bleached pulp, the rate of degradation at 20% RH is 8.7 times lower than at 65%. A proper adjustment of relative humidity during long-term storage might therefore lead to a considerable positive effect also for alkaline materials. However, only a full Arrhenius study and extrapolation to room temperature can provide an estimate of the actual effect at room temperature. Figure 7.14 provides an indication that the effect of humidity during ageing increases if the ageing temperature decreases. In the Arrhenius graph, there is a distinct discontinuity at approximately 100 °C. Degradation experiments at T > 100 °C are thus directly relevant for the low temperature region only in the absence of humidity. The apparent activation energy for the degradation process in dry conditions is (93 ± 2) kJ mol-1, while in humid atmosphere it is significantly different: (126 ± 12) kJ mol-1.
Ageing and stabilisation of paper 111 Chapter 7: Thermo-oxidative degradation
-14
-16
-18
-20 k -22 n l
-24
-26 0% RH 65% RH -28
0.0020 0.0022 0.0024 0.0026 0.0028 0.0030 0.0032 1/T (K-1) Figure 7.14: Arrhenius plot derived from degradation rate constants of sulphate bleached pulp aged at 0% and 65% relative humidity and various temperatures.1 The differences between dry and moist heat ageing are well known and were quantified for acidic samples,3 and the research of Erhard et al.50,51 on identification of degradation products indicated that the degradation mechanisms might be dissimilar. Samples containing CaCO3 were also already shown to degrade more slowly than those containing 52 53 MgCO3, at lower and at higher temperatures. This is also indicated in our studies (Figure 7.13). However, there is an apparent anomaly in ageing of MgCO3-containing samples in dry conditions (Figure 7.13), as the rate of degradation of all dry samples is higher than the rate of degradation at 20% RH. This is not the case of CaCO3. The reason for this behaviour might be in the strong promoting role of Mg2+ in electron transfer reactions.54 That complexation of enediols in solution is more favoured with Mg2+ than with Ca2+ is well known,55 and it is possible that this affinity persists in dry conditions, while as soon as some water is present, Mg2+ is probably rapidly hydrated.56
7.6 Thermo-oxidative degradation: Ca or Mg? The continual doubt of whether to use Ca-based or Mg-based deacidification solutions has been the subject of many studies.52,57,58 From the considerations outlined above, two effects might have to be taken into account: the effect of higher solubility of MgCO3 and thus 59 higher pH values of MgCO3-containing papers; and the effect of the cation. The much discussed effect during ageing at 80 °C can be explained in Figure 7.15. Each point represents the relative difference in DP vs. the relative difference in pH for the same pulp/paper sample: a positive y value means better behaviour of the MgCO3-containing sample compared with the same sample deacidified with Ca(HCO3)2. A positive x value means higher pH (determined at room temperature) of the MgCO3-containing sample compared with the one containing CaCO3. From the Figure it is obvious that there is little or no dependence of the degradation on pH of purified cellulose pulps and rag papers. The MgCO3-containing bleached pulps show a higher pH and a more pronounced degradation. Quite on the contrary, papers (with all additives and fillers) made of bleached chemical pulps exhibit different behaviour.
112 Ageing and stabilisation of paper Chapter 7: Thermo-oxidative degradation
20
10 rag papers filter papers
0 )
% papers made of (
a bleached pulps C -10 P D / ) a C -20 P D - g bleached M
P -30 pulps D (
-40
-50 -10 -5 0 5 10 15 (pH -pH )/pH (%) Mg Ca Ca Figure 7.15: The relative difference in decrease of DP during accelerated ageing (80 °C, 65% RH) of a variety of CaCO3 and MgCO3-containing pulps and papers vs. the difference of pH of the same samples. The index denotes the type of deacidification.24 It is obvious that the effect that the type of deacidification has on pH is more pronounced in pure pulps than in factual paper samples. Considering the data for DP, deacidification with Ca(HCO3)2 has a similar effect to deacidification with Mg(HCO3)2 for all samples but bleached chemical pulps.24 However, this was a study performed at 80 °C. As we discussed in Chapter 3, the ageing behaviour at one elevated temperature provides no information about the behaviour of the sample at room temperature. Therefore, we estimated the behaviour of samples from Arrhenius studies (Figure 7.16).
-23.5 CaCO 3 MgCO -24.0 3
-24.5
-25.0 k -25.5 n l
-26.0
-26.5
-27.0
-27.5 0.00275 0.00280 0.00285 0.00290 0.00295 0.00300 0.00305 1/T (1/K)
Figure 7.16: An Arrhenius study with Whatman filter paper containing MgCO3 or CaCO3 and aged at different temperatures, 65% RH. The data in Figure 7.16 show that activation energies for the two samples ageing in a moist -1 -1 environment are fairly similar: between 99 ± 5 kJ mol for CaCO3 and 102 ± 5 kJ mol for MgCO3, there is no statistical difference. The values of extrapolated rate constants for degradation of the three model samples at room temperature are still different. However, considering the experimental errors presented in Figure 7.17, we can conclude that there is no significant difference in ageing in atmosphere of 65% humidity. The only difference is
Ageing and stabilisation of paper 113 Chapter 7: Thermo-oxidative degradation indicated for Whatman paper: using a calculation according to (eq 3.16), the sample containing MgCO3 is likely to last approximately 9,800 years, while the one containing CaCO3 “only” approximately 4,400 years. On the basis of experiments with models, it is safe to conclude that if the present guidelines for long-term storage are adhered to,60 the type of aqueous deacidification is less important, concerning the changes in degree of polymerisation and the resulting mechanical stability. However, this conclusion should not be simplified to other deacidification types (e.g. mass deacidification) without reservations, as remains of degradation products (e.g. of acid-catalysed hydrolysis), an uneven distribution of the alkali or even partial deacidification may all further change the picture. In order to achieve stabilisation of mechanical properties, a study of changes of DP is relevant. However, for visual appreciation of a paper-based artefact, its colour is extremely important. As yellowing most often accompanies ageing, the CIE L*a*b* coordinate describing yellowness, i.e. b*, was used to evaluate the changes in colour during ageing. Again, colour changes at 80 °C will tell little about the changes at 20 °C – we would like to know how colour will change with time at the conditions of use. Considering that for the deacidified model samples the change of b* in time, i.e. b*/t, was constant, we can treat the so obtained “rate constants” of yellowing using the Arrhenius equation. All ageing experiments were performed at 65% RH. In Figure 7.18, we can clearly see that all the samples, containing MgCO3 will yellow at a faster rate than the samples with CaCO3.
1E-13 ) 1 - s
1 1E-14 - l o m
l o m (
3 9 2 k
1E-15 a g a g a g C C C M M M
A H C A H C S S W W Figure 7.17: Values of degradation rate constants extrapolated to 20 °C with the bars representing the calculated uncertainty interval ±SD. Denotation: WH – Whatman filter paper, SA – bleached sulphate pulp, C – cotton pulp, Mg – containing MgCO3, Ca – containing CaCO3. Undoubtedly, the decision for the most appropriate deacidification method has to involve both criteria: rate of degradation (change of degree of polymerisation) and colour change. If we compare the data in Figure 7.17 and Figure 7.18 for Whatman paper, we see that at room temperature it is likely to degrade more slowly, but it will yellow faster. For archival purposes, yellowing might be less important than mechanical stability, for graphic arts, the reverse might be true.
114 Ageing and stabilisation of paper Chapter 7: Thermo-oxidative degradation
1E-4 ) 1 - y a d
*
b 1E-5 (
3 9 2 k
1E-6 a g a g a g C C C M M M
A H C A H C S S W W Figure 7.18: Values of “rate constants” of yellowing extrapolated to 20 °C with the bars representing the calculated uncertainty interval ±SD. Denotation: WH – Whatman filter paper, SA – bleached sulphate pulp, C – cotton pulp, Mg – containing MgCO3, Ca – containing CaCO3.
7.7 Antioxidants As we outlined in Chapter 2, antioxidants are compounds that retard the process of oxidation. Considering the importance of the degradation process in all areas of human activity, a vast array of antioxidants is available for different purposes, ranging from biologically active vitamins to polymer processing stabilizers. Antioxidants for cellulose have to be hydrophilic in order to take part in reactions taking place in a water-containing cellulosic fibre and in order to exhibit some mobility. While reduction of carbonyls, as discussed on page 106, will reduce the rate of initiation, there are a number of strategies available once the autoxidation cycle begins, as indicated on Figure 7.19. Substances that react with free radicals more readily than the material itself are free radical scavengers. By reacting with scavengers, free radicals are eventually converted into non-radical products. A particular group of antioxidants are peroxide decomposers, which lead to reduction of hydroperoxides in a way that hydroxyl radicals are not produced. Metal deactivators are the third often used type – complexation of a transition metal ion may disable it for participation in the Fenton reaction. The oldest, and for cellulose quite possibly the most effective type of antioxidants are peroxide decomposers. Iodide is well known to inhibit the oxidation of cellulose,5,6,61 and its catalytic mode of action in conditions of low acidity is thought to be
– – 2I + H2O2 → I2 + 2HO ,
– – 2HO + I2 + H2O2 → 2I + H2O + O2, basing on the early work of Bray and Liebhafsky.62 In alkaline conditions the cyclic reaction
– – I + H2O2 → IO + H2O,
– – IO + H2O2 → I + H2O + O2,
Ageing and stabilisation of paper 115 Chapter 7: Thermo-oxidative degradation is another possibility. Similar reactions were considered for bromide63 and including the pseudo-halide thiocyanate, a study on the use of halides and pseudo-halides as antioxidants for paper-based cultural heritage was published recently,64 followed by a study on the use of tetrabutylammonium bromide for stabilisation of iron gall ink corrosion.65
Figure 7.19: The Bolland-Gee autoxidation reaction system as in Figure 7.2, with the possible points of interference of various types of antioxidants. The stabilisation of transition-metal-mediated degradation of cellulosic materials is a much- studied subject. As it is especially relevant for iron gall ink containing documents, to which another volume is dedicated,36 we shall touch the subject only briefly. Chelation of transition metals may result in inactivation only if two criteria are fulfilled: ― Stability constant for the higher oxidation state must be higher than for the lower oxidation state of the transition metal in question; ― There must be no free coordination sites on the transition metal. One of the most efficient natural antioxidants for iron-mediated oxidation of cellulose is phytate.66 Its use as an antioxidant in food systems is also widespread.67,68 Other chelating compounds have been tested for a possible antioxidant effect, e.g. EDTA, DTPA6 and in another study EDTA, DTPA, desferal, citrate and phytate.33 By far the most promising antioxidant is probably desferal, yet it is of little practical value as chelation of iron leads to orange colouration. DTPA is another promising chelating substance. The drawback of the chelating type of antioxidants is that they are metal- specific. It is well-known for EDTA, that it has antioxidant properties in the case of copper, while in the case of iron it has strong pro-oxidant properties.69 As we mentioned before (page 108), addition of compounds scavenging free radicals is not expected to be particularly successful. E.g., BHT was shown to have a very limited effect.6 On the other hand, a fairly common radical scavenger in paper is lignin, a well-known antioxidant.70 In an alkaline environment, it is rapidly oxidised,71 and effectively scavenges hydroxyl radicals and superoxide. Although it has been used to stabilise rubber72 and
116 Ageing and stabilisation of paper Chapter 7: Thermo-oxidative degradation polypropylene,73 the question remains, whether its presence in neutral or mildly alkaline paper is beneficial. While it has been demonstrated that lignin reacts with free radicals at a faster rate than carbohydrates,74 it may still accelerate their degradation in suspension.75 On the other hand, the beneficial effect of lignin on thermo-oxidative degradation of paper has been demonstrated.76,77 However, its fast degradation and formation of acids, may lead to a faster consumption of alkaline reserve after deacidification. It has been shown that oxygen consumption of deacidified lignin-containing paper is higher than that of non-deacidified one.78 Even if it does have a limited stabilising effect, formation of coloured products during oxidation of lignin (yellowing) is still an unwanted phenomenon. Some antioxidants may have compound effects, e.g. they may exhibit the role of a chelator and a radical scavenger. One such compound is gallic acid, which promotes the formation of hydroxyl radical during the Fenton reaction (page 108), yet due to its excellent properties as a radical scavenger, its presence in paper (as e.g. in iron gall ink) may be beneficiary if available in excess.79
7.8 Conclusions There is ample evidence suggesting the importance of oxidative reactions during degradation of neutral to mildly alkaline paper. Using a variety of new analytical methods, a number of reactive intermediates were identified and even quantified, e.g. hydroperoxides and hydroxyl radicals. The initiating role of aldehydes was demonstrated and it was shown that a higher content of carbonyls leads to faster degradation of paper. The knowledge of reaction pathways is important in order to devise new stabilisation strategies. Removal of products of acid degradation (aldehydes and carbonyls among them) during deacidification is thus extremely important, in order not to increase the rate of oxidation after the paper is mildly alkaline. To do so, the deacidification medium should be able to dissolve and wash out the low-molecular weight hydrophilic compounds. The important role of water was outlined: ― It is a reaction medium; ― It is a plasticizer; ― It facilitates dissociation; ― It is a reactant. Its effect on oxidative degradation is thus complex and it was shown that the degradation rate at 80 °C is lowest at 20% and 95% RH. Adjustment of the storage conditions to optimal RH might easily lead to a substantial stabilisation effect.
CaCO3-based and MgCO3-based deacidification will lead to comparable stability of cellulose and bleached pulp models in conditions of storage in the usual humidity range, with MgCO3 giving slightly better results. The alkaline sensitivity of some bleached chemical pulps is due to their high carbonyl content – such models should not be used for general studies of paper oxidative degradation. Further stabilisation may be achieved with stabilizers – their effect will be discussed in detail in Chapter 11.
Ageing and stabilisation of paper 117 Chapter 7: Thermo-oxidative degradation
7.9 References 1. D. Kočar, M. Strlič, J. Kolar, J. Rychlý, B. Pihlar, L. Rychlá, Chemiluminescence From Paper III. The Effect of Superoxide Anion and Water, Polym. Degrad. Stab., 2005, 88, 407-414. 2. L.F. McBurney, Degradation of cellulose, 1. Kinetics of Degradation Reactions, in: E. Ott, H.M. Spurlin, M.W. Grafflin (Eds.), Cellulose and Cellulose Derivatives, I, Interscience, New York, 1954, 99-130. 3. X. Zou, T. Uesaka, N. Gurnagul, Prediction of paper permanence by accelerated aging. Part I: Kinetic analysis of the aging process, Cellulose, 1996, 3, 243-267. 4. T.P. Nevell, S.H. Zeronian, Cellulose Chemistry and Its Applications, Ellis Horwood, Chichester, 1985. 5. J. Kolar, Mechanism of Autoxidative Degradation of Cellulose, Restaurator, 1997, 18, 163-176. 6. J. Kolar, M. Strlič, G. Novak, B. Pihlar, Aging and Stabilization of Alkaline Paper, J. Pulp Pap. Sci., 1998, 24, 89-94. 7. C.J. Knill, J.F. Kennedy, Degradation of cellulose under alkaline conditions, Carbohydr. Polym., 2003, 51, 281-300. 8. L. Audouin, V. Langlois, J. Verdu, J.C.M. de Bruijn, Role of Oxygen Diffusion in Polymer Ageing: Kinetic and Mechanical Aspects, J. Mat. Sci., 1994, 29, 569-583. 9. L. Reich, S.S. Stivala, Autoxidation of Hydrocarbons and Polyolefins; Kinetics and Mechanisms, M. Dekker, New York, 1969. 10. M. Celina, G.A. George, A Heterogeneous Model for the Thermal Oxidation of Solid Polypropylene from Chemiluminescence Analysis, Polym. Degrad. Stab., 1993, 40, 323-335. 11. M. Celina, G.A. George, D.J. Lacey, N.C. Billiingham, Chemiluminescence Imaging of the Oxidation of Polypropylene, Polym. Degrad. Stab., 1995, 47, 311-317. 12. L. Zlatkevich, Chemiluminescence and Oxidation of Polypropylene: Comments on the Heterogeneous Model, Polym. Degrad. Stab., 1995, 50, 83-87. 13. J.C.Jr. Arthur, O. Hinojosa, Thermal Initiation of Free Radicals in Cotton Cellulose, Text. Res. J., 1966, 36, 385-387. 14. N.-S. Hon, Formation of Free Radicals in Photoirradiated Cellulose. I Effect of Wavelength, Polym. Chem. Ed., 1975, 13, 1347-1361. 15. J.S. Arney, C.L. Novak, Accelerated aging of paper. The influence of acidity on the relative contribution of oxygen-independen and oxygen-dependent processes, TAPPI, 1982, 65, 113-115. 16. D.T. Sawyer, Oxygen Chemistry, Oxford Press, Oxford, 1991. 17. M.N. Schuchmann, C. von Sonntag, Radiation chemistry of carbohydrates. Part 18. Free radical induced oxidation of neutral aqueous solutions of D-glucose in the presence of oxygen, Z. Naturforsch. B., 1978, 33, 329-331. 18. B.H.J. Bielski, D.E. Cabelli, R.L. Arudi, Reactivity of HO2/O2– radicals in aqueous solution, J. Phys. Chem. Ref. Data, 1985, 14, 1041-1100. 19. P.J. Thornalley, A. Stern, The production of free radicals during the autoxidation of monosaccharides by buffer ions, Carbohydr. Res., 1984, 134, 191-204. 20. S.J.H.F. Arts, E.J.M. Mombarg, H. van Bekkum, R.A. Sheldon, Hydrogen Peroxide and Oxygen in Catalytic Oxidation of Carbohydrates and Related Compounds, Synthesis, 1997, 597-613. 21. M. Strlič, D. Kočar, J. Kolar, J. Rychlý, B. Pihlar, Degradation of pullulans of narrow molecular weight distribution - the role of aldehydes in the oxidation of polysaccharides, Carbohydr. Polym., 2003, 54, 221- 228. 22. M. Strlič, J. Kolar, B. Pihlar, J. Rychlý, L. Matisová-Rychlá, Chemiluminescence during thermal and thermo-oxidative degradation of cellulose, Eur. Polym. J., 2000, 36, 2351-2358. 23. D. Kočar, M. Strlič, J. Kolar, B. Pihlar, A new method for determination of hydroperoxides in cellulose, Anal. Bioanal. Chem., 2002, 374, 1218-1222. 24. J. Malešič, J. Kolar, M. Strlič, Effect of pH and carbonyls on the degradation of alkaline paper: factors affecting ageing of alkaline paper, Restaurator, 2002, 23, 145-153. 25. T.N. Kleinert, L.M. Marraccini, Aging and Colour Reversion of Bleached Pulps. Part II: Influence of Air and Moisture, Svensk Papperstidn., 1963, 66, 189-195. 26. L.M. Marraccini, T.N. Kleinert, Spectrophotometric Estimation of Peroxide in Cellulosic Materials, Svensk Papperstidn., 1962, 65, 78-80.
118 Ageing and stabilisation of paper Chapter 7: Thermo-oxidative degradation
27. B.H.J. Bielski, A.O. Allen, Mechanism of the Disproportionation of Superoxide Radicals, J. Phys. Chem., 1977, 81, 1048-1050. 28. W.H. Koppenol, Chemistry of iron and copper in radical reactions, in: C.A. Rice-Evans, R.H. Burdon (Eds.), Free Radical Damage and Its Control, Elsevier, Amsterdam, 1994, 3-24. 29. P. Wardman, L.P. Candeias, Fenton Chemistry: An Introduction, Rad. Res., 1996, 145, 523-531. 30. C. Walling, Fenton’s Reagent Revisited, Acc. Chem. Res., 1975, 8, 125-131. 31. C.J. Shahani, F.H. Hengemihle, The Influence of Copper and Iron on the Permanence of Paper, Historic Textile and Paper Materials, American Chemical Society, 1986, 387-410. 32. J. Kolar, M. Strlič, Stabilisation of ink corrosion, Postprints of the Iron Gall Ink Meeting, The University of Northumbria, Newcastle, UK, 2001, 135-140. 33. M. Strlič, J. Kolar, B. Pihlar, Some preventive cellulose antioxidants studied by an aromatic hydroxylation assay, Polym. Degrad. Stab., 2001, 73, 535-539. 34. M. Strlič, J. Kolar, V.-S. Šelih, D. Kočar, B. Pihlar, A Comparative Study of Several Transition Metals in Fenton-Like Reaction Systems at Circum-Neutral pH, Acta Chim. Slov., 2003, 50, 619-632. 35. R. Blattner, R.J. Ferrier, Effects of Iron, Copper, and Chromate Ions on the Oxidative Degradation of Cellulose Model Compounds, Carbohydr. Res., 1985, 138, 73-82. 36. J. Kolar, M. Strlič (Eds.), Iron gall inks. Manufacture, identification, degradation and stabilisation, Narodna in Univerzitetna Knjižnica, Ljubljana, 2005. 37. I.A. Shkrob, M.C. Depew, J.K.S. Wan, Time-resolved electron spin resonance study of radical species derived from naturally occurring carbohydrates, Chem. Phys. Lett., 1993, 202, 133-140. 38. S. Singh, R. Hider, Aromatic Hydroxylation: A sensitive method for detection of hydroxyl radical production by various iron complexes, in: C.A. Rice-Evans, B. Halliwell (Eds.), Free Radicals, Methodology and Concepts, Richelieu Press, London, UK, 1988, 61-90. 39. J. Kolar, M. Strlič, B. Pihlar, New colourimetric method for determination of hydroxyl radicals during ageing of cellulose, Analytica Chimica Acta, 2001, 431, 313-319. 40. J. Rychlý, L. Matisová-Rychlá, M. Strlič, J. Kolar, Chemiluminescence from paper I. Kinetic analysis of thermal oxidation of cellulose, Polym. Degrad. Stab., 2002, 78, 357-367. 41. E. Vittadini, L.C. Dickinson, P. Chinachoti, 1H and 2H NMR mobility in cellulose, Carbohydr. Polym., 2001, 46, 49-57. 42. V.S. Šelih, M. Strlič, J. Kolar, Catalytic activity of transition metals during oxidative degradation of cellulose, Intern. Conf. Chemical Technology of Wood, Pulp and Paper, Bratislava, 2003, 460-461. 43. A.M. Olsson, L. Salmen, The association of water to cellulose and hemicellulose in paper examined by FTIR spectroscopy, Carbohydr. Res., 2004, 339, 813-818. 44. J. Berthold, J. Desbrieres, M. Rinaudo, L. Salmen, Types of adsorbed water in relation to the ionic groups and their counterions for some cellulose derivatives, Polymer, 1994, 35, 5729-5736. 45. Y.I. Matveev, V.Y. Grinberg, V.B. Tolstoguzov, The plasticizing effect of water on proteins, polysaccharides and their mixtures. Glassy state of biopolymers, food and seeds, Food Hydrocolloids, 2000, 14, 425-437. 46. X. Zou, N. Gurnagul, T. Uesaka, J. Bouchard, Accelerated aging of papers of pure cellulose: mechanism of cellulose degradation and paper embrittlement, Polym. Degrad. Stab., 1994, 43, 393-402. 47. E.L. Graminski, E.J. Parks, E.E. Toth, The effects of temperature and moisture on the accelerated aging of paper, ACS Symposium Series: Durability of Macromolecular Materials, 1979, 95, 341-355. 48. C.J. Shahani, F.H. Hengemihle, N. Weberg, The effect of variations in relative humidity on the accelerated aging of paper, ACS Symp. Ser., 1989, 410, 63-80. 49. A. Barański, D. Dutka, R. Dziembaj, A. Konieczna-Molenda, J.M. Łagan, Effect of relative humidity on the degradation rate of cellulose - Methodology studies, Restaurator, 2004, 25, 68-74. 50. D. Erhardt, D. von Endt, W. Hopwood, The comparison of accelerated aging conditions trough the analysis of extracts of artificially aged paper, Preprints of Paper Presented at the Fifteenth Annual Meeting, The American Institute for Conservation of Historic and Artistic Works, Washington, D. C., 43-55. 51. D. Erhardt, M.F. Mecklenburg, Accelerated vs. natural aging: effect of aging conditions on the aging process of cellulose, Mat. Res. Soc. Symp. Proc., Materials Research Society, 352, 247-269. 52. A. Lienardy, P. van Damme, Practical deacidification, Restaurator, 1990, 11, 1-21.
Ageing and stabilisation of paper 119 Chapter 7: Thermo-oxidative degradation
53. M. Strlič, J. Kolar, B. Pihlar, J. Rychlý, L. Matisová-Rychlá, Initial degradation processes of cellulose at elevated temperatures revisited - chemiluminescence evidence, Polym. Degrad. Stab., 2001, 72, 157-162. 54. S. Fukuzumi, K. Ohkubo, H. Imahori, D.M. Guldi, Driving force dependence of intermolecular electron- transfer reactions of fullerenes, Chem. Eur. J., 2003, 9, 1585-1593. 55. J. Defaye, H. Driguez, Gadelle A., Stabilisation ïhexulosides et de l’oxycellulose en milieu alcalin par interaction avec des cations metalliques, Appl. Polym. Symp., 1976, 28, 955-969. 56. N.A. Smirnova, Phase equilibria modeling in aqueous-organic electrolyte systems with regard to chemical phenomena, J. Chem. Therm., 2003, 35, 747-762. 57. J. Kolar, G. Novak, Effect of Various Deacidification Solutions on the Stability of Cellulose Pulps, Restaurator, 1996, 17, 25-31. 58. H. Bansa, Aqueous Deacidification - with Calcium or with Magnesium?, Restaurator, 1998, 19, 1-40. 59. M. Strlič, J. Kolar, D. Kočar, T. Drnovšek, V.-S. Šelih, R. Susič, B. Pihlar, What Is the PH of Alkaline Paper?, e-PS, 2004, 1, 35-47. 60. ISO 11799 Information and Documentation - Document Storage Requirements for Archive and Library Materials, 2003. 61. J.L. Minor, N. Sanyer, Oxygen/Alkali Oxidation of Cellulose and Model Alcohols and the Inhibition by Iodide, J. Polym. Sci.: Part C, 1971, 73-84. 62. W.C. Bray, H.A. Liebhafsky, Reactions involving hydrogen peroxide, iodine and iodate ion. I. Introduction, J. Am. Chem. Soc., 1931, 53, 38-44. 63. W.C. Bray, R.S. Livingston, The catalytic decomposition of hydrogen peroxide in a bromine-bromide solution, and a study of the steady state, J. Am. Chem. Soc., 1923, 45, 1251-1271. 64. J. Kolar, M. Strlič, Stabilisation of alkaline cellulose with halides and pseudo-halides, Intern. Conf. Chemical Technology of Wood, Pulp and Paper, Bratislava, 422-423. 65. J. Kolar, M. Strlič, M. Budnar, J. Malešič, V.S. Šelih, J. Simčič, Stabilisation of corrosive iron gall inks, Acta Chim. Slov., 2003, 50, 763-770. 66. J.G. Neevel, Phytate: A Potential Conservation Agent for the Treatment of Ink Corrosion Caused by Irongall Inks, Restaurator, 1995, 16, 143-160. 67. E. Graf, Applications of Phytic Acid, J. Am. Oil Chem. Soc., 1983, 60, 1861-1867. 68. E. Graf, K.L. Empson, J.W. Eaton, Phytic Acid, a Natural Antioxidant, J. Biol. Chem., 1987, 262, 11647-11650. 69. T. Kocha, M. Yamaguchi, H. Ohtaki, T. Fukuda, T. Aoyagi, Hydrogen peroxide-mediated degradation of protein: different oxidation modes of copper- and iron-dependent hydroxyl radicals on the degradation of albumin, Biochim. Biophys. Acta, 1997, 1337, 319-326. 70. G.L. Catignani, M.E. Carter, Antioxidant properties of lignin, J. Food Sci., 1982, 47, 1745-1748. 71. G.P.B. Gellerstedt, Autoxidation of lignin, Svensk Papperstidn., 1980, 83, 314-318. 72. M.-A. De Paoli, L.T. Furlan, Sugar cane bagasse lignin as a stabilizer for rubbers: Part II – Butadiene rubber, Polym. Degrad. Stab., 1985, 13, 129-138. 73. C. Pouteau, P. Dole, B. Cathala, L. Averous, N. Boquillon, Antioxidant properties of lignin in polypropylene, Polym. Degrad. Stab., 2003, 81, 9-18. 74. M. Ek, J. Gierer, K. Jansbo, Study on the selectivity of bleaching with oxygen-containing species, Holzforschung, 1989, 43, 391-396. 75. K. Magara, T. Ikeda, Y. Tomimura, S. Hosoya, Accelerated degradation of cellulose in the presence of lignin during ozone bleaching, J. Pulp Pap. Sci., 1998, 24, 264-268. 76. L.R.C. Barclay, F. Xi, J.Q. Norris, Antioxidant properties of phenolic lignin model compounds, J. Wood Chem. Technol., 1997, 17, 73-90. 77. J.A. Schmidt, C.S. Rye, N. Gurnagul, Lignin inhibits autoxidative degradation of cellulose, Polym. Degrad. Stab., 1995, 4, 291-297. 78. J. Dufour, J.B.G.A. Havermans, The accelerated deterioration of deacidified paper by autoxidation, Proc. ARSAG Conf. Posters, 1997, 314-318. 79. M. Strlič, T. Radovič, J. Kolar, B. Pihlar, Anti- and pro-oxidative properties of gallic acid in Fenton-like systems, J. Agric. Food Chem., 2002, 50, 6313-6317.
120 Ageing and stabilisation of paper Chapter 8 Chemiluminescence of cellulose and paper
Matija Strlič, Jozef Rychlý, Olivier Haillant, Drago Kočar, Jana Kolar
8.1 Introduction Studies of chemiluminescence have been shown to be an extremely relevant method in polymer degradation studies.1 The light is usually, although not exclusively, a consequence of an exothermal reaction, such as oxidation reactions involving atmospheric oxygen. Studies of oxidative degradation have thus been in the focus of polymer degradation research using chemiluminescence ever since Ashby reported that polymers emit weak light during heating in air in 1961.2 As oxidative degradation is the dominant degradation pathway in alkaline cellulose and paper, chemiluminometry may well provide relevant information on kinetics of the process. Contemporary luminometers allow measurements at ambient and even lower temperatures.3 A measurement takes no longer than a few minutes. Since light is measured while the very reactions take place, the method is non-invasive and it can be, in principle, also non-destructive. Considering that the classical Arrhenius accelerated degradation studies are time- and work-intensive, benefits of the chemiluminometric approach are obvious. The potential of chemiluminometry in paper stability studies was recognized by Kelly et al. already in 1979.4 The authors cycled moist and dry air above a paper sample (containing 8% fillers, including 3% TiO2) at 70 °C (Figure 8.1), and obtained a repeatable behaviour: on each switch to moist atmosphere a sharp CL peak appeared and the signal then dropped below the level of CL in dry atmosphere. The hypothesis was put forward that water- absorption-induced swelling leads to bond-scission, and thus to enhanced oxidation. As we shall see later in this Chapter, the effect of water on chemiluminescence phenomena is fairly complex and most probably involves both physical and chemical phenomena. The authors also attempted a kinetic evaluation of the observed phenomena and despite the promising start the technique was not used again in cellulose degradation studies until
Ageing and stabilisation of paper 121 Chapter 8: Chemiluminescence of cellulose and paper late 1990s when its potential was reassessed5 and experimental work started again.6
Figure 8.1: The effect of cycling moist and dry air on CL of paper at 70 °C. Reprinted with kind permission from Kelly et al.4 Copyright (1979) American Chemical Society.
8.2 Sample morphology and size The size of sample for a chemiluminometric experiment depends on the instrument and varies between several millimetres and a few centimetres in diameter. In a recent study,7 it was shown that samples of diameter as small as 0.50 mm give comparable signals to samples of the usual size (ca. 1 cm). This enables chemiluminometric studies also of those samples, which are not available in big quantities or are too precious to be temporarily exposed to temperatures higher than, e.g. 60 °C. While cellulose crystallinity is not expected to influence the signal to a considerable extent, morphology of fibres (which can be changed by, e.g., pulp beating) is an important quantitative parameter, and is probably linked with the reactive specific fibre surface area.7 In Chapter 5 (page 81), an instrument is described, with which sampling can also be performed in a non-destructive manner. This enables assessment of oxidative instability of paper surfaces without the need for sample cutting. This is valuable in cases when oxidation phenomena are initiated at the edges of a sample that has been cut, and especially in cases of localised chemiluminometric investigation of oxidation "hotspots" in historical samples at low temperatures.
8.3 Chemiluminescence in inert atmosphere Chemiluminescence of cellulose in inert atmosphere is relatively intense, especially at elevated temperatures (Figure 8.2), although low when compared with the signal obtained in similar conditions in oxygen atmosphere. This phenomenon was shown to originate in transglycosidation reaction.8 The DP decrease was correlated with integrated CL signal during isothermal experiments at 180, 190 and 200 °C. However, the apparent activation
122 Ageing and stabilisation of paper Chapter 8: Chemiluminescence of cellulose and paper energy for the process, obtained by treatment of the DP data according to Ekenstam (eq 3.7) was 101 kJ mol-1, while the activation energy obtained for the chemiluminescence signal was 54-60 kJ mol-1. It is possible that the subsequent consecutive elimination reactions also lead to chemiluminescence, since some model monosaccharides (even levoglucosan, the primary degradation product of transglycosidation) give a chemiluminescence signal in inert atmosphere at higher temperatures.
4000 ) 1
- 3000
s o ( 200 C y t i s n e t n
i 2000
e
c o
n 180 C e c s e
n o i 1000 170 C m u l i m e h c 0
-2 0 2 4 6 8 10 12 14 16 18 20 time (h) Figure 8.2: Isothermal chemiluminometric experiment with a sulphite bleached pulp sample at the indicated temperatures in nitrogen atmosphere.
1200 B ) 1 -
s 1000 (
y t i s n
e 800 t n i
e
c A
n 600 e c s e n i 400 C m u l i m
e 200 h c
0
40 60 80 100 120 140 160 180 temperature (oC) Figure 8.3: Dynamic chemiluminometric experiments in nitrogen atmosphere with sulphite bleached pulp: (A) pre-oxidised in oxygen, 80 °C, 30 min; (B) irradiated under a 60 W incandescent light source at a 25-cm distance from sample; (C) without pre-treatment. In a later study,9 it was shown that the presence of alkali-earth metal carbonates affects the rate of transglycosidation: the higher the electronegativity of the alkali-earth metal, the higher the rate constant. The hypothesis that complex formation between the metal and glycosidic oxygen is formed, has been put forward. In the same publication, the authors demonstrated a measurable decrease of degree of polymerisation during an initial heating of cellulosic materials, sometimes referred to as weak links. This phenomenon was independent of water, metal or carbonyl content and it was estimated to be stress-induced.
Ageing and stabilisation of paper 123 Chapter 8: Chemiluminescence of cellulose and paper
Of considerable importance are dynamic experiments in inert atmosphere. In synthetic polymers, the obtained CL signal is usually attributed to decomposition of peroxides.10 In cellulose, however, several different processes were already identified (Figure 8.3).8 The signal obtained during heating depends on the particular pre-treatment. Without pre- treatment, an almost monotonously increasing signal is indicative of thermolysis (transglycosidation). If the sample is pre-oxidised, a peak appears with the maximum situated at approximately 120-150 °C. The peak area was shown to correlate with peroxide content in pulp, determined titrimetrically. The area can easily be calculated by deconvolu- tion of the signal supposing two independent phenomena, which is probably the case: decomposition of peroxides and thermolysis. This approach can be used for analytical determination of the relative content of peroxides in samples of similar origin. In Chapter 7 (page 105) we thus showed an example of peroxide determination in oxidised pullulan samples of different molecular weights. The curve C in Figure 8.3 exhibiting a peak with the maximum at approximately 85 °C represents a less well understood phenomenon and the hypothesis was put forward that recombination of charge-transfer complexes is responsible for it, supported by the low apparent activation energy of the process, 20.5 kJ mol-1.8 This line of study will be further discussed below and in Chapter 9. Other similar phenomena giving rise to elevated chemiluminescence at low temperatures in inert atmosphere can also be observed, especially after plasma or laser treatments.11,12
8.4 Chemiluminescence in oxidative atmosphere The heterochain character of the macromolecular backbone and the presence of hydroxyl groups in its structure are the reasons for susceptibility of cellulose towards ionic degradation. Thus, in acidic environment, cellulose degrades relatively rapidly via the acid- catalysed (cationic) mechanism, discussed in Chapter 6. The relatively complex and superimposed free radical and alkaline degradation mechanisms are typical for cellulose in a mildly alkaline environment, e.g. cellulose containing calcium or magnesium carbonate.13 The relevant degradation pathways were discussed in Chapter 7.
Figure 8.4: The repeating unit in a cellulose macromolecule with one 1,4--glucopyranosyl monomer with the usual notation of carbon atoms. The symbols and bonds of hydrogen and symbols for carbon atoms are omitted from the presented structure for the purpose of clarity. As shown, the pattern of oxidation depends on structural defects, such as carbonyl groups,14,15 which may be inherently present in cellulose, or introduced during degradation or production. Other points of oxygen attack are possible on the 1,4--glucopyranosyl monomer unit (Figure 8.4). Shafizadeh proposed that C1 and C4 are most susceptible, although an attack on C2 and C3, which are linked to hydroxyl groups, and C5 linked to
124 Ageing and stabilisation of paper Chapter 8: Chemiluminescence of cellulose and paper hydroxymethyl group cannot be excluded.16 The kinetic length of oxidation chain for cellulose appears to be rather short as the steady-state content of hydroperoxides is much lower than in polyolefins under comparable conditions.
Non-isothermal chemiluminometric experiments As discussed in Chapter 3, there are various ways of estimating material lifetime taking into account either mechanical (e.g. double folds) or chemical properties (e.g. degree of polymerisation). Using the chemiluminometric approach, we will try to demonstrate the relevance of “residual oxidisability” of a sample.
240
50000 220 )
1 200 - g
m 40000
180 1 - s ) (
C y 160 o t ( i
s
30000 e r n 140 u e t t a n i r
120 e L p
C 20000
m d
100 e t e z i l
a 80 10000 m r
o 60 n
0 40
20 0 20000 40000 60000 80000 time (s) Figure 8.5: Chemiluminescence during a temperature cycling experiment from 40 to 220 °C at the rate of 2.5 ºC min-1; sample SA, oxygen atmosphere.
0.0000050
0.0000045
0.0000040 ) 1 -
s 0.0000035 (
t n a t
s 0.0000030 n o c
e 0.0000025 t a r 0.0000020
0.0000015
0.0000010 0 2 4 6 8 10 cycle number Figure 8.6: Changes in the calculated degradation rate constant for sample SA determined for 120 ºC in oxygen atmosphere from the experiment shown in Figure 8.5. In this approach, experiments carried out under non-isothermal conditions of a pre- programmed steady increase of temperature and at a certain concentration of oxygen appear to be quite suitable (cf. Figure 5.8). Also, extrapolation of the data obtained under such conditions to other (ambient) conditions is possible. During the non-isothermal experiments in oxygen with sulphate bleached pulp (Figure 8.5) we can observe a reduction
Ageing and stabilisation of paper 125 Chapter 8: Chemiluminescence of cellulose and paper of the chemiluminescence peak at 220 ºC and also a reduction of the average rate constant of cellulose degradation calculated for 120 ºC (Figure 8.6), using the approach discussed on page 74.
The plots of DP0/DP vs. ageing time for sulphate bleached pulp (SA) show a discontinuation indicating a faster initial process and a slower one during the more advanced phase of degradation (Figure 8.7). Similarly, two degradation phases were observed also at higher temperatures (120-220 °C) in oxygen. A comparison of the rate constants of faster and slower processes determined from DP measurements for SA pulp is shown in Figure 8.8, along with the rate constants derived from chemiluminometric experiments. Notably, the rate constants obtained using determination of DP are well within the region of values of calculated rate constants at a given temperature.
1.3
1.2 P D / 0 P
D 1.1
1.0
0 5 10 15 20 25 time (days)
Figure 8.7: Plot of DP0/DP vs. time for degradation of sample SA at 90 °C in dry air.
-7
-8
-9
-10
-11 1 ) ' -12 k ( n l -13 3 -14 4 -15
-16 2 -17
0.0020 0.0021 0.0022 0.0023 0.0024 0.0025 0.0026 1/T (K-1) Figure 8.8: Values of k’ constants, cf. (eq 3.8), related to the faster and slower process obtained from non-isothermal chemiluminescence measurement (point series 1 and 2) and from DP decay (point series 3 and 4), sample: SA, oxygen atmosphere, temperature interval 120-220 °C. A similar agreement was reported also for Whatman cellulose.17 The rate constants were determined assuming that the chain scission processes are first-order processes. In chemiluminometric experiments, the introduction of a faster and slower process is based
126 Ageing and stabilisation of paper Chapter 8: Chemiluminescence of cellulose and paper on the observed declination of experimental curves of intensity vs. temperature in Arrhenius coordinates (Figure 8.9).
12
10 ln I = 36.36-12741/T air o 8 170 C
oxygen
) 6 I ( n l nitrogen 4
2
0 0.0020 0.0022 0.0024 0.0026 0.0028 0.0030 0.0032 1/T (K-1) Figure 8.9: Arrhenius plots of chemiluminescence intensities during oxidation of Whatman cellulose in different atmospheres. Rate of heating 2.5 °C min-1. A similar increase in the slopes of Arrhenius plots for oxidation of synthetic polymers (if the temperature is increased) is usually explained by an increase in mobility of macromolecular chains occurring at the temperature increases above Tg of the polymer. However, Tg of cellulose is not well defined and no shift to lower temperatures due to water was confirmed by DSC. Other sub-Tg transitions may be of interest, such as the so called -transition. It was observed in studies of temperature dependence of dielectric properties at approximately 170 °C, and appears to be related to migration of protons in cellulose. According to an IR study of thermally treated cellulose a reduction and subsequent increase of crystallinity also take place in this temperature region.18
800000 1 - WH; pH 6.6 700000 2 )
1 2 - WH + MgCO ; pH 9.8 - 3 s (
3 - WH + CaCO ; pH 8.4 3 y
t 600000 i 4 - WH + HCl; pH 3.5 s n e t 500000 n i
e c
n 400000 e 1 c s
e 3 n 300000 i m u l
i 4 200000 m e h c 100000
0
20 40 60 80 100 120 140 160 180 200 220 240 temperature (0C) Figure 8.10: Dynamic chemiluminometric experiments in oxygen with Whatman samples containing different additives as indicated, rate of heating 2.5 °C min-1. It is of interest, how initial pH affects the chemiluminescence intensity in oxygen atmosphere. For this purpose, Whatman paper impregnated with different additives was used (Figure 8.10), thus achieving different initial pH. It is immediately noticeable that a
Ageing and stabilisation of paper 127 Chapter 8: Chemiluminescence of cellulose and paper higher pH of the MgCO3-containing sample leads to a more intense signal. However, the generally accepted relation between chemiluminescence intensity I and rate of oxidation w: wI , (eq 8.1) does not allow us to directly compare the oxidisability of materials from absolute values of chemiluminescence intensities. This is due to the unknown proportionality constant, , which involves the formation of various emitters, presence and/or formation of quenchers, light absorbers, etc. However, some approximation can be made by normalising the maximum chemiluminescence intensity Imax at maximum temperature of the experiment (220 °C) to 1 by assuming:
wI maxmax (eq 8.2) and dividing (eq 8.1) by (eq 8.2) we obtain:
rel // wwIII maxmax . The parameter is thus eliminated and relative values of chemiluminescence intensity may be used for a qualitative comparison of sample stability. The above approximation is based on the simplification that the rates of oxidation of the same material (cellulose) at very high temperatures are comparable regardless of the presence of additives. A more detailed kinetic analysis of experimental curves to be discussed later indicates that such an approximation is quite acceptable. The transformation of experimental data shown in Figure 8.10 according to above equation is shown in Figure 8.11. It seems that the addition of MgCO3 significantly increases the stability of cellulose at low temperatures. Less efficient seems to be CaCO3 while the acidified sample degrades faster than the reference sample. These results are in agreement with the discussion on p. 114, based on accelerated degradation studies of the same set of samples.
1 1 - WH; pH 6.6
2 - WH + MgCO3; pH 9.8
3 - WH + CaCO3; pH 8.4 0.1 4 - WH + HCl; pH 3.5
0.01 x a m I / I 1E-3 4 2
1 1E-4 3
1E-5
20 40 60 80 100 120 140 160 180 200 220 240 temperature (oC) Figure 8.11: Data from Figure 8.10 replotted in relative units of chemiluminescence intensity and semi-logarithmic coordinates. Interpretation of the above observation may be based on a somewhat speculative initial step which involves an electron transfer between a glycosidic oxygen and oxygen molecule (Figure 8.12). Such direct interaction is probable in a neutral or slightly alkaline cellulosic
128 Ageing and stabilisation of paper Chapter 8: Chemiluminescence of cellulose and paper environment. Additionally, magnesium ions are known to promote electron transfer between potential electron donor and electron acceptor. Besides, pronounced chemiluminescence intensity is usually observed during oxidation of samples containing magnesium.9
Figure 8.12: Proposed reaction scheme involving a direct interaction between oxygen molecule and glycosidic oxygen, possibly leading to chemiluminescence. The glycosidic oxygen cation radical and superoxide anion radical in the proposed reaction scheme are attached, however, the slightly acidic hydroxyl group on C2 promotes the formation of dioxetane. Dioxetanes are unstable and decompose into an aldehyde in excited triplet state which may yield chemiluminescence. In parallel, hydroperoxides are formed which may lead to additional chemiluminescence during their decomposition, as well. The elementary step responsible for that is a disproportionation reaction of hydroxyl and alkyloxy radicals giving water and excited carbonyl groups.
Isothermal chemiluminometric experiments Isothermal chemiluminometric experiments in oxygen atmosphere typically exhibit a considerably higher intensity of light emission than in air.8 All curves for the purified cellulose sample (WH) in oxygen and in air exhibit a not very distinct maximum shortly after start of the experiment. Subsequently, light emission decreases slowly (Figure 8.13). It should be noted that samples oxidized in air show the maximum in chemiluminescence emission shifted to longer times than those oxidized in oxygen. This may indicate the importance of oxygen diffusion to the reaction sites at elevated temperatures. On one hand, degradation of cellulose, if followed by determination of DP, is slower in air than in oxygen atmosphere, on the other hand, the shift of chemiluminescence maximum indicates the dominance of the faster process. Also, more oxidised groups may form in oxygen; these may undergo further oxidation thus shifting the maximum of oxidation rate to longer times. An addition of maltose to sulphate bleached pulp (sample SA) increases the observed maximum considerably, while maltitol has no such effect (Figure 8.14). Due to the presence of carbonyl-group-containing low molar mass compounds, the intensity of chemiluminescence increases, which is in accordance with the previously discussed initiating effect of carbonyl groups during oxidation of polysaccharides. This reveals the potential of chemiluminometry in examination of paper ageing.
Ageing and stabilisation of paper 129 Chapter 8: Chemiluminescence of cellulose and paper
50000
O , 220 oC ) 2 1 - 40000 s (
y t i s n e t 30000 n i
e c n e
c o
s 20000 air, 220 C e n i
m o u
l O , 200 C i air, 200 oC 2 m 10000 e h c O , 180 oC 2 air, 180 oC 0
10 100 1000 10000 100000 time (s) Figure 8.13: Isothermal chemiluminometric experiments in oxygen atmosphere or air at different temperatures; sample: WH.
a - maltose 0.01 mol L-1 b - maltose 0.001 mol L-1 -1 )
1 20000 c - maltitol 0.01 mol L - s (
d - non-treated y t i s n
e b t n i
e c n e
c 10000 c s d e n i
m a u l i m e h c
0
0 5000 10000 15000 20000 time (s) Figure 8.14: Isothermal chemiluminescence runs (180 °C) for oxidation of sample SA immersed in solutions of maltose and maltitol as indicated, and dried.
Chemiluminescence in moist oxidative atmosphere Using the chemiluminometric instrument Lumipol 3, which allows the introduction of a controlled content of water in the atmosphere surrounding samples (Figure 5.11), it is possible to perform chemiluminometric experiments with humid oxygen. Since it was already discussed that water has a pronounced effect on oxidation of cellulose (page 109), its effect on chemiluminescence is of high interest in order to provide additional insight into mechanisms of degradation.
130 Ageing and stabilisation of paper Chapter 8: Chemiluminescence of cellulose and paper
120
1600
100
) 1400 1 - s (
r y e t
i 1200 l
80 a s t n i v e t e
n 1000
i h
u e m c 60 n
800 i d e i c t y s
e ( % n 600 i 40 ) m u l i 400 m
e 20 h c 200
0 0 0 1000 2000 3000 4000 5000 6000 7000 time (s) Figure 8.15: A chemiluminometric experiment in oxygen atmosphere of relative humidity as indicated, using four layers of sample C at 90 °C.19 In Figure 8.15, an isothermal experiment is shown, during which several layers of the sample were used in order to intensify the signal. Gaseous water flow was varied in order to obtain the desired levels of relative humidity at the temperature of study. Typical for such experiments is an immediate increase of light emission when water is admitted. This is followed by a relatively fast decay. The phenomenon was first reported by Kelly et al. already in 1979.4 A comparison of the peak intensities after the admission of humidity with intensities prior to water admission shows that the effect is chemical in nature, the presence of MgCO3 generally leading to higher peak intensities (Figure 8.16).
250000 SA Mg
r e t f 200000 a
WH Mg y t ) i 1 - s s n (
e t
n 150000 n o i i SA s e s c i n
m SA Ca e d c 100000 C Mg a s
r e e n t i WH Ca a m w u
l 50000 C Ca i WH m
e C h c 0
0 10000 20000 30000 40000 50000 60000 70000 chemiluminescence intensity before water admission (s-1) Figure 8.16: Correlation of the increase of CL intensity after water admission and the initial level of chemiluminescence emission for samples WH, C and SA, non-deacidified and deacidified using Ca(HCO3)2 and Mg(HCO3)2,the deacidified samples are designated by corresponding symbols; 150 °C, oxygen atmosphere.20 In an earlier study, it was already shown that water increases chemiluminescence intensity during oxidation of cellulose,9 both at T > 100 °C and T < 100 °C. The set of reactions
– – O2 + H2O → HO2 + HO , 2HO2 → O2* + HOOH,
Ageing and stabilisation of paper 131 Chapter 8: Chemiluminescence of cellulose and paper
O2* → O2 + h, was put forward as a possible explanation. The topic was further investigated19 using the superoxide scavenger tiron (4,5-dihydroxy-1,3-benzene disulphonic acid, disodium salt). It was demonstrated that the observed chemiluminescence may in fact originate from singlet oxygen, most probably from dimol emission, which can be detected by the photomultiplier used in the chemiluminometer and which exhibits the emission maximum at 634 nm.21 Superoxide is quite probably a long-lived species in the cellulosic environment as it has 22 been shown not to be able to abstract hydrogen from simple sugars. By tiron (TH2), it is converted to hydrogen peroxide, according to the following reaction:23
– + TH2 + O2 + H TH + H2O2. Hydrogen peroxide may, by reaction with traces of transition metals, lead to production of hydroxyl radicals in a Fenton-like reaction. This should lead to more extensive degradation of the macromolecular chain than in the case of cellulosic samples not containing tiron. This was in fact demonstrated recently.19 It was also shown that tiron reduces chemiluminescence intensity at 80 °C in paper containing approximately 5% moisture (Figure 8.17). This indicates that superoxide is an important precursor of the species giving rise to chemiluminescence during cellulose oxidation. According to the reactions outlined above, this experiment indeed provides further evidence that superoxide is converted to hydrogen peroxide before it is able to enter the proposed series of reactions leading to light emission.
18 A 16 ) 1 - 14 g m
1 -
s 12 (
y t i s
n 10 e t n i
L 8 C
d e
z 6
i B l a m r 4 o n
2
0 2000 4000 6000 8000 10000 time (s) Figure 8.17: Isothermal chemiluminometric experiments at 80 °C in O2 atmosphere. Samples: A – WH containing MgCO3; B – WH containing MgCO3 and tiron.19
8.5 Chemiluminescence following irradiation with light Irradiation of cellulosic and model carbohydrate materials with light has been shown to lead to intense chemiluminescence after the irradiation is discontinued, as shown by curve B in Figure 8.3.24 The phenomenon has already been noticed using carbohydrates other than cellulose,3 yet evidence on its dependence on oxygen present in the atmosphere surrounding the sample during irradiation was presented later (Figure 8.18).24 The species formed is short-lived even at room temperature (half-lifetime 10 days) and
132 Ageing and stabilisation of paper Chapter 8: Chemiluminescence of cellulose and paper decomposes at 120 °C in nitrogen in several minutes. The reactive centres sensitive to light are supposed to be carbonyl or glycosidic oxygen. Oxygen participates in the initiation process of light emission. A bimolecular termination of reacting particles is required at the same time. The hypothesis of charge-transfer complex formed between atmospheric oxygen and carbonyl and/or glycosidic oxygen or other chromophore in the material induced by light and subsequently destroyed by elevated temperature was put forward.8 Both atmospheric humidity and heating in oxygen atmosphere (page 140) lead to a fast decay of the luminescing species.
350000 A
300000 l a r g
e 250000 t n i
e c
n 200000 e c s e n
i 150000 m u l i B
m 100000 e h c 50000 C D 0 0 50 100 150 200 irradiation time (min)
Figure 8.18: Comparison of chemiluminescence integrals (N2, 120 °C) of Whatman paper (WH) after irradiation for the time period indicated in: A: O2, B: air, C: N2 atmosphere, D: blank experiment (air, no irradiation). A 60 W tungsten lamp was used at a distance of 25 cm from the sample at room temperature.
1.6 WH 1.5 SA 1.4 l a r 1.3 g e t
n 1.2 i
k
a 1.1 e p 1.0 e d i
x 0.9 o r
e 0.8 p
e
v 0.7 i t a l 0.6 e r 0.5
0.4
-100 0 100 200 300 400 500 600 700 800 exposure time (h) Figure 8.19: Chemiluminometrically determined content of peroxides in the cellulosic materials as indicated (Whatman paper and sulphate bleached pulp), following irradiation in Xenotest Alpha under the following conditions: filtered light source (Xenochrome 320 filter), energy density 70 W m-2, 60 °C, 30% RH. Peak areas are expressed relative to the peak area of the starting, i.e. non- irradiated material. The postulated charge-transfer species are rather short-lived and are rapidly replaced by peroxides. A steady-state content of peroxides during irradiation experiments is not achieved rapidly, if at all. In Figure 8.19, the content of peroxides determined from peak
Ageing and stabilisation of paper 133 Chapter 8: Chemiluminescence of cellulose and paper areas as in experiment A in Figure 8.3, was expressed in relative terms, i.e. according to the initial content of peroxides. The behaviour of the samples is dramatically different and might be a consequence of remaining traces of lignin in the sulphate bleached pulp, which should exhibit a pronounced sensitivity towards light.
8.6 Conclusions Chemiluminometry has been shown to be an extremely valuable additional tool in exploring degradation processes taking place in cellulosic materials. Several experimental methods exist that allow us to determine the content of reactive oxidation intermediates, such as peroxides or charge-transfer complexes. In order to obtain kinetic data, dynamic experiments have turned out to be most informative, however, the sample has to be heated up to 220 °C in oxygen atmosphere. If examination of precious samples or samples not available in big quantities is needed, it was shown that micro-sampling is a good option. Data obtained from such experiments, are valuable for comparison of stability at room temperature of samples of similar origin. While chemiluminescence in inert atmosphere is a consequence of thermolysis (transglycosidation), in oxidative atmosphere, there are several potential reaction intermediates which could lead to light emission during decomposition: ― Peroxides (including macromolecular peroxides); ― Dioxetanes; ― Singlet oxygen (in the presence of humidity); ― Charge-transfer complexes (formed during exposure to light). Using the typical dynamic experiments in inert atmosphere, it is possible to follow the kinetics of formation of some of these reaction intermediates, which is indispensable in studies of reaction pathways.
8.7 References 1. L. Zlatkevich, Luminescence Techniques in Solid-State Polymer Research, Marcel Dekker, New York, 1989. 2. G.E. Ashby, Oxyluminescence from Polypropylene, J. Polym. Sci., 1961, 50, 99-106. 3. G.D. Mendenhall, H.K. Agarwal, Slow Luminescence Emission from Polymers with a Hyperbolic Decay Law. A Survey of Commercial Materials With an Apparatus of Wide Aperture, J. Appl. Polym. Sci., 1987, 33, 1259-1274. 4. G.B. Kelly, J.C. Williams, G.D. Mendenhall, C.A. Ogle, The Use of Chemiluminiscence in the Study of Paper Permanence, in: G.B. Kelly, R.K. Eby (Eds.), Durability of Macromolecular Meterials, Advances in Chemistry Series, American Chemical Society, Washington, 1979, 95, 117-125. 5. J.L. Pedersoli Jr., The development of micro-analytical methodologies for the characterisation of the condition of paper, Preprint From the 9th International Congress of IADA, 1999, 107-114. 6. M. Strlič, J. Kolar, Stability of alkaline paper: can chemiluminiscence fortell the future?, PapierRestaurierung, 2000, 1 (suppl.), 69-74. 7. D. Kočar, J.L. Pedersoli Jr., M. Strlič, J. Kolar, J. Rychlý, L. Matisová-Rychlá, Chemiluminescence from paper II. The effect of sample crystallinity, morphology and size, Polym. Degrad. Stab., 2004, 86, 269- 274.
134 Ageing and stabilisation of paper Chapter 8: Chemiluminescence of cellulose and paper
8. M. Strlič, J. Kolar, B. Pihlar, J. Rychlý, L. Matisová-Rychlá, Chemiluminescence during thermal and thermo-oxidative degradation of cellulose, Eur. Polym. J., 2000, 36, 2351-2358. 9. M. Strlič, J. Kolar, B. Pihlar, J. Rychlý, L. Matisová-Rychlá, Initial degradation processes of cellulose at elevated temperatures revisited - chemiluminescence evidence, Polym. Degrad. Stab., 2001, 72, 157-162. 10. N.C. Billingham, E.S. O’Keefe, Chemiluminiscence from Oxidative Degradation of Polymers, in A.V. Patsis (Ed.), International Conference on the Advances in the Stabilization and Controlled Degradation of Polymers, Technomic, Lancaster, 1985, I, 1-8. 11. P. Rudolph, F.J. Ligterink, J.L. Pedersoli Jr., M. Van Bommel, J. Bos, H.A. Aziz, J.B.G.A. Havermans, H. Scholten, D. Schipper, W. Kautek, Characterization of laser-treated paper, Appl. Phys. A, 2004, 79, 181-186. 12. M. Strlič, V.S. Šelih, J. Kolar, D. Kočar, B. Pihlar, R. Ostrowski, J. Marczak, M. Strzelec, M. Marinček, T. Vuorinen, L.S. Johansson, Optimisation and on-line acoustic monitoring of laser cleaning of paper, Appl. Phys. A, submitted. 13. J. Kolar, Mechanism of Autoxidative Degradation of Cellulose, Restaurator, 1997, 18, 163-176. 14. J. Malešič, J. Kolar, M. Strlič, Effect of pH and carbonyls on the degradation of alkaline paper: factors affecting ageing of alkaline paper, Restaurator, 2002, 23, 145-153. 15. M. Strlič, D. Kočar, J. Kolar, J. Rychlý, B. Pihlar, Degradation of pullulans of narrow molecular weight distribution - the role of aldehydes in the oxidation of polysaccharides, Carbohydr. Polym., 2003, 54, 221- 228. 16. F. Shafizadeh, A.G.W. Bradbury, Thermal Degradation of Cellulose in Air and Nitrogen at Low Temperatures, J. Appl. Polym. Sci., 1979, 23, 1431-1442. 17. J. Rychlý, L. Matisová-Rychlá, M. Strlič, J. Kolar, Chemiluminescence from paper I. Kinetic analysis of thermal oxidation of cellulose, Polym. Degrad. Stab., 2002, 78, 357-367. 18. M.A. Yousef, A.A. Shabaka, A.M.A. Nada, K.N.A. Elnour, Physical properties and infrared spectroscopy of thermally treated bagasse paper sheets, Ind. J. Pure Appl. Phys., 1991, 29, 6-8. 19. D. Kočar, M. Strlič, J. Kolar, J. Rychlý, B. Pihlar, L. Rychlá, Chemiluminescence From Paper III. The Effect of Superoxide Anion and Water, Polym. Degrad. Stab., 2005, 88, 407-414. 20. J. Rychlý, L. Matisová-Rychlá, M. Lazár, K. Slovák, M. Strlič, D. Kočar, J. Kolar, Thermal oxidation of cellulose investigated by chemiluminescence. The effect of water at temperatures above 100 °C, Carbohydr. Polym., 2004, 58, 301-309. 21. E.A. Ogryzlo, C.W. Tang, Quenching of oxygen (1g) by amines, J. Am. Chem. Soc., 1970, 92, 5034. 22. M.N. Schuchmann, C. von Sonntag, Radiation chemistry of carbohydrates. Part 18. Free radical induced oxidation of neutral aqueous solutions of D-glucose in the presence of oxygen, Z. Naturforsch. B., 1978, 33, 329-331. 23. A.V. Peskin, Y.A. Labas, A.N. Tikhonov, Superoxide radical production by sponges Sycon sp., FEBS Letters, 1998, 434, 201-204. 24. M. Strlič, J. Kolar, B. Pihlar, L. Matisová-Rychlá, J. Rychlý, Chemiluminescence during thermal and thermo-oxidative degradation of cellulose, Eur. Polym. J., 2000, 36, 2351-2358.
Ageing and stabilisation of paper 135
Chapter 9 Photooxidative degradation
Jana Kolar, Matija Strlič, Jasna Malešič, Jacques Lemaire, Dominique Fromageot
9.1 Introduction The subject of this Chapter is the effect of light on cellulose. Natural light covers a wide wavelength interval, from 280 nm (ultraviolet radiation), through the visible light spectrum (the light visible to the human eye in the interval: 400-800 nm), and on to infrared radiation. Light is electromagnetic radiation exhibiting properties of both waves and particles. The particle of electromagnetic radiation is called a photon, whose energy is expressed as E = h, where h is the universal Planck's constant (6.63·10-34 J sec), and is the frequency. Photon wavelength is expressed as = c/, c denoting light velocity.
9.2 Photooxidative degradation of polymers It is a basic photochemical principle, that only light, which is absorbed by a molecule, may lead to chemical reaction. Pure cellulose absorbs visible light only to a small extent, while absorption in the near UV spectral region is more pronounced (Figure 9.1). It is therefore mainly light in the spectral range 300-550 nm, which induces most of the damage in cellulose during exposure to daylight.1 Light-induced degradation may be divided into two main groups of reaction pathways, direct photolysis and photosensitised degradation. Photolysis may occur when the energy of an absorbed photon is high enough to induce direct dissociation of a bond. Indoors, where sunlight is filtered by windows made of glass, the light spectrum contains wavelengths higher than 340 nm. Under such conditions, direct photolysis of cellulose does not take place.2 However, absorption of UV light in the spectral range below = 360 nm may result in homolytic decomposition of hydroperoxides yielding reactive hydroxyl radicals3: H2O2 + h → 2HO .
Ageing and stabilisation of paper 137 Chapter 9: Photooxidative degradation
Usually, the absorption of such a photon by other compounds leads only to excitation of electrons to a higher energy level. This occurs far more frequently than bond dissociation.4 In the process called photosensitised degradation (the chemical species absorbing a photon in this way being called a photosensitiser), the energy from electronically excited states is transferred to initiators, or oxygen, resulting in the formation of reactive species and subsequent photooxidative decay of cellulose.
100
90
80 ) % ( 70 R
60
50
40 200 300 400 500 600 700 800 (nm) Figure 9.1: Diffuse reflectance spectrum of cellulose pulp. The lower the reflectance, the higher the absorption of light in cellulose pulp. It is believed5 that the free radical mechanism of photosensitised oxidation of cellulose in the presence of sunlight proceeds through a chain reaction similar to that proposed by Bolland for thermal degradation of hydrocarbons in liquid phase,6 with the length of the photooxidation kinetic chain about 40 cycles. The application of the mechanism to cellulose was already reviewed7,8 and is discussed in Chapter 7. In the absence of light, it has been proposed that autoxidation in simple sugars is initiated by carbonyl groups, which react with oxygen, forming superoxide. It has recently been demonstrated that the rate of oxidative degradation of polysaccharides during thermal accelerated ageing depends on the amount of carbonyl groups in the starting material.9,10 In the case of light induced degradation, the most often discussed sources of radicals are ketones and hydroperoxides.3
Ketones Ketones may absorb radiation in the near-UV or visible region of the electromagnetic spectrum due to n-* transition, where an electron is excited from a nonbonding (n) into an antibonding (*) orbital. An excited ketone in a singlet state has the tendency to return to a lower energy triplet state. This may occur via photophysical (e.g. fluorescence, phosphorescence, energy transfer), or photochemical reactions. The most important processes include Norrish type I and II reactions (introduced in Chapter 2, page 16),11,12 named after the Nobel laureate R.G.W. Norrish. Radicals formed during homolytic fragmentation induce oxidative decay according to the Bolland mechanism, while the excited carbonyls produced during the heterolytic fragmentation act as photosensitisers. Thus, an excited carbonyl may transfer the energy to
138 Ageing and stabilisation of paper Chapter 9: Photooxidative degradation other molecules, resulting in electronically excited states of e.g. oxygen (singlet oxygen), which may add to double bonds thus leading to formation of peroxides. The importance of carbonyl groups in cellulose on degradation of irradiated cellulose is demonstrated by a study involving a variety of differently deacidified bleached chemical pulps with different contents of carbonyl groups (Figure 9.2).
0.00055
0.00050
0.00045
) 0.00040 1 - y a d
0.00035 1 - l o m
l 0.00030 o m (
k 0.00025
0.00020
0.00015
0.8 0.9 1.0 1.1 1.2 1.3 1.4 carbonyl content (mmol g-1) Figure 9.2: Rate constant of degradation of different bleached chemical pulps, deacidified using calcium bicarbonate, during irradiation in Xenotest with 75 W m-2 ( ≥ 340 nm) and 50% RH. The error bars represent SD (n = 3).
Hydroperoxides Hydroperoxides are considered to be the key compounds in initiation of polymer photooxidation. In addition to homolysis of peroxides mentioned above, transition metal ions in the reduced state, such as Fe2+ and Cu+, catalyse decomposition of peroxides via the Fenton reaction,13 which is believed to be the main source of hydroxyl radicals during thermal accelerated ageing of cellulose:14
2+ 3+ – Fe + H2O2 → Fe + HO + HO . However, in the presence of UV light and ferric ions, oxidation of organic materials is enhanced due to reduction of Fe3+ with water:
3+ 2+ + Fe + H2O + h → Fe + H + HO , yielding hydroxyl radicals and simultaneously producing Fe2+, which may enter into the Fenton reaction. These reactions are known to be the driving force behind photochemical self-cleaning in atmospheric and aquatic environments and are exploited for wastewater treatments.15 Changes in peroxide content during irradiation of cellulose in Xenotest (Xenochrome 320 filter, 70 W m-², 60 °C controlled with a black panel thermometer, at 30% RH) were followed using dynamic chemiluminometric experiments in nitrogen atmosphere (as described on page 123). The original non-irradiated cellulose sample exhibits a chemiluminescence maximum at ca. 130-140 °C (Figure 9.3, curve a), the phenomenon usually interpreted as decomposition of peroxides.16
Ageing and stabilisation of paper 139 Chapter 9: Photooxidative degradation
It is also known that by irradiation, even using an incandescent light source, a chemiluminescent species is formed (exhibiting an elevated signal in Figure 9.3, curve b, at T < 100 °C), which was interpreted as a charge-transfer complex between atmospheric oxygen and some appropriate structure in the polysaccharide, e.g. glycosidic or carbonyl oxygen.16 However, it is also interesting that only after 10 min of irradiation, some of the peroxides present in the original material are already decomposed (CL emission at 130 °C in curve b is considerably lower than in curve a). It is likely that this is due to homolytic decomposition of hydroperoxides. Enhanced radical activity thus induced leads to an elevated peroxide signal upon subsequent storage in darkness (Figure 9.3, curve c).17
75 ) 1 - g m
50 1 - s (
y t i s n e t n i
c
L 25 C
d e z
i a l b a m r
o 0 n
50 100 150 200 temperature (oC) Figure 9.3: Dynamic chemiluminometric experiments in nitrogen atmosphere for a – untreated; b – Xenotest irradiated sample for 10 min; and c – Xenotest-irradiated (10 min) and subsequently stored 580 min at 23 °C in darkness.17
9.3 Free radicals during irradiation of cellulose Free radicals were detected in irradiated cellulose with electron spin resonance spectroscopy (ESR). Phillips et al.2 studied the behaviour of cotton cellulose irradiated by light in the spectral region 325-400 nm, and demonstrated that most radicals were formed at 360 nm. However, Hon et al. detected only a trace amount of radicals in the presence of oxygen and light with wavelength higher than 330-349 nm.18 Further ESR studies suggested that various carbon-centred radicals are formed during irradiation with UV light, to which oxygen adds readily thus forming peroxyl radicals. Hydroperoxides are formed by abstraction of hydrogen by peroxyl radicals,19 which may decompose to hydroxyl radicals as discussed above. The formation of hydroxyl radicals during irradiation of Whatman paper containing calcium carbonate was followed using the N,N'-(5-nitro-1,3-phenylene)bisglutaramide (NPG) hydroxylation assay,17 as described in Chapter 7 (page 109). NPG reacts with hydroxyl radicals forming orange coloured ortho- and para- hydroxylated products, which may be determined using colourimetry. As a reference experiment, NPG was added to a fibrous glass filter sheet, which allows us to observe formation of its hydroxylated products formed in the absence of cellulose. Namely, it is known that aromatic substrates are hydroxylated when irradiated by light, the first step in their photooxidation being the reaction between aromatic hydrocarbons and
140 Ageing and stabilisation of paper Chapter 9: Photooxidative degradation singlet oxygen.20 From the increase of diffuse reflectance at 450 nm it is evident that hydroxyl radicals are formed even in the absence of cellulose, however, the reaction is much more pronounced in its presence, indicating the formation of hydroxyl radicals during irradiation of cellulose (Figure 10.4).
22 glass 20 WH CaCO 3 18 16 14
) 12 % (
10 0 5 4
R 8 6 4 2 0 -2 0 1 2 3 4 5 6 time (h) Figure 9.4: Decrease of diffuse reflectance at 450 nm (absorption maximum of hydroxylated derivates of NPG) plotted against duration of photoageing for calcium carbonate containing purified cotton paper (WH CaCO3) and fibrous glass filter sheets as reference material. The samples were irradiated using filtered xenon arc source ( 340 nm), with energy density 500 W m-2 in Heraeus Suntest CPS+ unit for up to 6 h.17
9.4 Depolymerisation of cellulose due to irradiation The effect of light as well as post-irradiation effects on cellulose properties may be evaluated using size exclusion chromatography (SEC)21 or viscometry. Viscometry may yield an underestimated DP of heavily oxidised cellulose22 although in such cases a reductive pre-treatment can be used. A further advantage of SEC is that it enables determination of distribution of molar masses. However, the technique is time-consuming and the data are less reproducible than viscometric data. Besides, SEC in LiCl/N,N-dimethylacetamide suffers from systematic errors which are difficult to evaluate23. In order to evaluate whether viscometry may be employed to study the degradation of cellulose during irradiation and after subsequent storage in darkness, it was necessary to establish whether a substantial amount of cross-linked material is formed during irradiation. Photooxidative reactions lead to gradual depolymerisation of cellulose, as demonstrated in Figure 9.5, where number-average molar mass distribution of polymer is presented. Degradation is not random, as a second peak with a lower number average distribution appears in photodegraded cellulose. However, no high-molecular weight fraction appears in the irradiated sample (Figure 9.6), meaning that cross-linking is not a dominant process. This justifies the use of viscometry for light-irradiated cellulosic materials, and the Ekenstam equation,24 discussed in Chapter 3 (eq 3.7) may be applied.
Ageing and stabilisation of paper 141 Chapter 9: Photooxidative degradation
Figure 9.5: Number distribution of Whatman cellulose containing calcium carbonate, irradiated in a SEPAP photoageing unit at 60 °C for A – 0 h, B – 333 h, C – 667 h, and D – 1430 h.17
Figure 9.6: Mass distribution of Whatman paper sample containing calcium carbonate, irradiated in a SEPAP photoageing unit at 60 °C for A – 0 h, B – 333 h, C – 667 h and D – 1430 h. 9.5 Natural and accelerated photoageing In order to conduct the natural-light ageing experiment, paper samples were placed behind a south-facing window for an extended period of time (2. Feb. – 10. Aug. 2004) and sampled in appropriate periods of time. Using a thermo-hygro-lux meter Telehum (Euromix, Slovenia), temperature, relative humidity, and light intensity were continuously followed (Figure 9.7). Neither temperature nor relative humidity were regulated during the exposure. Average daily relative humidity increased during the experiment, due to the change of season and ranged from 20% during winter, to above 70% in summer. During one day, conditions may vary considerably, as stronger illumination behind the window around noon time results in higher temperatures (up to 50 °C), leading to a decrease in relative humidity to a few percent only (Figure 9.8). The recently released ASTM D 6789-02 standard test method for accelerated light ageing of printing and writing paper by xenon-arc exposure apparatus requires glass filtration, irradiance of the lamp 765 ± 75 W m-2, determined in the 290 to 800 nm wavelength range, for 48.0 ± 0.5 h, subsequent storage in darkness for 24 h, after which brightness is determined according to ISO 2470 (diffuse reflectance at 457 nm) or yellowness (b*) according to TAPPI T 524. The temperature at paper surface during irradiation should be between 20 and 30 °C and humidity about 0.007 kg per kg of dry air, which equals 49%
142 Ageing and stabilisation of paper Chapter 9: Photooxidative degradation
RH at 20 °C and 27% at 30 °C.
80 30000000 relative humidity (%) illumination · time (lux h) 70 25000000 i l l u m ) 60 20000000 i % n (
a y t t i i 50 o d n i 15000000
· m
t u i m h 40 e e
v 10000000 ( i t l u a l x
e
30 h r 5000000 )
20 0 10 0 50 100 150 200 irradiation time (days) Figure 9.7. Exposure value (lux h) and relative humidity (%) during the experiment with natural light irradiation (2. Feb. – 10. Aug. 2004, south-facing window, Ljubljana, Slovenia).
60 35000
) 55 illumination
% 30000 (
50 temperature y t i relative humidity d
i 45 25000 m
u 40 h i l
l u e m
v 35 20000 i t i n a l
30 a e n r
15000 c d
25 e n
( a l
u ) 20 x
C 10000 ) o (
15 e r u
t 10 5000 a r
e 5 p
m 0
e 0 t 04:54 09:54 14:54 19:54 time of day Figure 9.8: Temperature, relative humidity and illumination vs. time on 05/12/2004, Ljubljana, SI. Two often used photo-ageing ovens, Xenotest Alpha by Atlas Material Testing Solutions and Suntest CPS+ by Heraeus Industrietechnik were used for the evaluation of the standard. Both have xenon arc light sources. While samples are laid horizontally on a tray below the arc in Suntest CPS+, they rotate around it in Xenotest Alpha, which assures a more uniform irradiation. In addition, relative humidity can be adjusted in Xenotest Alpha, but not in Suntest CPS+. In this study and throughout this Chapter, power settings refer to the respective photoageing unit type (cf. page 48). The corresponding energy density throughout the spectral range in question can thus be estimated. The temperature was set as low as possible (black panel temperature 50 ºC) and the relative humidity in Xenotest 50%. The ambient conditions and the ones achieved in the ageing oven are summarised in Table 9.1. Due to the rather high ambient temperature, which was largely the consequence of heating due to the operating photo-ageing chambers, the temperature inside both ageing chambers was slightly higher than 30 °C, which is the maximum allowed temperature according to the ASTM standard. The white standard thermometer was used to continuously record temperature close to the surface of a paper sample and was 32 ± 1 °C. The ambient relative humidity was quite low, 15% in the room with Suntest CPS+ equipment, which resulted in
Ageing and stabilisation of paper 143 Chapter 9: Photooxidative degradation an even lower relative humidity inside the ageing chamber (4%), which is far lower than the one prescribed by the standard.
Table 9.1: Ambient and chamber temperature and relative humidity during artificial photoageing experiments according to the standard ASTM D 6789-02 and in the evaluation study. ambient conditions chamber conditions T (°C) RH (%) T (°C) RH (%) standard - - 20-30 ca. 27-49 Suntest 26 15 33 4 Xenotest 32 28 33 46
The Xenotest Alpha unit met the relative humidity requirement as it allows its adjustment. Otherwise, the standard conditions can thus only be met if the ageing units are situated in an air-conditioned room with the conditioning system strong enough to compensate for the significant heating effect produced by the photo-ageing chambers. For Suntest CPS+ apparatus, the room should also have a possibility of relative humidity control. The SEPAP 12/24 photo-ageing unit, which uses a medium pressure mercury vapour arc light filtered with borosilicate glass was also included in the studies. Illuminance was not controlled, while the temperature of paper surface during the study was 60 °C.
9.6 Natural light and accelerated photoageing In a recent study, Whatman paper was exposed to irradiation in Xenotest Alpha and Suntest CPF+ at 70 W m-2 for 3 days. Relative humidity in Xenotest was set to 50% RH. The white standard thermometer, simulating the temperature of the irradiated Whatman paper during the experiment was 32 ± 1 °C.
89.4 WH 89.2 WH Ca WH Mg
89.0 )
% 88.8 (
s s e n t 88.6 h g i r b 88.4
88.2
88.0 -10 0 10 20 30 40 50 60 70 80 ageing time (hours) Figure 9.9: Decrease of brightness during irradiation of Whatman paper, and Whatman paper containing CaCO3 and MgCO3, in Xenotest (70 W m-2, 32 °C and 50% RH). The error bars represent SD.
144 Ageing and stabilisation of paper Chapter 9: Photooxidative degradation
6000 0.00030 y = (1.02 ± 0.03)·10-13 x - (6 ± 0.6)·10-6 5500 R = 0.9989 0.00025 5000
0.00020
4500 1 / D P
4000 x P -
0.00015 1 D / D
3500 P DP 0 1/DP -1/DP 0.00010 3000 x 0
2500 0.00005
2000 0 5000000 10000000 15000000 20000000 25000000 30000000 integrated illumination (lux h) Figure 9.10: Changes in DP of cotton paper sample (sample C) during exposure to daylight over time, and the same plot transformed according to the Ekenstam equation. According to the ASTM standard, brightness of paper after irradiation should be determined (diffuse reflectance at 457 nm). Irradiation of Whatman paper in Xenotest resulted only in a minor change of brightness. This is not surprising, as the chromophores, which are formed during irradiation, absorb visible light which may then lead to their decomposition (Figure 9.9). However, no visible damage does not imply that there is no degradation. DP of irradiated samples was determined viscometrically and the rate constant of cellulose degradation was determined using the Ekenstam equation (Figure 9.10).
3.5x10-11 WH -11 WH CaCO 3.0x10 3 WH MgCO 3
2.5x10-11 ) 1 - r h
-11 1 - 2.0x10 x u l
1 - l -11 o 1.5x10 m
l o
m -11 (
1.0x10 k
5.0x10-12
0.0 daylight Xenotest 50% RH Xenotest 10% RH Suntest
Figure 9.11: Decrease of DP during irradiation of Whatman paper and Whatman paper containing CaCO3 and MgCO3, by daylight behind a window, in Xenotest (70 W m-2, 32 °C and 50% RH), Xenotest (70 W m-2, 32 °C and 10% RH) and Suntest (700 W m-2, 32 °C and 5% RH). The error bars represent SD. Despite the fact that only a small increase of brightness is observed after irradiation, DP of the exposed paper decreased drastically, as evident from Figure 9.11. Using various ageing methods, similar degradation rates were obtained for untreated and CaCO3-containing Whatman paper. However, a marked decrease in stability of paper containing MgCO3 is observed during irradiation in Xenotest at 50% relative humidity. Namely, the paper deacidified by Mg(HCO3)2 degraded 3.7 ± 0.8 times faster than the one
Ageing and stabilisation of paper 145 Chapter 9: Photooxidative degradation containing CaCO3. Ageing in the SEPAP ageing oven, where the temperature was 60 °C and relative humidity below 5%, proceeded in a similar manner as in Suntest.17 The pronounced differences in stability of deacidified paper, exposed to irradiation in Xenotest at 50% RH, as opposed to irradiation by daylight, Xenotest at 10% RH or Suntest apparatus, are most probably caused by differences in relative humidity levels during the experiments. While RH can be maintained at 40-60% in Xenotest, it may decrease below 10% in Suntest. As evident from the Figure 9.11, conditions during natural-daylight irradiation may be closer to those obtained in Suntest, as stronger illumination behind the window around noon time results in higher temperatures (up to 50 °C), leading to a decrease in relative humidity to a few percent only (Figure 9.8). When artefacts are exposed to light during exhibition, light levels are usually controlled, and so are relative humidity and temperature, and degradation may be best approximated by ageing in Xenotest. The degradation rate constant for Mg(HCO3)2-deacidified Whatman paper in Xenotest suggests that if exposed to 100 lux hr for 12 h each day over a period of 100 days, its DP would decrease by 14%. An equivalent sample, stored in darkness for the same period of time, would not degrade to a detectable extent. However, it is important to keep in mind that the xenon arc, equipped with a glass filter, as prescribed by the ASTM standard, has a spectral energy distribution between 320 and 800 nm. This spectral distribution is broader than the one of indoor light present during exhibitions, where the UV portion of the spectrum (below 400 nm) is usually cut off by filters.
0.000040 thermal ageing only 0.000035 photo and subsequent thermal ageing
0.000030
) 0.000025 1 - y a d
0.000020 1 - l o m
l 0.000015 o m (
k 0.000010
0.000005
0.000000
WH WH Ca WH Mg Figure 9.12: Values of degradation rate constants during thermal ageing at 80 °C and 65% RH with the bars representing the calculated uncertainty interval ±SD. Sample designation: WH (untreated Whatman), WH Ca (Whatman containing CaCO3) and WH Mg (Whatman containing MgCO3). Exhibited objects eventually return into dark repositories. Here, the previously irradiated artefacts will continue to degrade faster than the unexposed ones (Figure 9.12). When Whatman paper, irradiated for 3 days at 70 W m-2, was exposed to subsequent thermal ageing at 80 °C and 65% RH, it degraded at a rate which was 40% faster than the unexposed sample. The effect was even more obvious in the case of deacidified paper, as the one containing CaCO3 was 30 times less stable and the one containing MgCO3 50 times, if compared to non-irradiated samples.
146 Ageing and stabilisation of paper Chapter 9: Photooxidative degradation
0.18
0.16
0.14
0.12
0.10 ) 1 - r 0.08 h
* b (
0.06 k
0.04
0.02
0.00
-0.02 WHp WHCap WHMgp WHt WHCat WHMgt Figure 9.13: "Rate constants" of yellowing during p – photoageing (Xenotest, 70 W m-2, 30 °C, 50% RH); and t – subsequent thermal ageing (80 °C, 65% RH) of: WH – untreated Whatman; WH Mg – deacidified using Mg(HCO3)2; WH Ca – deacidified using Ca(HCO3)2. The error bars represent SD. Although no significant change of brightness was observed, either during the irradiation, or during subsequent 12-day storage in the dark at room temperature, only 3 days of thermal accelerated ageing at 80 °C and 65% RH resulted in a dramatic decrease of brightness (Figure 9.13). The extent of change depended on the extent of depolymerisation induced by irradiation and is in accordance with the results presented in Figure 9.3 and the accompanying discussion.
9.7 Conclusions The importance of oxidative reactions during degradation of organic materials irradiated by light is well established. During light-induced degradation of cellulose, carbonyl groups play an important role, while evolution of hydroperoxides and hydroxyl radicals has also been demonstrated. Photooxidative reactions lead to gradual depolymerisation and it was shown that degradation is not random. However, no high-molecular-weight fraction appears in the irradiated samples, meaning that cross-linking is not a dominant process. This justifies the use of viscometry for characterisation of cellulosic materials irradiated by light. As the low- molecular-weight fraction is relatively non abundant, the Ekenstam equation may be used for determination of the rate constants. Results suggest that the recent ASTM D 6789 standard procedure for photo-ageing of paper may need revision: ― The standard requirements for temperature and relative humidity may often only be reached when ageing chambers are placed in a room with controlled temperature and relative humidity. Opting for lower irradiance, e.g. 500 W m-2, would have enabled easier control of the ageing conditions and at the same time increase the durability of the xenon arc.
Ageing and stabilisation of paper 147 Chapter 9: Photooxidative degradation
― Results suggest that relative humidity may play an important role. A narrower interval of the allowed relative humidity than prescribed by the standard (49% RH at 20 °C or 27% at 30 °C) may be required. ― Measurements of brightness after irradiation alone, as required by the standard, are not suitable, as the changes may be minimal due to simultaneous formation and degradation of chromophores. It is proposed that an additional step of thermal accelerated ageing, which leads to dramatic yellowing of paper, is included in the standard evaluation. To enable evaluation of the stability of paper-based materials during exhibition, a filter cutting off wavelengths below 400 nm should be used. In addition to brightness determination, DP or mechanical properties during ageing should also be followed.
9.8 References 1. J.S. Gratzl, Light-Induced Yellowing of Pulps - Cause and Prevention, Papier, 1985, 39, V14-V23. 2. G.O. Phillips, O. Hinjosa, G.C. Arthur, T. Mares, Photochemical initiation of free radicals in cotton cellulose, Text. Res. J., 1966, 36, 822-827. 3. M. Nowakowska, The Mechanism of Photosensitized Oxidation of Hydrocarbon Polymer Systems, In: A.V. Patsis (Ed.), International Conference on Advances in the Stabilization and Controlled Degradation of Polymers, Technomic, Lancaster, Basel, 1985, I, 123-135. 4. R.L. Feller, Accelerated Aging. Photochemical and Thermal Aspects., The Getty Conservation Insitute, USA, 1994. 5. S.I. Kuzina, I.A. Shilova, A.I. Mikhailov, Photochain free-radical oxidation of cellulose in pulp and paper materials, Pre-Symposium of the 10th ISWPC. Recent Advances in Paper Science and Technology, Korea TAPPI, Seoul, Korea, 1999, 265-268. 6. J.L. Bolland, G. Gee, Kinetic studies in the chemistry of rubber and related materials. II. The kinetics of unconjugated olefins. Trans. Faraday Soc., 1946, 42, 236-243. 7. J. Kolar, Mechanism of Autoxidative Degradation of Cellulose, Restaurator, 1997, 18, 163-176. 8. J. Kolar, M. Strlič, G. Novak, B. Pihlar, Aging and Stabilization of Alkaline Paper, J. Pulp Pap. Sci., 1998, 24, 89-94. 9. J. Malešič, J. Kolar, M. Strlič, Effect of pH and Carbonyls on the Degradation of Alkaline Paper. Factors Affecting Ageing of Alkaline Paper, Restaurator, 2002, 23, 145-153. 10. M. Strlič, D. Kočar, J. Kolar, J. Rychlý, B. Pihlar, Degradation of pullulans of narrow molecular weight distribution - the role of aldehydes in the oxidation of polysaccharides, Carbohydr. Polym., 2003, 54, 221- 228. 11. R. G. W. Norrish, C. H. Bamford, Photodecomposition of aldehydes and ketones, Nature, 1936, 138, 1016. 12. R. G. W. Norrish, C. H. Bamford, Nature, Photodecomposition of aldehydes and ketones, 1937, 140, 195-196. 13. H.J.H. Fenton, On a new reaction of tartaric acid, Chem. News, 1876, 33, 190. 14. J. Kolar, M. Strlič, B. Pihlar, New colourimetric method for determination of hydroxyl radicals during ageing of cellulose, Anal. Chim. Acta., 2001, 431, 313-319. 15. R. Bauer, H. Fallmann, The Photo-Fenton Oxidation - A Cheap and Efficient Wastewater Treatment, Res. Chem. Intermed, 1997, 23, 341-354. 16. M. Strlič, J. Kolar, B. Pihlar, J. Rychlý, L. Matisová-Rychlá, Initial degradation processes of cellulose at elevated temperatures revisited - chemiluminescence evidence, Polym. Degrad. Stab., 2001, 72, 157-162. 17. J. Malešič, J. Kolar, M. Strlič, D. Kočar, D. Fromageot, J. Lemaire, O. Haillant, Photo-Induced Degradation of Cellulose, Polym. Degrad. Stab., 2005, in print. 18. D.N.-S. Hon, Formation of Free Radicals in Photoirradiated Cellulose and related Compounds, J. Polym. Sci., Polym. Chem. Ed., 1976, 14, 2513-2525.
148 Ageing and stabilisation of paper Chapter 9: Photooxidative degradation
19. D.N.-S. Hon, Photooxidative degradation of Cellulose: Reactions of the Cellulose Free radicals with Oxygen, J. Poly. Sci., Polym. Chem. Ed., 1979, 17, 441-454. 20. T. Christ, F. Kulzer, P. Bordat, T. Basché, Watching the Photo-Oxidation of a Single Aromatic Hydrocarbon Molecule, Angew. Chem., 2001, 40, 4192-4195. 21. M. Strlič, J. Kolar, Size exclusion chromatography of cellulose in LiCl/N,N-dimethylacetamide, J. Biochem. Biophys. Meth., 2003, 56, 265-279. 22. M. Strlič, J. Kolar, M. Žigon, B. Pihlar, Evaluation of size-exclusion chromatography and viscometry for the determination of molecular masses of oxidised cellulose, J. Chromatogr. A, 1998, 805, 93-99. 23. M.Strlič, J.Kolenc, J.Kolar, B.Pihlar, Enthalpic interactions in size exclusion chromatography of pullulan and cellulose in LiCl/DMAc, J. Chromatogr. A, 2002, 964, 47-54. 24. A.M. Emsley, G.C. Stevens, Kinetics and Mechanisms of the low temperature degradation of cellulose, Cellulose, 1994, 1, 26-56.
Ageing and stabilisation of paper 149
Stabilisation
Chapter 10 Air pollution and its prevention
John B. G. A. Havermans, Ted A.G. Steemers
10.1 Introduction The role of air pollution in paper deterioration has been demonstrated by many researchers. Most of the work was related to artificial pollution of materials where Kimberly and Hudson did the pioneering work by demonstrating the effects of sulphur dioxide.1-4 Later, many other authors demonstrated the consequences of a combination of acidic and oxidative pollutants on the accelerated ageing of paper (e.g. sulphur dioxide, nitrogen dioxide, ozone).5-9 In general, these pollutants cause the so called acidification of paper, resulting in an increased brittleness of the stored material. This impairs its accessibility, as brittle materials may no longer be handled safely. Therefore, prevention measures have to be applied.10,11
It is well known that paper is able to absorb gaseous contaminants: sulphur dioxide (SO2), nitrogen oxides (NOx) and ozone (O3). Due to the presence of residual moisture in paper, these compounds lead to formation of acids, or react with paper components directly. This was demonstrated in several research projects.12-14 Reactions that may occur are various and 13 are summarised below. The first step is the uptake of SO2 by paper according to
SO2 (g) SO2 (ads) SO2 (aq). Subsequently, the following steps play an important role in formation of hydronium ions which lead to acid-catalysed hydrolysis of cellulose:
– + SO2 + 2H2O → HSO3 + H3O , pKa1 = 1.81,
– 2– + HSO3 + H2O → SO3 + H3O , pKa2 = 6.91,
– – HSO3 + O3 → HSO4 + O2,
– – 2– + 2NO2 + HSO3 + 4H2O → 2NO2 + SO4 + 5H3O . A common reaction of sulphur dioxide is formation of sulphuric(IV) acid, which subsequently dissociates in hydronium cations and hydrogen sulphate(IV) anions and subsequently to sulphate(IV) anion. These ions may be oxidised to sulphate(VI) ions in the presence of e.g. ozone, while in the presence of nitrogen dioxide hydronium ions will form
Ageing and stabilisation of paper 153 Chapter 10: Air pollution and its prevention simultaneously. The combination of acidic contaminants and moisture, which is normally present in cellulose fibres either as bound or free water, leads to acid catalysed hydrolysis of cellulose, resulting in chain scission and weakening of cellulosic fibres.15,16 This mechanism was already discussed in Chapter 6. In order to protect stored paper against the “attack” of acids, prevention measures are needed. One of them is air purification. While it cannot stop the process of acid catalysed hydrolysis, when endogenous acidic contaminants are present in the paper, it will protect paper against exogenous acids. This Chapter describes the effects of air purification on natural ageing of paper in the National Archives of The Netherlands. The evaluation started in 1994, when the archives put into use a new air conditioning and purification system. This work was made possible due to input of a European research project on paper deterioration, due to the establishment of Dutch guidelines on air purification in archives12,17 and due to funds from the Dutch Deltaplan for conservation.
10.2 Materials used in the study Among new model materials, three machine paper grades were selected: bleached sulphite softwood cellulose (P1), cotton linters cellulose (P2) and groundwood containing paper (P3). These reference materials were selected for evaluation of the role of air pollutants and the effects of deacidification on paper stability in the frame of a research project on the effects of air pollutants on ageing of cellulose based materials.12 P1 is a chlorine bleached sulphite softwood cellulose paper, with no additives, fillers or sizing. Fibre analysis showed 100% softwood cellulose. P2 is a chlorine bleached cotton linters cellulose paper, with no additives, fillers or sizing. Fibre analysis showed that it consists of 95% cotton cellulose and 5% softwood cellulose. P3 is a groundwood containing writing paper (acid mechanical pulp paper) that was alum rosin sized and coated with kaolin. It consists of 75% groundwood and 25% softwood cellulose. The model materials were prepared in the way schematically shown in Figure 10.1. The original paper materials were three sets of archival records that were made available for destructive testing. Preliminary research on these materials showed that they were suitable for the long-term study as the materials were available in a large quantity and the homogeneity of paper per volume was acceptable. The selected materials were: Nederlandsche Staatscourant (1872-1943), Handelingen der Staten-Generaal (1953-1964), and Handelingen der Staten-Generaal (annex, 1951-1971). The preparation of old materials differed from that of the new materials. Eighteen volumes (dating from 1872 to 1971) were made available. To obtain a distribution over the two archival storage rooms as homogeneous as possible, the heterogeneous quality of the paper grades and orientation during past storage had to be taken into account. The latter is very important since the top part of a volume has been in a more direct contact with the surrounding atmosphere than the bottom part. The distribution between both depots was done using an experimental design developed specially for the purpose.18 The scheme of sample cutting is given in Figure 10.2. All quarters of a volume were rebound after the cutting procedure.
154 Ageing and stabilisation of paper Chapter 10: Air pollution and its prevention
Figure 10.1: The cutting and distribution scheme for the new papers.
Figure 10.2: The cutting and distribution scheme for old papers. After cutting the four parts from one bound volume are rebound and placed in storage rooms. A variety of standard analyses were performed. Paper acidity was determined by measuring the pH of cold water extract.19 The total water-soluble ionic matter was determined by measuring conductivity of the water extract. Copper number, being a measure of the content of reducing groups in cellulose, was determined.20 In native cellulose, the only reducing group is the terminal one, and copper number is correspondingly low, while oxidation and hydrolysis result in an increased content of reducing carbonyl groups. Tensile strength is the maximum tensile stress developed in a test specimen before rupture under
Ageing and stabilisation of paper 155 Chapter 10: Air pollution and its prevention prescribed testing conditions.21 Folding endurance was also determined.22 In 1997, a slight difference in colour was found between the edges and the centre of pages of sample P3. Thus, colour measurements were also performed using Minolta Chroma Meter D-200. Evaluation of results and calculation of yellowness was done according to DIN 6167.23 From both, the bound and unbound new materials, one sheet of paper per volume was selected. This sheet was taken from the middle of a stack and stored in an acid free envelope for analyses. For the old bound materials, a sampling list was prepared in 1995. This list was based on a statistical approach – samples had to be taken randomly but each succeeding sample was not allowed to be the direct neighbour of the previously taken sample. The sheets were also stored in a neutral envelope until analyses were performed.
Figure 10.3: Material storage: left – the bound new materials; right – old materials and a small part of the wrapped loose materials.
10.3 Archival storage conditions and air purification The National Archives main building is located in the city centre of The Hague, The Netherlands. Two archival storage rooms are in use, of which one has air purification installed and the other is without it. The latter one is a separate closed storage room, located in the ground floor (volume about 143 m3). To this area, the air comes directly from outdoors through a by-pass system. There is air ventilation and circulation installed.
Table 10.1: Archival storage conditions according to Vosteen.17
Temperature 18 ± 2 °C Relative humidity 50 ± 5% Ventilation exchange rate 0.2 h-1 Circulation exchange rate 2.0 h-1 Minimum air velocity 0.01 m s-1
156 Ageing and stabilisation of paper Chapter 10: Air pollution and its prevention
The test storage room with air purification system has a separated area of about 165 m3 and is located on the second floor (total volume about 2500 m3). Based on previous research and the recommended guideline for air quality in archives, archival storage conditions were derived as shown in Table 10.1. The air purification used in the test storage room is based on a combination of purification steps. Air from both the outdoor source as well as from circulation (indoor air) is filtered through a system consisting of a dust filter (EU4), a Futura electro potential micro filter, a Purafil Chemisorbant filter, a Purafil Puracarb filter and again a dust filter (EU9).
10.4 Study of pollutants In both storage rooms, contents of different atmospheric pollutants were determined continuously: sulphur dioxide (SO2), nitrogen oxide (NO), nitrogen dioxide (NO2), total nitrogen oxides (NOx), and ozone (O3). For this purpose three analysers (Table 10.2) were placed in both storage rooms (API – Advanced Pollution Instrumentation Inc., San Diego). By means of a PC equipped with an A/D converter and Labtech Notebook software, data from the analysers were collected frequently, recalculated to the actual pollutant content and stored.24
Table 10.2: The instruments used and their principles of operation.
Instrument Analyte
API 100 Fluorescence SO2 Analyser SO2
API 200 Chemiluminescence NOx Analyser NO, NO2 (NOx)
API 400 Absorption O3 Analyser O3
On a monthly basis, the data were collected and evaluated using a Visual-basic program, developed specially for the purpose. The program allowed us to create weekly, monthly and annual reviews of pollutant contents per storage room with a peak content indicator. The data on outdoor air pollution was acquired from Landelijk Meetnet Luchtkwaliteit.25 10.5 Evaluation of pollution Before discussing the pollution in storage rooms, it is necessary to review the data relating to building physics, air and its ventilation and velocity. Two interesting conclusions can be made: based on technical data, both storage rooms are comparable. Only minor changes were found in the average air exchange rate, circulation rate and velocity (Table 10.3).
Table 10.3: Overview of data on air velocity, circulation and exchange rate (10-year averages).
Storage room Air exchange rate Circulation air exchange rate Air velocity [h-1] [h-1] [m s-1] With purification 0.27 1.60 0.12 Without purification 0.37 1.41 0.08
Ageing and stabilisation of paper 157 Chapter 10: Air pollution and its prevention
50 50 NO D1-NO x x SO D1-SO 2 2 40 O 40 3 D1-O3
) 30 ) 30 b b p p p p ( ( t t n n e e t 20 t 20 n n o o c c
10 10
0 0 1996 1998 2000 2002 2004 1996 1998 2000 2002 2004 year year
50
D2-NOx
D2-SO2 40 D2-O 3
) 30 b p p ( t n e t 20 n o c
10
0 1996 1998 2000 2002 2004 year Figure 10.4: Presentation of the average annual content of SO2, NOx and O3 in ppb. The upper left graph shows the outdoor contents, the upper right shows the contents in the storage room without air purification and the lower graph shows the contents in the storage room with air purification installed. Another relevant topic is related to construction of the storage room without air purification. The exhaust is much larger than the air supply. Air enters the non-purified archive through the door gab and an opening in the wall between the archive and a shaft. Based on our measurements, we conclude that ca. 2/3 of this air is transferred from the location where the common air inlet for all filter systems is located, and from other purified depots. It is estimated that only 1/3 of the air comes directly from the outdoors. Before evaluation, data processing and reduction was done, consisting of filtering and creation of weekly and monthly summaries. Subsequently, the data was evaluated regarding disturbances such as instrumental breakdown or baseline errors. In Figure 10.4, yearly average contents of the pollutants in ppb are shown.
Obviously, outdoor atmosphere is more polluted than indoor air regarding pollutants SO2, NOx and O3. For both NOx and SO2, the outdoor content decreased with time, which also resulted in the same pattern established for the storage room without air purification. Since 1999 the outdoor content of SO2 decreased to 2 ppb and is still decreasing. The non- purified storage room showed a yearly average of about 1 ppb while for the purified storage room the content of SO2 was often below the detection limit. For O3 only the first two years showed a significant difference between the two storage rooms. The content of O3 in the storage room with air purification was 50% lower than in the storage room without air purification (in 1997, 9 ppb and 4.4 ppb, respectively). The average annual outdoor content of O3 remains ~40 ppb. We also found a decrease of the outdoor content of NOx. In 1995 it was ~80 ppb and in 2003 only 10 ppb. However, in 2002 and 2003 an
158 Ageing and stabilisation of paper Chapter 10: Air pollution and its prevention increase was again registered, especially for NOx. For indoor air, this was the result of a delay in filter replacement. The filtering capacity for acidic and oxidative pollutants was significantly reduced. A closer look at the daily contents reveals that the pattern of pollution is comparable for outdoor and indoor air. The times, at which peaks occur outdoors, are identical to peak pollutant contents in indoor air, although certainly considerably reduced. In Figure 10.5, the upper graph shows outdoor content of NOx, whereas the lower one shows the same pollutant in the storage room with air purification, during the same period of time. It can be clearly seen that at high levels of NOx outdoors (250-300 ppb), the filtering capacity is somewhat reduced and approx. 25 ppb is contained in the storage room. For the storage room without air purification, contents are higher for a factor of up to 4.
300 a - outdoor NO content x b - indoor NO content x 250
) 200 b p p (
t
n 150 e t n o c 100
50 a 0 b
10 Jan 11 Jan 12 Jan 13 Jan 14 Jan 15 Jan 16 Jan 17 Jan date Figure 10.5: Peak outdoor contents of NOx result also in peak indoor contents of the same pollutant, albeit greatly reduced.
Based on results of the analyses of indoor air pollutants (NOx, SO2 and O3) we can demonstrate that filtering effectively purifies the atmosphere. If we compare actual outdoor and indoor contents, then a reduction of up to 90% is achieved. In case of high pollutant contents in outdoor air, the reduction is somewhat lower due to a decreased filtering capacity. A comparison of the pollutant levels in the storage rooms with and without air purification shows that the contents in the non-purified room, especially for NOx, are on average four times higher than in the purified room. Finally, we conclude that in our case, attention should be paid to NOx, because at peak NOx outdoor content, an increased content of NOx in the indoor air for both storage rooms was also found. This can probably be attributed to problems with capacity of the filter system at peak levels of NOx.
10.6 Study of stored materials The consequences of natural ageing during the period of study (8 years) might be negligible, especially in view of the research so far done on natural ageing of paper.26,27 However, the analyses have shown that air purification results in a cleaner indoor atmosphere. It is therefore of interest whether this has any effect on the stored objects.
Ageing and stabilisation of paper 159 Chapter 10: Air pollution and its prevention
Model materials The new, or model paper materials were used in two types of packaging: bound volumes (papers bound as a book, and placed directly on an archival rack shelf) and loose material (papers wrapped in alkaline archival covers and put in an archival box). It may be presumed that bound materials would deteriorate faster than loose materials.
0.5
P2 D1 I P2 D2 I 0.4 P2 D1 b P2 D2 b r
e 0.3 b m u n
r e
p 0.2 p o c
0.1
0.0 1995 1996 1997 1998 1999 2000 2001 2002 2003 year Figure 10.6: Average copper number determinations for cotton linters paper (P2) in storage rooms D1 (without air purification) and D2 (with air purification), b – bound volume, l – loose volume, throughout the years of study. An analysis of copper number determinations (Figure 10.6 and Figure 10.7), reflecting the content of reducing carbonyl groups, shows a trend of increase in both storage room D1 (without air purification) and D2 (with air purification). The trend seems to be clearer for groundwood containing paper P3, especially in the storage room without air purification. The trend also seems to be more pronounced for bound volumes than for loose volumes. The increased instability of groundwood containing paper can be attributed to lignin and hemicellulose fractions.
2.5
2.0 r
e 1.5 b m u n
r e
p 1.0 p o c P3 D1 I 0.5 P3 D2 I P3 D1 b P3 D2 b
0.0 1995 1996 1997 1998 1999 2000 2001 2002 2003 year Figure 10.7: Average copper number determinations for groundwood containing paper (P3) in storage rooms D1 (without air purification) and D2 (with air purification), b – bound volume, l – loose volume, throughout the years of study.
160 Ageing and stabilisation of paper Chapter 10: Air pollution and its prevention
20 P3 D1 I P3 D2 I P3 D1 b 15 P3 D2 b r o t c a f 10 g n i w o l l e y
5
0 1997 1998 1999 2000 2001 2002 2003 year Figure 10.8: Yellowing factors of materials calculated according to DIN 6167 for groundwood containing paper (P3) in storage rooms D1 (without air purification) and D2 (with air purification), b – bound volume, l – loose volume, throughout the years of study. The effect of presence of pollutants can further be demonstrated by means of colour change of the materials, especially of groundwood containing paper P3. As a measure of colour change, the yellowing factor was used, the value of which is obtained by first measuring the L*a*b* values of a specific point at the edge of a page and in the middle of the same page. The values were then converted to the yellowing value according to DIN 6167. Finally, the difference of yellowing values (yellowing at the edge vs. yellowing in the middle of a page) is calculated, providing the yellowing factor. In Figure 10.8, the values of yellowing factors per year and type of storage are presented. There is a significant difference between both types of packaging and also between both types of environment. Firstly, bound materials discolour more severely than the wrapped loose materials. Secondly, materials stored in a non-purified environment discolour more severely than materials stored in a purified environment. If we classify the way of storage basing on the yellowing factor, we see that the best way of storage is by wrapping loose papers in a storage room with air purification, while the worst way of storage is the form of bound volume put on an archival rack in a storage room without air purification. Based on these results, it is also remarkable that storage of wrapped loose papers in a non-purified storage room is preferable to storage of bound papers stored openly on a shelf in a storage room with air purification. However, before conclusions are drawn, light levels also have to be discussed. It is well known that under the influence of light, groundwood containing papers yellow faster. On the other hand, some pollutants are absorbed in lignin containing papers more readily, especially NOx. In the storage rooms, light intensity was evaluated. Both rooms were equipped with fluorescent light bulbs (type L36 W/20 Cool White). The measurements showed that in the storage room without air purification, the light was on approx. 25% of the time; the horizontal illumination was 50-100 lux while the vertical illumination was 30-60 lux. In the storage room with air purification, the lights were mainly switched on (90% of the time). Here, the horizontal illumination was approx. 90 lux, while the vertical illumination was approx. 60 lux. Considering that the yellowing factors are higher in less illuminated storage
Ageing and stabilisation of paper 161 Chapter 10: Air pollution and its prevention rooms, we can conclude that the effect of light is of lower importance than the effect of pollutants.
Taking the cumulative dose of NOx into account, we can correlate it with the yellowing factor found for the groundwood containing paper (P3) stored under the two types of environmental conditions as bound and loose wrapped materials. These results are presented in Figure 10.9. This correlation clearly points out the difference in the type of packaging. The bound materials discolour more severely than the wrapped loose materials. The cumulative annual dose <30 ppb was taken for the storage room with air purification, while the cumulative annual dose >30 ppb was obtained in the storage room without air purification. Due to the fact that the bound papers in the non-purified storage room showed a visual discoloration at a yellowing factor of 7, we consider this also as the LOAEL (lowest observed adverse effect level) criterion.28 If we take this value and apply it to the wrapped loose materials, we see that this yellowing factor is reached at a cumulative dose of 108 ppb NOx. We can now estimate the LOAED (lowest observed adverse effect dose). As the cumulative dose was calculated over the period of eight years, the LOAED is then 108 ppb divided by 8 years, i.e. 13.5 ppb. This can now be regarded as the NOAEL value (no observed adverse effect level) for a well- purified air in storage room. Taking the low accuracy into account we suggest that the annual cumulative dose of NOx in a storage room should not exceed 15 ppb. This is more than 40% lower than the current guideline according to the Dutch Archives Act, which is 25 ppb NOx.
16 bound materials loosely wrapped materials 14
12 r o
t 10 c a f
g 8 n i w o l l 6 e y
4
2
0 0 10 20 30 40 50 60 70 80 90 100 cumulative dose (ppb)
Figure 10.9: The effect of cumulative dose of NOx on the yellowing factor of the groundwood containing paper (P3), for bound and wrapped loose materials.
Old paper material Evaluation of old paper material is somewhat more complex, if compared to new materials. Although the old materials were selected with care, several factors remain unknown, e.g. (storage) history of the materials, and (in)homogeneity. Also, we should consider that these materials are already naturally aged for a period that was much longer than the duration of our research. And finally, we have to consider the low repeatability of the methods applied,
162 Ageing and stabilisation of paper Chapter 10: Air pollution and its prevention especially mechanical testing. In Figure 10.10, measurements of double folds are presented on a single page from the same book. As described above, all bound old materials were cut into four pieces and subsequently rebound in four sub-volumes: the front upper half (VB), front lower half (VO), the back upper half (AB) and the back lower half (AO). Two parts were placed in the archive with air purification and two in the storage room without air purification. In Figure 10.10, the effect of natural ageing for a selected volume is shown. As an example, let us consider the fold number data (Figure 10.10), one of the clearest cases in our study showing the effect of natural ageing. There is a clear difference corresponding to the sample origin: in all cases the front parts of the volume (VO and VB) exhibited a higher number of double folds than the back parts of the volume (AO and AB). The reason for this remains elusive. Although the correlation is less convincing due to data scatter, we see negative although not significantly different trends for all materials stored in the storage rooms with and without purification. Thus, no effects could be assigned to differences in storage.
900 1-3-VB-D2 1-3-VO-D1 800 1-3-AB-D1 700 1-3-AO-D2
600 s d l
o 500 f
e l
b 400 u o d 300
200
100
0 1994 1995 1996 1997 1998 1999 2000 2001 year Figure 10.10: The double folds of a stored old paper material (code 1-3) over time in storage rooms D1 (without air purification) and D2 (with air purification). Four rebound volumes from one original volume are presented. Two parts were stored in D2 (VB and AO) and two parts were stored in D1 (VO and AB).
6.0 6.0
5.8 5.8
5.6 5.6
5.4 5.4
5.2 5.2
5.0 5.0 H H p p 4.8 4.8
4.6 4.6
4.4 4.4 1-3-VO-D1 1-3-VB-D2 4.2 1-3-AO-D2 4.2 1-3-AB-D1
4.0 4.0 1994 1995 1996 1997 1998 1999 2000 2001 2002 1994 1995 1996 1997 1998 1999 2000 2001 2002 year year Figure 10.11: The average pH of cold water extract of four rebound volumes (code 1-2). Left: data and trends presented for the cut and rebound volumes of the lower part of the bound volume 1-2; right: data and trends presented for cut and rebound volumes of the upper part of the bound volume 1-2 materials, in storage rooms D1 (without air purification) and D2 (with air purification).
Ageing and stabilisation of paper 163 Chapter 10: Air pollution and its prevention
The data on pH also exhibit considerable scatter. If we assume that the original volume was always stored vertically with its top side upwards, two different plots can be drawn (Figure 10.11) for the rebound lower front and lower back parts (VO and AO respectively), and for the rebound upper front and upper back parts of the volume (VB and AB respectively). There seems to be an effect in the way of storage: materials stored in the storage room without air purification became somewhat more acidic than those in the storage room with air purification. In order to investigate whether the differences between the mean values of pH of the samples stored in the purified storage room and samples stored in the non-purified storage room are in fact statistically significant, we applied the t-test. We found that there is no significant difference in the pH of the cold water extract of the rebound lower parts of the volume in each storage room. On the other hand, a significant difference was found in the pH of the cold water extract of the rebound upper parts in each storage rooms. We can conclude that acidity of materials may increase during storage in non-purified atmosphere. By means of another statistical test, i.e. Wilcoxon test29 we can compare two paired groups. It calculates the difference between each set of pairs, and analyzes that list of differences at some significance level, e.g. = 0.05. In this way, we can calculate if the general pattern is significant or not. Simply said, we are able to prove if the dataset from the storage room without air purification generally show a lower pH of the cold water extract than the data obtained from the materials stored in the storage room with air purification. We applied this test to all data sets obtained for old materials. We were able to identify two significantly different sets of properties: acidity and tensile energy absorption measured in the cross direction of the materials (i.e. the total work done per unit area of a paper when testing to rupture). The pH of cold water extracts from papers stored in the storage room without air purification was lower than for papers stored in the storage room with air purification. This is reasonable, as the content of pollutants in the storage room without air purification is higher. Regarding tensile energy absorption (TEA) measured in the cross direction, we found that the TEA values for papers from the storage room without air purification were significantly higher. Thus, for papers stored under a more acidic environment it takes more energy to break fibre bonds.
10.7 Conclusions
The effects of air pollutants, SO2, NOx and O3, on natural deterioration of paper materials was demonstrated in a 8-years study combining measurements of material properties and analyses of indoor air quality. We can conclude that there is a measurable positive effect of air purification. The classical analytical methods we applied in our research yielded satisfactory data although interpretation was not straightforward. This was due to the relatively short duration of the study leading to small differences in the measured properties. However, the importance of a well-defined experimental set-up should be stressed. The NOAEL value (no observed adverse effect level) for a well-purified storage room has been estimated at 15 ppb, thus the annual cumulative dose of NOx in a storage room should not exceed 15 ppb. This is more than 40% lower than the currently valid guideline 31 in the Dutch Archives Act, which is 25 ppb NOx.
164 Ageing and stabilisation of paper Chapter 10: Air pollution and its prevention
10.8 References 1. A.E. Kimberly, A.L. Emley, Deterioration of book papers in libraries, Bureau of Standards, Miscellaneous Publications, 1933, 7, 140-144. 2. F. Lyth Hudson,W.D. Milner, Atmospheric sulphur and durability of paper, Journal of the Society of Architectorial Historians, 1961, 2, 166-167. 3. F. Lyth Hudson, R.L. Grant, J.A. Hocke, The pick-up of sulfur dioxide by paper, J. Appl. Chem., 1964, 14, 444-447. 4. F. Lyth Hudson, Die Sorption von Schwefeldioxid durch Druckpapiere und deren chemische Beständigkeit, Wochenblatt fur Papierfabrikation, 1967, 95, 660-663. 5. T. Iversen, J. Kolar, Effects of nitrogen dioxide on paper, FoU-projektet för Papperskonservering, Report No. 5, National Archives, Stockholm, 1991. 6. J.B.G.A. Havermans, Effects of Air Pollutants on the Accelerated Ageing of Cellulose-based Materials, Restaurator, 1995, 16, 209-233. 7. J.B.G.A. Havermans, Environmental influences on the deterioration of paper, Barjesteh, Meeuwes & Co, Delft, Rotterdam, 1995. 8. J. Palm, P. Cullhed, Deteriorating paper in Sweden : a deterioration survey of the Royal Library, Gothenburg University Library, Uppsala University Library and the National Archives, Riksarkivet, Stockholm, 1988. 9. H.J. Porck, J. v. Heijst, W.J.T. Smit, I. v. Leeuwen, Research Projects Within the Framework of a National Mass Conservation Plan in the Netherlands, Tappi Press, Atlanta, 1988, 74-80. 10. J.B.G.A. Havermans, T.A.G. Steemers, The indoor air quality of the Dutch State Archives: its purification, quality control and safeguarding the cultural heritage, Dobbiaco/Toblach, Italy, 25-29 June, 2004; http://www.asrm.archivi.beniculturali.it/CFLR/Dobbiaco/Dobbiaco.htm. 11. M. de Feber, J.B.G.A. Havermans, E. Cornelissen, The Positive Effects of Air Purification in the Dutch State Archives. Part I: Experimental set-up and air quality, Restaurator, 1998, 19, 212-223. 12. J.B.G.A. Havermans, J.P. v. Deventer, R. van Dongen, F. Flieder, F. Daniel, P. Kolseth, T. Iversen, H. Lennholm, O. Lindqvist, A.S. Johansson, The Effects of Air Pollutants on the Accelerated Ageing of Cellulose Containing Materials - Paper, Final report, EC/DGXII/STEP Project CT 90- 0100, TNO, Delft, 1994. 13. A. Johansson, Air pollution and Paper Deterioration, PhD Thesis, Göteborg University, Göteborg, 2000. 14. C. Fellers, T. Iversen, T. Lindström, T. Nilsson, M. Rigdahl, Ageing/Degradation of Paper, A literature Survey, FoU-projektet för papperskonservering, Report No. lE, National Archives, Stockholm, 1989. 15. E. Sjöström, Wood Chemistry. Fundamentals and Applications, 2nd ed., Academic Press, Boston, 1993. 16. D. Fengel, G. Wegener, Wood Chemistry, Ultrastructure and Reactions, 1 ed., Walter de Gruyter, Berlin, 1989. 17. R. Vosteen, Adviesrichtlijn luchtkwaliteit archieven, Ministerie van Volkshuisvesting, Ruimtelijke Ordening en Milieubeheer, Rijksgebouwendienst, 1994. 18. P. Defize, P. Marres, Experimental design for paper sampling (in Dutch), Internal TNO report from TNO TPD, Delft, The Netherlands, 1995. 19. TAPPI T 509; Hydrogen ion concentration (pH) of paper extracts (cold extraction method), 1988. 20. TAPPI T 430; Copper number of pulp, paper, and paperboard, 1988. 21. ISO 1924/2; Determination of tensile properties, 1985. 22. ISO 5626; Paper - Determination of folding endurance, 1978. 23. DIN 6167: Description of yellowness of near-white or near-colourless materials, 1980. 24. Labtech Notebook, 9.x - 13 ed., Andover, USA, 1995-2004. 25. Landelijk Meetnet Luchtkwaliteit, www.lml.rivm.nl, accessed 09/12/2004. 26. H. Porck, Rate of paper degradation. The predictive value of artificial ageing tests, European Commission on Preservation and Access, Amsterdam, 2000. 27. A. Barański,J. M. Łagan,T. Łojewski, The Concept of Mixed Control Mechanisms and its Applicability to Paper Degradation Studies, Proc. Durab. Paper Writing, Nov 26-20, Ljubljana, Slovenia, 2004.
Ageing and stabilisation of paper 165 Chapter 10: Air pollution and its prevention
28. J. Tetreault, Airborne Pollutants in Museums, Galleries, and Archives: Risk assessment, Control Strategies, and Preservation Management, Canadian Conservation Institute, Ottawa, 2003. 29. C. Chatfield, Problem Solving, A Statisticians Guide, University of Bath, Bath, 1988. 30. Anonymous, Van Regels naar Kennis. Regeling bouw en inrichting archiefruimten en archiefbewaarplaatsen, National Archives, The Hague, 2001 (text, in Dutch only, available at www.nationaalarchief.nl).
166 Ageing and stabilisation of paper Chapter 11 Stabilisation strategies
Matija Strlič, Jana Kolar, Jasna Malešič, Drago Kočar, Vid Simon Šelih, Boris Pihlar, Olivier Haillant, José L. Pedersoli Jr., Steph Scholten, Jozef Rychlý, Lyda Rychlá, Dominique Fromageot, Jacques Lemaire
11.1 Introduction Treatments leading to stabilisation of both model and real papers are discussed in this Chapter. The variety of procedures currently in use is overwhelming; however, our focus will not be to review them, but rather to evaluate their potential and identify the possibilities for development of practical treatments or processes. Also, we shall only focus on procedures using water as the solvent – the resulting distribution of the usual additives is quite possibly optimal using this medium. The use of non-aqueous solvents for introduction of alkalis or other additives may raise doubts about completeness of deacidification, washing out of harmful degradation products, side reactions with the particular solvent or its components, etc. These issues will not be addressed here. On the other hand, aqueous deacidification processes using Mg(HCO3)2, Ca(HCO3)2, or Ca(OH)2, will be reviewed regarding the rate of alkali deposition, achievable alkaline reserve and pH. Introduction of some of the better known stabilizers, such as phytate and halides will be studied. By application of the Arrhenius model to the data obtained from accelerated ageing studies, comparative data on degradation of differently treated papers at room temperature can be provided. This enables us to evaluate the suitability of treatments at the temperature of use.
11.2 Aqueous deacidification During paper deacidification, acids are either washed out of paper or neutralised. Using a polar solvent, such as water, at least a partial washing effect is easily achieved. Neutralisation of the remaining, even macromolecular acids can be achieved by introduction of alkalis. Since a surplus of alkali is desired (alkaline reserve), to prevent cellulose from acids that
Ageing and stabilisation of paper 167 Chapter 11: Stabilisation strategies build up or are absorbed during future degradation and storage, the pH of its saturated solution should be mildly alkaline, as this provides an optimal medium for cellulose. The pH of a solution of MgCO3 in equilibrium with atmospheric CO2 (which is expected also 1 in paper) at laboratory conditions (22 ºC) is 9.35, while that of CaCO3 is 8.35.
0.10 Ca(OH) Ca(HCO ) 0.35 2 3 2 -1 0.001 mol L 0.001 mol L-1 -1 -1 0.30 0.01 mol L 0.08 0.01 mol L -1 0.018 mol L 0.018 mol L-1 ) ) 1 1 - - g
0.25 g k k
l l 0.06 o o m m ( 0.20 (
e e v v r r e e s 0.15 s 0.04 e e r r
i i l l a a k k l 0.10 l a a 0.02 0.05
0.00 0.00 0 5 10 15 20 25 30 0 5 10 15 20 25 30 immersion time (min) immersion time (min) 0.40 0.30 Mg(HCO ) 3 2 Mg(HCO ) ; 0 min -1 0.35 3 2 0.001 mol L Mg(HCO ) ; 20 min 3 2 -1 0.25 0.01 mol L Ca(HCO ) ; 0 min 0.30 3 2 0.02 mol L-1 Ca(HCO ) ; 20 min
) 3 2 ) 1 1 -1 - -
0.04 mol L g Ca(OH) ; 0 min g 2 k k 0.25 0.20
l l Ca(OH) ; 20 min
o 2 o m m ( (
0.20 e e v
v 0.15 r r e e s s
e 0.15 e r r
i i l l 0.10 a a k k l l 0.10 a a
0.05 0.05
0.00 0.00 0 5 10 15 20 25 30 1 2 3 immersion time (min) repeated immersion Figure 11.12: Alkaline reserve in Whatman paper after immersion in solutions of differently concentrated Ca(OH)2, Mg(HCO3)2, and Ca(HCO3)2 for different periods of time. Some samples were repeatedly immersed in fresh solutions (all of concentration 0.01 mol L-1) with intermittent drying. The ratio of deacidification solution volume to paper was always 100 mL per 1 g. The error bars represent SD (n = 3).
Due to their low solubilities, the two carbonates are usually introduced as Mg(HCO3)2 or Ca(HCO3)2. On a small scale, preparation of solutions of hydrogencarbonates is non- demanding using bottles pressurized with CO2. During the procedure and on drying, the hydrogencarbonates release CO2 and precipitate in the form of respective carbonates. Contrary to Mg(OH)2, Ca(OH)2 is more soluble and preparation of a saturated solution -1 (0.018 mol L ) is possible. By reaction with atmospheric CO2, carbonate will form from the hydroxide during drying of the treated paper. An important difference between -1 solutions of Mg(HCO3)2, Ca(HCO3)2 and Ca(OH)2 is their pH, the pH of 0.01 mol L Ca(OH)2 solution being 12, the pH of hydrogencarbonate solutions with excess dissolved CO2 usually being 5.8-7.0. In the standard for permanent paper, the necessary amount of alkaline reserve is specified 2 to be 2% CaCO3. This specification is arbitrary – in general, a higher alkaline reserve will lead to better absorption of acidic gasses and will protect the material against endogenous acids. Due to its protective role, the amount of alkaline reserve should be one of the decisive parameters in optimisation of deacidification treatments, along with the achieved
168 Ageing and stabilisation of paper Chapter 11: Stabilisation strategies pH. As described in Chapter 3 (page 27), the standard method used to determine alkaline reserve is titrimetric, and it was extensively used in the study described here. To provide comparative data, the alkaline reserve is expressed here in mol kg-1; 0.2 mol kg-1 represents approximately 2% CaCO3 or 1.7% MgCO3. As seen in Figure 1.1, an adequate alkaline reserve is easily achieved in Whatman paper using concentrated solutions of Ca(OH)2 and Mg(HCO3)2, while with solutions of Ca(HCO3)2, the deacidification is less satisfactory. Longer times and higher concentrations are more favourable. Using more deacidification solution per mass of treated sample might still increase the effect. There is also practically no significant difference between the three solutions in the rate of precipitation, the final effect being achieved in 10 min in most cases. A repeated treatment increases the amount of deposited alkaline reserve considerably, a 3-fold immersion into a -1 0.01 mol L solution of Ca(OH)2 resulting in a 3.5% alkaline reserve.
11
10
9 H p 8
7 CaCO 3 MgCO 3
6 0.0 0.1 0.2 0.3 0.4 alkali reserve (mol kg-1) Figure 11.13: Dependence of pH of Whatman paper on the amount of alkaline reserve. The dependence of pH of Whatman papers with different amounts of alkaline reserve shows an interesting pattern (Figure 11.13). Paper pH was determined according to the new method of determination described recently,1 during which sample suspensions are rapidly equilibrated with atmospheric CO2. For both carbonates, the values initially increase very steeply: the more carbonate, the higher the pH. However, in the case of CaCO3, the curve levels off at the value of 8.5, the value being close to the one of a solution of CaCO3 saturated with CO2. As the maximum solubility is reached soon, only a small amount of CaCO3 is needed for the solution surrounding cellulosic fibres to become saturated. In the case of MgCO3, no plateau is apparently achieved partly due to its higher solubility, partly also due to the fact that a higher amount of alkaline reserve than 0.2 mol kg-1 was not achieved in the study. An analogous study with the paper sample S2, with initial pH 5.0, reveals a similar pattern of behaviour. However, since the starting pH is low, more alkali is needed to neutralise the acids and if the amount is insufficient, no deacidification is achieved at all, as demonstrated by negative values of alkaline reserve in certain cases in Figure 11.14. With Ca(OH)2, deacidification is easy to achieve even with more diluted solutions of the alkali, yet with a 0.001 mol L-1 solution, longer times are needed. The uptake of calcium and magnesium carbonates proceeds at a lower rate – the pH of treatment solutions (which still contain an
Ageing and stabilisation of paper 169 Chapter 11: Stabilisation strategies excess of CO2) is slightly less than neutral and in the acidic paper precipitation of carbonates is probably not favoured until neutralisation of acids is achieved.
0.55 Ca(OH) 0.08 Ca(HCO ) 0.50 2 3 2 0.001 mol L-1 0.001 mol L-1 0.45 -1 -1 0.01 mol L 0.06 0.01 mol L 0.40 0.018 mol L-1 0.018 mol L-1 ) ) 1 1 - - 0.35 g g 0.04 k k
l l o o 0.30 m m ( (
0.25 0.02 e e v v r r e e 0.20 s s e e r r 0.00
i 0.15 i l l a a k k l 0.10 l a a -0.02 0.05
0.00 -0.04 -0.05 0 5 10 15 20 25 30 0 5 10 15 20 25 30 immersion time (min) immersion time (min)
10.0 Mg(HCO ) Mg(HCO ) 3 2 0.15 3 2 9.5 -1 0.001 mol L-1 0.001 mol L 9.0 -1 0.01 mol L-1 0.01 mol L -1 0.10 0.02 mol L-1 8.5 0.02 mol L
) -1
1 -1 - 0.04 mol L 0.04 mol L g 8.0 k l
o 0.05 7.5 m ( e H v
p 7.0 r e 0.00 s
e 6.5 r i l a 6.0 k l -0.05 a 5.5
-0.10 5.0 4.5 0 5 10 15 20 25 30 0 5 10 15 20 25 30 immersion time (min) immersion time (min) Figure 11.14: Alkaline reserve in paper sample S2 after immersion in solutions of differently concentrated Ca(OH)2, Mg(HCO3)2, and Ca(HCO3)2 for different periods of time. The achieved pH of samples deacidified with Mg(HCO3)2 is given in the lower right graph. The ratio of deacidification solution volume to paper was always 100 mL per 1 g. The error bars represent SD (n = 3). As demonstrated in Figure 11.14, diluted solutions of the two hydrogencarbonates are not useful for the purpose, as no neutralisation is achieved even after long periods of immersion. This is also demonstrated by the pH of paper after treatment, which remains approximately 5. A positive effect is achieved only with more concentrated solutions, yet the deposited alkaline reserve is much smaller than in the case of Whatman paper: a 20-min -1 treatment with 0.04 mol L Mg(HCO3)2 provides 0.65% of MgCO3 after treatment, while -1 an analogous treatment with 0.018 mol L Ca(HCO3)2 leads to 0.50% of CaCO3. This may partly be due also to lower absorptiviry of the sized paper, as compared to the unsized Whatman paper.
The pH of papers treated with Mg(HCO3)2 reflects the deposited alkaline reserve. This is also true for sample containing CaCO3 (Figure 11.15). The observed behaviour is dissimilar to that of Whatman paper, demonstrated in Figure 11.13. Two features are particularly important: the pH of papers depends on the content of alkaline reserve apparently in the same way for both carbonates and levelling-off of the curve is achieved only at values higher than 0.2 mol kg-1. In the case of Whatman paper, this value is ten times less, i.e. 0.02 mol kg-1. Apparently, the remaining anionic compounds (sulphate, rosin, etc.) have a buffering effect in the rosin-sized paper sample S2, as discussed in ref. 3. No such effect is
170 Ageing and stabilisation of paper Chapter 11: Stabilisation strategies noticed in Whatman paper, which is composed of pure cellulose and its degradation products only.
10
9
8
7 H p
6
CaCO 5 3 MgCO 3
4 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 alkali reserve (mol kg-1) Figure 11.15: Dependence of pH of sample S2 on the amount of alkaline reserve.
0.6 0.12
0.5 0.10
0.4 ) ) 1
1 0.08 - - g g k k
l l o 0.3 o m m 0.06 ( (
e e v v r 0.2 r e e s s 0.04 e e r r
i i l l a 0.1 a -1 k k l -1 l Ca(HCO ) ; 0.018 mol L a Ca(OH) ; 0.018 mol L a 3 2 2 0.02 S1 S1 0.0 S2 S2 S3 0.00 S3 -0.1 0 5 10 15 20 25 30 0 5 10 15 20 25 30 immersion time (min) immersion time (min)
0.08
0.06
0.04 ) 1 - g k l 0.02 o m ( e
v 0.00 r e s e r
i -0.02 l a k l -1 a Mg(HCO ) ; 0.02 mol L -0.04 3 2 S1 S2 -0.06 S3
0 5 10 15 20 25 30 immersion time (min) Figure 11.16: Alkaline reserve in paper samples S1, S2 and S3 after immersion in solutions of 0.018 mol L-1 Ca(OH)2 or Ca(HCO3)2, or 0.02 mol L-1 Mg(HCO3)2 for different periods of time. The ratio of deacidification solution volume to paper was always 100 mL per 1 g. The error bars represent SD (n = 3).
Ageing and stabilisation of paper 171 Chapter 11: Stabilisation strategies
An important consequence of this effect is that irrespective of whether Mg(HCO3)2 or Ca(HCO3)2 is used, the resulting pH of deacidified rosin-sized papers will be similar. A study of behaviour of other real samples during deacidification treatments reveals similarities, but also important differences. The samples S1 (initial pH 5.7), S2 (initial pH 5.0) and S3 (initial pH 4.3) were used for comparison. The rate of neutralisation and uptake of alkali is practically identical for all three samples in -1 the case of 0.01 mol L Ca(OH)2. The more alkaline deacidification solution may also lead to fibre swelling and more rapid migration of acids and uniform neutralisation is thus achieved. This is not the case of hydrogencarbonate solutions, where the initial uptake is slower, especially in sample S2. This phenomenon does not depend on the initial pH of papers; sample S3 being the most acidic one with pH 4.3.
11.3 Thermo-oxidative stabilisation study using model paper Using the methodology as described in Chapter 3 (page 32), stability of differently treated samples can be evaluated at room temperature, provided that the Arrhenius model is valid. In the numerous studies described below, accelerated ageing of samples as separate sheets at 60, 70, 80 and 90 ºC provided no indication of Arrhenius graph curvature in any of the studied cases, it is therefore believed that extrapolation of the ageing behaviour to room temperature is generally possible and valid. Only a comparison of degradation rates at room temperature may provide a realistic estimation of the effects of stabilisation treatments.
140000
120000
100000 ) 1 -
l 80000 o m
J (
60000 a E
40000
20000
0 y y a 5 5 5 5 5 5 g 5 5 5 5 5 5 h h 0 0 0 0 0 0 0 0 0 0 0 0 C . . . M . . . 0 0 0 0 0 0 P P . 0 . 0 . 0 . 0 . 0 . 0
0 I 0 0 0 I 0 0 a g r r
N N I I r r K K B B C M
N N C C K K B B K K a g
C C S S K K a g C a g M S S K K C a g
M C K M K a g C M a g C M C M Figure 11.17: Apparent activation energies for ageing of Whatman paper containing CaCO3 (symbol Ca) or MgCO3 (symbol Mg). Deacidification was performed by immersion into 0.01 mol L-1 solution of Ca(HCO3)2 or 0.04 mol L-1 solution of Mg(HCO3)2, for 20 min. After drying, immersion into solutions of various potential antioxidants followed: KI, KBr, KSCN, phytate (Phy), some in two concentrations: 0.005 mol L-1 or 0.05 mol L-1. The error bars represent SD. A comparison of apparent activation energies for thermo-oxidative ageing at 65% RH is presented in Figure 11.17. The values are mostly similar, around 100 kJ mol-1, with a few exceptions. Especially the similarity of values of Ea in Mg-deacidified samples is interesting and speaks in favour of a common rate-determining step in all cases. As may be expected
172 Ageing and stabilisation of paper Chapter 11: Stabilisation strategies from the data in Figure 11.17, the corresponding values of degradation rate constants for thermo-oxidative ageing at 20 °C are much more variable in the case of CaCO3-containing samples than in the case of MgCO3-containing samples (Figure 11.18). The level of stabilisation of the sample containing CaCO3 and KI is particularly surprising: the samples will last 25-50 times longer than the one without the antioxidant. KBr and KSCN-treated papers are stabilized to a limited extent and will last up to 3 times longer than the one which was only deacidified.
1E-14 ) 1 - s
1 - l o m
l o m (
3 9 2
k 1E-15 y y a 5 5 5 5 5 5 g 5 5 5 5 5 5 h h 0 0 0 0 0 0 0 0 0 0 0 0 C . . . M . . . 0 0 0 0 0 0 P P . 0 . 0 . 0 . 0 . 0 . 0
0 I 0 0 0 I 0 0 a g r r
N N I I r r K K B B C M
N N C C K K B B K K a g
C C S S K K a g C a g M S S
K K C a g M C M K K a g C M a g C M C M
Figure 11.18: Rates constants of degradation of Whatman paper at 20 °C, 65% RH, containing CaCO3 (symbol Ca) or MgCO3 (symbol Mg). Deacidification was performed by immersion into 0.01 mol L-1 solution of Ca(HCO3)2 or 0.04 mol L-1 solution of Mg(HCO3)2, for 20 min. After drying, immersion into solutions of various potential antioxidants followed: KI, KBr, KSCN, phytate (Phy), some in two concentrations: 0.005 mol L-1 or 0.05 mol L-1. The bars represent SD.
Phytate (introduced using the treatment described in ref. 4) also leads to an approximately 9 times longer lifetime, which might indicate that even in purified papers, the role of -1 transition metals is important (CaCO3-containing Whatman paper contains ~20 mg g Fe). The addition of antioxidants to MgCO3-containing papers also has a pronounced effect, although significant conclusions are difficult due to the experimental errors associated with extrapolation. Stabilisation with iodides leads to a 2.5-3.5-times longer lifetime, similar to other halides and the pseudohalide (thiocyanate). There is no additional stabilisation effect of phytate. As discussed in Chapter 7 (page 114), stabilisation of samples in terms of DP is not the only goal we want to achieve. Sample yellowing can easily be evaluated on all the samples, as the plots b* vs. t are linear and the “rate constants” for yellowing (b*/t) can be calculated. By plotting them in Arrhenius coordinates, we can easily obtain apparent activation energies for the process (Figure 11.19). The differences are more pronounced than in Figure 11.17, and we can see that the error intervals are also wider. Nevertheless, the rates of yellowing, calculated at 20 °C, still allow us to draw some conclusions. The Whatman paper sample yellowing at the slowest rate is undoubtedly the one containing only CaCO3. Considering the high degree of stabilisation in terms of degree
Ageing and stabilisation of paper 173 Chapter 11: Stabilisation strategies of polymerisation (and the associated mechanical stability), achieved with iodide, it is unfortunate that the data in Figure 11.20 indicate a high instability towards yellowing. The rest of halides and the pseudohalide exhibit a similar tendency, phytate again with less associated yellowing. For the samples containing MgCO3, most additives will lead to a comparably intensive yellowing to that of the sample which was only deacidified.
140000
120000
100000 ) 1 -
l 80000 o m
J (
60000 a E
40000
20000
0 y y a 5 5 5 5 5 5 5 5 g 5 5 5 5 h h 0 0 0 0 0 0 0 0 0 0 0 0 C . . . M . . . 0 0 0 0 0 0 P P . 0 . 0 . 0 . 0 . 0 . 0
0 I 0 0 0 I 0 0 a g r r
N N I I r r K K B B C M
N N C C K K B B K K a g
C C S S K K a g C a g M S S
K K C a g M C M K K a g C M a g C M C M
Figure 11.19: Apparent activation energies for yellowing of Whatman paper containing CaCO3 (symbol Ca) or MgCO3 (symbol Mg). Deacidification was performed by immersion into 0.01 mol L-1 solution of Ca(HCO3)2 or 0.04 mol L-1 solution of Mg(HCO3)2, for 20 min. After drying, immersion into solutions of various potential antioxidants followed: KI, KBr, KSCN, phytate (Phy), some in two concentrations: 0.005 mol L-1 or 0.05 mol L-1. The error bars represent SD.
1E-3
1E-4 ) 1 - y a d
* b (
3 9 2
k 1E-5
1E-6 y y a 5 5 5 5 5 5 g 5 5 5 5 5 5 h h 0 0 0 0 0 0 0 0 0 0 0 0 C . . . M . . . 0 0 0 0 0 0 P P . 0 . 0 . 0 . 0 . 0 . 0
0 I 0 0 0 I 0 0 a g r r
N N I I r r K K B B C M
N N C C K K B B K K a g
C C S S K K a g C a g M S S K K C a g
M C K M K
a g C M a g C M C M
Figure 11.20: Values of “rate constants” of yellowing of Whatman paper, extrapolated to 20 °C, containing MgCO3 (Mg) or CaCO3 (Ca). Deacidification was performed by immersion into 0.01 mol L-1 solution of Ca(HCO3)2 or 0.04 mol L-1 solution of Mg(HCO3)2, for 20 min. After drying, immersion into solutions of various potential antioxidants followed: KI, KBr, KSCN, phytate (Phy), some in two concentrations: 0.005 mol L-1 or 0.05 mol L-1. The bars represent SD.
174 Ageing and stabilisation of paper Chapter 11: Stabilisation strategies
11.4 Thermo-oxidative stabilisation studies using papers for common use Already in 1998, a study on possibilities of stabilisation of alkaline paper was published,5 in which a number or treatments of notebook paper made from bleached pulp (60% magnefite spruce, 40% sulphate birch paper pulp) was described.
Figure 11.21: Limiting viscosity number of notebook paper during accelerated aging at 80 °C, 65% RH. Untreated sample was compared with the one deacidified with Mg(HCO3)2, the one additionally containing KI (1.4 mol g-1) and the one which was subjected to a reduction pre-treatment before deacidification.
As shown in Figure 11.21, apparently no degradation was noticed during a 40-day accelerated ageing experiment at 80 °C, 65% RH. The treatment took carbonyl groups -1 (page 106) into account and a reduction pre-treatment with NaBH4 (0.01 mol L , 2 h) prior -1 to deacidification with Mg(HCO3)2 (0.04 mol L , 30 min) and introduction of KI (iodide content in paper: 1.4 mol g-1) was the most favourable combination of treatments.
Sample S1 This sample from 1984, with a starting DP 1600 and pH 5.7, exhibits behaviour which is in many ways similar to that of Whatman paper. The values of apparent activation energies
(Figure 11.22) are mostly not statistically different from the Ea of the sample which was only deacidified, except for the sample containing KI. The resulting rates of degradation extrapolated to 20 ºC are all in favour of the use of halides and pseudohalide as antioxidants: especially in the case of Ca-containing paper, a decisively better stabilisation is achievable than if no antioxidant is used. In the case of MgCO3, standard deviation permits no significant conclusions, although iodide and thiocyanate both seem to have a stabilizing effect. Using the extrapolated rate constant of degradation for the untreated paper sample S1, the time it needs to degrade to DP 300, when a complete loss of mechanical properties is expected (page 38), can be calculated. At 20 ºC, it will take approximately 700 years, which seems to be a quite realistic prediction. The ratios of degradation rate constants at 20 ºC provide a factor of stabilisation and thus
Ageing and stabilisation of paper 175 Chapter 11: Stabilisation strategies enables us to compare the efficiency of different treatments (Table 11.4). During thermo- oxidative degradation, the most effective treatment among the ones studied, is deacidification with Ca(HCO3)2, and subsequent addition of iodide. Other antioxidants in combination with deacidification lead to 5-13 times longer lifetime. Mg-based deacidification performs approximately twice better than Ca-based deacidification.
140000
120000
1E-13 100000 ) 1 - s
) 80000 1 1 - - l l o o m m
l J
60000 o (
a m (
E 1E-14
3 9 2
40000 k
20000
0 1E-15 I I r r I I r r 6 a g N N 6 a g N N K K B B K K B B
C M C C C M C C
K K
a g K K
a g 6 S S 6 6 S S 6 C a g M C a g K K M
K K
C
6 M C 6 6 M a g 6
a g 6 6 6 C 6 M C
M
6 6 6 6 Figure 11.22: Apparent activation energies and rate constants of degradation at 20 °C for thermal degradation of sample S1 containing CaCO3 (symbol Ca) or MgCO3 (symbol Mg). Deacidification was performed by immersion into 0.01 mol L-1 solution of Ca(HCO3)2 or 0.04 mol L-1 solution of Mg(HCO3)2, for 20 min. After drying, immersion into solutions of various potential antioxidants followed: KI, KBr, KSCN. All concentrations of stabilizers in the impregnation solutions were 0.05 mol L-1. The error bars (left graph) and the bars (right graph) represent SD.
Table 11.4: Factors of stabilisation of paper sample S1 using a variety of treatments, compared to no treatment. treatment factor of stabilisation
Ca(HCO3)2 2.5
Ca(HCO3)2 + KI 58
Ca(HCO3)2 + KBr 4.5
Ca(HCO3)2 + KSCN 13
Mg(HCO3)2 4.5
Mg(HCO3)2 + KI 8
Mg(HCO3)2 + KBr 7.5
Mg(HCO3)2 + KSCN 12
While the uncertainty is high, a comparison of the "rate constants" of yellowing at 20 °C (Figure 11.23) reveals that, contrary to Whatman paper, the paper sample S1, containing CaCO3 and KI, is yellowing at the slowest rate. The data on degradation and data on yellowing at room temperature both speak in favour of this stabilisation treatment.
176 Ageing and stabilisation of paper Chapter 11: Stabilisation strategies
140000 0.01
120000
100000 ) 1 -
) 80000 1E-3 y 1 - l a o d
* m
b J
60000 ( (
a 3 9 E 2 k 40000
20000 1E-4
0 I I r r I I 1 a g r r N N 1 a g K K N N B B K K S B B C M C C S C
M K K C C g a
K K
a g 1 S S 1 1 S S 1 C a g M C a g S K K M S S K K S C 1 M 1 C
1 M a g 1
a g S 1 S 1 S 1 C S 1 M
C M S S
S S 1 1 1 1 S S S S Figure 11.23: Apparent activation energies and values of "rates constants" of yellowing at 20 °C during thermal degradation of sample S1 containing CaCO3 (symbol Ca) or MgCO3 (symbol Mg). Deacidification was performed by immersion into 0.01 mol L-1 solution of Ca(HCO3)2 or 0.04 mol L-1 solution of Mg(HCO3)2, for 20 min. After drying, immersion into solutions of various potential antioxidants followed: KI, KBr, KSCN. All concentrations of stabilizers in the impregnation solutions were 0.05 mol L-1. The error bars (left graph) and the bars (right graph) represent SD.
Sample S2 This sample provides the clearest example of potential of deacidification and further stabilisation through introduction of antioxidants. The sample originates from 1870, and has an initial DP 550 and pH 5.0.
200000
180000 1E-12 160000
140000 ) 1 - s
120000 1E-13 1 - ) l 1 - o l o
100000 m
l m
o J ( m
80000 ( a
3 E 9 1E-14 2 60000 k
40000
20000 1E-15
0 I I r r 2 a g N N I I r r K K B B 8 a g S C N N M C C K K B B K K a g
C
M C C 2 S S
2
K K a g C a g
M 8 S S S K K 8 S
C a g M C 2 M K K 2
a g
C 8 M S 8 2 S
2
a g C M
8 S 8 S C M 2
2
8 S 8 S Figure 11.24: Apparent activation energies and rate constants of degradation at 20 °C for thermal degradation of sample S2 containing CaCO3 (symbol Ca) or MgCO3 (symbol Mg). Deacidification was performed by immersion into 0.01 mol L-1 solution of Ca(HCO3)2 or 0.04 mol L-1 solution of Mg(HCO3)2, for 20 min. After drying, immersion into solutions of various potential antioxidants followed: KI, KBr, KSCN. All concentrations of stabilizers in the impregnation solutions were 0.05 mol L-1. The error bars (left graph) and the bars (right graph) represent SD.
Ageing and stabilisation of paper 177 Chapter 11: Stabilisation strategies
Table 11.5: Factors of stabilisation of paper sample S2 using a variety of treatments, compared to no treatment. treatment factor of stabilisation
Ca(HCO3)2 2.7
Ca(HCO3)2 + KI 5.4
Ca(HCO3)2 + KBr 4.5
Ca(HCO3)2 + KSCN 5.1
Mg(HCO3)2 58
Mg(HCO3)2 + KI 210
Mg(HCO3)2 + KBr 660
Mg(HCO3)2 + KSCN 400
The apparent activation energies for thermo-oxidative degradation of the sample, -1 deacidified with Ca(HCO3)2 are in the range of approximately 100 kJ mol , whereas those for Mg-containing samples are unusually high, up to 180 kJ mol-1 (Figure 11.24). The extrapolated rates of degradation with the surprisingly small uncertainty intervals show significant differences between the rate constants of untreated, deacidified and additionally stabilized samples. The calculated lifetime expectancy of 100 years can thus be significantly prolonged. The factors of stabilisation are listed in Table 11.5.
1E-3
1E-4
) 1E-5 1 - y a d
* b (
1E-6 3 9 2 k
1E-7
1E-8 I I r r 2 a g N N K K B B S C M C C a g
K K 2 S S C a 2 g M
S K K S C 2 M 2
a g S 2 S 2 C M
S S 2 2 S S Figure 11.25: Values of "rates constants" of yellowing at 20 °C during thermal degradation of sample S2 containing CaCO3 (symbol Ca) or MgCO3 (symbol Mg). Deacidification was performed by immersion into 0.01 mol L-1 solution of Ca(HCO3)2 or 0.04 mol L-1 solution of Mg(HCO3)2, for 20 min. After drying, immersion into solutions of various potential antioxidants followed: KI, KBr, KSCN. All concentrations of stabilizers in the impregnation solutions were 0.05 mol L-1. The error bars (left graph) and the bars (right graph) represent SD.
While deacidification of the sample with Ca(HCO3)2 leads to a 3-fold prolongation of lifetime, the use of Mg(HCO3)2 is even more recommended. The use of halogenides or the pseudohalogenide may lead to further pronounced stabilisation, especially in the case of
178 Ageing and stabilisation of paper Chapter 11: Stabilisation strategies
Mg-containing samples. Stability of colour during thermo-oxidative degradation is similar in most cases, while two samples exhibit a significantly pronounced stability. It is important to realize that no treatment leads to more rapid yellowing than that observed in the non-treated paper.
Sample S4 This sample from 1938 with a starting pH of 5.3 exhibits very similar behaviour to the previous one, although the more pronounced uncertainty intervals permit few definite conclusions. In the samples containing CaCO3, iodide and thiocyanate have a significant effect, while the effect of antioxidants in Mg-containing samples is only indicated (Figure 11.26). It is difficult even to evaluate the effect of deacidification; a possible reason might be low homogeneity of the sample.
120000
1E-12 100000 )
80000 1 - s
) 1 1 - - l l o o 1E-13 m m
60000 l J o (
a m ( E
3 9
40000 2 k
1E-14 20000
0 I I I I r r r r 1 a g 1 a g N N N N K K K K B B B B
C C M M C C C C
K K K K a g a g
1 S S 1 S S 1 1 C a g C a g M M
K K K K
C C 1 M 1 M 1 1
a g a g 1 1 1 1 C C M M
1 1 1 1 Figure 11.26: Apparent activation energies and rate constants of degradation at 20 °C for thermal degradation of sample S4 containing CaCO3 (symbol Ca) or MgCO3 (symbol Mg). Deacidification was performed by immersion into 0.01 mol L-1 solution of Ca(HCO3)2 or 0.04 mol L-1 solution of Mg(HCO3)2, for 20 min. After drying, immersion into solutions of various potential antioxidants followed: KI, KBr, KSCN. All concentrations of stabilizers in the impregnation solutions were 0.05 mol L-1. The error bars (left graph) and the bars (right graph) represent SD.
Sample S5 This sample represents a contemporary alkaline office paper and was included in this study to investigate the possibilities of further stabilisation using impregnation with antioxidants. The sample is has an initial pH 9.9 and CaCO3 content of 18%. The data obtained, i.e. apparent activation energies, rate constants of degradation and values of "rates constants" of yellowing are all comparable, i.e. no significant differences are indicated. No additional stability at 20 °C is achieved using antioxidants; however, the samples do not exhibit more pronounced yellowing either.
Ageing and stabilisation of paper 179 Chapter 11: Stabilisation strategies
140000
120000 1E-3
100000 1E-14 ) 1 - ) s 1 - ) 80000 1 y - 1 - l l a o o d
m * m
l b J
60000 o (
( a
m 3 ( E 9
2 3 9 k 2 1E-15
40000 k
1E-4 20000
0 I r I r I r 5 5 N 0 N N K B K K B B S 1 S
C C K C 5 K K 5 0 S
S S S 5 1 S 5 0 K
K K
S
1
S 5 5 0 S 1 S Figure 11.27: Apparent activation energies (left), rate constants of degradation (middle) and values of "rates constants" of yellowing (right) at 20 °C for thermal degradation of sample S5 containing 18% (m/m) of CaCO3. KI, KBr, KSCN were introduced by immersion into 0.05 mol L-1 solutions of the respective compound. The error bars (left graph) and the bars (two graphs on right) represent SD.
11.5 Stability of real-life paper samples during photodegradation Historical samples may be exposed to elevated light leves during exhibition. Evaluation of their stability during exposure to light as well as during subsequent thermal ageing is therefore needed whenever novel paper stabilisation methods are proposed. In accordance with the discussion in Chapter 9, acidic factual paper samples S1, S2 and S4 were exposed to irradiation in Xenotest at 70 W m-2, 30 °C and 50% relative humidity for three days. Results (Figure 11.17) are dramatically different from the ones obtained for Whatman paper (Figure 9.11), which was strongly destabilised by the addition of magnesium carbonate.
0.000012
0.000010 )
1 0.000008 - r h
1 - l o 0.000006 m
l o m (
k 0.000004
0.000002
0.000000 r r r r r r g g g 1 a 2 a 4 a B B B B B B S S S C C C
M M M
a g a g a g 1 2 4 1 2 4 C C C M M M S S S
S S S
1 2 4 1 2 4 S S S S S S
Figure 11.28: Rate constants of degradation during irradiation in Xenotest (3 days, 70 W m-2, 30 °C and 50% RH) of samples S1, S2 and S4. The samples were either left untreated, or deacidified using Ca(HCO3)2 or Mg(HCO3)2, denoted by Ca and Mg, respectively. Tetrabuthylammonium bromide was introduced by immersion into 0.05 mol L-1 aqueous solution (denoted by Br). The error bars represent SD.
180 Ageing and stabilisation of paper Chapter 11: Stabilisation strategies
Here, deacidification using magnesium hydrogen carbonate significantly stabilised historical paper samples S2 and S4. The difference in the behaviour of unsized Whatman paper and historical paper probably lies in the addtitives, which strongly decrease the pH of paper containing MgCO3 (Figure 11.4). Apart form paper no. 6, no significant stabilisation of paper during irradiation by light is achieved by the addition of bromide. During natural ageing, chromophores are formed, resulting yellowing of paper and thus an increase of b* values. As these chromophores absorb visible light, they may be degraded when exposed to light. This is observed (Figure 11.18) for samples no. 1 and 8, with initial b* values 14.6 and 17.5, where yellow component decreases during the three days of irradiation. Paper samples S1 from 1984 contains optical brighteners and has a starting b* 2.3. An increase of b* is observed after one day of irradiation, following by its decrease.
0.20 0.30 0.15
0.25 0.10
0.20 0.05 ) ) 1 1 - - h h
* 0.15 * 0.00 b b ( (
k k -0.05 0.10
-0.10 0.05
-0.15 0.00 t t t t t t t t t t p p p p p p p p p p
r r
r r 1 a g r r 2 a g r r 1 a g 2 a g B B B B S B B C S M B B C S M C S M C M a g
a g
a g 1 1 a g 2 2 1 1 2 C 2 M C S M C S S M C S S M
S S
S
1 1 2 2 1 1 2 2 S S S S S S S S
0.20
0.15
0.10
0.05 ) 1 - h
* 0.00 b (
k -0.05
-0.10
-0.15 t t t t t - p p p p p -
r r 4 a g r r 4 a g B B S B B C M S C M a g
a g 4 4 4 4 C M S C S M S
S
4 4 4 4 S S S S
Figure 11.29: “Rate constants“ of yellowing during: p – irradiation in Xenotest (3 days, 70 W m-2, 30 °C and 50% RH); and t – subsequent thermal ageing (2 days, 80 °C, 65% RH) of samples S1, S2 and S4. The samples were either left untreated, or deacidified using Ca(HCO3)2 or Mg(HCO3)2, denoted by Ca and Mg, respectively. Tetrabuthylammonium bromide was introduced by immersion into 0.05 mol L-1 aqueous solution (denoted by Br). The error bars represent SD. As with Whatman sample, subsequent thermal ageing resulted in strong yellowing. Unfortunately, due to a rather large error in determination of “rate constants“, no significant differences are observed between differently treated samples.
Ageing and stabilisation of paper 181 Chapter 11: Stabilisation strategies
11.6 Conclusions
A study of aqueous deacidification including solutions of Mg(HCO3)2, Ca(HCO3)2 and Ca(OH)2 showed that it is generally preferable if duration of immersion is long (>20 min) and concentration of the alkali high (>0.01 mol L-1). The alkaline reserve obtained after a single treatment of Whatman paper is up to 2-3% CaCO3, however, Ca(OH)2 precipitates much more efficiently than the two hydrogencarbonate analogues. The pH of Whatman paper, achieved after deacidification, does not depend on the content of CaCO3 if higher than 0.2%. The same study using historical paper samples showed that deacidification is not achieved as easily. Unless high concentrations of Mg(HCO3)2 or Ca(HCO3)2 are used (achievable in CO2-pressurised bottles), deacidification will not take place at all. The uptake of alkaline reserve seems to depend also on alkalinity of the deacidification solution. The pH after deacidification is similar for both Ca- and Mg-containing papers (with comparable alkaline reserves), indicating that the remaining non-cellulosic compounds in the rosin-sized real papers may have a buffering role. By ageing the samples as loose sheets at several temperatures, and by application of the Arrhenius model, a comparison of extrapolated rates of degradation at room temperature is possible. This provides comparative data on the suitability of a particular stabilisation treatment. A study using Whatman paper showed that the most favourable stabilisation is achieved using iodide in Ca(HCO3)2-deacidified papers, however, these samples are also expected to discolourate at a more pronounced rate. Bromide and thiocyanate also exhibited a stabilizing effect, although all samples are expected to yellow at a faster rate than the non- treated, slightly acidic sample. The study on historical samples showed a different picture. Generally, both deacidification treatments and all antioxidants are expected to perform decisively better than the non- treated sample. During thermo-oxidative degradation, the formation of colour proceeds at the same or smaller rate in deacidified and additionally stabilised samples than in the non- treated samples. A combination of iodide and CaCO3 mostly gave the best results, the factor of stabilisation (prolongation of expected lifetime) being up to 60. Other treatments mostly performed comparably. In one real-life paper, all samples containing MgCO3 were significantly better stabilised than CaCO3-containing samples, the achieved maximum factor of stabilisation was found to be 660. Contemporary papers containing CaCO3 as a filler do not benefit from an addition of antioxidant, however, no harmful effect was found either. Despite the strong degradation during irradiation of factual paper samples made from bleached chemical pulps, the yellow component decreased. The determination of b* or brightness after irradiation, as proposed by the recent ASTM standard is thus of no practical value. Instead, DP or mechanical properties have to be determined. In all samples, a subsequent thermal ageing step introduced strong yellowing of paper. It is proposed that such a step is included in standard evaluation of light stability of paper. While Whatman filter paper is an extremely valuable reference model, as it is generally available and of comparative quality, real-life paper samples may exhibit different properties. The differences in the ageing behaviour between unsized Whatman paper and historical paper may originate in addtitives, which decrease the pH of paper deacidified
182 Ageing and stabilisation of paper Chapter 11: Stabilisation strategies using Mg(HCO3)2. Any proposed stabilisation treatment must therefore be optimised and checked on a variety of such samples. The Arrhenius approach, although time-consuming, may provide comparative data on the suitability of treatments and their performance at room temperature.
11.8 References 1. M. Strlič, J. Kolar, D. Kočar, T. Drnovšek, V.-S. Šelih, R. Susič, B. Pihlar, What Is the pH of Alkaline Paper?, e-PS, 2004, 1, 35-47. 2. ANSI/NISO Z39.48-1992(R1997): Permanence of Paper for Publications and Documents in Libraries and Archives, 1997. 3. J. Kolar, G. Novak, Effect of Various Deacidification Solutions on the Stability of Cellulose Pulps, Restaurator, 1996, 17, 25-31. 4. B. Reißland, S. de Groot, Ink corrosion: comparison of currently used aqueous treatments for paper objects, Preprint from the 9th International congress of IADA, Copenhagen, 121-129. 5. J. Kolar, M. Strlič, G. Novak, B. Pihlar, Aging and Stabilization of Alkaline Paper, J. Pulp Pap. Sci., 1998, 24, 89-94.
Ageing and stabilisation of paper 183
Chapter 12 Outlook
Matija Strlič, Jana Kolar
12.1 Introduction Many of the processes of cellulose degradation bear resemblance to degradation of other polymeric materials. This has been demonstrated in the book numerous times. Still, cellulose is a notable exception among macromolecular materials: students of its degradation have materials on their disposal which already survived many centuries, even millennia. Proper stabilisation can thus be based on real historical examples and experience. In this final Chapter, we would like to provide a short summary of the state-of-the-art in paper stabilisation and related areas of research.
12.2 Analytical methodology There are still numerous possibilities for improvement of analytical methodology. Many methods used in paper research are outdated and give systematically wrong results, e.g. copper number, which might be used only on a selected set of very similar samples. Methods of determination of functional groups, essential in studies of mechanisms, are in need of a thorough revision. The recent advances in fluorescent labelling and chromatographic separation are a considerable step forward.1 Pedersoli stressed the need for miniaturisation of analytical methods in relation to studies of paper-based cultural heritage already in 1999.2 In the age of nanotechnology, this shouldn't be an insurmountable task. Miniaturisation of determination of pH was achieved,3 yet in life sciences it is possible to determine the pH in a single living cell so why stop at 1-mm electrodes? Application of non-destructive methods is equally sought for. An example of how beneficiary may be the introduction of a new technique to a long-studied subject, is chemiluminometry. Similarly, a number of techniques in degradation studies have been developed by polymer chemists, which are almost never in use in paper degradation studies, e.g. oxygen absorption. Another truly useful area of development is the recent advancement of imaging techniques. Spatially resolved information is extremely important in the studies of inhomogeneity of degradation processes, linked either with distribution of additives or stabilizers, or with migration and distribution of paper components.
Ageing and stabilisation of paper 185 Chapter 12: Outlook
12.3 Degradation mechanisms Due to well-known limitations, degradation mechanisms have to be studied by applying some external stress, e.g. temperature, irradiation or pollution. The relevance of studies of mechanisms is easily justifiable: we need to know the chemistry of a particular process before we can propose a suitable stabilisation treatment. Acid-catalyzed hydrolysis is a relatively well-researched area; however, evidence provided in this book suggests that in factual paper, it is influenced by a variety of paper additives. E.g. alum, the very source of acidity, may also influence simultaneous oxidation processes. Once made mildly alkaline, oxidation prevails over acid-catalysed hydrolysis, and studies of degradation mechanisms have provided knowledge needed for the development of stabilisation treatments using antioxidants. The superiority of such treatments over deacidification alone is well demonstrated in this book. An area of study, in need of profound further studies, is photooxidative degradation. Again, the general principles of polymer photo-degradation can be successfully applied to cellulose, yet it remains uncertain to what extent we can stabilize paper by deacidification, if at all. It is certain that it is not only degradation during irradiation, which is important, as the processes during subsequent storage may lead to pronounced discolouration. These processes need to be thoroughly understood before we can propose stabilisation treatments for frequently exhibited materials. The influence of paper components other than cellulose is another subject in need of research. A notable example is lignin – although a number of studies were published in literature related to papermaking, the behaviour of lignin-containing papers during both photooxidative and thermo-oxidative ageing remains elusive and effective procedures for stabilisation are missing. One reason for this is obvious: studies of ageing behaviour of lignin-containing paper are difficult due to the lack of appropriate analytical techniques. New methods of dissolution were proposed by researchers of lignocellulosic materials,4 and knowledge is accumulated in the area of wood research.5
12.4 Experimental approach to ageing The variety of ageing experiments in studies of mechanisms of ageing should certainly still be expanded due to the complexity of the processes. However, in studies of stabilisation treatments, it is preferable if researchers adhere to one of the standardized ageing procedures. The Arrhenius approach was extensively relied upon in this book. It is based on the established correlation between natural ageing and results of accelerated ageing provided by Zou et al.6 and based on chemiluminometric evidence. Although the studies are tedious and time-consuming, they are only capable of providing a realistic estimation of the behaviour of a particular paper at room temperature – which is the only relevant piece of information for the end-user, who wants to know the performance of a particular treatment at the temperature of use. Another important line of study was mentioned in the book, i.e. relation of single-sheet to stack or closed-vessel ageing. It is well-known that paper in closed books will degrade faster, but a general model to describe the behaviour has not been developed yet.
186 Ageing and stabilisation of paper Chapter 12: Outlook
Closed-vessel ageing is an attractive methodology not only because it provides an estimation of ageing behaviour of paper in closed books, but also because it enables the researcher to perform a variety of experiments in one ageing chamber simultaneously.
12.5 Standards on paper ageing A variety of standards on accelerated ageing of paper are in use and are referred to in Chapter 3. Among them, accelerated thermal ageing at 80 °C and 65% RH is probably most often used. Following a several-year project, three new ASTM7 standards were also introduced. The ASTM D 6819-02 standard on dry oven exposure utilizes the closed-vessel approach, and its relevance should be validated against the Arrhenius approach. Besides, the experimental difficulties discussed in Chapter 6, need further attention. The Arrhenius approach, based on our best understanding of the ageing processes, may provide the most relevant answer to the question of life expectancy of a particular paper at room temperature, and its use should be promoted. The ASTM D 6789-02 standard on accelerated light ageing might be in need of a revision, according to research presented in Chapter 9. Determination of brightness after irradiation does not seem suitable for evaluation of the effect of irradiation. The rather loosely defined experimental conditions (temperature, relative humidity, power) may lead to irreproducible research data. Furthermore, a thermal ageing step following light ageing should be introduced in order to demonstrate the vulnerability towards degradation proceeding in darkness after irradiation is discontinued.
12.6 Storage Chapter 10 is devoted exclusively to storage, and it provides a premium example of proper long-term care and an evaluation of the positive effects of removal of pollutants from storage area. However, an even lower limit of pollutant concentration is proposed than the currently accepted standards. Additionally, research on the effects of relative humidity on ageing of mildly alkaline cellulosic materials indicates that other guidelines might need revision, also. A lower relative humidity that the one currently prescribed by the ISO 11799 standard8 could turn out to be preferable. At 90 °C, relative humidity of 20% was found to be most favourable as the material was stabilized to a significant extent (up to 10 times) if compared to ageing at 65% RH. Experiments at lower temperatures are needed to confirm the trend at lower temperatures. In any case, significant stabilisation could be expected following the adjustment of a simple parameter, i.e. relative humidity. Another possibility is lowering the temperature during storage. Using the data presented in Chapter 11, a simple decrease of temperature from 20 °C to 15 °C will reduce the rate of degradation by 50%. In tropical climates, the effect of lowering the temperature from 30 °C to 20 °C would be even more dramatic: the rate of degradation would reduce by almost 75%.
Ageing and stabilisation of paper 187 Chapter 12: Outlook
12.7 Development of applications The development of applications for factual paper samples is work- and time-consuming. Numbers of different samples need to be tested under a variety of conditions of photo- and thermal ageing in order to assure that there is no adverse effect of a proposed conservation treatment on paper stability. In Chapter 11, a number of rosin-sized samples were discussed and the following general conclusions for aqueous deacidification could be drawn: ― If allowed, deacidification should proceed in as concentrated solutions as possible;
― The pH of MgCO3- and CaCO3-containing deacidified factual rosin-sized papers is comparable.
― MgCO3 provides better stabilisation than CaCO3, both alkalis provide a stabilisation factor of 2.5-60 at room temperature. Such stabilisation factors are probably not achievable by adjusting storage conditions alone; ― Application of antioxidants (peroxide decomposers) may multiply the stabilisation factor by a further factor of 2-10. The most favourable combination was usually iodide/CaCO3. ― Discolouration during ageing of treated rosin-sized factual papers compared to non- treated is not expected to be more significant. This behaviour is contrary to model papers. Other findings we would like to refer to are the beneficial effect of removal of degradation products during deacidification and the beneficial effect of reduction of carbonyl groups, which have been shown to initiate oxidative degradation. Many further findings described in this book might find application in stabilisation treatments. However, some items and problems are so particular or sensitive that no general treatment can be applied. Even in such cases, this book might be helpful in devising an appropriate individual approach to conservation.
12.8 References 1. J. Rohrling, A. Potthast, T. Rosenau, T. Lange, G. Ebner, H. Sixta, P. Kosma, A Novel Method for the Determination of Carbonyl Groups in Cellulosics by Fluorescence Labeling. 1. Method Development, Biomacromolecules, 2002, 3, 959-968. 2. J.L. Pedersoli Jr., The development of micro-analytical methodologies for the characterisation of the condition of paper, Preprint From the 9th International Congress of IADA, 1999, 107-114. 3. S. Saverwyns, V. Sizaire, J. Wouters, The acidity of paper. Evaluation of methods to measure the pH of paper samples, Preprints of the 13th Triennial ICOM-CC Meeting, ICOM committee for conservation, 2002, 2, 628-634. 4. F. Berthold, K. Gustafsson, R. Berggren, E. Sjöholm, M. Lindström, Dissolution of Softwood Kraft Pulps by Direct Derivatization in Lithium Chloride/N,N-Dimethylacetamide, J. Appl. Polym. Sci., 2004, 94, 424-431. 5. D.N.-S. Hon, N. Shiraishi (Eds.), Wood and cellulosic chemistry, M. Dekker, New York, 2000. 6. X. Zou, T. Uesaka, N. Gurnagul, Prediction of paper permanence by accelerated aging. Part II: Comparison of the predictions with natural aging results, Cellulose, 1996, 3, 269-279. 7. ASTM D 6819-02, Standard Test Method for Accelerated Temperature Aging of Printing and Writing Paper by Dry Oven Exposure Apparatus; ASTM D 6833-02, Standard Test Method for
188 Ageing and stabilisation of paper Chapter 12: Outlook
Accelerated Pollutant Aging of Printing and Writing Paper by Pollution Chamber Exposure Apparatus; ASTM D 6789-02, Standard Test Method for Accelerated Light Aging of Printing and Writing Paper by Xenon-Arc Exposure Apparatus. 8. ISO 11799 Information and Documentation - Document Storage Requirements for Archive and Library Materials, 2003.
Ageing and stabilisation of paper 189
Index Aluminium sulphate 93 Analytical techniques 25 A Antiozonants 18 Accelerated ageing 39,90,93,97 Antioxidants 18,70,115,173,186 Accelerated photoageing 50,144 Antioxidants, synergism 20 Acceleration of photooxidation 46,52,58 Antirads 18 Acetic acid 39 Aqueous deacidification 6 Acid-catalysed hydrolysis 85,88,90,105, Arrhenius equation 34,38,90,91,98,167 110,114,124,153 Arrhenius equation extrapolation 35 Acid-catalysed hydrolysis, dependence on Al Arrhenius equation confidence of prediction 36 content 96 Arrhenius plot curvature 94,127,172 Acid-catalysed hydrolysis, increase of acidity Arrhenius plot linearity 98 95 Archival box 155,160 Acidification 153 Archives 3,114,154
Acidity 6,89,91,155,164 Atmospheric CO2 27,168 Acidity of library collections 85 Attenuated total reflectance spectroscopy 30 Acids, migration 39,89 Autoxidation 13,70,76,102,138 Activation energy 11,34,54,75,90, Autoxidation, heterogeneous character 78,103 94,111,113,123,172 Autoxidation, initiation 13,77,103,115 Additives 102 Autoxidation, length of kinetic chain 125 Ageing 9,186 Autoxidation, propagation 16,75 Ageing, effect of moisture 46 Autoxidation, termination 19,77 Ageing, effect of water 46 Autoaccelerated reactions 66 Ageing in closed vessels 39 Autoacceleration 76 Ageing in dynamic conditions 39 Average molecular weight 28 Ageing in stacks 38,39,186 Ageing of encapsulated material 39 B,C Ageing of wrapped material 161 BHT 116 Ageing of loose sheets 38,186 Biodeterioration 4 Air purification 154,157 Bleached chemical pulps 102,106,112 Air ventilation & circulation 156 Bond scission 89 Aldehyde groups 102,105 Brightness 142 Alkali earth metal 105,123 Bromide 172 Alkaline degradation 101 Bursting strength 26
Alkaline reserve 6,27,167 Ca(HCO3)2 112,167
Alkali sensitivity 106 CaCO3 27,105,106,109,128,145,168
Alkalinity 27 Ca(OH)2 167 Alkoxyl radicals 72,73,78 Carbon arc light 45,47 Alkyl radicals 77 Carbon black 21 Alum 4,89,94 Carbonyl groups 15,66,88,102,123,
Ageing and stabilisation of paper 191 Index
129,133,138,155 Cupriethylenediamine 28 Carboxyl groups 88 Crystallinity 66,70,92,103,109,122,127 CED 28 Crystallization 18 Cellulose 15 Cellulose acetate 12,15 D Cellulose nitrate 15 Daylight 45 Chain-breaking antioxidants 19 Deacidification 6,27,102,108,112, Chain scission 57,74,75,108 114,117,154,167 Charge recombination 72 Degradation 9 Charge-transfer complexes 124,133,140 Degradation in inert atmosphere 96,122 Chemiluminescence, effect of humidity 80,130 Degradation, effect of humidity 39,90,93 Chemiluminescence, effect of irradiation 132 Degradation of linear polymers 32 Chemiluminescence, effect of molar mass 69 Degradation, post-irradiation 147 Chemiluminescence, proportionality constant Degradation products 114,167 128 Degradation, temperature cycling 125 Chemiluminometry 7,14,65,74,121,139 Degree of polymerisation 17,26,28,32,74, Chemiluminometry, dynamic experiment 90,93,103,125 104,124,125 Denaturation 9 Chemiluminometry, effect of sample Depolymerisation 16 morphology 122 Desferal 116 Chemiluminometry, effect of water 121 Determination of functional groups 25 Chemiluminometry, instrumentation 79,121 Determination of peroxides 107 Chemiluminometry, isothermal experiment 129 Diffuse reflectance FTIR 30 Chemiluminometry, non-destructive Diffuse reflectance spectrophotometry 31 sampling 80,122 Diffusion controlled reaction 103 Chlorophenol red 85 Diffusion phenomena inside books 97 Chromophore 14,32 Dioxetane 129 CIEEL mechanism 73 Dissociation energies of single bonds 11 Citrate 116 Distribution of molar masses 29 Closed vessel ageing 93 DSC 14,127 Coatings 5,102 DTA 14 Collection surveys 5,25,86 DTPA 116 Colour change during ageing 114 Durability 4 Colour measurement 31 Dyes 5,53 Colourimetry 31,108,114 Dynamic chemiluminometric experiments 74 Copper number 155 Corona discharge 10 E Corrosion 9 EDTA 116 Crack formation 17,53,58 Ekenstam equation 33,36,74,90,92,93,147 Critical photoproduct 57 Electric field 10 Cross-linking 9,14,28,74,141 Electronic data 3 Cumulative dose 162 Electron spin resonance 108,140
192 Ageing and stabilisation of paper Index
Electron transfer reactions 112,129 Hydrodynamic volume 29 Elementary fibrils 25 Hydrolysis 53,56 Endogenous factors of degradation 6,99,154 Hydrocarbons 138 End groups 10,66,71,75,102,104 Hydrogen peroxide 72,103,107,132 Environment 6,110 Hydroperoxides 13,16,56,66,68,70, Exogenous factors of degradation 6,89,99,154 77,107,125,129,137 Explosives 79 Hydroperoxyl radical 72,103,107 2-Hydroxybenzophenone 21 F Hydroxyl radicals 56,73,103,107,108,129,139 Factor of stabilization 176 Hydroxylation assay 108 Fenton reaction 13,21,108,115,117,132,139 Fillers 5,102 I,J,K First-order reaction 88,92 Imaging microscopy 55 Flame retardants 18 Incunabula 4 Fluorescence 16,32,138 Initiation of autoxidation 67 Fluorescent light 45,161 Information carrier 3 Fluorescent UV-lamps 46,50 Infra-red spectroscopy 30 Fluorophores 32 Inhibitor 14 Folding endurance 26,156 Inks 7 Food industry 7,79 Interfibre bonding 25 Formaldehyde 73 Intrinsic viscosity 28 -Fragmentation 12 Iodide 172 Free radicals 12,57,66 Iron gall ink 39,108,116 Free radical degradation 13 Isothermal chemiluminometric experiments 77 FTIR 28,30,99,127 Jagiellonian Library 85 Kaolin 154 G Ketone group 71,138 Gallic acid 117 Gelatine 4,26,29 L Glass transition temperature 15,103,127 L*a*b* system 32 Glucose 103 Laser cleaning 10,28,124 Glycosidic bond 32,88,123,128,133 Laser light scattering detector 29 Graphic art objects 31,114,146 Levoglucosan 123 Groundwood 4 LiCl/N,N-dimethylacetamide 29 Lifetime expectancy 176 H Lifetime prediction 38 Halide 173 Lifetime prediction uncertainty 35 HALS 21 Light, glass-filtered 137 Hemicellulose 102,160 Light-induced degradation 14 "Hollander" beaters 4 Light absorbers 20 Homolysis of peroxides 137 Light absorption 31,32 Humidity 39 Light filters 20,47,49,74,144
Ageing and stabilisation of paper 193 Index
Lignin 6,7,26,88,89,102,116,134,160,186 O,P Lowest observed adverse effect level 162 Outdoor weathering 45 Ozone 10,153 M Oxidation 7,13,98,121 Macrofibrils 25 Oxidation, role of water 109 Maltitol 129 Oxygen 128 Maltose 129 Oxygen depletion 39,55 Mark-Houwink-Sakurada equation 28 Paper acidity 5,26 Mass-average molar mass 29 Paper as a consumer material 3 Mass deacidification 6,85,114 Paper characterization 25 Mass transfer phenomena 93 Paper conservation 6 Mechanical properties 7,9,25,66,71,89,125,163 Paper consumption 3 Mechanisms of degradation 39,186 Paper lifetime 5,176 Mechanochemical degradation 10 Papermaking process 89 Mercury arc light 45 Paper mills 4 Mercury vapour arc light source 49,55 Paper production 4,86 Metal deactivators 115 Peeling off 12,33 Metal halide lamps 46 Peracids 13
Mg(HCO3)2 113,167 Permanent paper 27,168
MgCO3 106,109,112,128,132,145,168 Peroxide decomposer 20,115,115 Micro-drilling 27 Peroxides 104,111,124,133 Micro-sampling 27,81,134 Peroxides, kinetics of formation 107 Migration of degradation products 93,97,161 Peroxyacids 16 Mixed-control mechanism 85 Peroxyl radicals 13,65,68,70,73 Mixed-control kinetics 97 pH 5,26,91,93,105,112,164,168,185 Model samples 6,88 pH by cold extraction 26 Moisture 154 pH pens 27,85 Moisture content 91 Phosphites 20 Phosphorescence 138 N Photo-stabilizers 20
NaBH4 106 Photoacoustic detection 30 Natural ageing 37,90,163 Photoageing 45,137 Natural-light ageing 142 Photoageing units 46 Natural weathering 46 Photoageing, instrumentation 45 Neutralisation 172 Photoageing, effect of geometry 48 Newspaper 26 Photochemical reactions 45 Nitrogen oxides 39,89,153 Photochemical activation 53 N,N’-(5 nitro-1,3-phenylene)bisglutaramide Photodegradation 6,10,137,161,186 108,140 Photoinitiation 57 No observed adverse effect level 162 Photoionization 16 Norrish type I & II mechanism 16,138 Photolysis 10,16,53,55,137 Number-average molar mass 29 Photomultipliers 72,80,132
194 Ageing and stabilisation of paper Index
Photon counting 80 Rate constant 33,34 Photooxidation 10,55,66,137 Rate constant of degradation 11,88 Photooxidation, effect of water 54,56 Rate of degradation 74,76 Photooxidation mechanisms 52,54,57 Rate of yellowing 173 Photosensitisers 72 Recrystallization 9 Photosensitised degradation 137 Recycled paper 5 Phytate 116,172 Reducing groups 155 Pigments 53 Relative humidity 6,109,131,133,142 Plasma treatment 10,124 Reduction of carbonyl groups 106,175 Pollution 153 Residual oxidisability 125 Pollutants 39,89 Rosin-sized paper 170 Poly(-methyl styrene) 15 Rubber 13,18 Polyamide 15,52,56,67 Russel's reaction 71,75 Polyethylene 10,67 Polymer chain mobility 73,103,109 S Polymer degradation 7 -Scission 73 Polymer hydroperoxides 78 Secondary amines 19 Poly(methyl acrylate) 15 SEM-EDX 28 Poly(methyl isopropyl ketone) 15 Sensitizer 16 Poly(methyl methacrylate) 10, 12,15 Single-sheet ageing 97 Poly(methyl vinyl ketone) 15 Singlet oxygen 55,67,72,107,132,141 Polyolefins 68,125 Singlet dimol chemiluminescence emission Polypropylene 55,67,70 72,132 Polysiloxanes 14 Size exclusion chromatography 29,141 Polystyrene 10,15 Sizing 5,7 Polyurethane 54 Software 3 Poly(vinyl acetate) 17 Spectrofluorimetry 31 Poly(vinyl chloride) 12,16,52,79 Stabilizers 13,18,53,70 Poly(vinyl cyclohexane) 10 Stabilizers, synergistic effects 18 Pre-exponential factor 11,34,75,90 Stack ageing 97 Preventive antioxidants 19 Standard deviation of y-residuals 37,93 Pseudohalide 173 Standards for accelerated ageing 39,187 Pullulan 68,104 Stanford method 25,85,86 Pulping processes 4 Starch sizing 4 Star-Spangled Banner 4 Q,R Stereoregularity 69 Quantum yield of light emission 15,74 Sterically hindered phenols 19,70 Quenching 66,70 Sterically hindered amines 21 Quencher 16,20,74,128 Storage 4,97,99,111,146,154,168,187 Radical scavenger 108,115 Storage guideline 157 Radiolysis 10 Straw 4 Rag paper 6,30,112 Sulphur dioxide 39,89,153
Ageing and stabilisation of paper 195 Index
Superoxide anion 14,72,103,107,138 U,V Surface 26 Ultra-accelerated photo-ageing 52,52 Surface pH 27,91 Ultrasound 10 Surface temperature 48,55 Unzipping 17 -transition 127 UV-A 50 Vinegar syndrome 12 T Viscometry 28,89,141 Tearing resistance 26 Visual perception 31,53 Teflon® 16 Visual properties 58 Temperature 35 Volatile degradation products 93 Temperature cycling experiment 125 Tensile energy absorption 164 Tensile strength 26,52,155 W,X,Y,Z Termination of autoxidation 19 Washing 105,117 Tetrahydrofuran 29 Water 54,56,90,109 Textiles 7 Water absorption isotherm 109 Thermo-oxidative degradation 10,101 Water as reactant 111 Thermogravimetry 11 Water, plasticizing effect 109 Thermal degradation 10 Weak links 10,88,123 Thermolysis 10,124 Weathering 9 Thiocyanate 172 Weathering sites 46 Tidelines 26 Weathering by immersion 56 Time to 50% property loss 38 Xenon arc light 46,47,55,139 Tiron 132 Yellowing 106,114,147,173 Transglycosidation 122 Yellowing value 161 Transition metals 4,13,17,19,107, Zero-order reaction 93 109,116,132,173 Transmission FTIR spectroscopy 30 Triplet oxygen 71
196 Ageing and stabilisation of paper