Any idea how to measure the adhesion of a polymer to a metal...

[magomago]

Hey guys,

I'm faced with a smaller dilemma here. I'm trying to measure an adhesion force (easy enough, right?) of a polymer onto a metal.

But of course I have a few hurdles...

The first issue is that its a cylindrical shape - of course from what I've read using the peel method I can try to get a slice small enough such that I can approximate the curvature as a plane.

Here are two kickers:

a) The cylinder is cut into a mesh. Wheee- apparantly with over 10 mesh patterns I can smooth out over this issue as that mesh is repeatable enough not to hamper me too much (I got about 13 repeatable patterns)

and here is my main problem

b) my cylinder diameter is on the order of a few milimeters. About 3 mm....no more.


I tried to see if I could get my metal produced in a flat sheet and simply do a quick planar check but that isn't going to happen because the place I'm dealing with (as you can tell I'm doing my senior design project for materials science...of course measuring this adhesion force isn't the only thing I want ) can only produce cylindrical tubes of said shape.

So I'm trying to think of how I'm supposed to accomplish this. I looked into using an AFM (Atomic Force Microscopy) but old test data using a micro-scratch method (which is pretty unreliable and produces crazy results...hence my goal to find something better) shows a strength on the order of 10ish to 50ish mN. That automatically chucks AFM out of my options since it deals uN to pN....

So now I'm kicked back to a peel method (Which is going to be ridiculously hard on the scale I'm dealing with...) as my only option. Then I realized HT existed and I thought it might be worth it to solicit some opinions Does anyone have any advice? Perhaps another method?

Thanks all and I'll be actively watching and contributing...I'm not here for direct answers to all my problems, but something that can kick start me when I feel I've hit a dead end =)

 

[PolymerTim]

Coincidentally, I've been doing a little adhesion testing myself lately, but I must admit it hasn't been easy for me either. My problem is primarily that the adhesion is so low it is difficult to measure. Maybe you could add a couple more details for your problem. How long is the cylinder? What equipment do you have available?

 

[magomago]

Okay I have a shorter samples with me ATM but I'm looking at 15-18mm or so in length, and 3mm in diameter.

As for equipment...a lot of that is from the school. Beyond an AFM, there is apparantly a new tensile testing machine coming in that is sensitive enough for tiny (think 2cm long, .5cm wide t-bone style) samples. That is all that I think that is relevant.

Perhaps what may come into use indirectly are SEMs, DSC, TGA...wait let me stop with the acronyms. Differential Scanning Calorimetry, Thermogravimetric Analysis, an HPLC machine and another outfitted with GPC, Rockwell hardness machines (totally not sensitive to what I'm looking at ). But I can't see at the moment how these are useful for measuring the adhesion. For sure some of these will get use as I'm analyzing certain factors for later use...

 

[Billb2]

Use the polymer to "glue" two pieces together, then pull them apart.
The problem would be in how to deal with the area of polymer that gets separated from the metal. Probably different with each test.
Use a thin stripe of polymer of a known and repeatable contact area and hope that all the polymer separates when pulled apart.

 

[PolymerTim]

Originally posted by: magomago
Okay I have a shorter samples with me ATM but I'm looking at 15-18mm or so in length, and 3mm in diameter.

As for equipment...a lot of that is from the school. Beyond an AFM, there is apparantly a new tensile testing machine coming in that is sensitive enough for tiny (think 2cm long, .5cm wide t-bone style) samples. That is all that I think that is relevant.

Perhaps what may come into use indirectly are SEMs, DSC, TGA...wait let me stop with the acronyms. Differential Scanning Calorimetry, Thermogravimetric Analysis, an HPLC machine and another outfitted with GPC, Rockwell hardness machines (totally not sensitive to what I'm looking at ). But I can't see at the moment how these are useful for measuring the adhesion. For sure some of these will get use as I'm analyzing certain factors for later use...

I have a feeling that SEM, TEM, XPS, AES, GPC/SEC, LC-MS, FT-IR, NMR, DSC, TGA, WAXS, SAXS, RALLS, LALLS, and PALS, while excellent techniques in their own right, will do little to help you with adhesion. Now, DMTA, may come in handy since it is much more sensitive then typical Instron/MTS tensile testers.

Hmmm, sounds tricky with a sample that small. I have one idea for use with the tensile tester. I know there are techniques for adhesion such as roller drum peel test and climbing drum peel test that use drums as an aid to peel films off of rigid surfaces. I wonder if you could use a similar technique for your system. I don't know if there is enough material around the diameter for this though (only a little over 9mm circumference), but its worth a thought. Basically, I imagine slicing the coating down the length, initiating a couple mm of the peel, glueing the free end to a tab, and pulling the tab and the cylinder apart. For this to work of course, the cylinder would need to be mounted to a low friction bearing to allow it to freely rotate. It's a long shot, but it sounds possible. And since you're peeling normal to the cylinder surface, the curvature has no effect.

 

[PolymerTim]

Originally posted by: Billb2
Use the polymer to "glue" two pieces together, then pull them apart.
The problem would be in how to deal with the area of polymer that gets separated from the metal. Probably different with each test.
Use a thin stripe of polymer of a known and repeatable contact area and hope that all the polymer separates when pulled apart.

I think the trick is in what way do you pull them apart? Both sides would not even need to be the metal/mesh since only one side is being measured. It is also important to keep in mind that it is necessary to measure the force required to propagate the delamination, so the process needs to be slow enough that initial forces required to initiate delamination are allowed to settle and you get a measurable, smooth, steady-state delamination.

 

[Nathelion]

It wouldn't be very precise, mostly because it produces a different surface, but could you manually slice a cylinder in half, polish the cut surface to roughly the same quality as the outer surface of the cylinder, and then measure this flat surface?

 

[CycloWizard]

There are entire journals and textbooks dedicated to measuring this sort of thing. I happen to have a paper that describes your situation exactly (at least, as I understand your problem ). If you PM me I'll try to remember to send it your way when I'm in lab tomorrow. It's been a while since I looked at it, but it describes how to test adhesion in a circular cylinder of finite length, giving both test methods and an analytical solution to the elasticity problem (Navier equations). If you want a true solution to a problem where the cylinder is a "mesh" (which is a very nebulous term in this case, so perhaps you can describe a little more clearly what you mean by that), then you'll need to use a finite element method to model the situation. If you're just interested in the work of adhesion or something similar, then forget the modeling and the solution is pretty straightforward.

If I'm misunderstanding what you're doing, let me know. I should be able to help since I've spent the last few years working on things that are similar, though not directly related.

 

[wwswimming]

if you put a flat on 2 of the cylinders and then bond them at
right angles to each other using the test polymer - will this
configuration give you an useful answer ?

 

[magomago]

Originally posted by: Billb2
Use the polymer to "glue" two pieces together, then pull them apart.
The problem would be in how to deal with the area of polymer that gets separated from the metal. Probably different with each test.
Use a thin stripe of polymer of a known and repeatable contact area and hope that all the polymer separates when pulled apart.

Interestingly enough this idea came up as I was talking with someone, but I spoke to my professor and he said no :x Something about something (I don't want to say anything I'm not sure on since he was in a hurry and didn't really explain what he said)...

And PolymerTim...when I meant indirectly I meant using them for other analysis in general. Simply measuring the adhesion force would produce a poor design project

Actually I have a VERY thin layer for this stuff...I don't remember exactly how thick it is but I contacted my partner who was actually coating these things to get back to me....let us say that if it is thicker than a millimeter I'd be surprised.

Cyclowizard - awesome send the paper my way as per my PM When I mean mesh - lets just consider a laser cut hollow cylinder.

I've been stressed with some other stuff so I haven't had terribly much time (7 classes :\ ) but i'll look into all the methods stated today. Thank you all so much!

 

[magomago]

Originally posted by: wwswimming
if you put a flat on 2 of the cylinders and then bond them at
right angles to each other using the test polymer - will this
configuration give you an useful answer ?

What is a "flat"? Kinda confused there =)

 

[PolymerTim]

Hmmm, I'm curious what the total contact area is of the polymer and cylinder relative to a smooth cylinder. It occurred to me that your description sounds almost like a coronary stent. Is that what you mean by a mesh? If so, it really changes the way I view the problem from something more solid with a textured surface. If the cylinder is solid, are the gaps in the mesh deep enough that the plastic coating does not go into them, but rather is suspended over the gaps? I realize you may not be allowed to fully describe the problem, but the more you can describe, the more likely we are to be able to help.

With a thin polymer film I think you will definitely need to consider the elasticity of the coating. As CycloWizard mentioned, you can account for this in models, but it may be easier to use a technique where the film is not significantly deformed in the first place.

 

[magomago]

PolymerTim...you do have the right idea That is exactly what I'm talking about. And thanks for understanding I can't describe too much otherwise I wouldn't be allowed to get to play with some tool tools

But I can say that we can consider it like every other stent on the market with a coating on it - meaning that it only covers the surface of the metal.

That is what makes it so difficult because the actual surface area I have to play with it ridiculously tiny...

 

[PolymerTim]

Hehe, I do understand since this field is usually supported by biotech firms that really don't want to let their IP out. While I do not work in the bio area, I can say that I am intimately familiar with some of the research on stent coatings (I too will leave it at that). Heck, the company that sponsored the research I am familiar with has not even publicly announced they are attempting what has been in research for half a decade, even though most of their competitors are likely developing very similar technologies. There is a ton of literature out their on all aspects of the research, but it seems that everyone just pretends that they don't know what the application is, even though the acknowledgments give a pretty good clue.

So this setup is quite different from what I believe most of us were picturing. Not only do you have the complexity of the cylinder, but (and much more importantly) the coating is either wrapped around the wires of the mesh or possibly restricted to the abluminal surface. That brings us to the question of experimental design and what you are really trying to measure.

I'm not familiar with any adhesion testing in this area. Most people are more concerned with the toughness of the coating since, if no cracks are formed during expansion, then delamination is not an issue. I think it will be important for you to try to create an adhesion test that accurately replicates that type of failure you are trying to study in the actual system. Whatever mode of failure you are analyzing, I'm guessing that it won't be well predicted by a peel method. Also, given the small gauge of the wire for the mesh and the thickness of the coating, I'm concerned about even the possibility of a traditional peel test. Typical dimensions I have seen are wires in the mesh being less than 100 micrometers in diameter with the coating probably being less than 10 micrometers in thickness (although I'm sure coating thickness can vary widely).

So my thinking is that peel tests are out. It's probably not possible to "glue" to pieces of metal together with your polymer since the process of applying the polymer to the metal probably can not be reproduced between two pieces as opposed to having a free air surface. Due to the fragility of the coating, I think you will need a method where you do not directly handle the coating, but rather the metal itself.

Possibly you could perform a deformation on the mesh (or wire) itself. If you want to remove the effect of polymer toughness on the measurement, you would need to start with a precrack in the film. Then you would need to deform the mesh/wire in a controlled way to allow stress to transfer through the metal into the film. You should be able to visually verify adhesive failure. Getting to an accurate quantitative analysis of the adhesion would require some pretty in-depth understanding of stress transfer and the polymer mechanical properties and may be out of the scope of your project, but you should be able to develop a qualitative method of rating various samples. The ideal situation would be a simple, controlled, and consistent deformation to the wire that somehow results in a range of stress from high to low near your premade crack. Then you can rate the sample based on how far the delamination propagated before the adhesive force was strong enough to overcome the stress transferred to it by the deformed metal. This is similar to the conical mandrel test except that the conical mandrel test is primarily for determining ultimate strain (or flexibility) of coatings on panels.

I realize there are still some major hurdles in the above suggestion, but I hope it at least gives you some insight into the problem. Good luck with your project and keep us updated.

 

[Sho'Nuff]

If you have one available, I think you can do this with an Atomic Force Microscope.

 

[PolymerTim]

Actually, he mentioned in one of the posts that an AFM is available. I have heard these can be used for certain types of adhesion tests, but I am not familiar with them. If you are, could you elaborate on some of the techniques that might apply to his geometry?

 

[Sho'Nuff]

Some reading:

http://lib.tkk.fi/Diss/2003/is...845/isbn9512267845.pdf

http://www.freepatentsonline.com/6099939.html (the patent can be pulled from the PTO website, www.uspto.gov)

http://polymers.msel.nist.gov/...ject-detail.cfm?PID=73

 

[iwantanewcomputer]

shouldnt you be able to make this on a larger scale, and flat, and use the geometry of the curve to scale it down?

also, think about the speed and temp you want to test it at...the polymer may exhibit much different behaviour at high speed than low do to ductle/brittle transition

 

[CycloWizard]

Originally posted by: PolymerTim
Actually, he mentioned in one of the posts that an AFM is available. I have heard these can be used for certain types of adhesion tests, but I am not familiar with them. If you are, could you elaborate on some of the techniques that might apply to his geometry?

Again, it depends on what he's trying to measure. AFM can be used to measure adhesion, but only if you can find an AFM tip available in one of the materials you want to examine (e.g. bonding plastic to metal - find a metal tip and adhere it to the plastic, then pull away while measuring force and displacement). The nice thing about AFM in this case is that it relieves problems related to curvature sinceresults from indentation-based techniques are very weakly affected by curvature. They will be even less affected in this case since the probe dimension is far smaller than the radius of curvature of the sample such that you can treat it as a semi-infinite half space.

 

[silverpig]

Can't you just roll out the pattern into a new material, make a mold, then deposit your metal into the mold?

Roll your cylinder across a soft membrane to make an imprint that is as wide as the cylinder is long, and however long you want. Then find some way to deposit your metal onto this mold so it retains the pattern. Then use this flat version of the cylinder to do your tests. You wouldn't have to deposit several millimeters of metal... just enough to coat the mold. You could evaporate several microns of your material into the mold, then coat the back of the metal with some hard polymer with better adhesion than your test polymer. Then peel the mold from the metal side, coat it with your polymer and do your tests.