ABSTRACT—This article presents details of the rationale used in setting up a research project, describes the characterization of the 25 different papers and textiles selected for study, and gives an in-depth discussion of the purpose, advantages, and disadvantages of the 10 different analytical procedures used. The project investigated the effects of a fumigant, sulphuryl fluoride (Vikane), on cellulosic and ligneous materials. The data gathered by this project will be presented at a later date.

1 INTRODUCTION

THE EVALUATION of scientific results and recommendations in conservation literature is greatly aided by a good understanding of the rationale behind the selection of test materials and methods. Unfortunately, space rarely permits the authors to explain fully such important issues as how the chemistry of the system under study has influenced their approach to the selection of analytical methods; the extent to which these experimental procedures can be expected to yield the desired information; and how closely the scientific model approximates the “real life” situations that the conservator meets in his or her laboratory.

The thought processes used in setting up a major project depend upon the particular preservation issues being investigated as well as the materials under study. If the issues are explored in depth and the data are interpreted with insight, it is often possible to gain information about the materials that were not part of the original intent of the project. It is also possible that the problems addressed in the project may be relevant to other related materials. For example, cellulose is one of the main components of both paper and basketry; a chemically based study of one of these materials could be useful in predicting the chemical behavior of the other. Therefore, any detailed explanation of the rationale behind a project will be useful in understanding and evaluating the results of many other scientific studies.

It was with these ideas in mind that we decided to write a detailed account of the planning of a scientific project currently under way in the Conservation Processes Research Division of the Canadian Conservation Institute (CCI). This project is a collaborative research effort with the Getty Conservation Institute (GCI) and the Smithsonian's Conservation Analytical Laboratory (CAL); its aim is to investigate the effect of sulphuryl fluoride (Vikane) on materials commonly found in museum collections.

This article presents a detailed explanation of the planning of CCI's portion of this project, together with the pertinent information used in making important decisions. The numerical results of the project, along with interpretations and recommendations for the conservator, will be presented in subsequent papers. A literature review of the properties of Vikane as a chemical fumigant may be found in an article published by the three collaborating institutions (Derrick et al. 1990).

The specific role of CCI in the Vikane project is to examine cellulosic and ligneous fibers; GCI is responsible for studying pigments/stones, metals, resins, and oils/waxes; and CAL is investigating proteins and dyes. With all these materials, a project of this type could potentially have applications for paper, textiles, wood, paintings on canvas, and a very wide range of composite ethnographic and archeological artifacts. Yet encompassing so many areas is clearly not possible given a reasonable amount of time and resources. Therefore, the first important decision that needed to be made was how to limit the area of investigation. After discussions with conservators and scientists in CCI, we decided to limit our investigation to paper and textile materials. The cellulose and lignin in these particular fibers are present in a relatively pure form. This focus has the advantage of allowing for a straightforward analysis and interpretation of data. It is expected that information relating to the effects of Vikane on cellulosic and ligneous paper and textiles will have considerable application to other materials that also contain significant amounts of these fibers.

A very brief summary of the chemistry of Vikane as it relates to the study of cellulose and lignin is given in section 2. Understanding the chemistry of sulphuryl fluoride is necessary to decide what analytical procedures will be used to monitor the changes (if any) introduced by fumigation of the fibers by sulphuryl fluoride. This knowledge, coupled with information about the reactivity of the various components found in paper and textiles, permits production of some likely reactions between the sample material and the fumigant. Other considerations arise from possible impurities in the commercial product and/or special conditions imposed by the method of fumigation (concentration, temperature, RH, etc.). The conclusions reached after consideration of the pertinent literature were that degradation of the fibers by Vikane is most likely to proceed by acid hydrolysis.

Section 3 describes the materials selected for the project and discusses the criteria used, considers the important variabilities among the samples and their influence on sample selection, and presents an initial characterization of the samples. The analytical data for this initial characterization are given. For the results of this project to be of practical value to the conservation field, the samples selected have to bear some similarities to the type of artifacts found in museum collections. Therefore, considerable time and effort went into obtaining a wide variety of potential sample material.

The analytical techniques chosen for this study are outlined in section 4. A summary of the procedure is given as well as any pertinent chemical reactions. The benefits and disadvantages of the individual methods are discussed. As this project is the largest, in terms of the variety of paper and textile materials studied and the number of analytical techniques used, carried out by the Conservation Processes Research Division of CCI, it provides an unusually broad view of the materials and analytical methods that can be used in scientific investigations of its type. Although care is taken to comment on any special circumstances that relate specifically to the Vikane project, the vast majority of the information in the analytical techniques section will be of general use to readers who wish to know more about the procedures used by cellulose and lignin chemists. The information should also help those not familiar with the field to evaluate and understand many of the investigations in the technical literature.

The method of sample preparation will be outlined in future articles along with the analytical data arising from the experiments. The paper and textile materials samples were fumigated at the same time, according to the published experimental protocol (Derrick et al. 1990).

2 CHEMICAL CONSIDERATIONS

AN IMPORTANT starting point for this project was to survey the chemical reactions of sulphuryl fluoride and to relate this behavior to the known chemical and physical reactions of cellulose and lignin. An attempt was made to look specifically at the interactions most likely to occur under the conditions (e.g., temperature, RH, and Vikane concentration) used in the fumigation of museum collections. The most important aspects of the chemical behavior of the system under study are given below, along with our predictions of how the major components of paper and cellulosic textiles may be affected.

The solubility and reactions of sulphuryl fluoride with water have been found to be strongly pH dependent (Cady and Misra 1975). Rapid hydrolysis in basic solution and slow hydrolysis in more neutral media result in increased acid and fluoride (F−) content (Cady and Misra 1975). Buffering salts and nucleophiles such as amines are subject to attack by sulphuryl fluoride (Cady and Misra 1975; Padma et al. 1982). Consequently, any alkaline reserve in materials (e.g. deacidified paper) may be reduced during fumigation, and acidic residues may be deposited on the fibers. It is unclear whether any observed increase in the rate of degradation of deacidified paper by Vikane fumigation will be due mainly to the loss of buffer reserve or if there are other reactions that may lead to deterioration of the substrate.

An examination of the literature concerning the manufacture of sulphuryl fluoride indicates that hydrofluoric and hydrochloric acids may also be present as impurities in the commercial-grade of Vikane (Gustafson and Skinner 1976; Cook and Gustafson 1978). If present in any significant quantities, these acids will also contribute to depletion of buffer reserve and/or increase in the acid content of the fibers being fumigated. The manufacturer of Vikane is also investigating preparing this fumigant in a process that will reduce the final impurity concentrations (experimental-grade Vikane). Both the commercial-grade and experimental-grade Vikane were included in the GCI/CCI/CAL project.

The consequences of acidic impurities in the fibers can be considerable. In particular, cellulose is sensitive to acid in the presence of moisture. The 4%–9% moisture commonly found in fibers under common ambient conditions (30%–70% RH) is ample to allow acid-catalyzed hydrolysis to proceed, both during the fumigation itself and later, as acidic residues remain in contact with the fibers. Hydrolysis of cellulose most frequently occurs at the 1,4-glycosidic linkage. As hydrolysis proceeds over time, oxidation of the fiber (visualized by yellowing or darkening) and physical embrittlement of the substrate will be observed. Hydrolysis is dependent on time, temperature, and the quantity of water and acid available. Therefore, it was decided that the emphasis in this project should lie upon analytical methods that can detect acidic materials as well as follow any changes that may result from acid attack.

The chain breakage due to hydrolysis leads to a decrease in the average polymer length (also referred to as degree of polymerization, DP�). Therefore, some procedure that can monitor any changes in polymer length should be included in the experimental protocol.

Acid-catalyzed reactions of lignin have also been described as being of paramount importance in many aspects of lignin chemistry (Sarkanen and Ludwig 1971, 345). Therefore, one would expect that the lignin component of cellulosic fibers would also experience significant degradation, providing the acid content of the fibers increases considerably. The chemistry of the reactions between sulphuryl fluoride and lignin has not been well studied. However, numerous references exist in the literature concerning the chemical and physical interaction of sulphuryl chloride and lignin. It is likely that this information has considerable application to the fluoride derivative. As shown by the chemistry of sulphuryl chloride, the areas of potential reaction of fumigant with lignin are as follows:

1. halogenation of the aromatic nucleus (Telysheva et al. 1966). The hydrochloric acid which is present as an impurity in Vikane has been cited as a potential halogenation agent of lignin. (Rassow and Zickmann 1929);
2. esterification of side chains involving halogenation, demethylation, and degradation (Paschke 1922; Aparicio and Sastre 1969; Telysheva et al. 1966).

These reactions can degrade lignin by introducing chromophoric groups that may contribute to color changes and lead to reduction in polymeric length and physical strength.

It is unclear whether the fumigation process can have any effect on the state of oxidation of the cellulose and lignin. However, any study that seeks to follow fiber breakdown should include some method capable of determining the state of oxidation of the material. Similarly, it was considered essential to do some type of physical testing, especially since loss of physical strength is a parameter frequently used in scientific projects of this type. An added advantage of physical testing methods is that they give data of value in determining changes in both the cellulose and the lignin portion of fibers. As shown below, there are significant problems in monitoring changes in lignin by the chemical methods currently available.

The available literature suggests that cellulosic and ligneous fibers are not greatly affected by sulphuryl fluoride fumigation (Kenaga 1957; Gray 1960). However, these studies did not use analytical methods capable of detecting very small changes; nor did they include accelerated aging of treated samples or monitoring of materials over a long period of time. Accelerated aging is essential if we are to draw any conclusions about the long-term effects of fumigation with Vikane.

Much of the controversy surrounding fumigation includes the possibility of residues of the fumigant remaining with the fibers and slowly releasing over time (Meikle and Stewart 1962; Scudamore and Heuser 1971). Ethylene oxide has been especially criticized for this problem (Scudamore and Heuser 1971; Green and Daniels 1987; McGiffin 1985). Therefore, it was thought to be essential to investigate this question, particularly since it is such an important health and safety concern. It was also considered desirable to try and include some way of determining if there are residues of reaction products of sulphuryl fluoride associated with the fumigated samples. Detection of chemicals such as these would provide circumstantial evidence of a reaction between the fumigant and the samples and would support any finding of fiber degradation.

3 SELECTION OF MATERIALS

IN RECOGNITION of the fact that conservators will want to fumigate a very wide variety of materials, attempts were made to obtain many different types of paper and textiles. The criteria used in the initial collection of sample material were as follows:

1. availability of at least 100 g of fiber
2. absence of components that may greatly interfere with analysis of samples or interpretation of results (e.g., dyes on textiles, coatings on paper)
3. absence of large quantities of dirt, grime, etc.
4. a reasonably homogeneous appearance
5. ability to be dated
6. cost (for those materials purchased from dealers, secondhand stores, etc.)
7. the likelihood of appearance in museum collections with reasonable frequency.

More than 60 different types of paper and textiles were gathered. They were screened for obvious signs of similarity and lack of relevance to museum collections. The remaining 43 fiber types (see tables 1, 2) were subjected to a series of analyses that characterized the material further, thus permitting a more informed selection of the materials to be included in the investigation.

TABLE 1 Paper Samples Initially Characterized

TABLE 2 Textile Samples Initially Characterized

The analyses performed included:

1. spot test for starch size (Browning 1977, 90–91)
2. spot test for gelatin size (Browning 1977, 102–3)
3. spot test for lignin (Browning 1977, 72–73)
4. estimation of average degree of polymerization by viscometry (described in section 4).

Attempts were made initially to test papers for the presence of alum/rosin size through spot tests for rosin (Browning 1977, 78–87). However, the results of procedures currently available were difficult to interpret. Other methods were considered (e.g., atomic absorption, X-ray fluorescence) but were judged to be too time consuming for a simple characterization study.

After study of the characterization data, 23 papers and textiles were chosen for the investigation. The selection was based on the need to obtain a series of fibers that represents a wide range of ages, degrees of degradation, and concentrations of lignin (from barely detectable to heavily lignified). Attempts were also made to include samples that contain sizing material typical of their period (especially important for paper).

Obtaining suitable naturally aged materials was one of the greatest difficulties encountered in collecting materials for the project. In particular, we experienced problems finding authentic pre-Victorian textiles in quantities sufficient for the study. However, it was considered important to make every effort in this direction so that our data can be more confidently applied to the naturally aged artifacts that predominate in most museum collections. It is likely that naturally aged fibers will not be affected by Vikane fumigation in the same way as modern paper and textiles. To give conservators comprehensive guidelines, we needed to study both “old” and “new.”

The second most important variable in selecting material was the degree of degradation. This is a significant point, as the range of degrees of degradation should approximate those found in museum collections so that the results are applicable to museum collections. It is highly likely that the state of oxidation of a fiber, the presence of acidic material, and degree of crystallinity will greatly influence the chemical and physical interaction of sulphuryl fluoride and the fibers. From the point of view of the characterization and screening process, the estimation of average degree of polymerization (as calculated from intrinsic viscosity [η]i); lignin content (degradation tends to increase as lignin content goes up); and visual inspection for color change are most helpful in giving some estimate of the degree of degradation.

The presence of a size in paper and textiles could be a significant factor in determining the effect of sulphuryl fluoride on the fibers. Size affects the porosity of the artifact and thus absorption of the fumigant into the substrate. The type of size may also be a consideration, especially if it reacts with Vikane. The size most likely to react with sulphuryl fluoride is gelatin, a collagenous-based material commonly found in 18th- and 19th-century papers (Hunter 1978). The interaction between proteins and Vikane to form N-fluorosulphonyl derivatives (Meikle 1964) or undefined products (Osbrink et al. 1988) has been documented.

Deacidification buffering salts such as magnesium or calcium carbonate or zinc oxide could also influence the uptake and reactions of sulphuryl fluoride with cellulosic and ligneous fibers. Those buffering chemicals with an alkaline pH may promote the hydrolysis of the fumigant and may affect the quantity of fumigant reaction products that becomes associated with the samples after hydrolysis.

All of the above points concerning the possible effect of proteinaceous sizes and deacidification buffer salts pertain mainly to paper samples. Therefore, it was considered valid to limit experiments that specifically explore these questions to paper samples. A plan was developed to treat one paper, a ledger dated 1897 (paper 12) by an exhaustive washing procedure followed by sizing with 2% gelatin (paper 12B) or deacidification with magnesium bicarbonate at a concentration of 7.6 mg/ml (paper 12C).

These three experimental papers are listed in table 3, along with the 10 other papers chosen for the investigation. The 12 experimental textiles chosen are given in table 4. Altogether, 25 different paper and textile fibers were used in the project. Wheat starch in the form of dry powder as well as films cast from cooked starch was also included at the request of conservators (Weidner 1987).

TABLE 3 Initial Characterization of Paper Samples Selected for Study

TABLE 4 Initial Characterization of Textile Samples Selected for Study

Tables 1–4 contain the characterization data used in the final selection process. The age of the samples vary from 1622 to modern; the DP�s range from below 250 (aged groundwood papers) to more than 1,770; the lignin content is anywhere from zero to extremely lignified (jute textile and groundwood papers); the quantity and type of size varies with the type of processing a fiber has been subjected to, the age of the sample, and its past history.

4 SELECTION OF ANALYTICAL TECHNIQUES

ALL OF the analyses discussed in this section are being carried out on the 25 different fiber samples (tables 3 and 4). Fumigated samples are tested alongside analogous unfumigated (control) material. Suitable quantities of all the fumigated and unfumigated samples are subjected to accelerated thermal aging. Aging is allowed to proceed at 70�C and 50% RH for 8–12 weeks, the length determined individually for each fiber. The factors used in these decisions include:

1. age of sample
2. fiber content, especially the quantity of lignin
3. degree of degradation, as determined by average DP� and visual inspection for color change.

The main goal in the accelerated aging is to stress the samples enough so that significant chemical and/or physical changes result but not to the point at which the rate of degradation has almost ceased. This is a particular problem for many physical testing methods based upon the determination of the strength of fibers. In extreme cases involving naturally aged ligneous material, statistically significant physical strength measurements may be difficult or even impossible to obtain. As fibers reach the micro-crystalline level, the rate of change in molecular and chemical properties, such as DP� or carbonyl level, slows down considerably.

The chemistry of the fumigant and the two substrates, cellulose and lignin, indicates that the analytical procedures should be able to monitor changes in acid content, general deterioration, and fumigant residues. Indicators of degradation of fibrous polymeric materials include polymer length, degree of oxidation, color change, and physical strength. The techniques chosen to investigate all of these parameters are outlined below along with comments regarding anticipated benefits or disadvantages. Information is also given concerning the accelerated aging procedures used in this project.

4.1 DETERMINATION OF ACIDITY AND ALKALINITY

1. Surface pH(TAPPI 1982)A value is obtained for the surface pH of a substrate by putting a drop of water on the sample placing a flat-headed electrode on the wetted area and taking the reading from a pH meter.Benefits: for some samples, changes in surface pH (due to surface adsorption of acidic chemicals) may be differentiated from changes in pH that reflect absorption of acid into the substrate;procedure is relatively quick and easy to do;analysis gives good, relative, semiquantitative values;procedure has been used frequently in scientific projects in the conservation field.Disadvantages: many substrates are too thin or absorbent to give data that can be related to surface effects (especially problematic for textiles);many samples (especially ligneous or degraded papers) have low initial pH, and since pH is a log scale, significant changes may not be observed (similar problems may exist with highly buffered materials);many replicates are necessary to obtain statistically significant results.
2. Cold Extracted pH(TAPPI 1977b)Samples are cut into small pieces; suspended in pure water at room temperature, and incubated for a predetermined length of time. The pH is determined by combination electrode.Benefits: relatively large quantities of the sample are used, which tends to help “even out” inhomogeneities within the sample material;procedure is relatively quick and easy to do;analysis gives good, relative, semiquantitative values;procedure has been used frequently in scientific projects in the conservation field.Disadvantage: many samples (especially ligneous or degraded papers) have low initial pH, and since pH is a log scale, significant changes may not be observed (similar problems may exist with highly buffered materials).
3. Iodometric Total Acid(Nabar and Padmanabhan 1950; Slavik et al. 1967; Achwal and Murali 1985)Samples are pretreated with acid to ensure that all acidic functional groups are in the acid form (as opposed to a salt) and that any alkali present is neutralized. Excess soluble acid is removed by extensive washing. The total intrinsic acid (due to carboxylate and enediol functional groups in cellulose and lignin, as well as any water-insoluble materials associated with sizes, coatings, additives, etc.) is determined by an iodometric back-titration method. Samples are suspended in a solution of iodate/iodide in the presence of a known amount of thiosulphate. During a 48-hour incubation period, the acid in the sample catalyzes the conversion of iodate to iodine and the thiosulphate reacts with the liberated iodine. Residual thiosulphate is titrated with standard iodine. By subtracting the value obtained in the titration from the amount of thiosulphate originally added, the quantity of acid in the sample can be calculated. The reactions involved are as follows:Pretreatment: Incubation: Back titration: Benefits: data obtained are extremely accurate, even at low acid values;incubation is carried out at near neutral pH (note: the enediols are in equilibrium with carbonyl groups, which are sensitive to alkali, and so measurement of these groups is accurate only if the pH is kept in the neutral/acidic range during analysis);esterified carboxylates (i.e., lactones) are hydrolyzed during incubation, and the free carboxylate is liberated; consequently it is possible to obtain extremely accurate values for the total number of acidic groups present in the fiber.Disadvantages: analyses are very time consuming to perform;data do not differentiate among the various acidic functional groups without addition of other analytical procedures (e.g., sodium borohydride reduction of aldehydes and ketones).
4. Determination of Alkaline ReserveAlkaline papers are soaked in a standardized solution of weak acid. During this time the alkali in the fibers is neutralized by the acid in solution. After a 16-hour incubation time (to allow for complete neutralization), the excess acid is measured by a back-titration using neutrality as the endpoint (pH 7.0 measured by combination electrode). The quantity of acid that has reacted with the alkali can be used to determine the original amount of MgCO3 or CaCO3 present in the paper sample.Benefits: procedure is relatively quick and easy to do, and use of a pH meter to determine the titration endpoint eliminates errors associated with the use of color indicators;method gives a good measure of the amount of alkali present (alkaline substances like MgCO3 and CaCO3 are only sparingly soluble in water and cannot be accurately estimated by pH methods as described in (a) or (b));the long incubation time produces a more correct estimation of the total alkaline reserve on the fibers (in comparison to industrial test methods, which use a very short incubation time, ca. 1 hour).Disadvantage: measured value of alkaline reserve will differ from the value quoted by the manufacturer of the paper because different analysis methods are used.

4.2 DETERMINATION OF FIBER DETERIORATION

1. Viscometric Average Degree of Polymerization(Doty and Spurlin 1955)The cellulose is dissolved in a 100% solution of the aqueous solvent cadoxen (Donetzhuber 1960) and diluted to 50% with water, and the intrinsic viscosity is determined using a Canon-Fenske viscometer at 30�C. The calculation of average degree of polymerization (DP�) from the viscosity data is carried out using the following equation: The data obtained are estimates of the average polymer length.Benefits: procedure is extremely sensitive to very small changes in average polymer length;full DP� range of paper and textiles can be accurately determined by this method (note: cadoxen is able to dissolve the higher DP� paper and textiles, which are difficult to solubilize in other cellulose solvents such as cupriethylene diamine (CED) or cuprammonium hydroxide (cuoxam));monitoring DP� changes gives excellent data concerning changes in the general state of degradation of the fiber sample;cadoxen is an odorless, colorless, and relatively stable solvent (in comparison to the other commonly used aqueous cellulose solvents).Disadvantages: procedure is primarily of use in following changes in the cellulose portion of the fiber: lignin is insoluble in the commonly used cellulose solvents, including cadoxen, and interferes greatly with the calculation of intrinsic viscosity of heavily lignified fibers;very large quantities of cadoxen are required for projects of this scope, and the time and cost of synthesizing the amount needed are significant;cadmium oxide, used in the synthesis of cadoxen, is toxic and must be handled with appropriate precautions.
2. Carbonyl Functional Groups(Blair and Cromie 1972, 1977; Ermenlenko and Savastenko 1966)The carbonyl groups arising from ketone and aldehyde functional groups are determined by reaction of the fibers with 2,4-dinitro-phenyl hydrazine. The resulting hydrazone derivative is quantified by colorimetric analysis at 400 nm using two different methods: measurement of surface color by an integrated sphere reflectance spectrophotometerthe hydrazone is dissolved in 100% cadoxen and diluted to 60% with water and the optical density of the solution determined.Since aldehyde and ketone functional groups are important products of oxidative degradation of cellulosic fibers, the resulting data indicate the degree of oxidation of the fiber.Benefits: procedure is a direct method of analysis that gives good stoichiometric data;procedure differentiates oxidation on the surface of a substrate from oxidation that is more evenly dispersed throughout;method may be useful in following chemical changes in lignins, as high concentrations of carbonyls are found in lignified samples.Disadvantages: hydrazine reagent and hydrazone derivatives are suspected carcinogens;procedure is relatively time consuming;surface measurements may be difficult to obtain on very thin paper (e.g., tracing paper) or very thick rough-textured textiles (e.g., some linens).
3. Color Measurement(TAPPI 1972, 1977a, 1981a; Grum 1981)Reflectance measurements are made using a spectrophotometer fitted with an integrated sphere. The values are determined relative to an international white ceramic standard obtained from the National Research Council of Canada (Budde et al. 1982). The wavelengths used are 457 nm (Tappi brightness standard; TAPPI 1977a) and 416 nm (in order that data can be correlated to earlier work carried out in this laboratory). The wavelengths chosen are useful in following bleaching (% reflectance increases) or yellowing (% reflectance decreases) of samples. As fibers degrade, they tend to produce chromophores that cause yellowing or darkening of samples. Bleaching is also indicative of some chemical change taking place in the sample. Although color change may be observed and thus a change in the chemical structure of the fibers implied, a direct link between the quantities of functional groups and color has not been established.Benefits: method is non-destructive to sample;procedure is relatively quick and easy to perform;data relate to a physical property easily understood by non-technical personnel;integrated sphere facilitates accurate measurement of moderately rough materials (e.g., some linens);accurate data can be obtained for fairly inhomogeneous paper providing large areas of paper are covered in the measurements.Disadvantage: measuring samples that are extremely rough (e.g., some jutes) or crumpled is difficult.
4. Physical Strength Testing(TAPPI 1981b; ASTM 1982)Preconditioned samples (ASTM 1979; TAPPI 1970) as individual yarns (textiles) or one in wide strips of paper, are tested for tensile strength and percent elongation (TAPPI 1981b; ASTM 1982) using an Instron “constant rate of elongation” apparatus located in an environmentally controlled area. Tensile strength is the maximum strength of a material subjected to tensile loading. Percent elongation is a measure of sample ductility expressed as gauge length. These properties are related to the physical strength and durability of the sample.Benefits: results give easily understood information concerning the effects caused by the physical stress concurrent with handling and display of artifacts;after the physical measurements are completed, sample material can be used for other analytical procedures (e.g., acid measurements, etc.).Disadvantages: good data are difficult to obtain on very weak naturally aged materials (procedure is not sensitive in lower range of fiber strength, where it is still possible to get good data from many chemical methods of analysis);many replicates are necessary, because naturally aged material is difficult to measure due to inhomogeneities of substrate;data do not predict the behavior of materials subjected to sudden or repeated loading.

4.3 DETERMINATION OF FUMIGANT RESIDUES

1. Gas Chromatography/Mass Spectroscopy (GC/MS)(Hewlett Packard 1986)Air samples are withdrawn from sealed polyethylene bags containing shipping control and fumigated samples (separate) through a packed sampling tube for preconcentration. (Shipping control samples were not fumigated and were shipped along with the fumigated samples so that the only variable was fumigation.) Adsorbed gases are thermally desorbed and passed through a capillary gas chromatograph. A mass selective detector is used to recognize any molecular fragments characteristic of sulphuryl fluoride. The presence of fragments that can be related to the fumigant is a good indication that Vikane residues were present in air taken from the sealed samples.Benefits: procedure detects trace levels of gases in the original air samples;method is specific to the fumigant because ion fragments of any molecular sulphuryl fluoride as well as ion fragments of fumigant decomposition products will be detected (i.e., SO2F2, SO2F, and SOF).Disadvantage: fumigant residues that are adsorbed onto the samples cannot be detected.
2. Fluoride Ion-Specific Electrode(Levi et al. 1986; Speaker 1976)Samples are cut into small pieces suspended in pure water at room temperature buffered to pH 5 to 5.5 by the addition of TISAB and incubated for a predetermined length of time. (TISAB, a “total ionic strength adjustment buffer,” the Fisher Scientific Company (#SO-B-175), provides a constant ionic strength, decomplexes fluoride from iron and aluminum, and adjusts solution pH to between 5.0–5.5) The fluoride content is determined by ion-specific electrode. Since fluoride is a product of the breakdown of sulphuryl fluoride, the quantities of fluoride observed in the aqueous extract indicate the relative amount of fumigant residues present in the samples.Benefits: observed fluoride directly relates to fumigant residues because the natural background levels for fluoride are very low;solutions prepared for the measurement of cold extracted pH can also be used for fluoride determination (with the addition of suitable buffers after the pH measurement is completed);procedure is relatively quick and easy to perform.Disadvantages: fumigant residues that are not in the form of fluoride (F−) are not quantified;data values will be low if all the fluoride is not decomplexed from the sample substrate.

5 SUMMARY

THE VIKANE project is the most complex and broadly based projects ever carried out in the field of paper and textiles by the Canadian Conservation Institute. It involves 25 separate fiber types ranging in age from 1622 to modern. Ten different analytical procedures are being employed in monitoring the effect of sulphuryl fluoride on the long-term chemical and physical stability of cellulosic and ligneous fibers. The project has involved an unusual amount of time and effort to plan and setup. It has also drawn heavily upon our past experience with other scientific projects in cellulose and lignin. The detailed account of the initial stages in the investigations given in this article should be useful to conservators and other research scientists in the following ways:

1. it will provide a valuable reference for understanding and evaluating the data and conclusions that will be presented in the articles at the completion of the project;
2. background information concerning the chemistry of sulphuryl fluoride and its interactions with cellulose and lignin will be useful in defining how the conclusions from this project will relate to the Vikane fumigation of cellulosic and ligneous substrates other than those of paper and textile origin;
3. it outlines how sample materials have been selected to satisfy a) the constraints of the analytical methods used, and b) the need to have our results apply to the broad range of old and new material found in museum collections;
4. descriptions of how the chemistry of the system under study influenced decisions regarding what chemical and physical parameters would be monitored during the project should be of general use to anyone wishing to gain a better understanding of the behavior of cellulosic and ligneous fibers;
5. discussion of the benefits and disadvantages of the different analytical techniques used in this project relates to other scientific studies that make use of these same scientific procedures, irrespective of whether fumigation is involved in the investigation. For example, this research approach has been of vital importance in the planning and execution of another project in our laboratory that involves the short- and long-term effects of deacidification on paper (a joint project with the Canadian Council of Archives).

ACKNOWLEDGEMENTS

THE AUTHORS would like to thank J. C. McCawley, Chief, Conservation Processes Research Division of the Canadian Conservation Institute, for his help and encouragement during both the carrying out of the project and the writing of this paper. They also acknowledge the assistance of Mark Boyle, Assistant Conservation Scientist, Environment and Deterioration Research Services Division, CCI, in carrying out the GC/MS analysis; and Deb Rennie, Assistant Conservation Scientist, Conservation Processes Research Division, CCI, in setting up the fluoride ion-specific electrode analysis. The Canadian Conservation Institute acknowledges the financial assistance of the Getty Conservation Institute in carrying out the Vikane project.

An annotated bibliography of the pertinent literature of the Vikane project may be obtained by writing to the Scientific Department of the Getty Conservation Institute, 4503 Glenco Avenue, Marina del Rey, CA 90292

REFERENCES

Achwal, W. B., and R.Murali. 1985. Characterization of acidic groups in oxycelluloses, III: Effect of cation freeing and blocking on estimation of carboxyl groups and lactones by the iodometric method. Journal of Applied Polymer Science30:4297–4303.

Aparicio, F.J.L., and J.A.L.Sastre. 1969. [title not available]. Anales de Quimica55:191.

ASTM. 1979. Conditioning textiles for testing. ASTM Standard Test Method D 1776–79. In Annual book of ASTM standards. Philadelphia: American Society for Testing and Materials.

ASTM. 1982. Tensile properties of single textile fabrics. ASTM Standard Test Method D 3822–82. In Annual book of ASTM standards, pt. 32. Philadelphia: American Society for Testing and Materials.

Blair, H. S., and R.Cromie. 1972. Spectroscopic investigation of cellulose degraded celluloses, and their phenylhydrazine derivatives in relation to assessment of degradation. Journal of Applied Polymer Science16:3063–71.

Blair, H. S., and R.Cromie. 1977. Spectroscopic investigation of some derivatives of oxycellolose in cadoxen solution. Journal of Applied Chemistry and Biotechnology27:205–13.

Browning, B. L.1977. Analysis of paper. 2d ed.New York: Marcel Dekker.

Budde, W., W.Erb, and J. J.Hsia. 1982. International intercomparison of absolute reflectance scales. Color Research and Application7(1):24–27.

Cady, G. H., and S.Misra. 1975. Reactions of peroxydisulfuryl difluoride to give oxyfluorosulfates of metals in high oxidation states. Inorganic Chemistry13(4):837–41.

Cook, D. M., and D. C.Gustafson. 1978. Process for preparing sulfuryl-fluoride and chlorofluoride products. U. S. Patent #4,102,987.

Derrick, M. R., H. D.Burgess, N. E.Binnie, and M. T.Baker. 1990. Sulfuryl flouride (Vikane): A review of its use as a fumigant. Journal of the American Institute for Conservation29:77–90.

Donetzhuber, A.1960. Zur Herstellung de Cellulose-Losungsmittels Cadoxen. Svensk Papperstidning63(14):447–48.

Doty, P. M., and H. M.Spurlin. 1955. Determination of molecular weight and molecular weight distribution. In High polymers, vol. 5, Cellulose and cellulose derivatives, pt. 4, ed.E.Ott, H. M.Spurlin, and M. W.Graffin. 2d ed.New York: Interscience. 1133–1188.

Ermenlenko, I. N., and G. N.Savastenko. 1966. Microdetermination of carbonyl groups in cellulose materials by means of p-nitrophenylhydrazine using diffuse reflection spectra. Journal of Analytical Chemistry (USSR) 21(1):81–84.

Gray, H. E.1960. Vikane: A new fumigant for control of drywood termites. Pest Control28:43–46.

Green, L., and V.Daniels. 1987. Investigation of the residues formed in the fumigation of museum objects using ethylene oxide. In Recent advances in the conservation and analysis of artifacts,comp. J. Black. London: Summer School Press. 309–13.

Grum, F.1981. Whiteness and fluorescence. American Association of Textile Chemists and Colorists13(8):34–36.

Gustafson, D. C., and D.Skinner. 1976. Purification of gases containing hydrogen fluoride impurity. U. S. Patent #3,996,029.

Hewlett-Packard Company. 1986. NBS library and subsets software.

Hunter, D.1978. Papermaking: The history and technique of an ancient craft. 1943. New York: Dover Books.

Kenaga, E. E.1957. Some biological, chemical and physical properties of sulfuryl fluoride as an insecticidal fumigant. Journal of Economic Entomology50(1):1–6.

Levi, S., L.Zilberman, A.Frumin, and M.Frydman. 1986. Exposure to fluoride in the chemical industry. American Journal of Industrial Medicine9:153–58.

McGiffin, R. F.Jr.1985. A current status report on fumigation in museums and historical agencies. Technical report 4. Nashville: American Association for State and Local History. 1–16.

Meikle, R. W.1964. Fumigant decomposition: The fate of sulfuryl fluoride in wheat flour. Journal of Agricultural and Food Chemistry12(5):464–67.

Meikle, R. W., and D.Stewart. 1962. The residue potential of sulfuryl fluoride, methyl bromide and methane sulfonyl fluoride in structural fumigations. Journal of Agricultural and Food Chemistry10(5):393–97.

Nabar, G. M., and C. V.Padmanabhan. 1950. Studies in oxycellulose. Indian Academy of Sciences Proceedings. 31A:371–80.

Osbrink, W.L.A., R. H.Scheffrahn, and Nan-YaoSu. 1988. Sulfuryl fluoride residues of fumigated foods protected by polyethylene film. Journal of Agricultural and Food Chemistry36:353–55.

Padma, D. K., V. S.Bhat, and A.R.V.Murthy. 1982. Reactions of sulphuryl fluoride, sulphuryl chlorofluoride, and sulphuryl chloride with amines. Journal of Fluorine Chemistry20(4):425–37.

Paschke, F.1922. [title not available]. Cellulose-chemie3:19.

Rassow, B., and P.Zickmann. 1929. [title not available]. Journal f�r Praktische Chemie123(2):214.

Sarkanen, K. V., and C. H.Ludwig, eds.1971. Lignins: Occurrence, formation, structure and reactions. Toronto: Wiley-Interscience.

Scudamore, K. A., and S. G.Heuser. 1971. [title not available]. Pesticide Science2:80–91.

Slavik, I., M.Pasteka, and M.Kucerovs. 1967. Ethylene oxide and its persistent reaction products in wheat flour and other commodities: Residues from fumigation or sterilisation, and effects of processing. Svensk Papperstidning70(7):229–33.

Speaker, J. H.1976. Determination of fluoride by specific ion electrode and report of a fatal case of fluoride poisoning. Journal of Forensic Sciences21(1):121–26.

TAPPI. 1970. Standard conditioning and testing atmospheres for paper, board, pulp handsheets, and related products. TAPPI Standard T 402 os-70.

TAPPI. 1972. Spectral reflectance, transmittance, and color of paper and pulp (using spectrometer with reversed optics): TAPPI Standard—proposed revision and combination of T 442 m-47 and T 216 m-47 as a suggested method. TAPPI55(1):140–44.

TAPPI. 1977a. Brightness of pulp, paper and paperboard (directional reflectance at 457 nm). TAPPI Standard T 452 os-77.

TAPPI. 1977b. Hydrogen ion concentration (pH) of paper extracts—cold extraction method. TAPPI Standard T 509 os-77.

TAPPI. 1981a. Spectral reflectance factor, transmittance, and color of paper and pulp (polychomatic illumination). TAPPI Standard T 442 om-81.

TAPPI. 1981b. Tensile breaking properties of paper and paperboard (using constant rate of elongation apparatus). TAPPI Standard T 494 om-81.

TAPPI. 1982. Surface pH measurement of paper. TAPPI Standard T 529 om-82.

Telysheva, G. M., A. I.Kalnins, and V. N.Sergeeva. 1966. Latvijas P.S.R. Zinatnu Akademijas, Vestis Kimijas Serija (1966) 6:745; Chemical Abstracts 68 (1968): 14206m.

Weidner, M.1987. Personal communication to J. Druzik, Getty Conservation Institute.

Zeronian, S. H.1977. Heat-induced changes in the properties of cotton fibers. In Cellulose chemistry and technology: Proceedings of the American Chemical Society symposium 48, ed.J.Arthur, Jr.New York: American Chemical Society. 189–205.

AUTHOR INFORMATION

HELEN D. BURGESS graduated from the University of Lethbridge, Lethbridge, Alberta, with an honors B.A. degree in chemistry. She went on to obtain a M.Sc. in protein chemistry from the Chemistry Department of the University of British Columbia, Vancouver, and a Masters of Art Conservation from Queen's University, Kingston, Ontario, specializing in conservation science. In 1978 she joined the staff of Conservation Processes Research Division, Canadian Conservation Institute, where she is currently employed as a senior conservation scientist. Address: Canadian Conservation Institute, Department of Communications, 1030 Innes Rd., Ottawa, Ontario, Canada, K1A 0C8.

NANCY E. BINNIE graduated from Carleton University, Ottawa, with an honors B.Sc. in chemistry. She subsequently obtained a M.Sc. degree from Carleton specializing in Raman and fluorescence studies of chlorophyll a monomers and aggregates. In 1987 she joined the staff of the Conservation Processes Research Division, Canadian Conservation Institute, where she is currently working as an assistant conservation scientist. Address: Canadian Conservation Institute, Department of Communications, 1030 Innes Rd., Ottawa, Ontario, Canada, K1A 0C8.

Section Index