can this be difficult to tell whether the delamination was actually the cause?
Helisphere
Thank you for taking the time to read the paper.
It is possible by visual inspection to determine the type of bond failures present, and there are essentially two types: Cohesion, where the adhesive is fractured and residue remains on both surfaces, and adhesion, where the adhesive separates from one surface. Most structural film adhesives contain a "carrier cloth" which is a supporting cloth added during initial adhesive manufacture to aid handling. Cohesion failures usually fail through the plane of the carrier cloth because this is the weakest plane in the bond, provided the interface is strong. Cohesion failures require higher loads than adhesion failures, and in extreme cases adhesion failures can occur with very low loads.
Adhesion failure is caused by time-dependent degradation of the chemical bonds at the interface between the adhesive and the substrate (in this case the metal). The bond may start off with a high strength but over time the strength decreases, and usually the mechanism is that the surface of the metal forms a hydrated oxide and in the process the chemical bonds between the metal and the adhesive layer dissociate to enable hydration to occur and disbonding results. Failure naturally occurs at the weak interface, so all the adhesive remains on one surface and none on the other.
There is a third possibility; that failure occurs before the interface is fully hydrated, and this results in a mixture of cohesion failure and adhesiuon failure, commonly termed "mixed-mode" failure. The proportion of adhesion failure will increase with time as the interface hydrates, so if the structure is highly loaded at an early stage, the failure load will be higher and the proportion of cohesion failure will be higher. The converse is also true, such that the longer the bond is in service, the weaker it becomes, so failure will occur at a lower load and there will be a higher proportion of adhesion failure, and the locus of failure will migrate from the plane of the carrier cloth towards the interface.
The rate of strength loss depends upon a number of factors, the most important of which is the process used to prepare the surface during the factory bonding process. If the process produces chemical bonds which resist hydration. then the bond will last a long time. If not, then a shorter life will result. Other factors include higher operating temperatures and humidites, both of which will accelerate the hydration process and hence shorten the service life of the component.
Now, of all the tip disbonding examples I have seen there is a high percentage of adhesion failure.
Erosion by itself can not cause adhesion or mixed mode failure in otherwise sound adhesive bonds. Hence, the entire discussion about which tape to apply is meaningless unless the bond failures are confirmed as pure cohesion failures.
I have also observed adhesion failure on one crashed blade at a number of locations well inboard from the tip where erosion is not evident at all. (Release of the report from a certain Pacific country is being delayed internally.)
The real problem is that it is remarkably difficult to categorically point to a disbond and say that this was without doubt the absolute primary cause of the accident. This is because even though the bonds exhibit failure modes which are characteristic of low bond strength, the loads which caused that specific bond to fail may well have been associated with an event which occurred as a consequence of another causal event, but equally they may be the causal event. There is rarely any marker such as fatigue striations to give a definitive identification. Occasionally, there is discolouration of the adhesive due to the presence of liquid water, and this would positively indicate a pre-crash defect.
Hence, it is necessary to eliminate or ameliorate all other probable causes and to look for circumstantial evidence to support the hypothesis that bond failure caused the crash. For example, if the pilot had thousands of hours with that type, had a perfect medical record and post-mortem assessment confirmed the absence of alcohol or drugs, then it is reasonable to suggest that the probability of inappropriate pilot control input is extremely low.
Naturally, weather conditions would be taken into account, as would mission profile. If the aircraft was flying level on a set and commonly used transit route on a bright sunny day with no cloud and low winds, then it is reasonable to infer that weather and flight mission were not a factor.
Extraneous causes such as bird impact or impact with items departing the aircraft would also need to be eliminated.
Next you would examine other causes which may result in a loss of blade tracking. Most of these leave tell-tale evidence, so if that evidence is absent, then it is possible to suggest that the probability of those causes is low.
Next you would examine all bond failure surfaces. All cohesion failures can usually be discarded because the bond would have exhibited an adequate strength. Adhesion failures are important because they often indicate a defect which existed before the crash event. For mixed-mode failures, the proportion of adhesion failure AND the proximity of the failure surface to the interface are both strong indicators of weak bonds. (Often safety investigators use a scanning electron microscope to find the presence of ANY adhesive and incorrectly identify that as a cohesion failure.)
Next, you would look for any history of bond failures in the specific component, and in the case of R22/R44 there are a number of examples where aircraft have landed or crew have survived crashes with tip disbonds. If then, the component in question exhibits characteristics of bond degradation and hence a loss of bond strength, then the probability of bond failure causing the event increases. If the extent and distribution of bond degradation is high, then so is the probability that bond failure was the cause of the event.
Then it comes down to a balance of probability that bond failure was the cause and not the result.
Regards
Blakmax