some time ago when Steve Sando and I started testing adhesives, we started a few threads on it. I promised I would come back with some results and conclusions after it became too much to maintain all the threads. Here are the results I came up with, Steve is AWOL as he is a busy guy lately, so his results will follow but they are close to these I'm fairly sure. At the end of this is a link to an Excell spreadsheet with some additional data. I'd invite any comments if any one has any.
I learned several things from these adhesive experiments.
First, I had been told over several years by very good knife makers, that Devcon 2 ton epoxy was a great choice for knife making. They had been told, no doubt by other great knife makers that it was great adhesive. Devcon 2 part epoxy was one of the very first adhesives to fail both Steves tests and mine. It rarely lasted more than a few cycles through the dishwasher. The learning here was: test and check every thing youve been told even if it comes from a very good source. Nothing beats an opinion more than testing it yourself. Given this, you probably shouldnt believe another word that follows this one. I know I dont believe half of what comes out of my mouth some times.
The second learning:
Surface preparation is more important to a successful glue up than is the adhesive its self. Please dont underestimate surface preparation. A mediocre adhesive with an excellent surface preparation will out perform an excellent adhesive with a poor preparation by a large margin.
Steve diligently cleaned some of his test pieces with Simple Green, a soap that is often recommended for cleaning all kinds of things. The simple green apparently left a residue that prematurely failed all of his initial adhesive tests. Other surface cleaning solutions also left a residue that impaired adhesive performance.
My tests on a roughed up or scratched surface using a fresh 36 grit belt versus sand blasting gave exceptionally clear results. On ten test pieces I prepared, one end was sand blasted and the other scratched with a 36 grit belt, all of the scratched surfaces failed. Six of ten of the sandblasted surfaces survived. THIS IS HUGE. Again, routinely, high caliber and very competent knife makers recommended scratching or roughing up the surface of a knife to provide a better purchase for the adhesive. Certainly this is true but sandblasting a surface vastly out performs a scratched surface to the point where I recommend a knife maker absolutely invest in at least a minimal sandblasting setup to prepare a surface for glue up. Watch chemical cleaners for residue. Test this by allowing you cleaning chemical to dry on a piece of glass and have a look. I use isopropyl alcohol myself.
Third learning:
Read up on your adhesive. With the Internet, there really isnt a reason you cant completely understand what you are working with. Gorilla glue was being used or at least considered for use in full tang knives since it tended to expand as it cured. I thought this was a great feature of this particular adhesive. It also seemed to make this adhesive perfect use for a full tang knife. Fill the hollow in the handle material part way, put it together and let the expansion fill the rest of the hollow. The problem is that the expanded (foam looking) adhesive has no mechanical strength at all. This was documented fairly clearly on the Gorilla glue web site but had not been reported to others on the forums as a caution. On the other hand, you cant believe every thing you read. JB Weld failed most dishwasher tests and is reportedly good up to over 400 degrees Fahrenheit. They must mean a dry heat dishwasher.
Fourth learning:
There is no one perfect adhesive. There are some that work really, really well at some things but not at all things. Gorilla glue outperformed every adhesive in my destructive tests when gluing wood to metal. It failed miserably when gluing metal to metal or metal to Micarta. Loctite Speedbonder 324 held a stunning 100lbs of weight when bonding metal to metal but failed early gluing metal to wood.
Fifth learning:
Thicker glue is easier to use than thin glue. Viscosity ratings are impossible to relate to so Id suggest the perfect viscosity is butterscotch pudding. Peanut butter consistency is too thick, syrup is to runny and messy. This makes such a difference that if a couple of adhesives are close in performance, the hands down choice would be to go with the butterscotch stuff. I hope Im not being too technical.
Sixth learning:
There are some amazing adhesives out there. Amazing adhesives. Our recommendations can and probably should fall to the wayside over the next several years as adhesives technology continues to improve. They glue cars together now instead of welding or rivets. Cars!
Glued together! What the hell? Keep an eye open for new adhesives but make sure and test it.
Seventh learning:
Testing wasnt all that hard. If you find a new adhesive you want to test, glue some wood to metal and throw it in the dishwasher for a half a dozen cycles. If it survives that, you might have something. If it doesnt, dont even consider it. If it makes it through the dishwasher tests, boil it for an hour or so. If it makes it past that test, you are really on to something. From there make something else up and let us know what youve found as you are probably on to a good one.
Eighth learning:
Industrial adhesives are generally better than adhesives packaged for retail sales to the general public. There are exceptions of course but it generally holds true. K&G epoxy, which is a recommended adhesive from our tests, is industrial epoxy repackaged and marketed to knife makers and gunsmiths.
Ninth learning:
Gorilla glue shouldnt have held up like it did. It was supposed to fail early and was just meant to bulk up the tests. It kicked most every other adhesives ass all over the block. Keep an eye out for things that arent supposed to work and try them out.
Tenth learning:
Steve found some adhesives I never would have. I found some he didnt. Collaboration and sharing information helps the process. Not really a new learning but its worth repeating and it was clearly demonstrated here.
Eleventh learning:
There is a type of adhesive that cures by a lack of oxygen. Id never heard of anything like it before. Its called anaerobic curing adhesive and wont cure well when the material is porous like wood.
AND NOW FINALLY
.
For wood handle scales to metal:
Recommended:
Gorilla Glue beat out every other adhesive in every test
Loctite E-120hp
K&G epoxy
Golf shafting epoxy by Golfsmith
Acraglass
Avoid:
Anaerobic curing adhesives
JB Weld
Full tang handles:
Acraglass which was designed almost exactly for this kind of application.
Avoid:
Gorilla Glue or any other polyurethane expanding type glue.
For metal to metal
Recommended:
Loctite Speedbonder 324
JB Weld
PC7
Avoid:
Any CAs, ie superglues. These have no place in knife making anywhere except to fill flaws in handle material.
For metal to man made materials (ie, G10, micarta, etc)
Recommended:
Loctite 324 Speedbonder or any of the overall use adhesives.
For overall use:
Loctite E-120hp
K&G Epoxy
Golfsmith Golf shafting epoxy
Acraglass (Acraweld is different and was not tested. Acragel was tested and failed)
adhesive spreadsheet
