| dc.description.abstract | Metal objects have traditionally not been repaired or constructed by adhesives, but
primarily by soldering, welding, and mechanical assemblies. Soldering has many
disadvantages for the conservator such as lack of reversibility, problems to solder
a corroded surface, corrosive flux, heating causing damage to the patina, crystal
structure and more. Bonding of historical metal objects with adhesives is here
examined. A literature survey is done. Another goal with the work is to get an
indication of the suitability of various adhesives by means of simple test methods.
Two test objects (candle holders) of tin, were chosen as basis for the investigation.
Six adhesives were selected for testing: Paraloid B-72, (a thermoplastic acrylic
copolymer), two different epoxies, (Araldite 2011 A + B and Bostik Epoxy Rapid)
an ethyl cyanoacrylate product, (Loctite Super Glue Precision), Bostik contact, a
contact adhesive based on polychloroprene rubber, Dana limstänger, (a hot melt
adhesive consisting of an EVA copolymer). These adhesives were subjected to
simple tests, a pH test, a so-called Oddy test and a simple tensile test with
specimens of tin to investigate strength. In addition, a test for removability of the
adhesive and a very simple contact test were carried out.
pH tests showed that many of the adhesives are quite acidic when wet. After
curing the pH of the epoxies are about pH 9-10. The others were more neutral
after curing. Oddy test showed that all but Paraloid B-72 affected lead slightly.
”Epoxy Rapid” affected silver and copper with a dark colour. The results of
strength tests were somewhat difficult to interpret. The strength differed greatly
between the adhesives. For example the specimens bonded by the hot-melt
adhesive fell apart by themselves. The cyanoacrylate was so strong that in one
sample the joint was not broken at all, instead the tin specimen broke. When the
tests were done the two candlesticks were treated. Only some indicating
conjectures can be made by these results. | sv |
