Testing for the ultimate adhesive

[tmickley]

The adhesive tends to almost always stay on the wood in the failed wood/metal joints. In the metal to metal joints, the epoxy is fairly even on both surfaces if both were sand blasted. If metal to metal and one surface was 'scratched' and one sandblasted, the sandblasted surface clearly had a better purchase.

I am using 70 grit AO in my sandblasting. When my current stuff wears out, I'll be looking for a bigger grit AO.

The 324 Speedbonder adhesive is anaerobic curing (a lack of oxygen cures the adhesive) and I think the wood allowed some oxygen through the pores of the wood and didn't allow it to fully cure. The failed joint had an odd, slick texture. I am going to try the 324 with micarta and metal and see what happens. I think it will hold much better there.

The wood I am using is stabilized donated by Chuck Bybee. I don't know what kind it is. I used some of my own stabilized in the first test.

[fitzo]

Thanks, Tracy. I *thought* the adhesives usually stayed with the wood , but couldn't tell in the last foto. It led me to wondering whether the repeated dishwasher cycles had allowed water to soak the wood and hold moisture up against the adhesive and eventually affect the bond. Those rust spots under adhesives are very intruiging. I would never have thought these type of adhesives would absorb that much water.

This has grown to be one of those experiments where every set of data answers some questions and generates more! But, that's no surprise. There are polymer chemists that make a career doing testing like you two have done. It also explains why there are a plethora of choices out there.

You and Steve have given us a lot of information. Thanks, again, guys!

[SteveS]

Mike,

On my 2 week dishwasher test (I used a smaller sample and zip tied it under the glass rack - very steathy - below the wife radar) Most of the adhesives that failed stayed attached to the wood, leaving the steel fairly clean. The SCARY! part was the rust underneath. Therefore, you can't say "well my adhesive isn't the stongest, but I only use it to seal the scales." I had rust under blocks that held on - and it was CM-154!

I had 4 glues on there each twice. Both samples would fail under the same conditions - meaning the results were pretty accurate.

Tracy, I'm glad that K&G is still on there. Mine failed in the 2-week dishwasher test. I think it's because I didn't use a scale to mix it.

Steve

[tmickley]

First a comment on dishwashing for testing. Every time I tinker with these tests I wonder if dishwashing is 'fair' as a method for accelerated failure. Every time I decide it is. I can't imagine putting a $100+ knife through the dishwasher but I bet people do. I don't believe the temperature range in a dishwasher is too high to be considered unfair either. Put a black knife on the dashboard of your car some time in the hot sun. Dishwashing seems to have turned into a substantial test method so it deserves some examination.

I've had this rust show up under every single failure. I'm not sure if it comes from water migrating through the adhesive, if it is trapped moisuture or if it is from loosened or cracked joints. I'll try to pay more attention to this going forward....

[tmickley]

I went back and had a look at the failed joints on the metal. What I found was what I will call 'creeping failure'. It seems the failed joints didn't just give it up all at once, they failed starting at the edges. This failure creeps toward the center of the joint or the area farthest away from the seams. You can see (barely, it's as good as I could photograph) in this picture where the rust creeps in around the square outline towards the middle. In the center of each failed joint, marked by the X and Y, it is basically rust free and that is where the remaining adhesive was still doing it's job hanging on for dear life when it was knocked off by the hammer. The Devcon joint (not pictured) that simply slid off in the dishwashwer was rusted almost entirely.





Here is a picture of the back of the blocks. You can see the blocks marked X and Y have clear areas in the center that kind of match the clear areas on the metal.

Mike got me to go back and have another look at the 324 Speedbonder. All of the failed joints left the bulk of the epoxy on the wood. It has a better purchase. On the 324 all of the failed epoxy was left on the metal. That explains the odd texture I felt previously when I had a look at it. The texture I felt was actually a reverse casting of the wood grain. I believe now the 324 didn't fully cure. I believe this stuff partially cured and gave some degree of adhesion but no where near it's full effectiveness. I think the wood is porus enough that the oxygen in the pores prevents the anaerobic adhesive from a complete cure. Based on this, I would stay away from using an anaerobic adhesive to join porus wood (and what wood isn't porus?) to metal. I have a piece of micarta and metal curing now that I will be testing to verify this. I know this to be an exceptional adhesive when joining metal to metal pieces and I suspect the micarta to metal will be the same. I think this test eliminates 324 or most any other anaerobic curing adhesive (a stretch I know but it seems to be reasonable) from being a good choice for glueing wood to metal on a knife handle.



respectfully submitted
Douglas C. Neidermeyer
Sargeant at Arms
Alpha Phi Kappa

[fitzo]

Good obsesrvations, guys! This is the type of stuff where information analysis can start to create knowledge!

I've got another off the wall question, Tracy: was the 324 a two part? IE, coat one surface with activator and the opposite with the resin and then clamp? If so, is anything specified about what goes on what? I have no idea what I'm getting at other than wondering if this could possibly make a difference since the stabilized wood already has an acrylic impregnated into it and could have a different chemical reaction with the activator or resin.

One other question: have either of you guys noticed any obvious signs that the adhesives are absorbing moisture, like a change in the translucency or whatever? As mentioned, the question of the mechanism for the rusting is important here, whether water is getting under the adhesives because of broken bonds or absorbing into them and the rusting being actual cause of bond failure. If the first possibility is the case, that could be due to differing rates of expansion/contraction in dissimilar materials (wood vs metal) from the heating/cooling cycles.

Jeesh, there's just never a polymer chemist around when we need one.

[SteveS]

Mike,

I had two rust failures on my samples. One of the rust that developed under mine was like Tracys - it started at the edges and appeared to work in.



The odd one was U-05FL. That is a flexible adhesive. It's designed for differing expansion rates. Now, on my samples I finished the edges like you would a knife. So barely any adhesive is visible there. The dishwasher action appeared to eat away at the U-05FL. The rust was pretty bad there. Suprisingly there were bad spots in the middle too. It must have wicked moisture all the way thru.

I thought maybe the dymond wood got saturated and it got in that way. Nope, the Acraglas Gel sample didn't have any rust under it.

As much as I like 05FL (holding power, impact resistance, flexible bond) I'm not going to use it because of the rust thing. Acraglas Gel doesn't hold as well, but it seals better.

Steve

[Jerry Shorter]

Tracy: What are the attributes of air that cause anerobic adhesive to not cure? I wonder if a wood to metal joint using 324 would cure in a vaccum?. Have you tried it?.

Jerry

[TexasJack]

324 contains peroxides. One side of the surfaces to be joined has a catalyst spread on it that causes the peroxides to decompose, forming free radicals necessary for the acrylic resin to polymerize. Loctite 324 does not depend on exposure to the air in order to harden.

[Jerry Shorter]

Originally Posted by TexasJack

324 contains peroxides. One side of the surfaces to be joined has a catalyst spread on it that causes the peroxides to decompose, forming free radicals necessary for the acrylic resin to polymerize. Loctite 324 does not depend on exposure to the air in order to harden.

Well that answers that question thanks TexasJack

Jerry

[tmickley]

Mike the 324 is two part. A spray on activator and then the adhesive. The directions are ambiguous, you coat either one or both pieces with the activator, wait a few minutes until the solvent has evaporated and then apply the adhesive. I sprayed both parts.

I have not be able to determine if water is taken up into any of the epoxies. I'd like to think it isn't but I bet it does. I'll shoot some into some cups, let it harden, soak it and see what it looks like.

Jerry, this anaroebic stuff requires a total lack of oxygen to kick off. It comes in a bottle that is only 3/4'ths full so there is oxygen in the container all the time. The activator does something too. I know that if you spray activator and then run a bead of this stuff and let it sit with out another piece 'closing it up', the stuff will not cure. I can even spray some more activator on top of it and it won't set up. It requires a lack of oxygen also.

[tmickley]

Jerry, missed your comment earlier on what about curing in a vacuum? I don't have the equipment to pull a vacuum that I think would be needed. A vacuum bag might be just the thing though. I don't know how much available oxygen is needed to prevent curing.

Now to think out loud for a minute and I'd invite any comments either way on this. The 324 cures in the absence of oxygen but at the edges, it doesn't cure to a depth of about 1/16". Now picture a full tang knife and you have it profiled to where you want it, you glue on the scales and let it cure. The scales are always over size so you now have to grind down the scales to match the profile of the handle. Since you all ready have the handle (tang) just about exactly where you want it, you grind the scales down to match the tang and then maybe a couple thousands here and there to smooth it all out. The problem is the anaroebic adhesive doesn't cure at the edge where it is exposed to air so now you have adhesive right at the edge of your knife handle that isn't cured. This to me, is a big problem.

In looking at this edge not cured issue, the uncured edge is approximately 1/16" deep so I suppose you could make it a habit to profile your handle 1/16" or more oversize and then grind it down but the problem is you generate enough heat in grinding this much down to melt or char any epoxy unless you dip in water or go insanely slow.

This makes the 324 a deal breaker for me.

[justice]

yea 1/16th is alot to get out of the way when you are at that point. i think some kind of vacume pump set up might work. but it seems like alot of work!!


by the way im lost are you guys still testing gorilla glue?

...justin

[Bernez]

Just seal the edges with tape

Bernez

[SteveS]

RE: 324

I've heard from one fulltime maker that he uses and loves 324. Sounds like you have to have tight fits to get it to work properly. But it sounds like a good choice to me.

Justin,

The gorilla glue tests are buried back in the thread here. We'll have the detailed in the final report. Basically, it does very well - IF on of the surfaces is a bit porus. Wood, even dymond wood, on steel works really well. Steel on Steel forget it. Also it's not a good choice for gap filling - it doesn't get real hard. Don't use it on hidden tang knives. But, it's really handy and simple. Personally, I don't like the mess it makes. So I won't be using it.

Steve