DRV-Conference Berlin 2000
Temporary protection of sensitive surfaces
About the usage of volatile binding agents
Hans Michael Hangleiter
More than 5 years passed since I have developed, together with Mr. and Mrs. Jäger, the first procedures for the usage of volatile binding agents in 1995. (1) We already described various possibilities for the usage of volatile binding agents in our first publication in 1995. Much to my delight some colleagues have also taken up this topic as a result of which some very important and interesting articles were published. (2)
After more than 5 years of further experience, it is the goal of this lecture to inform about this experience with volatile binding agents and, at the same time, to point out the dangers involved.
Physical and chemical characteristics
Of the available materials cyclododecane (CCD) surely is the most important offering the widest range of applicability. In addition, its chemical features are nearly no risk when applied to works of art. Cyclododecane is a saturated, alicyclical hydrocarbon. It is one of the most stable representatives of this family, and its capability of chemical reaction inertia is comparable with the saturated open chained hydrocarbons.
The wax-like material is meltable at 60°, a temperature which is acceptable for many works of art, and it regains it firmness while cooling off.
It is soluble in non-polar solvents and it also reverts to its solid form after the evaporation of the used solvent. In water and most of the polar solvents it is not or almost not soluble.
It features a relatively high vapor pressure, so it can dissolve directly from the solid into the gaseous phase. This means, it sublimates at room temperature.
Even more than its chemical and physical key data the question of density and crystal structure, the most important features of a CCD-film, is one of central importance for the success of any application.
The insolubility in polar solvents like water, ethanol, isopropanol or acetone is of central importance for the application in the field of restoring. It is only through this characteristic that a volatile sealing or adding of hydrophobic features as a protection against water or other polar solvents becomes possible. This protection is not reached because of the insolubility o cyclododecane in polar systems alone. The characteristics of the film are decisive. They are differing so much, that although the thickness of the film is the same, polar solvents may penetrate without any problems or may not be able to do so at all. The formation of the film and its characteristics are the most critical factors for the production of the film.
These characteristics are determined exclusively by the form of the application.
Formation of the film and its characteristics
The way cyclododecane forms a film is best compared with a crystallization. Basically it does not form an absolutely homogeneous film, neither when solidifying from the melt nor when precipitating from a solvent. It forms needle-like crystals, positioned close or less close to each other. As the density of the films very often decides whether or not the procedure will be successful, the conditions during the crystallization are very important.
Key rule for crystallization from a melt
The slower the temperature sinks from above the melting point to a temperature below it, the more distinctive the felt formed by the crystals will be. A quick cooling off on the other hand will result in significantly more homogeneous and denser films.
SLIDE 01 - Crystallization of the melt
The slide of a crystallization of a 200 litre drum of melt cooling down to room temperature shows this quite nicely. The piece in slide 1 shows the contact area to the drum casing. Here an about 1,5 cm dense film is formed, behind which there is an area of differently sized crystals, only loosely fitted into each other.
In practice this slow cooling off and forming of a crystal felt from a pure melt only takes place when casting big pieces. Films of less than 3mm will cool of that fast at room temperature, that a crystallization is not possible.
Now I want to go into the problem of what films are actually possible and how the material is being used in practice.
Films from a melt
Having a melting point of about 60°C does not mean that you can apply the melt at this temperature onto any surface using a brush. Instructions proposing 65°C are totally useless in practice. Should you try this with a brush, the material will harden immediately, forming incredible lumps. Evidently it has to be used at a much higher temperature. I am using a water bath at water temperatures between 90° and 100°C.
However, using a brush to apply the pure, unthinned melt should be avoided in practice. An exception to this are only totally impermeable and absolutely non-absorbent surfaces. The high melting point and the fast cooling down of the material very often leads to unsatisfying results.
For small, local consolidations or seals with CCD-melt I use a CCD-coated foil as a transfer paper. For this purpose a hostaphan (polyester) foil is coated using a spray can and cut into strips of usable length. Using a heatable spatula you can then apply it where needed. This method proved to be especially effective for very small areas of a painting.
Films from a melt with added solvent
One gets really good results with solvent added. On a non porous ground like glass a felt of crystall needles is formed. On every porous ground the coating becomes more compact.
As an explanation the followings seems sufficient
When reaching the consolidating temperature two processes are taking place.
The fluid CCD, being homogeneous at a higher temperature, disintegrates into two phases. A relatively temperature-independent fluid phase consisting of a saturated solution of cyclododecane in the solvent and a temperature-dependent phase of pure cyclododecan. The presence of the fluid phase on a non-absorbent surface enables the pure cyclododecane to form crystal needles during solidification.
This is quite different on an absorbing surface. As soon as the separation into two phases starts with sinking room temperature, the fluid phase is being absorbed by the surface and separated from the solidifying cyclododecan. The forming of the film is similar to the pure melt. A very dense film is being formed. (This phenomenon is being researched more thoroughly at the moment)
To be able to apply a melt by using a brush in a practicable manner at all, a certain amount of a solvent should be added. By doing this you lower the melting point, at the same time you are avoiding a premature solidification of the material on the brush. From experience I can say that the hotter the mixture is being applied, the better the characteristics of the film will be.
The boiling point of the solvent seems to have hardly any influence on the characteristics of the film. However, only the melting ranges (of petroleum spirits) from 60-80°C as well as 100-140°C have been checked. Solvents with a lower melting range evaporate too fast from the hot melt, a point against a higher melting range is the long time of waiting after the application.
The following recipes proved themselves to be very successful
On well absorbing surfaces adding of
10% petrol ether 60-80°C
or
10% petrol ether 100-140°C
to the melt of cyclododecane.
Films from a solution of CCD
For non absorbing surfaces the key rule for the crystallization also applies
The slower the temperature sinks from above the melting point to a temperature below it, the more distinctive the felt formed by the crystal needles will be. A fast transition on the other hand leads to rather more homogeneous and denser films.
This leads to select a solvent that is highly volatile and has a low melting point in order to end up with films as dense as possible.
This key rule does not apply to porous, absorbing surfaces.
Solutions of cyclododecane in petroleum ethers (3) with melting ranges between 40°C and 140°C result in dense films without any significant differences in their characteristics. There are, however, significant differences as to their penetration performance. Generally these solutions penetrate the deeper, the slower the solvent evaporates. The waiting time for the evaporation needs to be extended accordingly. The production of solvents is quite easy. At 20°C 40 parts of CCD dissolve into 60 parts of petrol ether. Both components can be kept in a container for 2 to 3 days without stirring. (A PE-bucket with a lid will do). Of course, for a couple of days a magnetic mixer can be used, too. If, on the other hand, you need to produce a ready to use solution rather fast, you melt CCD in a water bath. The solvent ( in this case only materials with a melting point above 60°C) is heated in the water bath to a minimum of 60°C and mixed with the melt.
Films from a spray can
In contrast to the applications described so far, melt and solution, the usage of cyclododecane as a spray represents a third form. Inside the spray can cyclododecane is present in a solved form. In this case the propellant is the only solvent. Additional solvents are not present.
The propellant, being extremely volatile, also determines the major characteristics of the cyclododecane film.
When producing sprayed films you need to keep some rules, which differ from the standard usage of a spray can. The most important is the distance.
Because the propellant dissolves rather fast from the fluid into the gaseous state, the cyclododecane , dissolved at first, precipitates in a solid state. That part of the gas that is still fluid in the sprayed fog when exiting the spray can gets less and less with growing distance to the nozzle, until only the pure cyclododecane dust is left over.
To end up with a film that is as dense as possible, the distance between nozzle and object should be as short as possible.
To get a film that does not rub off a distance of 3 to 4 cm is recommended.
You will get a soft but very even film with a distance between 6 and 10cm.
Larger distances will lead to films that come off easily. They also lead to high loss.
To be able to describe the protective, consolidating or sealing characteristics of the just described films in more details, we examined the sealing feature of various CCD-films in contrast to polar fluids in more detail.
Cyclododecane was being tested alternatively as a melt or as a solution on absorbing surfaces, on recycled paper and on plaster.
As agitating fluids we used water, water with a wetting agent as well as ethanol. As a marker eosine was added to act as a colouring for the polar fluids. It dissolves very well in water and in spirit, but it does not colour cyclododecane. At the same time I wanted to check on the connection between the volatility of the solvent and the thickness of the film. To achieve this, chemically similar solvents with different melting ranges were selected.
SLIDE 02 - samples with CCD in a solvent
This slide shows the samples with CCD in a solvent.
CCD from a spray can was added to this group. The solvent is always a saturated one. The solvent used here was petroleum spirit.
The left half on a green background shows samples on recycled paper, those on blue background are on plaster.
The left column on green background shows the front page of the paper sample, the right column shows the same samples from the back.
The left column on blue background shows the front page of the plaster sample, the right column shows the same samples in a longitudinal section.
As a result the following can be generally said: Alcohol, as well as water thinned with an interlacing agent, is capable of penetrating a film which was produced from a solution. The samples on plaster led to the same results.
In a more detailed analysis the films on the paper samples showed different reactions to alcohol and water thinned with a wetting agent. The film does not seem to prevent alcohol from penetrating it. Water thinned with a wetting agent at least penetrates the paper that far, that a slight soaking with a corresponding distortion can be noticed. The small red dots on the back of the paper show that the fluid has gone all the way through. The results on the samples on plaster do show that spirit as well as water thinned with a wetting agent penetrated the film almost without any problem at all. The longitudinal section of water thinned with a wetting agent shows this even more clearly than on paper.
The reaction opposed to pure water ought to be watched with special attention. At first the samples on paper show good protective effects. Never the less we can notice minor distortions of the paper on the samples with petrol ether. On the other hand the sample with a film from the spray can does not show any distortions. The samples on plaster do highlight this result again. The small distortions have already hinted at this. After a drop of water had been on the surface of the plaster for about 2 hours, the water had deeply penetrated the plaster in three areas. Only the film from the spray can had been able to resist without any problems.
There is no problem to stop water from penetrating by adding a thickener . (0,5% of tylose 30.000 are already sufficient).
SLIDE 03 - samples with CCD as a melt
This slide shows the samples with CCD as a melt
Here the pure melt was compared to melts containing a 10% solvent as an additive. The influence of the melting range on the thickness of the film was to be examined, too. As a solvent petrol ether with a melting range above that of cyclododecane was used. One at 60-80°C and another at 100-140°C. Like before, the left half on green background shows samples on recycled paper, those on blue background show samples on plaster .
The results can be summarized as follows: All three films resisted the penetration without problems. None of the three solutions could penetrate into the material. Likewise no difference between the two used types of spirit was found.
Another important result to be noted is, that on porous surfaces no difference in the characteristics of the film was found in relation to different melting points of the solvent.
However, the time needed for the solvent to dry must not be shortened.
The effect of not waiting long enough can be shown by comparing two samples of the same material.
 |
|
SLIDE 04 - Sample after 2 hours of drying |
SLIDE 05 - Sample after 15 minutes of drying |
Slide 4 shows the back of the sample already presented in the summary: CCD in petrol ether 100-140.
The next slide shows the back of a sample of CCD coated in the same way with petrol ether 100-140.
Slide 4 was taken after 2 hours of drying, the next after only 15 minutes of drying. The importance of a sufficient drying period is shown quite clearly in slide 5 where a mixture of water and a wetting agent penetrated nearly without any problems at all.
Of special interest for the application of CCD are of course the evaporational features, i.e. its sublimational characteristics.
Basically the evaporation depends on two factors
1) Temperature
2) Ventilation
SLIDE 06 - Characteristics of evaporation
The dependency of the speed of the sublimation on the temperature is not linear, i.e. not simply proportional to the rising temperature.
To be able to compare three equally big hostaphan foils were coated with CCD-spray. One each was exposed to air at 10°C, 20°C and 30°C. The material was weighed in one-hour intervals and the loss of material was checked. After two hours the loss of material was twice as high at 30°C compared to the sample at 10°C. The 20°C sample was only 1,75-times as high. The difference after 9 hours was quite blatant. Here the loss of material of the 30°C -sample was 10-times higher compared to the 10° sample. The 20°C was only just twice as much.
Right now we are conducting even more detailed tests on this topic, the results of which will be made available to all colleagues via my homepage.
The most important statement is already evident by now. It is not only the increasing of the temperature leading to a much faster sublimation, the increasing of the temperature should take place in the decisive temperature range.
I believe that realizing this is one of the significant aspects of applying volatile binding media.
The second factor influencing the speed of the sublimation is the ventilation. Of course it is not enough to take into account the airflow only, you also have to consider the geometry of the surface, which again influences the airflow.
 |
|
SLIDE 07 - Briquettes of brown coal, coated with CCD-spray |
SLIDE 08 - Evaporation process on these briquettes |
Both slides show a simple setup for testing. One pack of brown coal briquettes was coated with CCD-spray. The process of sublimation was watched in a closed room without any noticeable draft. It clearly shows the difference in sublimation, caused only by the flow of the air across structured surfaces.
When we talk about the characteristics of evaporation and sublimation, we also talk about the characteristics in different structures or textures. We talked about porous structures in detail and so far we only addressed temperature and ventilation. We actually do know a lot about the CCD-vapor permeability of lime mortar. At this point I have to mention the outstanding article of Nicole Riedl. (4) Mrs Riedl has studied the sublimation process in the pores of lime mortar and is able to show this with fascinating pictures, taken on a cryo-rem. Mrs. Riedl clearly shows that the sublimation process also takes place deep in the structures of the pores. However, significantly slower than on the surface. In addition she quite clearly shows that the rate of the sublimation depends on the density of the film.
The question of CCD-vapor permeability of denser materials like oil- or resin-bonded films of paint is very important for panel paintings and sculptures. Only if you have answered the questions of how to safeguard the transport of paintings or sculptures, and especially the volatile consolidation of layers of paint during straining jobs involving a scalpel, only then can you carry out these jobs and still have a clear conscience about it. You don't have this problem when the painting is on canvas, because the backside doesn't normally pose a problem to speak of for CCD-vapor. Wooden panels or sculptures are much more of a problem. You have to visualize that at first the CCD penetrates the craquelée in a fluid form, but then it only has a very small surface left for vaporization.
To get an idea myself I invented the following setup: An oil-painting on canvas was to act as a cover for a CCD-film. The idea was to watch if CCD sublimes under this cover and if, then as a gas, it penetrates the resin-bond films.
SLIDE 09 – Painting as a test area
SLIDE 10 - The CCD-film lies between the
hostaphon foil and the paintings
SLIDE 11 - Quartz sand as a weight guarantees
no evaporization to the sides
SLIDE 12 - Details after the coating
On a hostapan foil measuring 29,5 x 21cm an area of about 12 x 13,5 cm was coated with CCD-spray. The CCD-coated side of the foil was then placed on the surface of the painting, using quartz sand as a weight in order to prevent direct contact and to eliminate evaporation to the sides.
The back of the painting was exposed to air. To accelerate this process a hot air blower was used from below. The air directly underneath the painting was exactly 30°C.
The results did surprise me.
DIA 13-16 - Sublimation progress during
the diffusion through the painting
The weight decreased by 0,1g every three hours. The process began with a craquelée hardly visible in the beginning. At the same time the thickness of the film decreased, being indicated by the growing transparency.
These slides show the sublimation process in a sequence. I don't yet want to conclude from these results, that oil or resin-bonded films are not a vapor barrier for CCD-films. On the contrary, I recommend to carry out object related tests before faring on such tasks.
For me the time until it has totally disappeared very often is a reason not to use CCD to solve a preservation problem. It simply very often is an obstacle hindering the progress of work. So
you very often have the problem of how to speed up this process. Again the thickness of the film plays a major part. The denser the film, the slower the sublimation process. This relation is always true!
This alone is reason enough to carefully plan ahead what you want to achieve and what density the film must have for the desired effect. (Of course it doesn't make any sense to apply a thick melt, staying ten times as long than a film from a spray can, to have a surface protection for a short period of time.)
The following steps speed up the rate of sublimation
Increasing the temperature
As already pointed out simply increasing the temperature is not enough. If possible it should be an increase into a range over 30°C.
Ventilation
If increasing the temperature is accompanied by good ventilation, the speed of the sublimation can be accelerated in a way that you need between half and one tenth of the time.
Solvent
Like always the density of the film also plays the major part when using a solvent. Therefore it is a time-consuming, laborious task, to remove a melt using petrol ether. I can only discourage anyone to do so. You remove a film from a spray can by simply spraying the film with petrol ether.
Footnotes: (0) Deutsche Restauratoren Vereinigung
(1) H.M. Hangleiter - Elisabeth Jägers - Erhard Jägers: Flüchtige Bindemittel, in Zeitschrift für Kunsttechnologie und Konservierung 1995/2 S. 385 - 392
(2) Gudrun Hilby: Das Flüchtige Bindemittel Cyclododekan, in Restauro 1997/2 S.96 - 103
H.M. Hangleiter: Erfahrungen mit flüchtigen Bindemitteln Teil 1, in Restauro 1998/5 S.314 - 319
H.M. Hangleiter: Erfahrungen mit flüchtigen Bindemitteln Teil 2, in Restauro 1998/7 S.468 - 473
Georg Hilbert - Nicole Riedl: Cyclododekan im Putzgefüge, in Restauro 1998/7 S.494 - 499
Cornelia Bandow: Cyclododekan in der Papierrestaurierung, in Restauro 1999/5 S.326 - 330
Gudrun Hilby: Cyclododekan als temporäre Transportsicherung, in Restauro 1999/5 S.358 - 363
(3) Petroleum distillates with a boiling range between 40° and 60°C are petroleum ethers, those with a boiling range above 60°C are petroleum spirits (or petroleum benzines)
(4) Nicole Riedl : (Title?), in Restauro 1998/7, pp.