Conclusions on Aluminum Adhesive Bonding Tests

[GESchwarz]

I'd like to point out that many people in this forum have supplied key information, the absence of which would have stalled or killed this learning project. I won't name names because they are all here for all to see. In the last several days my attention has turned to the bonding of PVC foam ribs and formers to aluminum skin. It was tough getting started finding information, a lot of which seemed to have been bottled up in Europe. But there have been key breakthroughs provided by members of this forum. As a result I have accumulated enough data to begin a new test program to determine the equivalency between rivited aluminum ribs and bonded PVC foam ribs. There are several different aircraft designs and hundreds, if not thousands of planes in service, mostly beyond our borders, that are constructed with bonded foam ribs.

I have sample material being delivered to me as I speak. Looking forward to seeing how this stuff performs. From what I've learned already, this construction method has some distinct advantages over rivited ribs. I'm not going to talk about it until I have proven those things out first-hand. As with riviting, there are unique skills and processes required for bonding too.

 

[lr27]

Can't argue with that. Obviously a bunch of careful tests like Gary's are worth a lot!

Would point out that, regardless of the amount of stuff I remember about it, Strojnik's bonding was tested in real life. As was Boeing's, the Tiger's (I see several at my local airport), etc.

Still, at the moment I think we have more detail on what Gary has done, which makes it more valuable.

Hello, Ir27;

snip

It seems to me that Gary's work has pointed in the direction of an adhesive that should perform well when used in a home workshop. His work is not based upon conjecture or vague memories of what someone might have done. His testing has produced real data under conditions which he can reproduce.
snip
MalcolmW

 

[GESchwarz]

The following are some findings from preliminary testing with PVC foam and applies to bonded joints in general.

Rules for Bonded Joint and Foam Design


Prevent stress concentrations at the edges of the bond joint. To perform to it’s potential, stress must be distributed across the width and length of the bond joint.

Bond lines should form a closed loop. Open loop bond lines are vulnerable to peel because cantilevered loading is introduced to the joint. Closed loop joints maintain the entire bond line in shear. A closed loop prevents peel at any particular location, by shear strength of the joint adjacent to it.

Open loop bond lines or bond lines that otherwise have terminations are unprotected from cantilever or peel load vectors, will fail at the termination.

The substrate material, by geometry or thickness, must be stiff enough to not distort, deflect, twist, bow, or bend under design loading. Such movement directly results in localized stress and strain concentration on the joint and/or foam.

 

[BBerson]

What is "open or closed loop bond".
I have not heard of this.
BB

 

[dino]

The generic method used for preparing aircraft structure for bonding is phosphoric acid anodizing. The Boeing spec is BAC5555 and is similar to ASTM D 3933

Dino

 

[GESchwarz]

By "loop" I mean that there should be no bond line terminations that are vulnerable to peel. Take for instance a nose rib. If a nose rib is bonded continuously around its circumference from leading edge, to top skin, to spar, down to lower skin, and back to the leading edge; that is a closed loop bond line. If the upper and lower skins continue aft and are well attached aft of the spar, then the closed loop bond line described above is protected from peel under design loading, provided that the other rules are followed, namely that the substrate has sufficient stiffness.

Another typical loop are the bond lines that describe the individual wing rib bays... Each loop takes the load applied within that circumscribed area. A skin wrapped around from trailing edge to leading edge and back to trailing edge has no opportunity to introduce a peel vector to the rib caps except along the trailing edges where you would install rivets. The rivets close the loop of an otherwise open loop bond line.

An open loop, or a bond line that otherwise has terminations, is vulnerable to peel. Take for instance a common "L" angle bracket. Tensile forces applied along one edge of the bond line makes this joint prone to failure. A better design than an "L" bracket would be a "T" bracket which better distributes the load from the center of the bond outward, whereas the bottom of the T is the cantilevered part. Again, the material must be thick enough to not distort under design loads.

The idea is to design the joint so that the load is distributed as evenly as possible across the total bond area. United it stands, divided it fails. Poor joint design results in a stress concentration that easily peels a good bond.

Tensile deformation (stretching) of the aluminum substrate at the edge of the bondline is typically where failure begins. It's this stretching that the adhesive and foam cannot tolerate. And it happens in a zipper like fashion. Design your joints so that can't happen. A skin attached as described above cannot peel and unzip, therefore it is a viable bond joint design.

 

[BBerson]

OK, I see what you meant by closed loop. Is that normal terminology used by the bonding industry?
BB

 

[GESchwarz]

Sorry, but it's not. It's a term that came to mind that I thought helped describe what I'm thinking.

I do have an industry term that I just learned yesterday while watching one of several programs on the Discovery channel. This one program was titled, I think, "Accidents That Changed Flying". It gave several examples of innovations that were born in the ashes of airline crashes. Well anyways, this term or name is Scrim. Scrim is just what it sounds like and it is used in adhesive joints to maintain a minimum bondline thickness. It is a very open weave material of a certain thickness that simply acts as a spacer without getting in the way of the adhesive so it can do it's job. Earlier in this post I wrote of using "Tule" a fabric for the same purpose.

 

[MalcolmW]

Hello, Gary;

Yes, I saw the same ‘Discovery’ program and was dismayed at the adhesive failure – shouldn’t have happened. However, a point of clarification: The use of ‘scrim cloth’ in structural adhesives has nothing to do with bond thickness or holding the two substrates apart.

When high-performance adhesives, such as epoxy-novolac, are made (as used on the F-111 fighter-bomber), they are ‘B-Staged’ or partially cured to the ‘gel’ stage. At this point, their viscosities are very high which makes them difficult to spread. So, the adhesive manufacturer spreads the adhesive on a fiberglass scrim cloth and passes it between rollers. This produces a thin film of adhesive, which is contained between two sheets of polyethylene film. The adhesive is immediately chilled with dry ice to arrest further curing.

Shipped cold to the manufacturer, the adhesive is warmed to room temperature for use, cut to the precise size of the adhesive joint, stripped of the polyethylene film and applied. Epoxy-novolacs require an elevated temperature to cure and bond very strongly to properly prepared aluminum (covered in earlier posts) and have a very long service life. However, they are not the easiest adhesives to use, and I don't recommend them for home-built aircraft.

So, the scrim cloth is a 'carrier' and has nothing to do with bond thickness.

Keep up the good work, and keep posting.

All the best & fly safe,
MalcolmW

 

[GESchwarz]

Boy, if anyone could catch that, it'd be you Malcolm. Thanks for the correction. So what is used to hold minimum bond thickness...viscosity, microballoons?

 

[orion]

The composites industry uses a couple of mechanisms to maintain bond line integrity. The scrim sheet is actually one, especially in applications where you have significant clamping pressure where the contact magnitude could actually squeeze out all the available bonding agent. This however is not common since it has been shown that the scrim material can actually increase several detrimental mechanisms that would decrease the bond line integrity.

In our case here we use the epoxy viscosity to assure ourselves of a particular bond line thickness, coupled with vacuum bagging where the amount of vacuum pulled is limited so that excessive pressures are not generated. The epoxy is mixed slightly warmed so it is spreadable - a notched squeegee is used in order to create a finite buildup on both surfaces. As the mixed epoxy contacts the surface it cools, increasing its viscosity to a very thixotropic state.

The parts are then mated and a vacuum bag is applied. The initial vacuum is allowed to reach about 22" (or about 11 psi). This is maintained for about 5 - 10 minutes after which point the vacuum is decreased to between 10 and 15 inches, depending on part geometry and the gauge of the mating components (too much pressure and the parts could distort during cure). In this way we maintain an average bond line of about .005" to .010". The only drawback to thin bond lines is that they do take a few days to reach a dependable handling strength so the parts must not be removed from the tool too soon or damage could occur.

 

[GESchwarz]

One of those detrimental mechanisms is that the scrim itself is displacing adhesive...it adds nothing to the strength of the bond and could even serve as a built in stress riser. It's interesting what you said about cure rate in thin bond lines. i have noticed that. Is it due to the fact that very little chemical heating is able to be generated to facilitate cure. Is cure rate always proportional to temperature?

 

[orion]

Yes, it seems to be. If you look at a lot of room temperature cure epoxies, handling strength cures can vary significantly as a function of environmental temperature. What may take up to three days at 70 deg. can be sped up to only a couple of hours at 160 deg.

And full strength is even more significant. Most epoxies and even Vinylesters used for secondary bonding may require several months to reach full cure. But take those resins and cure them at 200 degrees and you might be able to achieve that full strength in an hour.

But the chemical heating can also be a secondary indicator, being a function of the rate of molecular attachment - as it was explained to me some years back, the thin bond lines will have a tendency to have fewer free molecular chains so the mechanism of the physical attachment just takes longer. The pressure of the contact surfaces also seems to slow the reaction, although that effect is not as significant as that of the thin bond lines and cooler temperatures.

 

[MalcolmW]

Hello, Gary, Orion;

A very interesting discussion of adhesion & bonding agents.

First, the curing of adhesives is a chemical reaction and rates of chemical reactions roughly double for each 10 degrees Celsius rise. So, Orion, your comments are on the mark and consistent with my experience testing adhesives. A slow rise in temperature to the optimum cure temperature usually produces a slightly stronger bond.

Now it gets a little more complicated. The thickness of the bond line – of a well-designed adhesive joint – has little to do with its ultimate yield. A uniform bond line on two substrates that have been properly cleaned and passivated is stronger than one of varying thickness. This has to do with uneven stress loadings and susceptibility to vibration, etc. of uneven bond lines. Scrim cloth has no material role in the strength of a well-designed joint – its role is one of convenience, i.e., to make the application easier and more consistent.

When it comes to molecular bonding, the thickness of the bond line is massively larger than the molecules which bond (think Van Der Waals forces) the two substrates together. A good adhesive match for maximum strength depends upon chemical compatibility, (this topic is beyond the forum on homebuilt aircraft). Suffice it to say that certain adhesives bond more strongly to certain substrates than others. In the case of aluminum, there are a number (identified by Gary) that have more than adequate strength for aircraft construction. For example, polyurethane adhesives work exceptionally well with polyester composites, for which there are sound chemical reasons.

As for the pressure upon the joint, I seem to remember that hydraulic presses were used, but at what specific applied surface pressure I don’t remember. In the case of bonding aluminum skins to an aluminum honeycomb core, the pressure was fairly high. Lap joints were subject to lower pressures – but what, I don’t remember.

As I said previously, I have no expertise in aeronautical engineering, but do have some experience with adhesives.

I hope that helped.
All the best & fly safe,

MalcolmW

 

[Cazador]

Hello Gary and everbody.

I would like to know if you tested the Araldite®AW106/HV953U combo and (if you effectively test it) which are the conclusions you obtained from those tests. Araldite is a two component epoxy paste adhesive a multipurpose, room temperature curing, paste adhesive of high strength and toughness made in Switzerland. I'm interested in the strength and durability of a bonded aluminium alloy - Divinycell H100 joint. Any info you can provide will be welcomed. Thanks in advance.

Edgardo.-

 

[GESchwarz]

I haven't looked into that combo, but I will. Thanks for the tip.

I did just receive samples of four different densities of the PVC foam. I hope to be able to put some coupons together in the next few days.

 

[MalcolmW]

Hello, Cazador; (is that 'hunter' in English?)

as to your question about bonding / joining aluminum to vinyl foam, that is a question about two materials which have quite different coefficients of thermal expansion. With temperature cycling, any adhesive which has a significantly different coefficient of thermal expansion (glass transition temperature differentials) between the two substrates will be subject to significant stresses, perhaps sufficient to cause the bonding material to delaminate.

Consequently, when bonding two quite different material, the preferred adhesive is one which is quite elastomeric, i.e., 'stretchy' and is not subject to fatigue or hysteresis aging. One such adhesive class could be polyurethane adhesives, particularly polyester polyurethane adhesives.

As for epoxy adhesives, I believe that their glass transitions temperatures may be too high to withstand the expansion differentials of such dissimilar materials and may succumb to 'fatigue' and develop cracking and bond failure.

In addition to polyurethanes, you might also consider the polysulfide (ProSeal) adhesives / caulks which provide a durable, elastomeric bond to both aluminum and vinyl.

Have you ever considered using aluminum honeycomb ribs in place of vinyl foam? There are vendors of aluminum honeycomb sandwich panels which have tremendous stiffness and light weight (and are used in the manufacture of commercial and military aircraft), and provide very long performance lives.

As alway, as Gary Schwarz has done, do some testing to verify that these adhesive meet the performance requirements of your proposed application. Do be careful, for flying in an aircraft that falls apart would certainly ruin your day... right?

As always, fly safe,

MalcolmW

 

[john14]

Hi Gary,
Excellent work. On the French Cri-Ci, the aircraft uses the exact same aluminum to rib structure you are looking at doing. Type of bonding is Araldite 420 A/B - Epoxy Adhesives - Kirkside Products Perth WA
In the prep for aluminum PDF, I believe that decreasing / abrading the surface is all that is needed. This is also what is recommended when building the Cri-Ci, which has +20 years of field history.
Is your conclusion for applying the Partite 7350 the same (decreasing / abrading)?

 

[MadRocketScientist]

Speaking of the CriCri, I have been keeping a close eye on this thread as I am in the process of building one. The CriCri manual recommends that the aluminium be scoured manually. This is then primed with a primer that is compatible with the epoxy being used. It also goes on to say the alodining should not be used where the parts are epoxied together due to the lower bond strength between the alodined surface and the primer.

Shannon.

 

[flyoz]

Colomban also suggests using nylon/cotton as a scrim spacer when he uses pultruded carbon fiber to reinforce the alloy with an epoxy joint . Allowing the CF to contact the alloy is not a good idea .
Flyoz