Ah! This is what I call "Grandma's cup syndrome"! They once glued the handle back on Grandma's cup and sometime later the handle fell of because the glue failed. Therefore that is evidence that ALL bonds are susceptible to failure. Regrettably the history of poor bond performance in real structures backs up this assertion. Therefore, bonded repairs must be designed on the basis that the bond will some time fail.

The essence of the problem is how to verify that an adhesive bond will sustain its strength throughout the life of the part. Traditionally, the verification of adhesive bond performance has been based on strength or fatigue tests. Indeed, the FARs are structured in such a way that airworthiness is demonstrated by static strength tests, fatigue tests and damage tolerance analysis. However these type of tests have failed to prevent disbonding in bonded structures in later service.

The reason why current methodologies have not prevented bond failures is that the tests do not interrogate the predominant mechanism for in-service disbonding. Adhesive bonds depend on two critical factors: (1) the ability of the bulk adhesive material to sustain static and repeated dynamic loads within the operating environment and (2)the ability of the interface between the adhesive and the substrate to also sustain those loads whilst resisting environmental degradation. Current tests only address the first factor.

The most common mechanism for interfacial degradation for metal bonds is hydration of the surface oxides duri9ng service environmental exposure. For example Al2O3 formed during surface preparation has an affinity to form a hydrated oxide Al2O3.2H2O. For that reaction to occur, the chemical bonds between the adhesive and the surface oxides formed at the time the adhesive is cured will dissociate so that hydration can occur. Hence, the bond fails at the interface. Strength and fatigue tests do not interrogate the resistance of the interface to hydration. Worse yet, because the interface strength is degrading, the results of damage tolerance testing and analysis is meaningless because the fundamental implications of damage tolerance is that the surrounding bond sustains pristine strength and that is simply not true if the interface is hydrating.

The ONLY way to prevent metal bond strength degradation at the interface is to use a preparation process that provides resistance to hydration. A very effective method for evaluating surface preparation processes is the wedge test ASTM D3762, but the acceptance criteria need to be tightened (see http://www.tc.faa.gov/its/worldpac/techrpt/artn065... ). The FAA has a program aimed at amending the standard.

Now, there are four ways a bonded repair can fail: (1) the repaired damage may propagate (2) the repair doubler itself may fail and (3) the bulk adhesive may fail through static or fatigue loading or (4) the interface may fail. The first two cases are easily handled by structural analysis and design methods already in wide use. The last case has been addressed above, so the remaining issue is the design requirements to prevent failure of the bulk adhesive. What is not widely recognised is that it is possible to design bonds for thinner structures (typical of those being repaired using adhesive bonds) such that the adhesive is never the critical element in the joint. In other words, the structure will be weaker than the bond. This approach is very powerful for repair design, because if the interface is hydration resistant, and if the bond is always stronger than the structure, then the locus of failure will NEVER be through the adhesive bond.

Hence, if a few fundamental rules are followed, then there is absolutely no reason why adhesive bonded repairs need to be considered purely cosmetic or simply applied as a fatigue enhancement.

Regards

Blakmax

I don't know the application (and hence the requirements), but the spirit of the approach is that any surface that requires preparation before bonding may be subjected to such limitations. This is the statement from AC20-107B, which is for composites, but also discusses metal bonds:

- Structural Bonding. Bonded structures include multiple interfaces (e.g., composite-to-composite, composite-to-metal, or metal-to-metal), where at least one of the interfaces requires additional surface preparation prior to bonding.

- For any bonded joint, § 23.573(a)(5) states in part: "the failure of which would result in catastrophic loss of the airplane, the limit load capacity must be substantiated by one of the following methods—(i) The maximum disbonds of each bonded joint consistent with the capability to withstand the loads in paragraph (a)(3) of this section must be determined by analysis, tests, or both. Disbonds of each bonded joint greater than this must be prevented by design features; or (ii) Proof testing must be conducted on each production article that will apply the critical limit design load to each critical bonded joint; or (iii) Repeatable and reliable non-destructive inspection techniques must be established that ensure the strength of each joint."

Brian
www.espcomposites.com

Bazzo,
Process, process, process. Every step of the repair requires an effective process that accomplishes what it is intended to do, and the sum of all steps must guarantee a reliable joint. Any missing steps will be a chance to allow the repair to fail. Many of those steps will not be explicitly documented in the existing aircraft repair books, meaning you will have to develop your own (unless you already have this data available, and suitable equipment, environment, and personnel capable of accomplishing the processes you need).
Contaminants are the enemy, and if you aren't able to control them (water, solvents, oils, salts, dust, etc) then you can't accomplish what you need to do. Which is what typically what leads to failures, or "grandpa's cups" as they were called.

Today, and in the past, mechanical fastening methods are preferred and bonds are fraught with difficulties, but consider the circumstances: Shop environments were impossible to control so strictly, and the training required to make a processor competent enough to produce reliable bonded parts was unthinkable. As time goes by, though, we find ourselves competing against designers and fabricators who push the limits of weight and performance, forcing us to raise our own standards. Like all other technological advances, eventually, the ways to make bonds reliable and long-lasting will sink in to engineers' heads, be used more regularly in their skills, and be taught more widely to students.


Blakmax:
I read your paper last month. Slowly. It's starting to sink in. Thank you.
I think I've committed half of the sins you wrote about.


EPSComposites:
You've quoted just the part I wanted to see. So, the FAR's still don't directly address the process control when adhesives are used to bond joints. It is still inferred from paragraphs like 25.605, 25.613 and so on, and from there you have to read through advisory circulars and a few policy memos before building up a coherent picture.

STF

Brian,

I am well aware of AC 20-107B and I am also well aware of its limitations. Indeed AC 20-107 was updated from A status to B status as a direct result of A08_25_29_Recommendation.pdf and I can claim some credit for that work. Unfortunately the redraft of AC 20 107 only concentrated on adhesion (interfacial) failures. The real crux of the problem is not limited to Cohesion failure (fracture of the adhesive) OR Adhesion failures (pure interfacial failures). There is a third failure mode, which is mixed-mode failure, where there is a combination of adhesion and cohesion failure, adhesion where the interface has degraded, and cohesion where the remaining bond can not sustain the load so the adhesive fractures. It is the transitional zone which is of importance because that is where the interface is degrading but has not reached the extend of degradation that would result in adhesion failure.

Summary: Cohesion failure is strong, Adhesion failure is very weak (no strength?) and mixed-mode is the transition from cohesion to adhesion where the strength is degrading but the level of strength loss depends totally upon the level of degradation. (For metals read Degradation as Hydration).

I say again, that the regulatory framework and AC20-107 if read in isolation beside the FARs still presents the situation where it is possible to certify an adhesive bond which has a definite probability of failure in later service. Static strength testing, fatigue testing and damage tolerance analysis simply do NOT interrogate the resistance of the interface to degradation. It is only when you read the FAA Policy Statement PS-ACE100-2005-10038 that you actually encounter the recommendation for the use of the wedge test to interrogate interfacial resistance.

Spar Web. Thanks for reading the paper. However, the major issue with Grandpa / Grandma's cup is not just contamination. It is almost exclusively driven by selection of a process which has absolutely no chance of survival in a real environment. Changing the adhesive simply changes the colour of the disbond, not the result.

I have since 1997 been advocating a Rule Change (see http://www.abebooks.co.uk/products/isbn/9780958531...). However, I ama aware of the difficulty the FAA faces in even suggesting a rule change. I believe that the current preferred method for addressing this issue is to draft another AC which specifically addresses the issue of bond degradation.

I have published a number of papers on this topic at my web site. Google Adhesion Associates and you will find it. Follow through the resume until you find the list of papers.

I say again, if the issue of hydration is addressed and validated correctly, and if the bond is designed such that the adhesive bond is always stronger than the parent structure, then the "grandma's cup" syndrome has no place in a truly scientifically based bonded repair technology.

The real problem is to get over the grandma mind-set. For goodness sake in the latter part of my career I had to deal with a blinkered philistine who insisted that bonded repairs should never be applied over active cracks. The cracks must always be removed, because no aircraft is certified with known cracks. (OK, so Damage Tolerance assumes that the structure is cracked, doesn't it?) When presented with test data that showed that "stop drilled" cracks performed marginally worse than unmodified cracks, and samples where cracks were routed out produced the SHORTEST fatigue life, the blinkers were turned on and "but we don't certify structures with cracks" became the mantra. (I'll elaborate on the mechanisms behind this data if requested.)

So, if I can demonstrate that my bond is resistant to hydration, and I can demonstrate that my bond will always be stronger than the parent material, and I can demonstrate that if my bond is hydration resistant then this is the ONLY case where damage tolerance is valid, and my design precludes conditions where the adhesive stress levels lead to fatigue, then my bond should never fail under any circumstances.

So AC 20-107B and the FARs need to be amended to recognise these conditions, or there needs to be a RELIABLE AC to provide valid guidance in adhesive bonding technology such that Grandma or Grandpa never have to worry about their cup.

I am happy to elaborate on any point.

Regards

Blakmax

Max, I agree with your comments about AC 20-107B. But that can be quite an involved discussion, as you mention. I was pointing out that the current certification approach is related to the prepared surface and not the adherend itself. In that light, AC20-107B helps to provide guidance. Hopefully the FAA can make things more clear in future revisions.

Brian
www.espcomposites.com

Thanks for the comments Gentlemen, the AC link was a big help.
Exactly, the problem lies in the process, bonding is like baking, the result varies and depends on the baker, repairs done by an Airline in deep Timbuktoo will produce another result than a “professional” MRO following lab standards. Until we can even that out we have to design and justify safe for any SRM repair. The key was as you all said test pieces, which operator will go to the trouble of producing test pieces and keeping his a/c on the ground until it’s been evaluated?
I cannot see In-Service bonding ever being Structural.

Quote:

Bazzo...I cannot see In-Service bonding ever being Structural...

Under certain circumstances, it can be. Very tightly controlled circumstances.
I was once shown the operation of a heavy structural repair shop that was fully capable of refurbishing, repairing, or almost completely rebuilding composite structural panels used in a common type of helicopter. Bonded structural joints in primary structure are not a new idea. They have been flying successfully since the '60's and many are still in the air. The example I'm referring to are "Hueys" in common speech, but hardly the only example.
The facility was equipped with the necessary equipment, and a crew of people who knew their jobs very well, and were all part of developing the procedures they need for each style of bond, series of parts, and degree of repair required. Of course, the paperwork put ISO9000 to shame, but that's to be expected when the degree of scrutiny on every step is so high.

STF