Third Post:
Thinner is better:
There has been some discussion about not clamping glued scales too hard, for fear of “squeezing out all the adhesive.” My point here is that you should not NEED to worry about that. The first point here is that, if the surface has been suitable roughened to produce scratches, “nooks” and “crannies”, those surface features should act as “reservoirs” – holding the adhesive and protecting it from being squeezed out of the interface. Sure, if the surfaces are really smooth, you can squeeze out most of the adhesive, but you do not WANT the surfaces to be that smooth.
The second point here, which is initially counter-intuitive but makes sense after it is explained, is that the thinner the layer of adhesive, the stronger the joint between the two surfaces. The reason for this is that in the realm of failure mechanisms, and this is a very widely applicable statement, failure of anything (metals, plastics, adhesives, etc etc), the initial point of failure always begins at a defect. These “defects”, depending on the situation, could be a scratch on a surface, an inclusion (i.e. foreign object) in the material, a void (i.e. a “bubble”) in the material, etc. It is for this reason that airplane engine turbine blades are regularly inspected for surface defects. It is why you can “scribe” the surface of a piece of glass and then break the glass cleanly by bending it along the line of the scribe, etc.
Here is the key point in adhesion: defects are always present in the adhesive layer (bubbles, dust and other foreign particles, etc) … and the less adhesive present, the fewer the total number of defects (statistically speaking) are likely to be present … on the other hand, the more defects that are present the greater the probability that one of them is just in the wrong place (in terms of the stresses placed on the joint), and also the greater the probability that one of the defects is particularly “bad” in some way (much bigger, more irregular with sharp corners, etc).
Ever heard the phrase “the weak link”? This comes from the workings of a chain, which is made of independent links strung together. When pulled, all of the links see the same amount of stress. However, not all links are created equal: some are stronger, some are weaker, and that strength characteristic follows a statistical distribution of some sort. All such distributions allow for a small number of elements that are way-way out there. The longer the chain, the greater the raw number of links are in it, and the greater raw likelihood that one of those really, really weak links gets included in the chain. Thus, the longer the chain, the weaker it is overall – because the chain WILL break at that weakest link.
The same is true with adhesive layers: it will likely fail at the weakest point – and by making the layer thinner, you reduce the likelihood of having a really, really bad defect present in the layer. So go ahead and clamp aggressively – just make sure you have present the surface roughness so that the adhesive does not actually all get squeezed out.
Mode of failure:
The strength of an adhesive joint is extremely dependent on the stresses placed on it. Let me say that again: the strength of an adhesive joint is extremely dependent on the stresses placed on it.
Typically, we can think about three basic ways of stressing an adhesive joint: “normal” (pulling straight up perpendicular to the surface). Lateral or “shear” (pushing sideways), and “peel” (i.e. pulling up only on one edge). Depending on the joint configuration, these three modes of failure, for a single joint, can have very different strengths.
Consider, for example, a surface that has been roughened by sanding only in one direction, and in a way where few or no undercuts. If you try to pull “straight up” on one side, there are basically no mechanical interactions between adhesive and surface resisting that pull – and you are only left with whatever chemical interactions were created (which are relatively weak). That interface could be pulled apart fairly easily. Now, consider that you try to shear the interface (push from the side), but do so in the direction that the sanded grooves run. Again, there are few or no mechanical interactions to resist that stress, and the joint will fail fairly easily. NOW consider that you shear (push from the side), but do so perpendicularly to the directions of the sanded grooves. In this case, you have adhesive inside of the groove being pushed against the side of the groove – they “bump” against each other. This is considerable mechanical interaction – and the joint will be much stronger against failure with this mode of stress.
On the other hand, a joint with well developed mechanical interactions, and especially “undercuts”, if you just try to pull the surfaces apart (normally), the force of that pull is distributed among all of the “undercuts” (they are in “parallel” to the stress), so that each undercut has a much smaller force placed on it – with the result that no one of them is likely to exceed its individual strength. In this situation, the joint is extremely strong against perpendicular stresses.
The same type of argument can be made to sideways (shear) forces: if (and I really do mean “if”) the material being bonded is rigid enough, then the sideways force is distributed among all of the activated mechanical interactions, so that each mechanical interaction only needs to withstand a pretty small stress. This is NOT true of the bonded material is flexible. In this case , the applied force gets concentrated to the mechanical interactions present near the edge being pulled on – and thus are more likely to fail.
This latter example is most akin to “peel” – where only one edge of the thing has a normal stress applied to it. In this case, ALL of the normal stress is placed on the few mechanical interactions in the immediate vicinity of the edge – and so each of these individual mechanical interactions have a great amount of stress placed on them, and so are likely to exceed their strength. When they fail, the concentrated stress is then passed on to the mechanical interaction next in line along the surface of the interface, its strength is exceeded, and it fails, and so on. The interface “unzips” from one edge to the other. The ability to withstand “peel” forces is the weakest strength of any adhesive bond.
How might this be applied directly to knife handles? A few points come to mind…
Warpage of handle material (after being applied to the tang) will likely apply a normal stress to the adhesive joint. The joint will be able to withstand some of that – but as Ben will tell you, the developed stresses in wood due to moisture loss are extremely large – hence comes the common knowledge that you should get your wood stable (in terms of moisture content) before you attach it to a tang. Loss of moisture will also cause general shrinkage of the wood – which will cause a lateral stress on the adhesive joint. Same point and resolution as above for warpage.
But also again, if the joint is suitable prepared with sufficient surface texture and the adhesive layer is suitably thin, you will increase its strength to hopefully better withstand any stress that will (not may) be placed on the joint due to handle expansion or shrinkage.
I said above that the adhesive joints are weakest when “peel” type stresses are placed on them. For a working knife, when you use the blade, you will apply some level of “bending” stress to the blade. This will be seen at the location of the handle as a “bending” right at the junction of the handle and blade – which is identical to a “peel” stress. In this case, the thinner the blade material, the more bending, and the greater the amount of peel stress applied to that edge of the handle. How to address this? Use the blade more gently (reduce the stress), or use thicker blade material. Another thing that occurs to me after Natlek’s question is that you can protect from this stress by using a metal bolster that is securely attached (typically soldered?) to the blade. Because the back side (handle side) of the bolster is typically perpendicular to the blade, if the handle material is securely bonded (per everything said above) to the backside of the bolster – the stresses at that location will be in shear (not peel) between the handle and the bolster. Because shear resistance is much stronger than peel resistance, this will then protect the handle against adhesive failure due to bending of the blade.
Ok … I think this last post ends up completing all the points I initially stated I would try to cover. Hopefully you (and future knifemakers) will find this understandable and helpful. Please do chime in with any questions.
). Even had your boot stuck in the mud? same idea. Adhesion overall is COMPLICATED.... )
Actually I ask about polyurethane glue because I never used that . Thanks again for explanation....
