No occurrence in finishing is worse than having layers of finish peel or chip from each other or from the wood, especially after a job has been delivered and installed.
Peeling and chipping are more common now than ever with the increased use of high-performance and water-based finishes, neither of which bonds as well, layer upon layer, as do nitrocellulose and CAB-acrylic lacquers.
Nor do high-performance or water-based finishes bond as well over decorative steps that are left thick, such as glazes or stains that are “dirty wiped” (not wiped off entirely). And water-based finishes often don’t bond well over oil-based products, including stains and paste wood fillers.
High-performance, high-solids finishes can even bond poorly to wood sanded to a very fine grit. This has led some manufacturers to discourage sanding to finer than 220 grit on tight-grained woods such as maple or birch because their finishes don’t bond as well to the finer scratches. (I’ve never experienced this problem, but someone has or the manufacturers wouldn’t be issuing these cautions.)
Even traditional nitrocellulose and CAB-acrylic lacquers bond poorly to sanding sealer that is applied thick, and to thickly applied stains and glazes.
Unless you are using nitrocellulose or CAB-acrylic lacquer alone or over a wiped-off stain, you need to have a method of testing to be sure your finishes are bonding well.
Several easy methods exist for testing adhesion. These methods have been established by the American Society for Testing and Materials (ASTM) and are used by all coatings manufacturers and by the furniture industry.
To conduct an “official” test, you need to use specified tools, which are fairly expensive, and you need to abide strictly to the ASTM standard. But you can perform a perfectly adequate test in your shop using nothing more than a razor blade (box cutter) and firm-sticking masking or Scotch tape.
The test is commonly referred to as the “cross-hatch” or “cross-cut” test.
You should perform it on a separate panel, of course, not on a cabinet or piece of furniture that will be delivered to a customer. And you should perform it in exactly the same manner with the same wood, sanding steps, products, application steps and drying times you intend to use on the actual job.
The non-official test is very simple. Make six to eight straight, parallel cuts about an inch long and approximately 1/16 inch apart into a finish that has dried for several days or weeks. Cut through to the wood, but not deep into it because the wood itself could separate (especially in the deep grain of a wood like oak) and give you false results.
You can use a straight edge so the cuts are perfectly straight, but there’s no need to do this (and I never do). You could also use a rule to make the distance between the cuts identical, but this isn’t necessary, either (and I never do it). You won’t have any problem making adequate cuts by eye.
Once these parallel cuts have been made, make six to eight similarly spaced cuts about an inch-long perpendicular to the first set to create a latticework, or “tic-tac-toe” pattern. Then apply the tape over the cuts, press it down to create a good bond and pull it up in a quick motion.
Finally, check the rectangles for adherence and the tape for finish that has been pulled off. You may need to do several tests on different areas of a panel to get the full picture of the bonding strength, especially in situations of varying grain patterns.
If you are comparing different finishing products for their bonding characteristics, be sure to use the same drying times and the same tape. Not being consistent can yield differing and questionable results.
What you’re looking for
The ASTM standards categorize the degree of adhesion (see chart). Clearly the top two latticework drawings indicate a fairly good bond. The middle drawing indicates some possible problems. And the lower two are unacceptable.
Use these drawings as a guide to help determine how good a bond you’re getting. After you’ve performed this test several times in different situations, you’ll get a good feel for the strength of a bond.
Also, look closely at what’s stuck to the tape. The damage to the latticework will show clearly when the finish is coming off wood, but it may not show as clearly when the finish is separating between layers. But this will be obvious on the tape.
If it isn’t clear at which step in a multi-step finish the separation is occurring, make a step panel and test at each step. A step panel is wood or veneered with each layer of finish or decorative step taped off in a small section before the next is applied to create a stair-step progression of your finishing process.
Keep in mind that adhesion and scratch resistance aren’t the same thing. For example, a finish such as conversion varnish should be very resistant to scratching, but may not bond well over a glaze or thickly applied stain.