September 2000 Volume 22 Number 3

Cyclododecane and Modern Paper: a note on ongoing research

Introduction

Since its introduction to the field of conservation around 1997, there has been an increasing interest in the use of cyclododecane in our field. Cyclododecane (C₁₂ H ₂₄) used as a volatile binding media has great potential as a completely reversible temporary fixative, consolidant, blocking, and/or masking agent. The published results of previous tests and anecdotal information from colleagues has left me uncertain as to its effectiveness in the field of paper conservation.

In order to further investigate this material and its use, and with the hope of understanding when it can be safely and effectively applied, my studio has undertaken a series of simple practical tests. These tests were intended to duplicate situations in treatment where cyclododecane would be used. Since the studio works mostly with modern and contemporary art, our tests concentrated on the use of cyclododecane with modern papers and media.

The following notes are really just an "interim report" which we have put together in order to share with our colleagues what we have learned so far. The study is incomplete and none of the chemical analysis has been done. We are hoping that others studying the use of cyclododecane will let us know what they are doing and suggest improvements, changes, or additions to our project.

The effectiveness of cyclododecane as a fixative has been approached in the past mainly from three angles: the application method; the type of medium; and the characteristics of the paper. Cyclododecane can be applied to paper molten (it has a melting point of approximately 65^°C) or it can be dissolved in a variety of non-polar or aromatic solvents. (The saturation of a concentrated solution will vary depending on the solvent used). Experiments with a wide variety of application methods, reported in recent publications and tested in our project, have led to the conclusion that in general, molten cyclododecane provided better protection than solvent-based application. When solvent based application is used, the slower evaporating solvents result in the formation of a larger crystal with a more open network. Faster evaporating solvents create a smaller crystal and denser network but are more difficult to get to penetrate deeply into the paper.

An uneven application or coverage of the cyclododecane film (as would result from an open latticework of crystals) as well as bad penetration of the cyclododecane into the paper will decrease its effectiveness. This is especially true when used as a fixative for water sensitive media during aqueous treatment.

A literature survey shows that several general observations have been made concerning the role of the type of paper in determining the effectiveness of cyclododecane as a fixative. All authors seem to agree on the fact that cyclododecane protects sized paper better than unsized, and that papers with short, strongly beaten fibers, thick papers, and dimensionally stable papers are less susceptible to penetration of water after fixation.

Several issues of concern regarding the use of cyclododecane on paper have already been raised by conservators in previous studies: cyclododecane could cause changes in the fiber structure during the crystallization process; the difference in response to aqueous treatment of fixed and unfixed areas might result in tensions within the paper; the treated area could cause formation of tide lines; the heat from molten application could cause accelerated aging of the treated area. For the most part these problems are not much different than they would be for any other local treatment.

Most of the studies seem to focus on possible problems of fixing various media. We found no studies investigating the possible interaction of cyclododecane - applied with or without solvents - and modern papers.

The number of undefined fillers, sizings, and brighteners that can be found in modern papers makes them very unpredictable. From experience we know that treatments proven safe and effective for years on a variety of papers have failed miserably on many common modern papers. (BFK Rives and its reaction to certain solvents would be an excellent illustration of this point). Unforeseen reactions with cyclododecane or one of its solvents therefore cannot be ruled out . Possible impurities in the cyclododecane might also have an effect on sizings or fillers.

For these reasons it seemed to us that before we launched into a study of cyclododecane as a fixative we needed first to look at its reaction with various papers. In order to be able to suggest safe treatments involving the application of cyclododecane and to identify areas of possible future problems, a two part research project was designed.

Experimental Outline--Part One

In the first part of the project, possible interactions of cyclododecane or one of its solvents with the paper were investigated. Seven different types of papers were chosen. They included:

• 4 mouldmade papers
  • Rives BFK (smooth surface,100 % cotton, soft sized)
  • Arches Cover (100% cotton fiber, cold pressed,
  • calcium carbonate (?) buffer reserve, internal size)
  • Fabriano Ingres Cover (laid paper, cold pressed)
  • Arches MBM Ingres (75% cotton linter, gelatin (?) tub sized with a buffer reserve)
• 2 machine made papers
  • Canson Mi Tientes (65% cotton, gelatin sized (?), buffered)
  • Stonehenge (100% cotton, buffered w/ calcium carbonate (?), cold pressed)
• 1 hand made paper
  • HMP Handmade Paper (by the Koller's)

These papers were selected by surveying what most frequently came into my studio as well as a number of other studios, by asking several artists for their preferences, and by surveying paper suppliers as to what sold best.

All experiments were executed at a relative humidity of 42% and a temperature of 75^°F.

Two sets of samples were kept as a reference: one set untreated; one set humidified in a Gore-Tex-envelope for 30 minutes previous to a 30 minute float wash, followed by 5 minutes of immersion washing.

Three main application methods of cyclododecane were chosen:

• saturated solution of cyclododecane in isooctane;
• saturated solution of cyclododecane in n-heptane;
• molten cyclododecane

Cyclododecane can be dissolved in non-polar or aromatic solvents. From a literature survey it appears best always to use a saturated solution; the concentration of a saturated solution changes with the solvent selected.

The solvent choice is a determining factor in the size and shape of crystals formed and in the density of the crystal latticework. Solvents such as benzine 100-140, isooctane, and n-heptane form larger crystals but apparently a more homogeneous film. These solutions also appear to penetrate the paper better than solutions in a very volatile solvent such as petroleum ether 30-40. It appears that "coverage" of an area and the penetration into the paper is less effective when the cyclododecane is in a fast evaporating solvent simply because the evaporation rate is too rapid to allow the material to travel through the paper. The crystal formation is therefore on the surface only. The solvents selected for our project were chosen accordingly.

One set of samples was prepared with a saturated cyclododecane/n-heptane solution in the following way. The samples were humidified in a Gore-Tex-envelope for 30 minutes. Each sample was taken out, n-heptane was dropped on the paper with a pipette, then cyclododecane/n-heptane was applied with a brush from recto and verso, and the sample was put back between the Gore-Tex.

Thus, the cyclododecane was applied to the already expanded paper, and the chance of tensions developing within the paper between wet and hydrophobized areas was minimized. (It was also a more realistic approach since most paper conservators will humidify a work before washing it). N-heptane was applied prior to the cyclododecane/n-heptane to increase penetration of the solution into the paper.

After all samples were prepared in this way, they were float washed in deionized water for 30 minutes and then immersed for 5 minutes. After air-drying, the samples were left exposed to air inside the fume-hood to increase the pace of sublimation of the cyclododecane.

On a second set of samples, cyclododecane was applied molten, following the same procedure as described above. Cyclododecane was melted in a double boiler on a hot plate and was applied with a brush on the humidified samples. After solidification, cyclododecane was driven further into the paper with a heated spatula to ensure total penetration. In this area, the paper appeared translucent and lost its flexibility. After aqueous treatment, the samples were air dried, and the cyclododecane sublimed within a few days.

Results

After sublimation of the cyclododecane, the samples were closely examined in daylight. No changes of the paper texture or tide lines could be observed. However, since the importance of observations under ultraviolet illumination has been stressed by several authors, the samples were also examined in UV light.

On Rives BFK, Arches Cover, Canson, and the handmade paper a dark purple fluorescing tide line developed, accompanied in the latter two papers by a second, bright yellow line inside the purple tide line. The tide lines were visible from recto and verso.

Comparison with samples on which only n-heptane had been applied with a pipette showed that the tide line was not, or at least not entirely, caused by the solvent.

On samples of Fabriano Ingres Cover and Arches MBM, no tide line could be observed in ultraviolet illumination.

Samples prepared with molten cyclododecane showed somewhat different fluorescence patterns under UV. No tide lines could be found on either Rives BFK or Arches Cover. Fabriano Ingres Cover had a very pronounced dark purple tide line, as did Canson, which also showed a bright purple fluorescence within the tide line. The handmade paper exhibited strong tide lines as well.

On the latter (HMP), a pattern of dark purple fluorescing spots was observed in addition to the purple and yellow double tide line. Comparison to samples that were washed without cyclododecane application indicated that this phenomenon was not caused by the aqueous treatment. Hand made papers would by their nature be somewhat uneven so it is not possible for us to interpret this data without further instrumental analysis.

Neither molten nor solvent-combined cyclododecane seemed to cause fluorescing tide lines in Arches MBM.

Conclusions

Several conclusions can be drawn from the previously described experiments.

Since Rives BFK and Arches Cover both seem to respond to a solvent/cyclododecane combination with fluorescing tide lines, yet do not show any changes in ultraviolet light after application of pure cyclododecane, the tide line can perhaps be explained by the assumption that cyclododecane retains n-heptane in the paper. The n-heptane would thus interact with components in the paper for an increased amount of time. (The tide lines appear much stronger than solvent alone produces on these papers, thus the idea that the amount of time they remain in the paper is crucial. This point is to be investigated further so we can better compare solvent alone and solvent with cyclododecane).

Since the presence of solvent does not necessarily seem to be a deciding factor in the formation of tide lines in Canson and HMP, the effect could be caused by interaction of cyclododecane itself with certain components in the paper. Another explanation could be the transport of water-soluble components in the paper during aqueous treatment, resulting in a concentration of these materials at the wet/dry boundary created by the cyclododecane. This phenomenon was described comprehensively by Eusman (see bibliography), who also pointed out the possible future damage caused by this process.

However, Eusman detected tide line formation after local aqueous treatment in almost every paper he tested. This leads to the question why, if tide lines are a result of local aqueous treatment, Arches MBM does not show any tide lines, and other papers show tide lines only after being exposed to a cyclododecane/solvent solution or pure cyclododecane. These points will be the subject of further study.

Experimental Outline--Part Two

In the second part of the project, different media were applied to the papers in order to investigate the effectiveness of cyclododecane as a fixative during aquaeous treatments and to test for possible interactions between it and some modern media. The media were chosen to represent the most common problems encountered during wet treatment in conservation of contemporary art and included following types: modern black ink (Pelikan); modern brown ink (Higgins Calligraphy); red and green stamp ink; red and blue color pencils (D�rer, Staedtler, Derwent, Prismacolor); and copy pencil.

One set of samples was left untreated as a reference, a second set of samples underwent aqueous treatment without fixation as described in the first part of the project.

Before application of cyclododecane in a solvent, the media were tested for solubility first. All color pencils, and the green ink stamp were found to be soluble in both isooctane and n-heptane and were therefore not treated with cyclododecane dissolved in either of these.

On the first set of samples, cyclododecane was applied molten onto the recto of the humidified paper and driven in further with a heated spatula. During the following 30 minute float wash, most of the media bled. It appeared that the migrating components were "trapped" by cyclododecane and caused a more dramatic bleeding effect along lines than in the unfixed samples. In the unfixed samples, lines changed color and became paler because color components were washed away more freely, but there was less discoloration along the lines. This phenomenon was especially pronounced with black and brown ink and red stamp ink on Stonehenge and HMP Handmade paper.

Under UV light, strong dark tide lines could be observed along black and brown ink lines recto and verso on Fabriano Ingres Cover, Canson, and HMP Handmade paper. D�rer and Prisma red color pencils showed a yellow halo along the lines visible recto and verso. Since the same halo appeared on the washed, unfixed samples, they are likely to be caused by water rather than the fixative.

On a second set of samples, molten cyclododecane was applied twice, once as described, but with the second application left on the surface recto and verso rather than driven into the paper. In these samples, cyclododecane provided sufficient protection during float and immersion washing.

However, during the second application of cyclododecane, green and red stamp ink and blue and red color pencil started bleeding. Color components migrated both laterally and through the paper to the verso, changing especially the green stamp ink to a paler, more bluish green. This effect was strongest on Arches MBM and Stonehenge papers.

On a third set of samples, cyclododecane was applied as a saturated solution in n-heptane as described in part I. Since all the media in this set had been tested insoluble in n-heptane, bleeding due to the solvent could be ruled out. During wet treatment, bleeding of almost all media occurred. The only exceptions were brown and black ink on Canson and copy pencil. However, bleeding was much less dramatic than in unfixed samples.

Under ultraviolet light black and brown ink on Rives BFK and Arches Cover showed dark halos at the reverse along the lines. These did not necessarily coincide with the bleeding seen in normal light.

A solution of cyclododecane in isooctane applied as described in part I protected the media less efficiently than cyclododecane in n-heptane. Bleeding of all media on all paper occurred during wet treatment, although the migration of color components was less than in washed, unfixed samples.

Results

Molten cyclododecane protected all tested media on all paper during aqueous treatment when it penetrated the paper and also covered the media recto and verso as a film on the paper surface. However, components of certain media such as stamp ink and color pencil seem to be heat-sensitive and are transported by molten cyclododecane in both lateral and vertical direction during application of molten cyclododecane. These components bleed and sink into the paper.

Conclusion

Since the variables in this test series are limited, it is difficult to predict the efficiency of a certain application method of cyclododecane as a fixative or its reactions with a specific media or paper. Any decision about treatment has to be made according to the object and has to be based on pretesting results.

Molten cyclododecane seems to provide best protection for most media. However, interaction of cyclododecane--molten or in combination with solvents--with the paper or media can occur, and long term damage cannot be ruled out. .

This paper does not include our charts which are at present incomplete. These charts make it fairly clear that often it is the paper which will determine the success of the treatment.

We were struck repeatedly by the fact that the same media applied in the same way, with the same cyclododecane application, gave completely different results, from disastrous bleeding of media to no visible changes in media. (For example two of the color pencils bled very strongly on Stonehenge, not at all on Canson or HMP papers). In other cases such as the brown ink samples, the media itself was very water soluble, but we got no bleeding on any of the paper samples.

These observations are confusing since we had originally assumed that the type of paper (mouldmade, machine made or hand made) would be the determining factor. This assumption appears to be incorrect. While the paper type is clearly a determining factors in the success or failure of the treatment, it is not at all clear what aspects of the paper characteristics are responsible for the resulting effect.

Remaining questions and further research

As stated earlier none of the chemical analysis has been done nor have we done SEM of the paper samples before and after treatment. Looking at the fiber structure before and after cyclododecane application would be important in order to see if the papers are being structurally damaged by the treatment. Gas chromatography of cyclododecane itself (to determine purity) and of treated and untreated samples might give us an indication of changes in paper after treatment and possibly explain some of the more bizarre tide line formations we are seeing.

At present we feel that the factors that determine whether cyclododecane can be used successfully in a particular treatment are not clear, and therefore it cannot be used by us safely.

We are now looking into modified uses of the material in situations such as temporary facings or as a blocking material around a poultice or a local stain removal. We would also like to see if it can be used to replace the methyl cellulose layer in a system of temporary masks as described by A. Dwan.

It appears that once again conservators have come across a material promising to solve major treatment problems and have ended up with a material which can in rare cases help modify a problem.

A more comprehensive well-charted, graphed, and photodocumented article will follow this report. We would be grateful if other conservators who have used or experimented with cyclododecane would share their thoughts and results with us.

BIBLIOGRAPHY

Bandow: Cyclododecan in der Papierrestaurierung, Restauro 5, 1999.

Br�ckle, Thornton, Nichols, Strickler: Cyclododecane: A Technical Note on some Uses in Paper and Object Conservation JAIC 38, 1999.

Choi: unpublished study at the Library of Congress, 1999.

Dwan: Temporary Masks for Aqueous Paper Treatments. The Book and Paper Group Annual, Vol.17,1998.

Eusman: Tideline Formation in Paper Objects, Conservation Resources, 1995.

Hangleiter: Fl�chtige Bindemittel, Kunsttechnologie und Konservierung 9, Heft 2, 1995.

Hangleiter: Erfahrungen mit fl�chtigen Bindemitteln, Restauro 5, 1998

Hangleiter: Erfahrungen mit fl�chtigen Bindemitteln, Restauro 7, 1998

Hiby: Das fl�chtige Bindemittel Cyclododecan, Restauro 2, 1997

Hiby: Cyclododecan als tempor�re Transportsicherung, Restauro 5, 1999

Wimmer, Haberditzl: Neue Fixierverfahren im Praxistest, Restauro 7, 1999