Why a Cutting Board Is Supposed to Scratch

This page contains affiliate links. If you buy through one we may earn a commission, at no cost to you.

It is normal, and the scratches are not wear. They are the record of the one thing a cutting board is for. Elm is soft, so it gives way early and visibly, and every groove in it is a contact your edge did not have to absorb.

The useful version of that is arithmetic rather than reassurance, because hardness is not a score. It is a pressure: the pressure at which a material stops resisting an indenter and starts flowing under it. So when a steel edge meets a board, the two hardnesses are two pressures, and the ratio between them is the ratio in which they share the contact. Elm gives way about 165 times more readily than the steel in a good kitchen knife. That is why you can see it.

Put every board material on one scale

Hardness scales do not convert cleanly into each other, which is why this is worth doing carefully. Our own page on why you cannot drill out a broken tap already establishes the right move and this page will borrow it rather than re-argue it: work in Vickers, where the number is linear in indentation pressure, and multiply by 9.807 to get megapascals.

Do that to everything a kitchen board is made of and the spread is four decades wide. American elm comes out near 42 megapascals. High density polyethylene is 64. Hard maple is 74. Flat laminated bamboo is about 78, and strand woven bamboo roughly doubles that to 152. Compressed wood fibre composite is near 294. Commercially pure titanium is 1,422. Cold rolled 304 stainless sheet is about 1,765. Tempered soda lime glass is 5,394.

And the edge. A three piece plastic, bamboo and stainless steel cutting board set is the most useful single listing in this whole argument, because it puts three rungs of that ladder in one box and publishes the thickness of each: the bamboo board at 15 by 10.2 by 0.82 inches, the stainless at 13.4 by 9 by 0.32, the plastic at 12 by 7.2 by 0.32. Its own drawbacks say the "steel surface can be harder on knife edges", which is the belief this page is going to test, and it is the only listing we carry that lets you test it without changing brands.

A sharpened edge is a wedge, and a wedge sinks until it fits

Here is the mechanism, and it needs one equation. An edge is a wedge of some half angle. When it lands on a board it penetrates until the width it has opened up, multiplied by the length in contact, multiplied by the board's indentation pressure, equals the force you are pushing with. So the depth is the force divided by twice the pressure times the contact length times the tangent of the half angle.

Take a firm push of 50 newtons, about 11 pounds, with 50 millimetres of edge touching down, at the 15 degree half angle one of our knives publishes. The depths follow directly. On elm the edge sinks 44.3 micrometres. On hard maple, 25.4. On polyethylene, 29.3. On flat bamboo, 23.8, and on strand woven bamboo 12.3. On composite fibre, 6.35. On titanium, 1.31. On 304 steel, 1.06.

Now the part that matters, and it is about the tip rather than the wedge. A freshly sharpened kitchen edge has an apex radius of roughly half a micrometre. Compare that with the depths above. On elm the apex ends up buried 89 of its own radii below the surface, which means it is not carrying the load at all: the flanks of the wedge are, spread over 44 micrometres of contact. On titanium it is buried 2.6 radii. On steel, 2.1.

That is the whole protection mechanism, stated as a number. A soft board does not cushion the edge in any springy sense. It swallows the apex so that something other than the apex takes the force.

An extra large 18 by 12 inch composite wood cutting board with a juice groove sits in the interesting middle of that range. Compressed wood fibre is roughly four times harder than the hardwood it imitates, so it records a 3.4 micrometre groove where a maple board records 13.6 and an elm board records 23.7. Its record is candid that "composite fiber behaves differently from solid hardwood, and knife feel is a matter of taste", and the arithmetic says what that difference is: it is four times less willing to take its share. Its drawbacks also note there is "no listed weight or thickness figure", which matters here only because thickness is the one board property that has nothing to do with the edge.

The thesis I started with was wrong, and the metal boards are the reason

I expected this page to end up as a case against metal boards. It does not, and the arithmetic is not close.

The rule the broken tap page sets out is that a cutting tool needs around one and a half times the hardness of what it is cutting, in Vickers. Turn it round and it becomes a rule about boards: an edge is safe indenting anything at least 1.5 times softer than itself. The edge in question is 6,963 megapascals, so the threshold sits at 4,642.

Titanium at 1,422 is 4.90 times softer than the edge. Cold rolled 304 at 1,765 is 3.94 times softer. Both clear the threshold comfortably. Tempered glass at 5,394 does not clear it at all: the edge is only 1.29 times harder, which puts the two materials inside each other's margin, and that is the one genuine failure in the category.

So titanium is the softest metal anyone makes a board out of, and it is softer than the stainless it is usually lumped with. A large double sided 100 percent pure titanium cutting board states exactly that, and it is the only comparative hardness claim in the whole category: its record says the surface is "said to maintain blade sharpness longer than stainless steel and reduce dulling". That claim is correct. Pure titanium runs near 145 Vickers against roughly 180 for 304 sheet, so it is about 20 percent softer, and it will take about 24 percent more penetration per contact. The record also says fine scratches are normal and cosmetic rather than flaking, and the arithmetic agrees for a different reason than the one given: the groove is 0.70 micrometres wide, which is right at the limit of what scatters visible light, so you are looking at the shallowest mark that can be seen at all.

The record that needs correcting is the one that argues with itself. A Homaider pure titanium double sided cutting board lists among its strengths a "knife-friendly surface per maker" and then, among its drawbacks, that "metal feels harder than wood", with its own performance text adding that titanium "is hard, which can be less forgiving on knife edges". Both halves cannot be right, and the arithmetic says the strength is the true one. Titanium is the gentlest metal on this ladder by a clear margin. What the drawbacks are actually describing is stiffness and sound, not hardness, and those are real: this board publishes 0.06 inches of thickness, which is 1.52 millimetres of metal, so it rings and it does not feel like a block. Hardness and feel are different properties and the listing has merged them.

Glass is the one that fails, and it fails by leaving no mark

A leopard tempered glass cutting board, 16 by 12 inches is the honest end of this. Its own record says the surface is "more scratch resistant than plastic, per the maker" and, two lines later, that "hard glass can dull knives faster". Both are true and they are the same sentence said twice, because scratch resistance and edge wear are the same transaction seen from opposite sides.

Work the depth for glass and you get 0.35 micrometres, which is less than one apex radius, so the tip never gets buried. Except the figure is worse than that, because glass does not do what the equation assumes. Metals and woods flow plastically under an indenter, which is what makes the calculation meaningful. Tempered glass does not flow. It holds elastically and then cracks. So there is no groove to form and no share for the board to take, and the full 50 newtons stays on the apex.

What happens next is that the apex itself yields, because it is the only thing in the contact that can. Press 50 newtons through 50 millimetres of edge and the steel stops flowing once the contact has flattened to 0.144 micrometres wide, which is 29 percent of a fresh apex radius. That is a single full force contact.

And here is where this page stops, deliberately. I am not going to turn that into a blunting rate, because the honest version needs the force of every individual contact and no listing publishes any of it. A chopping stroke, a slicing stroke and a scrape are three different forces and they differ by more than an order of magnitude, so a strokes to dull figure derived from one of them would be fiction. The thing you can measure is on the board: elm records 23.7 micrometres, titanium records 0.70, glass records nothing. The mark is the measurement, and no mark is not the board winning.

The most honest record in the category is a steel one

A RIZIZMVG oversized 304 stainless steel cutting board says the thing outright instead of hedging: "chopping directly on it dulls your knife edges noticeably faster. That is the nature of any metal board, not a flaw here." It then recommends the correct remedy, which is to use it as "a baking and pastry surface, for raw-meat hygiene, and as a counter protector" and keep something softer for knife work.

That is right, and the arithmetic adds one thing to it. Steel at 3.94 times softer than the edge is not in the glass regime. It is in the regime where the board takes its share, just a quarter as much of it as composite fibre does and a twenty fourth as much as maple. It is a board you can cut on and a board that will cost you sharpenings, and those are compatible statements.

The record also publishes a real figure almost nothing else here does: a 14 gauge core, which is about 1.9 millimetres of stainless, against the "flimsy 18 or 20-gauge sheets that warp, dent and ring". That is a stiffness and ringing figure rather than a hardness one, and it is the correct answer to the complaint the titanium listing mislabelled as hardness.

Bamboo, where our own page has the right answer and the wrong reason

Our page on why kitchen knives get dull so quickly gets the practical advice entirely right, and the four habits it names, cut on wood or soft plastic, hone regularly, hand wash and dry, store so the edge touches nothing hard, are the four that matter. Two of its explanations need correcting, and both corrections come out of the ladder above.

The first: it says "Glass, granite, marble, and ceramic are harder than steel". Tempered glass is not. At 550 Vickers against 710 for a 60 HRC edge it is about 23 percent softer. The conclusion that glass wrecks edges is correct, but the reason is not that glass outranks the steel. It is that glass is close enough to be inside the 1.5 times margin and, unlike every other board, will not deform at all.

The second: it says "Bamboo is popular but noticeably harder and more abrasive than most wood, so it dulls knives faster than a soft maple or walnut board." Harder is the wrong word and abrasive is the right one, and they are separate mechanisms. Flat laminated bamboo sits at roughly 78 megapascals against hard maple's 74, which is a five percent difference and not worth a sentence. What bamboo actually has is opaline silica grown inside the plant, and our page on why a juicer auger shows wear after one use already puts the number on it: Mohs 5.5 to 6.5. Converted, that is roughly 600 to 1,100 Vickers, which brackets the edge and in its upper half exceeds it. The ratio at the middle of that band is 0.835, meaning the grain is harder than the steel and abrades it rather than the reverse. So bamboo is not a hard board. It is a soft board with hard particles in it, and the particles are the only thing on this entire page that beats a knife outright.

Which makes a small glue free bamboo cutting board, 14 by 10 inches and 0.8 inches thick an interesting test case rather than a contradiction. Its record claims a "pre-oiled surface resists knife marks" and that it "helps protect blade sharpness during use", and on bulk hardness it is right: the board is maple's equal and will swallow an apex 47 radii deep. The oiling claim is also doing something real, because a filled surface keeps grain exposure down. The listing and our own page are each half right, and the half each gets wrong is the other one's half.

What the knife brings to it, and why the geometry pulls the other way

Every figure above used an edge at 6,963 megapascals and a 15 degree half angle, and both of those come from a product rather than from me. A NOVKSO 7 inch santoku in high carbon steel publishes "HRC 60 to 61" and "a hand-polished 15-degree edge per side" on a 2 millimetre blade, which is two of the three numbers this page needs and more than most knives state.

The angle has a consequence that runs opposite to the one you would expect, and it is worth spelling out because our own page on why a knife edge fails on geometry, not hardness establishes the other direction so firmly. That page shows apex strength goes as the square of the tangent of the half angle, so going from 20 degrees to 15 costs 45.8 percent of the edge's capacity. True, and this page takes it rather than re-deriving it.

But penetration depth goes as one over the tangent of the same angle. So the same move from 20 degrees to 15 makes the edge sink 1.358 times deeper into any board that will yield, which buries the apex 36 percent further and hands 36 percent more of the contact to the wedge flanks. A more acute edge is weaker and better protected at the same time, and which of the two dominates is decided entirely by the board. On elm, where the apex is already 89 radii down, the protection is total and the acute edge is free. On glass, where it is 0.7 radii down, the protection is nil and the acute edge pays the full 45.8 percent penalty with nothing to show for it. That is the real argument for matching an acute knife to a soft board, and it is stronger than the usual version.

Our comparison of wood, plastic and stainless steel boards sets the three materials out properly in words and this page is the arithmetic underneath it. One thing to add to it: polyethylene at 64 megapascals comes out softer than hard maple at 74, so on indentation alone plastic protects an edge slightly better than maple does, by about 15 percent of penetration depth. It loses on other grounds, and a plastic board's grooves hold what they catch in a way wood's do not, which is a hygiene argument rather than an edge one.

The honest limit, and what to do about the elm

The soft inputs here are all about contact and all stated. The 50 newton push is representative of a firm chop and nothing publishes it. The 50 millimetres of edge in contact is representative. The half micrometre apex radius is a materials figure for a freshly sharpened kitchen edge, and a dull one is four to ten times that, which changes the burial counts and not the depths. Every board hardness is a materials reference rather than a product figure, because not one board listing we carry publishes a hardness, and the wood values in particular are Brinell numbers scaled by Janka ratings, which is a reasonable conversion and not an exact one. The wedge model also ignores everything about the bevel above the apex, and real edges are not perfect wedges.

What survives all of that is the ladder. The depths are all proportional to the same force and the same contact length, so every ratio between two of them is unchanged by any of those assumptions. Elm will always take about 1.74 times the penetration of hard maple, because that is the Janka ratio. Titanium will always take about 24 percent more than 304, because that is the Vickers ratio. And tempered glass will always sit inside the 1.5 times margin, because 5,394 and 6,963 are what they are.

So: the elm board is working. If the scratching bothers you aesthetically, sand and re-oil it, which is a thing you can do to elm precisely because it is soft. If it bothers you because you think it signals damage, it does not, and the board that would have reassured you by staying unmarked is the one that would have been quietly spending your edge instead. The only change worth making is the one nobody sells: put the soft board under the knife and the hard board under the pastry, the hot pot and the raw meat, and let each surface take the job it is physically suited to.