Schematic description of cellulose chains altered by oxidation. Carbonyl groups (circles) and carboxyl groups (rectangles) are introduced along the molecule chain. The cellulose chain is not shortened by oxidation, but has suffered from a modifi cation that eases further attacks of the molecule. 

Context in source publication

Context 1

... metallo-gallate inks are typically composed of two main ingredients, an iron sulphate salt and a tanning agent that was mostly extracted from gall apples. The stability of iron gall inks in terms of permanence and durability has been found to rely on the ratio between iron ions and tanning agent (Neevel 1999). For preparation of iron gall ink, acid components are used preferably and on formation of the iron gall ink complex, sulphuric acid is additionally deliberated (Neevel and Mensch 1999). Ideally, iron ions should be bound in colour-forming complexes provided enough chelating agents, i.e. tannic acid available. Due to impure source materials this cannot always be assumed. Free transition metal ions are expected to act in re- dox reactions leading to Fenton type radical formation (Emery and Schröder 1974; Neevel 1995; Strlic, Kolar et al. 2003). A broad variety of available recipes for iron gall ink (Zerdoun Bat-Yehouda 1983) and various probable impurities in historic source materials for ink preparation, among them other transition metal ions such as copper salts render the investigation of this writing and drawing material diffi cult. Iron gall ink has to be considered as a hardly reproducible system. When considering conservation treatments of objects that suffer from iron gall ink corrosion, the question of the driving force behind degradation grows very important. A synergistic effect has been suggested for both, historic and model papers. According to this model, acid hydrolysis caused by sulphuric acid and autoxidation due to iron ions degrades objects. Consequently, treatments should inhibit both degradation mechanisms. Translated into conservation treatment of objects that suffer from iron gall ink corrosion, a deacidifi cation and a chelating step is needed for successful treatment. This approach was investigated in depth (Neevel 1995) and an aqueous calcium phytate/calcium hydrogen carbonate treatment was de- veloped and further studied (Reißland 1999; Kolar, Strlic et al. 2003; Kolar, Šala et al. 2005). However, many conclusions are based on rather unspecifi c descriptors of the ongoing process, such as pH and burst strength, both of them indicating merely the result of degradation not its cause. In order to clearly evaluate the impact of the combined phytate treatment two main characteristics are indicative for the success of the treatment. The oxidation of the cellulose molecule is monitored by the development of the carbonyl groups while the hydrolytic chain scission is refl ected in a decrease of molecular weight. A successful treatment has to suppress the formation of new carbonyl groups and the ongoing cleavage of the cellulose chains. To study the effect of calcium phytate on these two parameters, model papers and historical paper samples were included in the present research. An unbalanced iron gall ink was applied on Whatman fi lter papers and on modern rag test papers made of historic linen textiles. Additionally, historical samples were used to evaluate the effectiveness of the treatment on naturally pre-aged papers. All of the samples underwent accelerated ageing after the treatment to observe their long-term stability. The impact of the accelerated ageing on the treated samples and those without treatment is monitored by changes occurring on the cellulose molecule. Cellulose degradation occurs via oxidation and acid hydrolysis. During cellulose oxidation, the amount of oxidized functionalities is increased; carbonyl and carboxyl groups are introduced (see Fig. 1). Due to the different chemical structure of the two groups of oxidized functionality, selective labelling and thus differentiation between the two types can be achieved. Cellulose chains that have been altered by acid hydrolysis will also develop more carbonyl groups as new reducing end groups will be formed (see Fig. 2). Each cellulose chain has one reducing end group. If hydrolytic cleavage occurs, a new end group is generated. However, this new group cannot be directly measured as there has no method available so far that can distinguish between reducing end groups and other carbonyl groups along the cellulose chain. Nevertheless, the amount of reducing end groups can be calculated from the chain length distribution. Besides the weight average molecular weight (Mw) the chosen analytical set-up yields also the number average molecular weight (Mn) of cellulose. This value (Mn) represents the average number of cellulose chains and hence is equal to the amount of reducing ends. The reciprocate of Mn in g/mol times 10 6 represents the amount of reducing ends in mmol/kg. This calculation is based on the assumption that the end groups are present in the form of an aldehyde and not already oxidized. Since this value is calculated indirectly it is referred to as “theoretically formed reducing ends”. The main characteristic of acid hydrolysis is the decrease of the molecular weight of cellulose. As described above, the amount of newly introduced reducing end groups correlates with the decrease in molecular weight. No additional carboxyl groups will be introduced due to acid hydrolysis (see Fig. 3). In order to detect oxidative and hydrolytic damage of cellulose samples, fl uorescence labelling in combination with gel permeation chromatography (GPC) and multiple detection were employed (Röhrling, Potthast et al. 2002). For this analytical approach, the cellulose sample has to be dissolved. A very common and well-studied cellulose solvent for GPC analysis is N,N -dimethylacetamide/lithium chloride 0.9% (v/w) (DMAc/ LiCl) (Dupont and Mortha 2004). Fluorescence labelling is performed before the cellulose sample is dissolved. In GPC columns the cellulose molecules will be separated according to their hydrodynamic volume, i.e. the size of the dissolved molecule in its solvent. This hydrodynamic volume can be translated into molecular weight by multi angle laser light scattering (MALLS) and a refractive index detector (RI). Additionally, a fl uorescence detector quantifi es the amount of oxidized functionalities. With this analytical set-up it is possible to determine the content of oxidized functionalities and at the same time the molecular weight distribution of cellulose plus the averaged molecular weight parameter (Mw, Mn). As the labels are especially designed to bind covalently to a specifi c oxidized functionality, i.e. either carbonyl or carboxyl groups, a very sensitive means for the detection of oxidization in cellulose is available. A signifi cant advantage of this method is the small amount of sample material needed; analysis may be performed with 5 mg of cellulose on a routine base. Additional paper analysis was performed by means of direct laser ablation induc- tively coupled plasma mass spectrometry (LA-ICP-MS). The surface of the paper under investigation is ablated by an Nd: YAG 213 nm laser system. The amount of ablated material is related to the amount of cellulose removed from the surface measured by the 13 C signal which originates from organic material, i.e. cellulose. The ablated material from the paper surface is transported by a helium gas fl ow directly into the ICP-plasma and analyzed for its metal ion content (Prohaska, Latkoczy et al. 2002). Test papers were prepared of Whatman fi lter paper no. 1 containing almost pure α cellulose without additional sizing or fi llers. Even though it does not entirely refl ect the characteristics of rag paper, the changes of the cellulose in Whatman fi lter paper can be measured without having effects originating from hemicellulose, residual lignin and other naturally occurring paper components. Test papers were prepared of modern handmade paper composed of historic linen textiles from fl ax (linum usitatissimum) without fi llers and with gelatine surface sizing. Flax fi bres are composed of 76-80% of cellulose, 2-3% of lignin, 12-17% of hemicellulose and 5% of other components (wax, pectin and fats). In historic pa- permaking, some of these impurities originating from the raw material fl ax would have been removed during fi bre processing, but residues of them are present in the test paper. The impurities make the test paper more similar to original rag papers. This type of model paper is intended to link the fi lter paper to historic handmade papers. A set of naturally aged papers carrying iron gall ink was used for testing the phytate treatment on naturally aged test objects. It is very likely that these papers are made out of linen fabrics, ...