A Food-Safe Finish for Bowls and Cups

Nearly every wood finish on sale is safe against food once it has fully cured, and the position that Bob Flexner argued decades ago has not been seriously contested since: the solvents and driers that make a finish hazardous are gone or locked into the cured polymer by the time the finish has hardened. The questions worth spending time on are whether it is genuinely cured, whether anyone eating from it has a nut allergy, and whether the object will hold something hot.

That reframing matters because the phrase "food safe finish" sells a lot of products that are no safer than the ordinary ones, and because it distracts from the failure that actually happens, which is a finish that was rushed, never hardened, and came off into the food as a sticky residue with a smell.

The four classes and what each is for

Non-drying oils never cure. Food-grade mineral oil, sometimes sold as liquid paraffin, sits in the wood, keeps it from drying out, and wipes off over weeks. It is the correct choice for a chopping board precisely because it is renewable, and the wrong one for a display bowl because it will always feel slightly oily. Vegetable and olive oils belong in this class too, and they go rancid. Do not use them.

Drying oils cure by oxidising into a solid. Raw and boiled linseed, pure tung, and walnut oil are the common ones. Cured, they are inert. Uncured, they are a sticky film with a strong smell, and the difference between the two states is measured in weeks, not hours. Tung comes from a nut and walnut oil obviously does, which is a real consideration if you sell your work and cannot know who will drink from it.

Waxes on their own are not a finish for anything that gets washed. Beeswax and carnauba, usually blended into an oil, give a pleasant hand and a slight sheen, and they melt or wash off at a temperature well below that of hot coffee. Treat a wax blend as a dressing that gets renewed, not as a barrier.

Film finishes, including varnish, polyurethane and shellac, are safe cured and are genuinely the most protective option for a salad bowl that lives in a dishwasher-owning household. Their failure mode is worse than the others: instead of wearing thin, they chip, and a flake of polyurethane in a salad is not what anyone wants. Shellac is safe and beautiful and dissolves in alcohol, which rules it out for anything holding spirits.

Cure is the whole question

A drying oil is cured when it no longer smells. That is the test, and it is more reliable than any schedule on a tin. Rub the inside of the bowl hard with a clean thumb, warm it with friction, and smell it. Linseed or tung that is still curing gives off an unmistakable oily smell under warmth; cured oil smells of nothing.

Give a thin coat at least a week in a warm, ventilated room, and give a full oil schedule on a thick-walled cup a month. Cure is a chemical reaction with oxygen, so it needs air, warmth and time, and a bowl stacked inside another bowl in a cold garage is not curing. Anything applied to an interior surface that cannot breathe will take far longer than the same coat on a flat board.

Where a finish is being used on work that is sold, the sensible practice is to finish, cure, wash in hot soapy water, dry, and only then pack. Washing removes whatever has not bonded, and if a wash reveals tackiness, the piece was not ready.

Cups are a different problem from bowls

A cup that holds hot liquid is the hardest case in the whole category. Hot coffee at 80°C drives oil out of the surface, softens wax, and puts the inside of the wall into a wet, expanded state while the outside stays dry. Any finish you apply will be partly removed within a dozen uses.

This is why the Nordic carved cup is traditionally left unfinished inside, or close to it. The old practice was seasoning rather than finishing: brining in salt water, or simply filling the new cup with strong coffee or hot milk and letting the wood absorb it, repeatedly, until the surface stabilised and stopped tasting of raw wood. It works, it is free, and it produces a surface that repairs itself with use rather than one that degrades. If you make cups, try it on one before you argue with it.

For a cup that will hold cold drinks, or a decorative one, a thin schedule of pure tung or polymerised linseed inside and out, fully cured, is fine. For a dry-goods bowl, anything at all. For a salad bowl, which sees oil, acid and washing, a wiping varnish is the honest recommendation even though it is the least romantic.

What the wood itself contributes

Curly birch is diffuse-porous with a fine, even texture, so there are no large open vessels for food residue to sit in, which is part of why birch has been the carving wood of choice for eating vessels across the north for as long as there have been eating vessels. The dark inclusions in the figure are bark and parenchyma laid between the twisted fibres, not voids, and they do not open up into holes on a cleanly cut surface. More on the structure is under curly birch.

Two practical points follow from how it is supplied. Slabs and boards ship at 13 to 15% moisture content, which is drier than green but not at equilibrium with a heated kitchen, so a cup turned or carved to final wall thickness straight out of the box will move afterwards. Logs go out green or air-dried as you ask, and a green one moves a great deal more. Rough it oversize, let it sit, then finish it. And no finish will save a piece that goes in a dishwasher: the combination of 60°C water, detergent and a long drying cycle will check a wooden cup regardless of what is on it. Hand wash, dry standing up, keep it out of a windowsill.

Wall thickness matters more than finish for durability. A cup carved thin in the wrong place cracks at the point where end grain meets the rim, whatever it is coated with.

What to do next

Decide first what the object will hold and how it will be washed, then pick from the four classes, then give the finish twice as long to cure as the tin suggests and test it by smell. If you sell your work, say on the label which oil you used and whether it is nut-derived, because that is the one piece of information a buyer cannot work out for themselves.

If you want to test finishes before committing the good piece, keep the offcuts from your own cutting rather than testing on another board. Figure density and absorption vary along a log, and only wood from the piece in hand predicts the result. Log and board sizes as we cut them are listed under products, and a quote can be asked for against a finished dimension rather than a stock size.