"Could you send me a quote for SiC powder?" That single line guarantees at least two more rounds of email before anything can be priced. Silicon carbide is a material whose name covers an unusually wide spectrum: powder at one dollar a kilogram and powder at a thousand dollars a kilogram come back in the same search result. A steelmaking deoxidizer and a crystal-growth feedstock share the same chemical formula. So an SiC inquiry that opens with "how much" will stall; "what exactly" has to be settled first. This article organizes the seven specifications worth fixing on your own side before the RFQ goes out.
1. Fix the Grade First — One Decimal Place Moves the Price by an Order of Magnitude
What sorts SiC powder into grades is, above all, purity.
Moving from 99.5% to 99.99% is not "a 0.49% improvement." It changes the manufacturing route itself, and the price moves by an order of magnitude.
Checkpoint
Do not write only "high purity" in an RFQ. Write a numeric band. SiC ≥ 99.5% and SiC ≥ 99.9% are entirely different conversations.
For reference, a representative composition for a green abrasive fine grade looks like this:
SiC 99.25% / free C 0.05% / SiO₂ 0.35% / free Si 0.15% / Fe₂O₃ 0.04%
Purity is only one axis. Morphology — crushed angular powder, spray-dried granule, plasma-spheroidized bead — is an independent axis that also has to be named, and the three grades are separated in SiC Powder Grades Explained.
2. Particle Size: "400 Grit" Does Not Belong in an RFQ
Sizing is where SiC sourcing generates the most disputes. The same number means a different thing in a different standard.
JIS #400 ≈ 30 µm / FEPA F400 ≈ 16.5 µm
The same "400" differs by roughly a factor of two in D50. A quote written with only a grit number and no standard named is almost guaranteed to be mis-compared.
A Chinese W-number is not a D50
In Chinese fines standards (GB/T 2481.2), the W-number is the upper limit of the basic-grain band (基本粒). It is not a median.
The Implied D50 column is simply the midpoint of the band. GB/T 2481.2 specifies what fraction of the powder must fall inside the band, not a median, so a compliant W28 whose distribution is skewed coarse sits above that midpoint — use the column to put quotes on a common footing, not as a number to write into a specification.
On that same assumption — a distribution roughly centred in its band — reading W28 as "D50 28 µm" overstates the real size by about 15%.
The measurement method shifts the number too
- Sedimentation (沈降法) — the traditional method in the Chinese fines industry. Reports a hydrodynamic equivalent diameter.
- Laser diffraction (ISO 13320) — the Korean and Japanese standard. Reports a light-scattering equivalent diameter.
The same sample can read 8 µm by sedimentation and 10 µm by laser diffraction. Sedimentation generally reads finer.
Write it this way
D50 30 ± 2 µm (equivalent to JIS R6001 #400) — laser diffraction per ISO 13320, wet dispersion
Without that one line, three suppliers' quotes cannot be laid side by side and compared.
3. Do Not Stop at D50 — Distribution and Top Cut Decide Performance
Two powders with the same D50 can be entirely different materials once D10, D90, and span (= (D90 − D10)/D50) are compared.
- A broad distribution → packing behaviour and sintering shrinkage drift, and lot-to-lot reproducibility falls away
- An unmanaged top cut (粗粒) → scratches in lapping and polishing, and defect initiation sites in the sintered body
Ask for D10 / D50 / D90 together with the top cut. A COA that reports D50 alone is a signal that the distribution is not being managed at all — which is why incoming inspection pairs particle-size analysis with morphology and composition checks, as set out in Powder Quality Verification Methodology.
4. Iron (Fe): Align the Units Before Discussing Price
Supplier COAs mostly report Fe₂O₃ (iron oxide), while buyer specifications mostly write Fe (element). The conversion factor is 0.6994.
The last row is the one that costs money. If you asked for Fe 0.05% and accept material certified to "Fe₂O₃ 0.05%," you are paying for roughly 30% of headroom you never requested. A single unit mismatch moves the price.
5. Titanium (Ti): Not a Negotiation, a Pass/Fail Gate
Specifications in ceramic, fine-ceramic, and semiconductor-adjacent applications increasingly exclude Ti explicitly. There are four reasons worth understanding.
It is electrically active inside the crystal lattice
In 4H-SiC, Ti forms acceptor levels at roughly E_C − 117 meV and E_C − 160 meV below the conduction band, corresponding to the hexagonal and cubic lattice sites. Note what those numbers do and do not say. In a 3.26 eV band gap they are shallow levels, roughly 1.5 eV away from mid-gap, and Shockley–Read–Hall recombination is dominated by near-midgap states — so Ti is a poor recombination centre in SiC, and the defect recognised as lifetime-limiting in 4H-SiC is the carbon-vacancy-related Z₁/₂ level near E_C − 0.65 eV, not Ti. The objection to Ti here is that it is electrically active at all: it introduces trapping and luminescence centres into the lattice at a density that tracks an impurity nobody specified. In silicon, Ti does sit close to mid-gap and has been a textbook "lifetime killer" for decades, which is why device makers habitually treat it as a substance to exclude rather than one to bound — but that is silicon's mechanism, not a conclusion these SiC levels support.
It creates uncontrolled second-phase TiC
SiC–TiC is an excellent composite when it is designed deliberately, improving both toughness and electrical conductivity. The problem is that Ti arriving as an impurity was designed by no one. TiC precipitates whose size and location are uncontrolled produce local thermal-expansion mismatch, then microcracking, then lot-to-lot scatter in properties.
The reason a buyer specifies "Ti-free" is not that Ti is unconditionally harmful, but that in this form its effect cannot be made reproducible.
Ti moves as one body with iron
In the Acheson process, Fe- and Ti-bearing oxides in the raw material are reduced to a metallic melt at high temperature and then, on cooling, precipitate as an Fe–Si phase and an Fe–Si–Ti alloy phase enclosed within it. Ti sits trapped inside the iron phase.
A plant that cannot control iron cannot control titanium either. The two specifications are physically aimed at the same impurity.
Ti starts at the silica mine, not at the furnace
Ti arrives attached to the raw silica sand as heavy minerals — ilmenite, rutile, and similar. So the question to put to a supplier is not "do you remove Ti?" but the following three.
Three questions for supplier qualification
- What instrument and method do you use for Ti, and what is the detection limit in ppm?
— "Not detected" from an instrument with a 100 ppm detection limit carries no information.
- How do you specify and control the origin of the raw silica sand and its TiO₂ content?
- What are your magnetic-separation stages and field strength, and your acid-leach conditions (acid type, concentration, time)?
A supplier who cannot answer those three concretely cannot guarantee Ti at the lot level. Whatever the price, that is not a candidate to proceed with.
6. Why 0.5 / 10 / 30 µm Are Bought as a Set: Trimodal Packing Design
Buyers often ask why three particle sizes are purchased together. This is not coincidence; it is a multi-modal packing design.
The maximum packing fraction of a single-size powder is about 60–64% — but that figure is random close packing of spheres. A single-cut angular SiC powder packs below it, because angular particles interlock and bridge, so treat 60–64% as an optimistic reference rather than the baseline the real material starts from. Add mid-size particles to fill the voids between the coarse particles, then fine particles to fill the voids between those, and the packing fraction climbs to 75–85%. The size ratio is typically 1/3 to 1/10 at each step.
As packing fraction rises:
- Green-body density ↑ → sintering shrinkage ↓ → dimensional accuracy ↑
- Sintering-aid consumption ↓ → final purity ↑
- Porosity ↓ → strength, thermal conductivity, and plasma resistance ↑
So buying a three-cut set is not "using several sizes"; it is a single design. That is why the purity and impurity specifications of all three items must be held at the same level, and why it is preferable to take them from the same plant and the same raw-material lineage. Coordinated cuts across submicron, mid, and coarse fractions are how the SiC powder range is put together for exactly this reason.
7. Contract and Logistics: Where a Trader Actually Draws the Line
The last item matters most. With that clause in the contract, a supplier does not send questionable material in the first place. Transferring quality risk through the contract up front, rather than through a claim after the fact, is the core of sourcing practice. The customs and classification side of the same shipment — HS code, FTA preferential rate, K-REACH — is covered in the SiC Powder Import Guide for Korea.
Pre-RFQ Checklist
- Has the grade been fixed as a numeric band (in the form
SiC ≥ 99.5%)?
- Have you written D50 plus tolerance plus standard name plus measurement method, all four?
- Have you also asked for D10 / D90 / top cut?
- Have you specified Fe on an elemental basis and noted the Fe₂O₃ conversion alongside it?
- Have you requested Ti together with the detection limit in ppm?
- Have you asked for three or more production-lot COAs, not a sample COA?
- Have you put Incoterms, payment terms, C/O, and the return clause on the table at the initial quotation stage?
Frequently Asked Questions
Why can't a supplier quote from "SiC powder, high purity" alone?
Because purity alone spans metallurgical grade at 88% to crystal-growth grade at 99.9999%, and the manufacturing route — and therefore the price — changes by orders of magnitude across that range. A quotable RFQ needs a numeric purity band, a D50 with tolerance, the sizing standard, and the measurement method.
Is a Chinese W-number the same as a D50?
No. A W-number under GB/T 2481.2 is the upper limit of the basic-grain band, not a median. W28 means a 20–28 µm band, so if the distribution is roughly centred in that band the D50 lands near 24 µm and reading W28 as "D50 28 µm" overstates the actual size by roughly 15%. The standard sets band fractions rather than a median, so the band midpoint is a comparison aid, not a spec value.
How do I convert Fe₂O₃ on a COA into elemental Fe?
Multiply by 0.6994. So Fe₂O₃ 0.07% corresponds to Fe 0.049%, which passes an "Fe ≤ 0.05%" specification. Requiring "Fe₂O₃ ≤ 0.05%" instead is over-specification — it corresponds to Fe 0.035% and means paying for headroom you did not ask for.
Why is titanium treated as a pass/fail item rather than a negotiable limit?
Because Ti is electrically active in the lattice and, as an unplanned impurity, also precipitates uncontrolled TiC. Its acceptor levels in 4H-SiC (about E_C − 117 meV and E_C − 160 meV) are shallow rather than mid-gap, so they act as trapping centres rather than as the lifetime killer Ti is in silicon — the recognised lifetime-limiting defect in 4H-SiC is the carbon-vacancy-related Z₁/₂ level. Neither effect can be made reproducible lot to lot, so the practical question is whether the supplier can demonstrate detection limits, raw-sand TiO₂ control, and separation conditions — not what discount is available.
Why buy 0.5, 10, and 30 µm cuts together?
Because packing fraction is a design variable. A single-size powder caps out near 60–64% — that being random close packing of spheres, with angular powder lower still — while a multi-modal blend with roughly 1/3 to 1/10 size steps reaches 75–85%, which improves dimensional accuracy, final purity, and sintered properties. For that reason all three cuts should carry the same purity and impurity specifications, ideally from the same plant and raw-material lineage.
References (Public Sources)
- Phys. Rev. B 55, 13618 — Electrical properties of the titanium acceptor in silicon carbide
- J. Iron Steel Res. Int. (2024) — Impurity formation mechanism of silicon carbide crystals smelted by Acheson process
- Ceramics International (2021) — Occurrence forms of major impurity elements in silicon carbide
- Ceramics International (2025) — Packing density optimization of multi-modal SiC powder feedstocks
- JIS R6001 / GB/T 2481.2 / GB/T 3045 / ISO 13320
Reference values in the tables above are indicative and intended for cross-checking specifications; actual determinations depend on the standard invoked and the measurement conditions.
None of the seven items above is optional — a quotation given against a loose specification is not a price, it is a different material each time. Nami Tech Solutions (NTS) puts one written specification to mills directly, in their own language, so the quotations come back comparable, then reads each production-lot COA against that same specification before the lot ships.
