A “cut point” is a single number standing in for a curve. D50 is the median — the particle size at which half the mass reports to the underflow (removed) and half to the overflow (kept). D100 is the size above which nothing passes — total separation. They describe the same separator differently, and which one a spec quotes changes the number you read — on a shaker screen, on a cyclone, on a centrifuge.
Separation isn’t a clean cut
No real separator removes everything above a size and nothing below it. Plot the probability that a particle reports to the removed stream against its size and you get an S-shaped grade-efficiency (partition) curve, not a vertical step. Small particles mostly stay; large particles mostly go; in between is a band where it’s a coin toss. A single “cut point” is just one chosen point on that curve.
D50 — the median (equiprobable) size
D50 is the size with a 50 % chance of going either way — the equiprobable size. It is the natural “average” cut and the one most often quoted for hydrocyclones and centrifuges (a desilter’s ~15 µm or a centrifuge’s ~2–7 µm are D50-type figures). It tells you where the curve is centred, but nothing about how steep it is.
D100 — the size nothing passes
D100 is the conservative end: the size above which the separator removes essentially everything. For the same device, D100 is a larger micron value than D50, because you have to go coarser before separation becomes total. It answers a different question — not “where is the average cut?” but “what is the largest particle that could still get through?”
Sharpness — the part a single number hides
Two separators can share a D50 and behave nothing alike if their curves have different slopes. Sharpness — often expressed as the ratio d75/d25 (closer to 1 is sharper) — measures how steep the partition curve is. A sharp separation has D50 and D100 close together; a sloppy one has a wide gap between them. This is why quoting only a D50 can flatter a poor separator: the median looks fine while the curve is so shallow that coarse particles still slip through.
Why it matters for shaker screens
This is exactly where the standard changed. API RP 13C labels a screen by its D100 cut point — the size at which nothing larger passes — while the older API RP 13E rated screens on a D50 basis. Same screen, stricter reference: a panel once sold as a “200 mesh / D50” screen carries a coarser-sounding API number on the D100 basis. If you compare a D100 number from one source with a D50 number from another, you are comparing two different points on two different curves.
Side by side
| D50 | D100 | |
|---|---|---|
| Definition | 50 % reports to underflow | Nothing larger passes |
| Question it answers | Where is the average cut? | What’s the largest that gets through? |
| Micron value | Smaller | Larger (for the same device) |
| Typically quoted for | Cyclones, centrifuges; old RP 13E screens | API RP 13C shaker screens |
Key takeaways
A cut point is one point on a separation curve, and the basis matters. D50 is the median — half removed, half kept. D100 is total separation — nothing larger passes — and it is always a coarser number than D50 for the same device. Sharpness (d75/d25) tells you how far apart they sit. Before you compare two cut points, make sure they are the same D — API RP 13C screens are D100, where the old standard was D50.