Abrasive Grain Selection by Material: Which Grain for Which Metal

Quick Answer

Match the grain to the metal, not to its hardness. Use aluminum oxide for carbon steel and wood, silicon carbide for non-ferrous metals, stone and glass, zirconia alumina for heavy steel grinding under pressure, and ceramic alumina for stainless, titanium and hardened alloys run hard. Diamond never grinds steel; CBN does.

Why hardness is the wrong starting point

The intuitive rule "use the hardest abrasive" is the most expensive mistake in abrasive selection. Diamond is the hardest known material and is the worst grain for ordinary steel, because carbon is thermodynamically unstable against iron at the temperature a grinding contact reaches. Above roughly 700 °C the diamond's carbon dissolves into the iron and the grain reverts toward graphite, dulling in minutes (American Machinist, 2024; CutterMasters, 2026). This is chemical wear, a separate failure axis from hardness.

So grain selection turns on two questions, in order: what is the workpiece made of, and how hard can the tool push. Hardness, toughness and friability all matter, but the first filter is chemistry, the second is the pressure the operator and the tool can actually apply. Get those two right and the grit size and bond follow.

This guide collapses the common rules into one matrix, then walks each metal. Whitby Abrasives is a value-tier Canadian distributor, and the honest position is that the right grain for the job is usually one or two rungs down from the most expensive one. Where a premium grain genuinely pays for itself, the case is made on cost-per-cut, not hardness.

The abrasive grain chart: grain by material

Best = first choice. OK = workable or a blend. Avoid = wrong tool, with the one-word reason.

A note on what "Avoid" means: it is almost always a chemistry or wear failure, not a physical impossibility. Diamond can abrade steel briefly, it just destroys itself doing it; silicon carbide will cut steel, only quickly and wastefully. Treat the matrix as the default value-correct fit, then confirm pressure, bond and grit before locking a spec.

Best abrasive grain for steel (carbon and mild)

Aluminum oxide (Al₂O₃, fused alumina or corundum) is the general-purpose default for carbon and mild steel. It is chemically stable on steel, it does not react with iron, so its limit is mechanical dulling, not chemical wear. It is also the toughest and cheapest of the common grains and is the single largest grain by demand, around 39–40% of the market by grain type. Brown fused alumina (BFA, around 94.5–98% Al₂O₃) is the everyday grain for general carbon-steel cutting, grinding and sanding.

Step up to zirconia alumina the moment the job becomes heavy steel stock removal or weld grinding under firm pressure. Zirconia is pressure-activated: its self-renewing micro-fracture only kicks in when the operator can push hard. On a high-power grinder it outlasts aluminum oxide and cuts faster, which is why the trade calls it the standard grain for steel fabrication. On a light die grinder or a feather touch it glazes and underperforms a cheap aluminum-oxide disc while costing two to four times as much.

Grain for stainless steel

Stainless is heat-sensitive and contamination-sensitive, which changes the answer. Two grains lead.

Ceramic alumina is the first choice when the tool can push hard. Made by the seeded sol-gel process, it is micro-crystalline and self-sharpens by shedding sub-micron fragments rather than dulling, giving a faster, cooler, burn-free cut that protects the heat-affected zone and prevents heat tint. Nadolny's 2014 review of microcrystalline sintered corundum documents exactly this mechanism: these grains wear by micro-scale fatigue fracture rather than bulk fracture or glazing, continually shedding dulled layers to expose fresh sharp edges, which is the source of their extended life over white fused alumina (Nadolny, 2014).

Where iron contamination is the worry, white fused alumina (WFA, ≥99.5% Al₂O₃, iron-free) is specified instead of brown, because brown aluminum oxide carries Fe₂O₃ that can leave rust spotting and weld defects on stainless. For deeper coverage of heat tint, contamination and galvanic rust, see our guide to the best abrasives for stainless steel.

Caution: ceramic only out-performs when run hard and fast on tough, heat-sensitive metals. On a light-pressure hand application or a low-power tool it will not fracture, so it dulls and glazes like ordinary aluminum oxide while still costing three to five times more.

Grain for aluminum, non-ferrous, glass and stone

Silicon carbide (SiC) owns hard, brittle and non-ferrous work. Produced by the Acheson process, it is harder and sharper than aluminum oxide (Mohs ~9.1–9.5, Knoop ~2,480–2,600) but more brittle, and its very high thermal conductivity, around 135 W·m⁻¹·K⁻¹, lets it cut cooler on heat-sensitive substrates. Soltys and colleagues' 2023 review confirms the property base: SiC combines high thermal conductivity with a low thermal expansion coefficient, giving it stability under thermal stress that oxide grains lack (Soltys et al., 2023).

That makes SiC the standard grain for glass, tile, ceramics, stone and non-ferrous metals such as aluminum, brass and bronze. Aluminum is the classic mixed case: it benefits from a hard, cool-cutting grain, so SiC, often blended with aluminum oxide, is common there. The trade-off against aluminum oxide is wear on tough ferrous steel, where SiC fractures and wears fast. The honest rule is SiC for hard, brittle or non-ferrous and fine finishing, aluminum oxide for ferrous and heavy removal where life dominates. For the full comparison, see aluminum oxide vs silicon carbide.

The grain ladder runs from economy aluminum oxide up to superabrasive diamond and CBN; cost rises with cut rate and life, but the right rung is the cheapest grain that does the job well.

Cast iron, titanium, superalloys and hardened steel

The hard end of the chart splits cleanly along chemistry.

• Cast iron takes silicon carbide well (sharp, cool) and zirconia for durable heavy removal; CBN is the superabrasive answer for hard iron.
• Titanium and nickel superalloys are heat-sensitive and want ceramic alumina run hard, or CBN at the top end. Aluminum oxide burns them.
• Hardened tool steel above the reach of oxide grains uses CBN, not diamond. CBN is chemically inert to iron and nickel, stays thermally stable to roughly 1,000 °C, and stays harder than diamond above about 800 °C, which is why hardened-steel and tool grinding uses it.

Diamond's place is the non-ferrous hard materials: tungsten carbide, stone, glass, ceramics and composites. Never steel.

The cost axis: match the rung to the job

Grain type sets a cost tier. The right metric is cost per unit of metal removed, not price per disc, because a higher grain price buys faster cutting and longer life.

Two cautions. Raw grain cost is only a fraction of finished-disc price; bond, backing, conversion, freight, duty and margin dominate, so a two-to-four-times grain-cost gap rarely shows up as a two-to-four-times shelf price. And the ladder is a value ladder, not a quality ranking: the cheapest grain that does the job well is the best choice. A ceramic disc on light rust removal is wasted money; economy aluminum oxide on all-day stainless weld grinding is false economy. The market share data tracks the same logic: aluminum oxide is around 39–40% of grain demand, zirconia about 28%, and ceramic about 18% and growing fastest. For the head-to-head life comparison, see ceramic vs zirconia vs aluminum oxide.

A final reminder on the other axis: grain type is not grit size. Coarse grit (24–60) is for stock removal, fine (120–P2000) for finishing, and the FEPA-P and ANSI/CAMI scales diverge at fine grits, so a "600-grit" claim means different things on a European-spec versus a North-American-spec disc. See the abrasive grit chart for the full conversion.

The Whitby Abrasives recommendation

Most value-tier buyers default to aluminum oxide for everything, which is right for general steel and wood and wrong for heavy stainless, hard alloys, and stone or non-ferrous work. The upgrade path is specific, not blanket: steer heavy-stainless and hard-alloy jobs to zirconia or ceramic alumina run hard, route stone, glass and non-ferrous to silicon carbide, and leave the economy aluminum-oxide disc on the light and DIY work where it wins on value. We source and specify our discs to the grain grade we state, and the certification and test-data wedge is what makes a per-material life or cut claim credible rather than marketing.

The obvious objection is that a cheaper grain must be lower quality. It is not; over-specced ceramic on a light tool just glazes and wastes the premium, while a correctly chosen value-tier grain finishes the job for less. Start with our Best Sellers for the proven general-purpose picks, browse Metal Fabrication Essentials for steel and stainless work, or compare grains directly in our flap disc range.

Frequently asked questions

What is the best abrasive grain for steel?

For general carbon and mild steel, aluminum oxide is the default: it is chemically stable on iron, tough and the lowest cost. For heavy steel stock removal or weld grinding under firm pressure, step up to zirconia alumina, which cuts faster and lasts longer when the tool can push hard.

Which abrasive grain is best for stainless steel?

Ceramic alumina run hard is the first choice, because it cuts cool and burn-free and protects the heat-affected zone, preventing heat tint. Where iron contamination matters, use iron-free white fused aluminum oxide rather than brown. Avoid letting either glaze on a low-power tool.

Why can't you use diamond on steel?

Diamond is the hardest abrasive but reacts with iron above roughly 700 °C: its carbon dissolves into the steel and the grain reverts toward graphite, wearing out in minutes. For hardened ferrous work the correct superabrasive is CBN, which stays chemically inert to iron and thermally stable to about 1,000 °C.

What grain should I use on aluminum and other non-ferrous metals?

Silicon carbide. It is harder and sharper than aluminum oxide and its high thermal conductivity lets it cut cool on non-ferrous metals, glass and stone. On aluminum specifically it is often blended with aluminum oxide. It wears fast on tough ferrous steel, so it is the wrong all-purpose metal grain.

Is a more expensive grain always better?

No. The right metric is cost per cut, not price per disc, but the cheapest grain that does the job well is the best choice. Premium ceramic on light rust removal or a low-power tool glazes and wastes the three-to-five-times premium. Match the grain to the metal and to the pressure the tool can apply.

Does the grain or the grit matter more for finish?

They are separate axes. Grain type sets chemistry and cost tier; grit size sets cut versus finish. Use coarse grit (24–60) for stock removal and fine grit (120–P2000) for finishing, and check whether a disc is graded to FEPA-P or ANSI/CAMI, because the two scales diverge at fine grits.

Sources

• Grain Cost-Performance Ladder — internal reference, 2026 — cost multiples (SiC ~1.2–1.5×, zirconia ~2–4×, ceramic ~3–5× vs aluminum oxide), life and cut-rate tiers, "cost per cut not price per disc."
• American Machinist, All About Abrasives (2024) — https://www.americanmachinist.com/archive/features/article/21893227/all-about-abrasives — diamond graphitises/reacts with steel; iron lowers the threshold.
• CutterMasters, Diamond vs CBN (accessed 2026-06-28) — https://cuttermasters.com/diamond-vs-cbn/ — CBN stable to ~1,000 °C and harder than diamond above ~800 °C; ferrous vs non-ferrous split.
• Krzysztof Nadolny (2014). State of the art in production, properties and applications of the microcrystalline sintered corundum abrasive grains. The International Journal of Advanced Manufacturing Technology. DOI: https://doi.org/10.1007/s00170-014-6090-2 — micro-scale fatigue fracture and self-sharpening behind ceramic-alumina longevity over white fused alumina.
• L.M. Soltys et al. (2023). Synthesis and Properties of Silicon Carbide (Review). Physics and Chemistry of Solid State. DOI: https://doi.org/10.15330/pcss.24.1.5-16 — SiC extreme hardness, corrosion and thermal-shock resistance, high thermal conductivity with low thermal expansion.
• Standards referenced: FEPA-P (coated) and ANSI/CAMI B74 (North American) grit scales; FEPA-F (bonded) grading.