I got a bunch of different answers to my question about, ‘What do you find is the best adhesive for gluing loops, pins and studs to cabochons, brooches and/or earrings etc?
http://www.frit-happens.co.uk/forum/index.php?topic=45455.0 Everyone seems happy with their particular choice and I know there isn’t a problem but I thought I’d do my own trial with the adhesives that I have in the workshop.
I did four trials using two different glues, I used Araldite Rapid adhesive for two tests and I used Evo-Stik Contact Adhesive for the other two tests.
I cut four strips of 0.1 mm thick copper foil, each 6 mm wide by about 30 mm long and following the manufacturer’s instructions I glued the four strips to the side of an old OXO gravy jar. The contact/glued length was about 10 mm long in each case.
I slightly abraded the copper strips but not the glass. I cleaned everything with acetone and made sure everything was dry before gluing.
DSCF8765 A by
, on Flickr
The first two strips glued in place.For the first test I did a ‘pull test,’ that is pulling the copper strip with the force in-line with the length of the copper and the second test was to check the peel strength by pulling the copper foil away from the jar at an angle of 90 degrees.
You might say why bother to do this? Well, I find it very interesting and when messing about like this I usually learn something unexpected which helps in other projects.
I hadn’t got any silver loops or bails to play with so I used the copper strip for the tests. I’ve been working with copper in industrial applications for over forty years and yesterday I learnt something new about it.
I soldered copper wire loops onto the strips to give me something to hook onto to pull, and I used a baggage scale with a maximum weight indicator on it to record the readings as the bond broke apart.
DSCF8766 B by
, on Flickr
In the pull test for the Araldite the copper strip broke but read on.The Araldite Rapid is a two part epoxy and really needs twenty four hours to attain full strength; I'm impatient and tested it after an hour.
The Evo-Stik Contact Adhesive has to be applied to both surfaces and left to dry for a while, and I gave that 15 minutes before placing the strips on the jar, and again I tested it after an hour of curing.
In the ‘in-line’ pull test with the Araldite Rapid the scale registered 22.5 kg and that’s when the copper strip broke, so on the face of it produced a very strong bond and I had a respectable reading of 12 kg when the Evo-Stik joint failed.
DSCF8771 C by
, on Flickr
At first sight from the pull test the Araldite looks good with copper, but it’s not. Surprisingly in the ‘peel test,’ the strip held with the Araldite peeled away with less than a 0.1 kg reading, it was very poor and it needed very little force to pop it off. The peel reading for the Evo-stik was a respectable 4 kg but read on!
In all four tests the adhesives mainly remained firmly attached to the glass and that struck me as odd. With a compatible adhesive I would have expected the joint to fail with some glue on the copper and some glue left on the glass, but here very little of the glue remaining on the copper, it nearly all remained on the glass.
As I say I’ve worked with copper for many years, I’ve machined it, soldered it, brazed it, drilled it and used it in sheets, tubes and blocks, but I’ve never glued it before and this is where the problem lies.
Because I hadn’t got any silver bails for the test I used copper. I’ve now found out that although the glue manufacturers say their products are suitable for ‘glass and metal’ what they don’t tell you is that many of the adhesives don’t work at all well with copper or aluminium (and possibly brass.)
It turns out that the hardener in the two part epoxy reacts particularly badly with copper and I guess that had I have left it longer for that negative reaction to take place then the in-line pull test would have been much weaker.
Now I think about it I do remember seeing green/blue verdigris on copper that may have had glue on it and simply accepted it as moisture getting under the surface or into the joint. I know that silicon glue produces acetic acid (smells of vinegar) and will always turn copper green but I didn’t realise that other glues were also problematic with copper.
DSCF8767 D by
, on Flickr
Peel test set up.I will buy some silver bails and carry out further trials, not that I need to prove anything, it’s just that I do enjoy pushing things to the limit (or in this case pulling them) and then working safely within those limits.
I used to design and manufacture special purpose machines and I once carried out trials for an American aircraft company who wanted to glue sheets of aluminium together using a thermal curing adhesive. The test panels were about three foot long each with about a 1 inch overlap of metal where the glue was and there was a rivet at each end of the overlapping joint. When I asked what the rivets were for they said it’s a ‘Chicken Rivet.’ I still didn’t know what they meant so they explained, that when a glued joint like this fails it is almost certainly going to be as a result of the joint been ‘peeled’ apart rather than being ‘pulled’ apart laterally and the chicken rivets are for the scaredy-cats who don’t trust the glue. In other words the failure usually starts at one end not in the middle and the rivet stops it from starting to pull apart. And yes, they really do glue some parts of aeroplanes together but they use some very sophisticated adhesive, (and chicken rivets.)
DSCF8771 C by
andy.c.davies@btinternet.com, on Flickr
Glue left from pull test
DSCF8770 E by
andy.c.davies@btinternet.com, on Flickr
Glue left from peel test Anyway, I hope you found that interesting. It’s 2nd of December and it's been a bit too cold to sit in the workshop with the doors open, (must fit an extract fan sometime,) so I thought I’d share this with you. I didn’t find out what I really wanted to know about the glue but hopefully it might help someone. By the way there are glues available specifically for use with copper.
For my next test I'll try silver bails and hopefully that’ll be more useful.
Just one more point that can confuse things further, glass jars like the OXO jar that I used are made to have a ‘tamper evident’ peel off foil lid and part of the manufacturing process of these jars involves a thermo/chemical process that helps that peel off foil to bond to the glass so I guess for a more realistic test I'll just use a bit of plain old glass next time.
Thanks for the interest.
Kind Regards . . . Andy