TECHNICAL GUIDE · INKJET DISPERSION QUALITY
D50, D90 and D99: How Particle-Size Distribution Affects Inkjet Ink Development
A small median is a useful starting point—not a complete qualification. Learn to read the distribution, identify the measurement basis and investigate the coarse tail before selecting an aqueous inkjet pigment dispersion.
Read the median and the tail together
An inkjet colorant contains a distribution of particle sizes. D50 describes its midpoint; D90 and D99 describe progressively higher percentiles. Two nano pigment dispersions can share the same D50 while having substantially different coarse tails. A single median cannot reveal that difference.
| Parameter | Meaning | What it does not establish |
|---|---|---|
| D50 | The diameter at the 50% cumulative point: the median. | The absence of larger particles or the arithmetic mean. |
| D90 | The diameter at the 90% cumulative point; approximately 10% lies above it on that basis. | A maximum size or a nozzle-compatibility guarantee. |
| D99 | The diameter at the 99% cumulative point; approximately 1% lies above it on that basis. | The largest particle present in the batch. |
High-end percentiles need particular care: rare coarse material, sample representativeness and measurement noise can influence them. HORIBA explains why extreme percentiles are less robust than central values. Do not substitute D99 for a validated oversized-particle detection method. Source: HORIBA particle-size specification guidance.
Dv, Dn and Di: the suffix changes the interpretation
A volume-based Dv50 means 50% of the measured particle volume is below that diameter. A number-based Dn50 refers to particle count. An intensity-based Di50 refers to the cumulative scattered-light intensity assigned by the analysis. These are different distributions, not interchangeable names for the same result. Confirm the notation in the instrument software. Source: Malvern Panalytical particle-characterization guide.
Before comparing suppliers, ask: “D50 on which basis, obtained by which method?” A percentage of light is not a percentage of particles. Converting optical distributions into volume or number distributions involves assumptions; a software conversion does not create an independent measurement.
DLS Z-average is not D50
Dynamic light scattering derives particle motion from fluctuations in scattered light. Its Z-average is a cumulants-derived average hydrodynamic size—not a median, a maximum, or the arithmetic average of a plotted peak. NIST research demonstrates that distribution-derived means can differ from the cumulant result. Source: Farkas and Kramar, Journal of Nanoparticle Research, 2021.
DLS can also display derived distributions and percentiles, but their interpretation requires care. Malvern Panalytical cautions against treating its DLS-derived volume percentiles as quantitative control values without appropriate method selection. Source: D90, D50, D10 and span for DLS.
Laser diffraction instead fits an optical scattering pattern to a particle-size distribution. The result uses an equivalent-sphere model; it need not equal a hydrodynamic diameter or a microscope dimension. Relevant current method frameworks are ISO 22412:2025 for DLS and ISO 13320:2020 for laser diffraction. Referencing these standards does not establish that an AIC report was tested in accordance with them.
Record the conditions before trusting a comparison
For laser diffraction, document the optical model, particle and medium refractive indices, absorption index, concentration or obscuration, circulation and measurement duration. Source: Reporting reproducible laser-diffraction measurements. For DLS, include temperature, dispersant viscosity, concentration and analysis settings. Source: NIST SP 1200-6, DLS measurement protocol. A reported diameter without a reproducible method is incomplete.
Take representative samples after an agreed mixing procedure. Record dilution ratio, diluent, waiting time and any sonication, filtration or centrifugation. Preparation can change the dispersion: removing coarse material or breaking agglomerates answers a different question from examining the supplied product. Wet-method development should distinguish primary-particle sizing from assessment of the existing agglomeration state. Source: Wet-dispersion method development.
AIC PR122: a supplier-data example, not a universal limit
The existing AIC-Magenta Red 1015 materials report D50 = 132 nm, D90 = 181 nm and D97 = 214 nm. The public COA example identifies lot 26122-C-065. Review the original particle-size evidence and chart links alongside the batch record.
These values illustrate one reported distribution. This article does not assign an unconfirmed volume, number or intensity basis, or measurement principle, to that report. No D99 is supplied: it cannot be inferred from D97. Do not extrapolate these numbers to other pigments, every shipment or a finished ink containing binders and additives.
Connect the coarse tail with filtration and jetting trials
Oversized particles and agglomeration can affect printhead operation and printed-film quality. A manufacturer application note describes laser diffraction as a tool for monitoring inkjet dispersion and milling. Its examples are not a universal acceptance limit for every modern printhead. Source: Laser-diffraction analysis of inkjet inks.
- Characterise the delivered dispersion. Retain the full curve, percentiles, method, lot and replicate results—not only D50.
- Recheck the complete ink. Compare before and after formulation, agreed storage exposure and relevant circulation conditions.
- Test filterability explicitly. Record membrane material, rating, area, sample mass, pressure and time. Check retained material and pigment loss where relevant.
- Qualify the printer. Evaluate drop formation, nozzle recovery, idle behaviour and sustained printing under the actual operating conditions.
A suitable distribution alone does not guarantee clean jetting or freedom from clogging. Consider viscosity, conductivity, pH and surface tension together with the ink vehicle, waveform and maintenance system.
White TiO₂ dispersions need a separate evaluation: light scattering and opacity are central objectives, so “smaller is always better” is not an adequate rule. Do not apply a CMYK nanometre threshold automatically to white ink. Source: Titanium-dioxide particle-size analysis.
Frequently asked questions
Does D50 below 150 nm mean every particle is below 150 nm?
No. D50 is the median on the stated basis. Inspect the tail and full distribution.
Can D97 be relabelled as D99?
No. They describe different cumulative points. Request the actual D99 result if your validated method requires it.
Can published literature replace a product test?
No. Literature supports interpretation and method design; a product claim requires data for the actual grade, batch and test conditions.
Compare AIC colourant grades and TDS, then request a current COA and the measurement details needed for your development programme.