Abstract
Resolve inconsistent particle-size reports by checking sampling, dispersion and distribution basis. This note provides a sequence for comparing D50 values without assuming that the powder has changed. A median alone does not describe the full distribution or its agglomerates.
Check what the instrument weights
Number-based and volume-based distributions weight the same population differently. A comparison of sizing methods illustrates why method and sample preparation belong with the reported result [1]. Do not convert one median into another using a simple scale factor.
Original example: consider 999 ideal spheres of diameter 1 µm and one sphere of diameter 10 µm. The large sphere has the same volume as 1000 small spheres. It therefore represents about 50.0% of total particle volume despite being only 0.1% of the particle count. A rare large particle can strongly affect a volume-weighted result.
Also compare the complete curve. Two powders can have the same D50 and different coarse tails. A delivery or surface-finish problem may follow the tail or occasional clusters while the median stays within its usual range.
| Proposed check | Hold constant | Interpretation |
|---|---|---|
| Repeat one aliquot | Instrument method | Short-term measurement variation |
| Measure fresh aliquots | Preparation protocol | Sampling and segregation |
| Change dispersion energy | Powder and medium | Agglomerate sensitivity |
| Compare with images | Representative sampling | Shape and large-particle check |
Build a comparable measurement
1. Reconcile the report fields
Confirm wet or dry measurement, distribution basis, optical model where applicable, dispersant and energy input. Record D10, D50 and D90 with the curve. A method identifier is useful only when it resolves to the actual preparation and analysis settings.
2. Use an energy ladder cautiously
Compare the initial dispersion with one or more controlled increases in energy. A decreasing coarse tail may indicate deagglomeration, but continued size reduction can also indicate particle damage. Examine morphology or another integrity measure before declaring the smallest result the true size.
3. Test a representative sample
Mix and split the bulk powder using a suitable reproducible procedure. Compare aliquots from the same split before comparing supplier and customer laboratories. If within-container variation dominates, changing the instrument will not resolve the underlying sampling problem.
Development decision
Agree a common method and retained reference sample before narrowing the D50 limit. If the functional failure tracks D90 or an image-based defect population, include that metric rather than relying on the median alone.
References
[1] Whiting JG, Garboczi EJ, Tondare VN, Scott JHJ, Donmez MA, Moylan SP. A comparison of particle size distribution and morphology data acquired using lab-based and commercially available techniques: Application to stainless steel powder. Powder Technol. 2022;396:648-662. doi:10.1016/j.powtec.2021.10.063.
https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=932029