I’m trying to refinish two large windowsills in my living room. I sanded them down to the bare wood, applied Minwax oil-based stain, then two coats of Minwax oil-based polyurethane with a foam brush, following the instructions on the cans for drying and sanding between coats, and stirring the cans periodically. The result looked great except for a couple areas that had bumps from bubbles. I figured it would be easy to sand those areas smooth and apply a third coat there. But for some reason, those areas then took on a cloudy or milky appearance. I read that that’s most often caused by moisture, so I sanded those areas and tried again, keeping the windows closed and the air conditioning on to ensure dry conditions. That gave the same result, so then I suspected I had contaminated the can of poly by using a different stir stick that had white paint on it. I bought a new can of poly and a plastic stir stick and tried again, with the same results. I tried using a hair dryer to remove moisture trapped under the poly, as I read on several websites, but that had no effect, possibly because this is an oil-based product and those posts were dealing with water-based. Then I waited three days, hoping the cloudiness would clear up on its own, but it didn’t.
Next, I decided that at least a uniform appearance across the whole piece would be better than a few isolated patches of cloudiness, so I sanded the whole thing down to the stain and started over with two new coats of poly, being extra careful to follow all instructions. I noted an indoor temperature of 65 degrees and a humidity of 61% when I started. This time, the whole thing was cloudy from the first coat, but in patterns following my brushstrokes — with a striped appearance — so not an improvement.
But weirdly I noticed a few lines of clear, good-looking finish, which corresponded to shadows falling onto the surface from the leading in the windows. This was on the south window, which gets more sun, but if the sun caused this problem I would expect it to be less on the north window. Instead, the north window has the same amount of cloudiness, also in a pattern matching the brush strokes. The only clear spots were where the shadows fell onto the south windowsill. If the problem is sunlight, I could sand again and apply a final coat at night. But I’m pretty sure I did one of the spot recoats at night and it still ended up cloudy. Also, the shadows from the big wooden pillars between the windows didn't create clear areas, so I’m not sure what to make of this.
I’ve been thinking about why the first two coats were clear and what changed since then. It was similarly sunny when I did those first coats, but warmer outside, so probably around 75 inside. Another variable is that I had the windows open and fans running initially, but after the cloudiness appeared I read that drying too fast could be a problem, so for my last attempt, I kept the windows closed and left the fan off so it would dry more slowly. Another variable is the thickness of the application. I was putting it on pretty thick initially, but when I tried patching the bubbled areas, I probably went thinner because it was just a finishing touch. Then I read that thin coats are better and have tried to do that since then. But maybe a thicker coat is actually the way to go.
And I read just now that too much sanding between coats can cause a cloudy appearance. I’ve been using a random orbital sander with 240 grit mesh (the instructions recommend 220 grit) then wiping with a lint-free cloth. Although I’ve done more sanding on the later attempts to try cleaning up the problems of the earlier attempts, the cloudiness clearly matches my brush strokes, not the variations visible after sanding. Maybe some sanding dust is left behind and is then getting dragged across the surface by the brush, creating the patterns of cloudiness? (BTW I was always brushing with the grain until the last attempt pictured here because I realized it was easier to get a consistent application by going the “short way” across the windowsill.)
To summarize:
- Sunlight: a shadow falling on the surface left a clear area on my last attempt, but otherwise I haven’t seen that the presence or absence of direct sunlight makes a difference.
- Temperature/humidity: I think the temperature was warmer when the problem didn’t occur.
- Ventilation while drying: doesn’t seem to make a difference.
- Brushing pattern and direction: no difference.
- Thick coats: I think I was applying thicker coats when the problem didn’t occur.
- Sanding: I think I sanded less between coats when the problem didn’t occur.
My current theory is that drying too fast causes the problem; that explains why the problem didn’t occur with thicker coats or in that strip of shade. Does anyone have another explanation? And can I fix it with one more sand/coat cycle or would I need to sand all the poly off again, or even sand off the stain to start completely over?
Here's one of the windowsills I'm refinishing:
Closeup of the problem. The large white areas are glare from the sun ... the pale white streaks are the problem. Although the appearance changes with the viewing angle, this is pretty consistent across the whole project area:
Closeup of the clear patches left by shadows from the glass leading. It still looks a bit cloudy in the photo but in real life it's a pretty distinct difference. The other dark areas in this photo are shadows from the current sun position:
These circles were from bubbles that I popped with the corner of my brush after applying the last coat. I think this created a small "crater" with a ring of thicker poly. That would support the theory that slower drying avoids the problem:
UPDATE 1
I tried the suggestions here to sand lightly with 400 grit, wipe with a damp rag and dry, then apply in the direction of the grain, with the same result. Then I tried several other things based on other theories of what's going wrong: thinning with mineral spirits, wiping on instead of brushing, applying very thick or very thin, applying with heavy pressure or light pressure, applying at night (no sunlight), buying a new can of poly from a different store, etc. Everything led to the same problem: pale white cloudiness in the finish, following the pattern of the brush strokes. Fortunately each coat erased the problem from the previous coat, so I didn't have to keep sanding everything off.
Finally I found a forum discussion describing how flattening agents in polyurethane work (I'm using Minwax "warm satin") and how they could collect in the brush and be applied inconsistently, so I bought a can of gloss polyurethane and applied that, and it's perfectly clear. It's also ridiculously glossy -- it's dry but still looks wet!
The humidity was around 60% when I did the previous attempts but it dropped to 47% today. So possibly the additional 13% of humidity was the problem. But given how the cloudiness always follows my brush strokes, and is erased by later coats, it seems more like an issue with the flattener. Does anyone have a more informed opinion on that? And if the problem is with the flattener, how to avoid it? I thought I was mixing adequately and stopped to mix again every 15-20 minutes while applying.
Right now I don't know whether to try a final satin coat, which might take me right back to the original problem, or try one of the methods I've read about for scuffing a gloss finish into a flat finish, which seems like it could create a whole new problem.
UPDATE 2
I found a tutorial on rubbing a satin finish that seemed straightforward, so I gave the glossy coat a month to cure and tried it. I don't know what happened, but the results were not good: the shine didn't really go away but the finish became pale and dull and obscured the wood color.
Shortly after that, I noticed the humidity had dropped to 47%, so I made another attempt with the satin polyurethane, this time with a can of Behr polyurethane from Home Depot instead of the Minwax I'd been using. That looked even worse than before, so the problem is neither the Minwax brand nor the humidity.
