POLISHING · Technical note

D50 sets the finish, D99 sets the scratches

Two polishing powders can share a D50 of 1.5 μm and behave nothing alike. One gives a clean mirror; the other produces scratches that only appear after the wipe. The difference lives in the coarse tail, and a D50-only specification hides it.

Updated November 2026 · 6 min

A particle size distribution has two ends, and they do different jobs. The middle sets how fast the surface comes down; the top end sets how often it comes down damaged. Buyers compare D50 because it is easy to quote. Yield is decided by D99 and by the count of particles above 15 μm.

Read both ends of the distribution

ParameterWhat it controlsWhy it shows up on the line
D50Removal rate and the Ra level you can reachOne step finer buys finish and costs cycle time
D99 / oversize countScratch rateA single 20 μm particle scores a cover glass
DmaxPump, valve and seal safetySets the filtration you need in the loop
BET and morphologyHow the powder disperses and stays suspendedSame D50, visibly different removal rate
CeO₂ / TREOChemical driving force behind the cutSets the price band and the over-etch risk
Fluorine contentActivity and etch aggressivenessHigh F cuts faster and attacks the glass sooner
pHSuspension stability and cleaning loadDrifts as the bath is used, taking the rate with it

Laser and sedimentation do not agree

The same powder measured by laser diffraction and by sedimentation can differ by up to a factor of two in D50, because the methods weight particles differently. This is the most common reason two powders both sold as “1.5 μm” do not compare. Ask which method, and whether the figure is D50 by volume or by weight. Our tables state the method per grade, and the COA repeats it.

Where scratches actually come from

  • Oversize in the powder — the tail you did not specify.
  • Dried slurry on the pad or in the pipe — flakes behave like 50 μm particles.
  • Contaminated water or a loop that was never flushed after a grade change.
  • Pad debris and broken conditioner diamonds.
  • Handling: a contaminated brush, a dirty tray, an open drum in a dusty bay.

Diagnose in that order. Filter a sample and look at the residue; SEM the scratch to see whether it is a gouge or an embedded particle; then run a blank pad with fresh DI water. If the scratch disappears, the powder was not the source and no amount of grade changing will fix it.

Powder or slurry

Slurry is powder plus water plus dispersant. You pay freight on the water, and in exchange you get a fixed dilution, no make-up equipment on site, and a concentration that only drifts if circulation stops. High-volume lines with their own slurry plant normally stay on powder. Automated lines with frequent changeovers normally prefer slurry, and the one-to-one numbering between our powder and slurry grades means no new purchasing code and no new approval file.

CeO₂ is not TREO. TREO is total rare-earth oxide; the CeO₂/TREO ratio is what separates a high-ceria grade from a mid-ceria one, and it moves both the price band and the etch behaviour. A quotation that gives “99.9% purity” and nothing else cannot be compared with another quotation.

All technical notes

Send us the application — the grade and the parameters come back together

Material, machine, target result: three lines are enough. Grade recommendation within 48 hours; standard grades in stock, samples in 3–5 working days.

Get a grade recommendation peter@ef-global.cn