Current developments in the preservation of archaeological wet wood with melamine/amino resins at the Romisch-Germanisches Zentralmuseum
See this text in
A. Haas and H. Müller-Beck of the Bern Historical Museum developed a method to preserve archaeological wet wood with water-dilutable melamine resin over forty years ago. It became known as the » Arigal C method« 1. When the production of Arigal-C was stopped in the seventies, the manufacturer2 offered comparable melamine resins under the trading name » Lyofix« that have been successfully used by us, among others3. At the beginning of the nineties, A. Kremer (Roman-Germanic Central Museum - RGZM) was instrumental in establishing contacts with BASF scientists 4 who, since then, have helped us in every possible way to optimise the preservation of archaeological wet wood with amino resins. The name of the melamine resin currently in use at our Institute is » Kauraminâ impregnating resin 800« 5.
The melamine resins that were first introduced in Germany in 1936 are used in a broad range of industries (e.g. wood, furniture, fabrics, or as leather auxiliaries and soil conditioners). The resins are well adapted to the preservation of archaeological wet wood due to their dilutability in water, their low viscosity (150-200 m Pa.s), the small size of their molecules
(1 melamine molecule = 5 Ångström) and thus, their good penetration behaviour, among their most important properties. Hardening, or more accurately polycondensation, is the three-dimensional irreversible crosslinkage of the resin. It is sped up by acid, i.e. a low pH-value and increased temperature. The resin dry content is approximately 70 %, its active ingredient part 90 %6. Kauramin 800 differs from the Arigal and Lyofix types among other things through the complete alcohol etherification which slows down the increase in viscosity and prolongs the life of a soaking bath up to 14 months. In practical terms, this means that the soaking baths no longer turn bad and that spontaneous resin deposits on the wood surface no longer occur, as was to be feared with Lyofix. Usual amino resins become from brittle to hard when they harden. Kauramin 800 has been modified with ethylene glycol to give the resin more elasticity. As a result, the preserved wood pieces show less transverse cracks due to brittleness than was the case with the Lyofix preservation.
3. Preservation procedure at the RGZM
The following preservation procedure has given good results at the Roman-Germanic Central Museum in the course of over 35 years of experience in the use of melamine resins in the treatment of wet wood.
The object is put into deionized water after the wood surface has been mechanically cleaned, in order to remove organic and inorganic acid rests (mostly in the case of oak and sewage wood). This is necessary because Kauramin 800 is a melamine resin that hardens in acid;
impurities in a soaking bath (e.g. humic acid rests in wood) reduce its life and trigger the hardening process too early. The vats are covered with black film in order to protect the cleansing bath from light and air. This simple but efficient method, coupled with regular renewal of the deionized water, renders an addition of biocides superfluous. The cleansing process is monitored a) by measurement of the pH-value (the cleansing bath should tend as much as possible to a neutral pH-value) and b) by measurement of the conductivity (the cleansing process should be carried on as long as soil salts can be extracted). The duration of the cleansing process depends on many factors: wood type, wood size, deterioration, composition of the ground, duration of storage in the ground, relation between wood volume and water quantity and frequency of water renewal. 1 - 5 months are usually necessary.
The subsequent treatment with melamine resin takes place at room temperature (heatable vessels are not necessary). Every watertight vessel can be used. It is possible to build plywood vats corresponding exactly to the treated object and line it with a watertight sheet. This reduces solution consumption in the case of large objects.
Kauramin 800 is supplied with a 90 % active ingredient part and is diluted at room temperature with deionized water to obtain a 25 % concentration. Tests carried out at the beginning of the nineties using solutions with concentrations of 10%, 15%, 25%, 35%, 45% and 90% have shown that concentrations over 20 % give good results if used in combination with careful drying procedures (see under 4. Drying).
Approximately 0.5 % triethanolamine (in relationship to the active ingredient part) are added to the solution. Triethanolamine is a base used to neutralise the acids that are present or formed in the preserving bath without chemically reacting with the solution, thus lengthening solution life to approximately 12 - 14 months. Additions should be no more than 1 % TEA, otherwise hardening might be thoroughly prevented. In the case of small objects with accordingly short soaking times, it is also possible to work without any triethanolamine addition. Solution life is then reduced to about 2 - 4 months.
Approximately 5 % carbamide are also added to lower the solution viscosity and thereby improve the penetration power of the synthetic resin solution. Carbamide binds free formaldehyde (health protection) and builds a urea-formaldehyde resin giving additional strength.
The wooden object ist placed in the preservative solution and carefully covered with a film. This prevents water and formaldehyde evaporation and reduces oxygen supply. Oxygen reacts with formaldehyde to formic acid; it functions as a catalyst, reducing the life of the preservative solution.
Because prepolymeric melamine resins have a small molecule size (of 5 - 30 Ångström) and the preservative baths a low viscosity, the duration of this preservation process is short in comparison to other preservation methods, ranging from a few weeks for smaller objects to approximately one year for very large ones.
The initial viscosity of Kauramin 800 (150-200 m Pa.s) is reduced by a factor of approximately 20 to approx. 10-11 m Pa.s/20° ‚ø C through water dilution and carbamide addition. The diffusion rate is accordingly high and can only incidentally be increased by a temperature rise.
Viscosity 7 in relation to temperature:
|temperature in ° ‚ø C||viscosity in m Pa.s|
The good diffusion properties and penetration behaviour can be established and proved with the nitrogen estimation method. Core samples of preserved woods (or woods placed in the preservative solution) are dried for 15 hours at 105° ‚ø C. The wood bulk and hardened resin are what remains after the drying process. The nitrogen content is then determined by elementary analysis and the resin concentration calculated (see footnote 8). We have chosen to illustrate this process with a Roman ship found in the Danube near Oberstimm in Bavaria9. Here is some information:
Sample A from rib 4 (6 cm thick oak)
|Drilling depth||N g/100g|
|0-10 mm||12.3 g|
|10-20 mm||10.8 g|
|20-30 mm||4. g|
Sample B from plank 1 (4 cm thick pine wood)
|Drilling depth||N g/100g|
|0-10 mm||13.8 g|
|10-20 mm||13.3 g|
Sample C from keel ( 8 cm thick oak)
|Drilling depth||N g/100g|
|0-10 mm||13.1 g|
|10-20 mm||12.9 g|
|20-30 mm||12.2 g|
|30-40 mm||12.0 g|
The unpreserved wood contains 1 to 2 % nitrogen that has to be deducted from each result. If we take sample A 0-10 mm as an example, this means: 12.3 minus 2 = 10.3 % N. Pure Kauramin 800 contains 40 % nitrogen when it is completely hardened (i.e. 10.3 : 0.4 = 25.75). 12.3 % N corresponds theoretically to a resin concentration of 25.75 %. The nitrogen present in carbamide and triethanolamine explains these results that are superior to the initial 25 % concentration. After 3 further months impregnation time, core samples were again taken from the same wood pieces but in other parts places (this explains the differing nitrogen concentrations in e.g. samples B/B2).
Sample A2 from rib 4 (6 cm thick oak)
|Drilling depth||N g/100g|
|0-10 mm||13.2 g|
|10-20 mm||13.3 g|
|20-30 mm||5.4 g|
Sample B2 from plank 1 (4 cm thick pine wood)
|Drilling depth||N g/100g|
|0-10 mm||13.5 g|
|10-20 mm||12.8 g|
Sample C2 from keel ( 8 cm thick oak)
|Drilling depth||N g/100g|
|0-10 mm||13.8 g|
|10-20 mm||13.3 g|
|20-30 mm||13.0 g|
|30-40 mm||12.2 g|
The final evaluation shows that impregnation with melamine resin is quick and uniform. The strikingly low nitrogen concentration of rib 4 (samples A and A2) at a depth of 20 - 30 mm is possibly due to the fact that the ribs are made out of warped branch timber.
d). Monitoring the preserving solution
A solution sample is taken on a weekly basis to determine the pH-value that slowly decreases from its original value of 8 - 9. A few drops of the soaking solution are also put into a test tube with some clear water each week. When the mix no longer appears clear and transparent but looks milky and turby (at a pH-value of approx. 6 to 7), the resin is about to harden. Detecting this » ‚point of turbidity« ‚@ is an excellent method of diagnosing inceptive polycondensation at an early stage, a long time before the preservative solution becomes turbid as a whole and before undesirable resin deposits could occur on the wood surface as a result of turbidity. Now is the time to change the preservative solution or to stop preservation and remove the object from the bath.
e). Preservation end and hardening of resin
When the object is completely impregnated but the bath remains alkaline and the point of turbidity has not yet been reached, one can either wait for a further decline of the pH-value or add a weak acid or an acidifier like glycerol diacetate (1 to 5 %). Glycerol diacetate slowly decomposes to acetic acid, glycerine and water, pulling down the pH-value and therefore accelerating resin crosslinkage.
The wood is taken out of the bath once the turbidity point has been reached, is rinsed under running water, carefully packed in wet cellulose, then packed in film (e.g. in a grip-bag or a tubular film) and treated at 50° ‚ø C in a warming cupboard. The purpose of the cellulose is to absorb the synthetic resin that might ooze out of the wood during cupboard treatment, thus preventing resin deposits on the wood surface (caution! cellulose should not dry out during the heat treatment as it is difficult to remove from wood surface when dry!). A sample of the preservative bath is also put into the warming cupboard to indicate when the hardening process is completed. At 50° ‚ø C, this process normally lasts 7 to 14 days (depending on the pH-value), sometimes longer. A higher temperature accelerates the hardening process but makes the use of heat-resistant films necessary. Temperatures below 50° ‚ø C prolong hardening times. The wood also hardens at room temperature at an accordingly slower pace (this is useful in the case of big objects that cannot be put into any warming cupboard).
After the resin has hardened, the wood is still saturated with water and must be dried. To obtain good preservation results (without cracks, deformation or shrinkage), a careful drying process is decisive in the case of wet wood that has been treated with melamine resin in order. This shows that the concentration of the preservative solution and soaking time are not the only factors having an influence on the result; drying is even more important (see picture 1).
The easiest drying method that has proved successful over the decades is the very slow drying of wood in PE-film permeable to steam. The wood is tightly wrapped in a PE film (approx. 10 m ) or put into grip-bags. It is unwrapped from time to time to remove the condensed water that has formed. The water diffuses through the film in steam form. The drawback of this method is its long drying times (weeks, months, for smaller objects, or even years for large ones), which is set off by the excellent results obtained. Wood shrinkage or warp are thoroughly avoided11, formation of hairline cracks transversal to the fibre can be limited to a minimum or also completely avoided. Numerous tests were carried out at the beginning of the nineties compensate for this drawback and accelerate the drying process. 150 core samples with a wet weight between 0,5 and 6 kg were taken from strongly deteriorated oak (U max. 650-850), preserved in a solution of melamine resin at a concentration of 25 %, hardened and dried using the following methods:
The samples that were freeze-dried did not show any warps. About 10% of the samples showed a shrinkage in their lengths of up to 1% and a volume reduction of up to 3 %. The form of the other 90% remained unchanged. A few but noticeable transversal cracks appeared at -5° ‚ø C and 4 mbar. At -40° ‚ø C and 0.12 mbar, the number of transversal cracks increased but these were small and hardly noticeable. The duration of treatment was twice as long as at 4 mbar. Wood pieces that were microwave-dried showed less transversal cracks but approximately 5 % of the samples shrank in their lengths by about 2% and showed a volume reduction of up to 4 %. These wood samples were warped by the quick drying (alterations always took place in the critical phase of the drying process above fibre saturation, when humidity reaches about 40 %). In the overall evaluation of both drying methods, the microwave-dried wood samples performed somewhat better than the freeze-dried ones (see picture 2).
Other positive aspects of microwave-drying are:
We combine the advantages of the microwave-drying method and those of the in-film drying process to avoid warps and transversal cracks and still shorten the duration of the drying process. The wooden objects are pre-dried in a microwave oven at 30-40° ‚ø C (the wood must be stored in such a way that air reaches the surface from all sides; if one fails to do so, the upper side dries faster than the bottom side, tensions arise and the wood is deformed) until wood humidity is just above fibre saturation (40-50 %). The object is removed from the oven, carefully wrapped in film and the drying process slowly carried on until wood humidity reaches a level of about 15 %.
5. Subsequent treatment
The surface of treated and dried woods is light-coloured, comparable to bleached driftwood, sometimes soft (like balsa) and calcareous. A subsequent treatment improves colour and better reveals the wood structure. In addition, it strengthens the surface and protects the (very absorbent) wood against humidity variations. One can use waxes or synthetic resins12 that are solvent-dilutable or even drying oils13.
6. Properties of preserved wood
Preserved and dried wooden objects are protected against infestation with micro-organisms on a long-term basis through the presence of a small amount of formaldehyde, they reveal the finest details on their surfaces and show traces of workmanship and wear. Annual rings can be very clearly seen, making dendro-chronological analyses possible even after preservation. The light weight of wood preserved with melamine resins makes the setting-up and display of large constructs such as ships easy. It is possible to work with glue and make additions, using all the products suitable for dry, absorbent surfaces. Preservation of archaeological wood showing organic or inorganic rests like pitch, leather, fabrics, non- water soluble inks, iron or stone is unproblematic14 (see pictures 3 and 4). Copper base alloys lose their patina in the slightly basic solution and must be removed or carefully covered.
Disadvantages of preservation with melamine resins: the molecules of the melamine resins are duroplastic after polycondensation, i.e. they are irreversibly crosslinked. Only a careful, time-consuming drying process, starting from the moment fibre saturation is reached, can prevent the apparition of cracks (quick drying provokes the occurrence of cracks and volume alterations).
7. Putting back archaeological wet wood into form during preservation with melamine resins
When the need arises to do so, it is best to put back into form archaeological wet wood awaiting treatment with Kauramin 800 after it has been cleansed but before it is soaked in the preservation bath. The objects are pressed back into their original forms using water-resistant templates, and are preserved and dried with these. Corrections are also possible after drying the object (e.g. using steam).
8. Protection of health
Working with MF resins is subject to compliance with the (German) decree on handling dangerous products, as restorers normally do. In its concentrated form, Kauramin 800 contains about 1.5 % free formaldehyde. It is possible, not only to respect but even to lie much under the limits set by the German professional associations15 (max. concentration at the working place = 0.5 ml/m³ ‚ ) thanks to carbamide addition (binds free formaldehyde), to dilution to 25 % and to the careful covering of the preservation vats.
The highest formaldehyde emissions occur during heat treatment. It is therefore recommended not to remove the objects at the end of the treatment, while they are still warm, but to let them cool down in the warming cupboard after switching it off.
- 7 -
9. Waste disposal
The discharge of sewage originating from the treatment with condensation resin into a public sewage system is ruled by local by-laws that give information on the permitted nature of the said sewage. Standard values for sewage16 in a public sewage system:
Meeting these values is not a problem. After use, the preservation baths contain formaldehyde in concentrations that are easily decomposable by bacteria present in the sewage system and sewage treatment plants.
Interdisciplinary co-operation and the help of many have been necessary to improve the method of treatment in a number of ways and rationalise the preservation stages. I am particularly indebted to my friend and colleague Albert Kremer, Dr. Günther Matthias, Dr. Eberhard Pfütze, Dr. Walter Pitteroff and Dr. Wolfgang Eisele who have given us advice and help with so much unselfishness, generosity and friendliness.
Current developments in the preservation of archaeological wet wood with melamine resin are described here, using Kauramin impregnating resin 800. The work was carried out at the Roman-Germanic Central Museum.
The recipe of the preservative solution has been altered in order to improve life and viscosity of the impregnating solution. Nitrogen estimations have established the levels of resin concentration inside the wood. The drying process can be accelerated through microwave treatment at temperatures of approximately 40° ‚ø C until fibre saturation is reached, thus reducing the duration of the drying process. Treated wood has a stable form and is light in colour. The finest details on the surface remain unaltered.
This strongly deteriorated piece of oak heart dated AD150 was soaked for 9 months in a preservative bath at a concentration of 25 % (nitrogen estimations confirmed that synthetic resin distribution inside the wood sample was homogeneous). After treatment in a warming cupboard, the wood was cut in the middle and the halves were dried in different ways. Left half: was dried at room temperature without using a PE film; shows extreme longitudinal and transversal cracks as well as a strong volume alteration. Right half: was dried in a PE film at room temperature. No shrinkage, no volume alteration, two small cracks transversal to fibre. Scale = 1 : 4.
A late Roman piece of plank (oak) that was preserved in a synthetic resin solution with a concentration of 25 %. After treatment in a warming cupboard, the wood was cut in the middle and the halves were dried in different ways. Left half: was microwave-dried for three hours at 40° ‚ø C. The wood form did not change, no transversal cracks (the dot-like marks originate from several humidity measurements with electrodes). Right half: was dried for five days at 0.12 mbar and -40° ‚ø C. The wood form did not change but a lot of cracks appeared. Scale = 1 : 2.
Detail of a late Roman plank fragment (oak) that was preserved together with iron and textile rests (impregnating solution with a concentration of 25 %). Scale = approx. 1 : 1.
A Roman tablet (fir) showing writing in ferro-gallic ink and a branded seal. It was preserved with an impregnating solution at a concentration of 45 % (without triethaloamine). The writing is easily readable after preservation through the light-dark contrast. Scale = 1 : 1.