Finish Flaking Off at Butt Joints

Competing theories on why finish has failed at the joints on a cabinet door. December 8, 2012

I would like some opinions on the photos below. These are Shaker cabinet doors with 3" stiles and rails, made of hard maple. The photos show the joints of the stiles and rails, and as you can see the finish is flaking off. It is typical to see cracking at the joints due to the expansion and contraction of the wood, but I have not seen the finish actually coming off like this. I contacted the finisher and of course he tells me that it's the door manufacturer's fault - that the joints are loose.

There is no indication that they are loose - the joint is tight with no visible gap. We outsource both our doors and our finishing, so I will not be offended by any response. Regardless of whose fault it is we will need to address this for our customer, so I would also be interested in hearing opinions as to what can be done, whether it can be fixed or will need to be stripped and refinished.

Forum Responses
(Finishing Forum)
From contributor R:
It appears to be a finish problem. It's always a tough call, but my guess would be the finish was applied too quick after the stain. The stain will slip down into that joint and get sucked up the end grain, it dries slower there.

My second guess would be (and this might be the case if they've been in use for a long time), the normal breaking of that joint through wood movement (just a small crack in the finish) that would allow moisture to get in if someone used a wet rag often enough. One would see it on the end grain of the doors and the panel joint as well.

From contributor G:
I had this happen to one door out of about 50 on a kitchen job. It was only at the glue joint. It was the strangest thing, I have never seen it before. I was the finisher and the doors were outsourced. I never really figured out what it might have been but I guessed that somehow it was the glue. It removed everything right down to the raw wood. This was an extensive finishing schedule for the maple. I conditioned the wood, stained it, applied three coats of shading lacquer to develop the dark color, and then two coats of clear.

From contributor R:
I'm sorry I never got around to the fix it part. How many doors? If it's more than just a few and it's happening to most I think it's either the case of stripping it or new doors. Fixing a few wouldn't be a problem, however the customer has seen it so they'll always see the fix too if they are picky. No way to perfectly fix it in my opinion but if the customer had never see them I think there are several fixes that would have done the job satisfactory.

From contributor U:
Any info on the products used and/or the finish schedule? Was this a spray or wipe stain? Was the stain mixed in with the clear sealer or lacquer? Also, who was responsible for sanding the doors?

From contributor E:
I have seen this kind of stuff happen on hard maple. Sometimes the wood gets sanded too smooth and actually becomes polished, which causes adhesion problems. Sometimes you will see it happen when the edge gets cut, and when you have a real sharp edge the finish just can't hang on. That's why it's important to break sharp edges off anything that gets finished. In this case you had a solid film covering the joint, and then when the joint "popped" apart it took the finish with it.

From contributor T:
I am a finisher. That said I am going to say that this problem breaks 50/50 between the two parties. First off, Contributor E is right on with breaking edges on materials before finishing. The reason he is correct is that finishes are a "film" when cured. When sprayed on a broken or rounded edge the film has a better chance of maintaining its thickness as it wraps around a rounded or broken edge.

A sharp or unbroken edge will give the absolute minimum film build on the edge and that is where the finish will breakdown first, regardless if it's cut or not. However, that is not what happened here. In the photos you have provided it appears that all the failure happens at perpendicular joints. It also appears that the finish is a multi-step spray finish. In my opinion, what happened here is that those particular areas were not sanded absolutely flat in relation to the connecting piece. It does not take a lot of deviation here. The most critical part of finishing is sanding. The most critical part of sanding is touch, plain and simple. Your fingertips can pick up the slightest deviation in joinery, believe me.

So when it was finished there may have been some glue residue or the joinery was off just enough for the film to bridge but ultimately not to last. It's true if those cabinets are wiped down and wetted over time that it will break down the finish at that joint.

I cannot tell from the picture, but those doors do not appear to be brand new. If it were me I would have your finisher try a repair and touch-up at the joints and present it to the client for approval. If the doors are newer then I would have to say sand them back and refinish.

From contributor O:
If the finish is flaking off only at the joint then most likely it is wood movement that is causing the problem. The best bet is that the joint does not have a continuous glue line. Where there is no glue the joint is free to move with changes in humidity or the introduction of moisture from something like a cleaning liquid. Wood tends to move more tangentially across the grain of a flat sawn piece of wood than radially, and in this case the thickness of the wood, so the joints may still feel flush even though it is changing in the width dimension.

As the wood moves microscopic cracks develop in the finish and allow more moisture into the wood which then accelerates the wood’s movement. Before long the finish can no longer keep up with the movement of the wood and you have finish failure. The only true way to test this hypothesis is to break open a joint where the finish has failed. If it is true that there is not a continuous glue line then the doors will continue to fail and any touchups will also fail in the future.