The treatment development for the wooden components of the jet is guided by a few simple principles. As this is a conservation treatment, our primary goal is to preserve as much original material as possible. The following statements helped to shape the treatment protocols.
Given the requirement to return the Horten to a level of robustness in keeping with the jets original construction, our team made some difficult, but well considered decisions regarding how much of the deteriorated plywood could be saved through consolidation. In some areas, fungal attack of the wooden veneers was such that all that remained were thin, friable adhesive layers, interspersed with dirt and blackened organic matter unrelated to the deteriorated wood. This material was so far removed from the original plywood material in both function and appearance that it was regrettably deemed beyond recovery. In cases of severe deterioration the unrecoverable material was pared back to reveal material suitable for conservation treatment.
It was clear from the outset that our team should investigate wood consolidation treatment options given the level of deterioration of much of the Horten plywood components. Whilst a large body of conservation literature exists relating to conservation of wooden objects, very little has been written about the conservation of plywood. Our literature search into treatment options for plywood produced a paucity of information, but what was discovered about plywood treatments was not applicable to our needs. For example, modern plywood by virtue of its low production costs and widespread use as a construction material is considered by many a 'throwaway' or readily replaceable material. A small number of conservation publications on the material relate to fine furniture and typically describe aesthetic surface treatments. Panel painting conservation treatments typically involve the complete removal of the plywood backing as this material is considered secondary and not historically relevant in comparison to the painted surface. Wooden aircraft repair literature is driven by the requirement to return the aircraft to a serviceable condition and as such the replacement of deteriorated wooden components is mandatory. Our treatment choices therefore developed out of a period of experimental and empirical testing.
The choice of consolidant was driven by the treatment protocol guidelines described above. While offering extremely high levels of strength, the requirement for treatment to be reversible ruled out the use of thermosetting epoxy and PMMA resins. Our team considered a range of acrylic resins based on the available conservation literature before settling on polyvinyl butyral as a suitable consolidant. We considered Butvar B98 the most suitable candidate for its strong binding and adhesion, low viscosity and therefore good penetration, long-term stability, and importantly, it's solubility parameters. Butvar B98 has a long history of use in the consolidation of powdery wooden artifacts with proven results of strengthening the cellular walls of the wood (Nakhla, S.M. 1986, Sakuno, T., and Schniewind, A.P. 1990, Schniewind, A.P. 1985, Spirydowicz, K.E et al. 2001, Knauer et al 2013).
Investigation into long held conservation concerns about cross linking in Butvar B98 found that the accelerated longevity testing that lead to cross linking in Butvar B98 was carried out in excess of the glass transition (Tg) temperature (72°C-78°C) for this material. Our team concluded that in the unlikely event of the aircraft undergoing these extreme temperatures, cross linking of Butvar B98 consolidant would be the least of our concerns. The concerns about cross linking became less relevant when we considered that the original material choices of urea and phenol formaldehyde are both cross linked resins.
Our choice of solvent carrier was primarily influenced by the requirement to not disrupt the surface coatings or paint layers. Of the range of solvents compatible with Butvar B98, ethanol was found to least affect the paint layers and was considered to help destroy any remaining fungal activity within the veneer layers without significantly dehydrating the wood. Ethanol was also considered to evaporate at a slow enough rate to avoid carrying consolidant to the plywood surface, yet quickly enough to avoid long setting times or residual solvent retention. Butvar B98 was prepared in ethanol in concentrations ranging from 5% to 25% w/v to span the range from light consolidant to viscous adhesive.
Our initial investigations into consolidant application methods identified two potential barriers not addressed in the wood conservation literature. It has long been established that con solid ants move more efficiently through wood in the longitudinal direction (along the grain) than across the grain both radially and tangentially. The cross laminate structure of plywood suggests that consolidants applied along an exposed plywood edge will more efficiently permeate the longitudinally oriented veneers than those arranged at right angles to this orientation, leading to an uneven distribution of consolidant through the plywood overall. In addition, our team discovered that the phenol formaldehyde adhesive layers between the veneers were impermeable, reducing the consolidant application options to exposed plywood edges only.
Our team experimented with submersion and vacuum treatments for deteriorated plywood. We found that whilst submersion and vacuum was very effective for ensuring penetration throughout the plywood, the large volume of consolidant, solvent, and large vacuum processing tanks required to scale up to aircraft size components was cost and resource prohibitive. Submersion methods also made targeted application more complex and were deemed more likely to disrupt surface treatments and paint layers than localized application methods.
Our team concluded that injecting the consolidant into the plywood using needle tipped syringes offered a solution that balanced targeted application in deteriorated areas with the least risk to surface coatings and paint layers. Our initial concerns over differential consolidation in relation to grain direction were for the most part unfounded as the wood in the areas we injected with consolidant was typically so degraded that penetration was not a problem. The application process is somewhat long winded however, as the consolidant must be applied between each impermeable inter-veneer adhesive layer, with the majority of the panels having 20 or more individual ply layers! The area to be consolidated is typically flushed with ethanol prior to the first application of low concentration consolidant (5% w/v). The concentration of consolidant is increased with successive applications as necessary, up to approximately 20% w/v. Silicon release Mylar is applied to the consolidated area, and cushioned clamps are applied to help retain the particular component shape until the consolidant had cured (typically 24 hours). A polyethylene wrapping can be applied around the consolidated area to slow solvent evaporation at lower concentrations.
As described previously, some of the Horten plywood panels have undergone severe deterioration. A combination of poor storage and handling since the time of the aircraft's capture has contributed to areas of plywood loss on most of the panels. The edges around the areas of loss are complex in shape, are comprised of multiple layers, and the plywood in these areas is often delaminating and deteriorated. As before, our solution was driven by the requirement to preserve as much original material as possible and to meet the bullet pointed requirements outlined above.
Our loss compensation material choices were informed by the materials technical study completed at the beginning of the project. Given the proposed close integration of repair with original wooden material, we felt that the repair materials should respond to environmental changes and have similar mechanical and physical attributes to those originally selected by the jets designers. The technical study indicated that the plywood components throughout the jet were formed by combining two basic material units:
While European and American Beech plywood and veneer have very similar mechanical properties and appearance, sourcing plywood and veneer cut to metric units proved difficult within the Imperial (U.S. customary) unit within the United States' plywood market. Despite a long search for 5-ply aircraft grade plywood to match the plywood used in the jets original construction, we were unable to find a "like for like" replacement. We sourced a replacement 3-ply, phenol formaldehyde bonded, rotary cut, European Birch (Betula pendula) plywood from an American aircraft plywood specialist that closely matches the mechanical properties, thickness, and appearance of the original material.
Replacement veneer was rotary cut to our specification—1 mm thickness, defect free European Beech—from locally sourced logs by a small plywood producer based in Northern Germany. The veneer is of high quality and of the exact species to that originally used on the Horten Ho 229 V3.
We devised a method for repairing areas of loss using the Beech veneer and Birch plywood described above. Given the layered nature of the Horten plywood panels, we started with the premise that an area of loss was in fact made up from several strata, each with a slightly different shaped loss.
In order to accurately produce veneer or plywood fills, we first traced the outline of the loss area for each layer onto Mylar sheets. The Mylar tracings were scanned on a flatbed scanner to produce images of the loss area and processed in Adobe Photoshop to remove scanning artifacts. The processed images were imported into Adobe Illustrator and measured using the on-screen measurement tools to ensure the image size was true to the original tracing. The Adobe Illustrator 'live trace' tool was used to turn the image bitmaps tracing into vector based line drawings, and the results exported as Encapsulated PostScript (EPS) files compatible with the Museum's laser cutter software. The files were uploaded to the laser cutter and cut as shown in the video below. Cutting the fills in this manner produces a level of detail that would be extremely difficult and time consuming to achieve with traditional methods. The laser cuts equally well along and across the grain of both plywood and veneer.
The fills are adhered in place, one layer at a time, using a 25% w/v Butvar B98 in ethanol solution. The combination of overlapping individual fill layers produces a structurally stable and well bonded fill. When viewed in cross section, the fill visually integrates with the original multi-layer plywood structure. In the example shown below, the outside edges of the fill extend beyond the edge of the original panel to allow for a small amount of final finish after gluing.
From the first layer to the final fifth layer, these images show how a detailed plywood fill is created one layer at a time.
The process was repeated for a single layer fill using 1 mm thick Beech veneer described above. The fill shown in the video below was cut in 36 seconds to a level of accuracy that would be extremely difficult and time consuming to achieve using traditional methods.
The laser cutting method described above is well suited to large areas of loss but is difficult to scale down to smaller sizes. For small areas of loss (less than 5 cm x 5 cm) not adversely affecting structural stability, our team experimented with combinations of different fillers and acrylic resins, the results of which are shown in the table below.
Butvar B98 with cellulose powder was selected as the most suitable fill material as it was considered easier to handle and apply than the other materials tested and dried to a hard strong finish. The crumbliness described in the table above was eliminated by extending the drying period prior to carving. To produce the fill material, 15% w/v Butvar B98 granules were added to 5:1 ethanol:acetone. Acetone was added to more effectively solubilize the Butvar granules, but owing to concerns surrounding object surface coating solubility, was allowed to vent after mixing by stirring the solution in a fume hood for 30 to 40 minutes. Cellulose powder was added to the resulting solution to form a putty like material.
The plywood along the edges of the loss area is typically consolidated with Butvar B98 in ethanol prior to application of the fill putty. The edges of the loss are lined with Teflon tape and a small dam is constructed around the loss. The putty is applied with spatulas and shaped with dental tools. Depending on the size of the fill area, a veneer layer can be added to the upper and lower edge of the fill putty to more closely integrate with the original wooden surface. Once dry, the fill is removed from the loss area, the Teflon tape removed, and the closely fitting fill adhered in place with 25% Butvar B98 in ethanol.