
D50 is useful, but it is only the midpoint of a measured particle size distribution. Two electronic-grade TiO₂ powders can share the same D50 and still differ sharply in fine particles, coarse particles, hard agglomerates, dispersion response and ceramic-processing behavior. For purchasing and R&D teams, a defensible evaluation therefore needs the complete distribution, the measurement basis and the sample-preparation protocol—not one number copied from a TDS.
Quick answer: Evaluate electronic-grade TiO₂ with at least D10, D50, D90 and Span, then confirm whether the results are volume-, number- or intensity-based and how the sample was dispersed. Add microscopy when primary particles, morphology or hard aggregates matter. Compare suppliers only under the same method and preparation conditions, and finally validate the powder in mixing, calcination and sintering trials. D50 alone cannot reveal a broad coarse tail, a fine-particle fraction, bimodality or persistent agglomerates.
1. What does D50 mean in a TiO₂ particle size report?
D50 is the particle diameter below which 50% of the reported distribution lies. In a volume-based laser-diffraction report, it is more precisely written as Dv(50): 50% of the measured particle volume is attributed to particles with an equivalent diameter smaller than that value, and 50% to larger particles.
The phrase equivalent diameter matters. Laser diffraction does not place a ruler across every titanium dioxide particle. It measures a light-scattering pattern and uses an optical model to calculate the distribution of spheres that would produce a corresponding pattern. Irregular particles, porous clusters and agglomerates are therefore represented by an equivalent-sphere size.
D50 is valuable for tracking the center of a repeatable distribution. It can support lot-to-lot trending and process control when the material, method, instrument settings and dispersion procedure remain aligned. It becomes misleading when buyers treat it as a complete description or compare values generated by different methods.
D50 does not tell you:
- How much very fine material is present.
- Whether a small but important coarse tail exists.
- Whether the distribution is narrow, broad or bimodal.
- Whether the measured objects are primary particles, fused aggregates or loose agglomerates.
- Whether the sample dispersed completely under the chosen test conditions.
- Whether two reported medians are based on volume, number or intensity.
- Whether the result predicts performance in a specific ceramic formulation.
Avoid a common wording error: do not call a laser-diffraction D50 the “average primary-particle size” unless the supplier has demonstrated that the measured objects are individual primary particles and has stated the reporting basis. “Median particle size under the stated test method” is more accurate.
2. Primary particles, aggregates and agglomerates are not the same

Powder terminology is not perfectly consistent across industries, so every report or specification should define its terms. A practical ISO-style distinction for ceramic raw materials is:
- Primary particle: the smallest discrete particle identified by a stated technique, normally electron microscopy. The identification method must be named because the apparent boundary can depend on image resolution and preparation.
- Aggregate: a group of strongly bonded or fused particles that is difficult to separate by routine dispersion. Neck formation, crystal growth or strong chemical bonding may make the group behave as one persistent unit.
- Agglomerate: a more loosely bound collection of particles or aggregates. It may break into smaller units when exposed to an appropriate dispersant, shear or controlled ultrasonic energy.
Why does this matter for electronic-grade TiO₂? A laser-diffraction result obtained after strong wet dispersion may describe a deagglomerated state. The same powder fed dry into a mixer may enter the real process as much larger clusters. Conversely, a mild test may leave removable agglomerates intact and make the material appear coarser than it becomes after production milling.
The difference between a soft agglomerate and a hard aggregate is therefore operational, not merely semantic. Ask how much energy was needed to change the distribution and whether the result stabilised. A dispersion-energy study—measuring PSD after progressively longer or stronger treatment—can reveal whether the apparent coarse fraction is easily dispersed or persistent.
3. How to read D10, D50, D90 and Span

For a cumulative distribution:
- D10 is the diameter below which 10% of the distribution lies. It helps describe the fine side of the powder.
- D50 is the median diameter. It locates the center of the distribution.
- D90 is the diameter below which 90% of the distribution lies. It is often more sensitive than D50 to a coarse tail or incomplete deagglomeration.
The distribution width is frequently summarized by Span:
Span = (D90 − D10) / D50
A lower Span generally indicates a tighter distribution, but it is not automatically “better.” A formulation may need a broader or deliberately bimodal distribution to improve packing. Span is also not portable across unlike methods: it is meaningful only when the same size basis, instrument model or verified equivalence, optical parameters and dispersion protocol are used.
Report the basis with the percentile
Write Dv(10), Dv(50) and Dv(90) for a volume distribution when possible. Number-based values can be dramatically smaller because every small particle counts equally, while a small number of coarse particles can dominate a volume distribution. Intensity-based dynamic light-scattering results apply another weighting. Converting between intensity, volume and number introduces model assumptions and can magnify errors, especially in broad or multimodal samples.
4. How can two powders with the same D50 behave differently?

| Illustrative result | Sample A | Sample B | What D50 misses |
|---|---|---|---|
| D10 | 0.32 μm | 0.12 μm | Sample B has a much larger fine fraction. |
| D50 | 0.54 μm | 0.54 μm | The medians are identical. |
| D90 | 0.82 μm | 3.20 μm | Sample B has a pronounced coarse tail. |
| Span | 0.93 | 5.70 | Sample B is far broader under this test. |
These numbers are constructed examples, not XC-258 data. They show why a median can remain stable while both tails change. In production, a broad distribution may reflect intentional packing design, mixed populations, incomplete dispersion, contamination or hard agglomerates. The curve shape and the process context determine which interpretation is reasonable.
Look beyond three percentile values when the application is sensitive
D10, D50, D90 and Span are compact summary statistics. They can still hide a secondary shoulder, two distinct modes or a very small oversize population. Ask for the full differential and cumulative curves when coarse particles could cause local stoichiometric variation, screen blockage, surface defects or residual phases. A coarse-tail control such as D95, D99 or an oversize limit may be useful, but only after the method demonstrates acceptable repeatability in that tail.
5. Which particle-size method should you use for electronic-grade TiO₂?

| Method | What it is useful for | Main limitation or control point |
|---|---|---|
| Laser diffraction | Routine D10/D50/D90, distribution shape and coarse-tail monitoring over a broad range | Equivalent-sphere model; results depend on wet/dry dispersion, optical inputs, concentration and instrument settings |
| SEM or TEM with image analysis | Primary-particle size, shape, necking and aggregate morphology | Small field of view and sample-selection bias; preparation and counting rules must be defined |
| Dynamic light scattering | Hydrodynamic size of sufficiently dispersed submicron suspensions | Large particles strongly influence intensity; converted volume/number results are model-dependent |
| Sedimentation or centrifugal methods | Size distribution inferred from settling behavior | Requires density and shape assumptions; convection and aggregation can distort results |
| Sieving or air-jet sieving | Coarse contaminants, granules or oversize material | Not suitable for characterizing submicron primary particles |
| BET surface area | Accessible surface area and a complementary indication of surface evolution | BET-derived diameter assumes particle shape, density and nonporosity; it is not a direct PSD |
| XRD crystallite size | Coherently diffracting domain size and phase information | A particle may contain several crystallites; crystallite size is not particle size |
For supplier qualification, a practical package is often laser diffraction for the bulk distribution, SEM for morphology and cluster structure, and application testing for the final decision. TEM can be added when nanoscale primary-particle information is genuinely needed. BET and XRD help explain changes in surface area and crystallinity but should not be substituted for a PSD measurement.
6. Why do test conditions change a laser-diffraction result?
ISO 13320 treats laser diffraction as a method that must be controlled and validated. A comparison is only fair when the following details are aligned or shown to be equivalent:
- Wet or dry dispersion: wet testing can use dispersants and ultrasound; dry testing depends on feed conditions and air pressure. They may represent different powder states.
- Dispersion medium and chemistry: liquid, pH, ionic strength and dispersant dosage affect whether TiO₂ clusters separate or re-form.
- Ultrasonic energy and time: insufficient energy can leave soft agglomerates; excessive treatment can heat the suspension, damage fragile structures or introduce a test state unlike production.
- Stirring and circulation: settling, foam or poor suspension uniformity can change the sampled population.
- Optical model: the assumed refractive index, absorption and the choice of Mie or another approximation influence calculated fine-particle results.
- Concentration or obscuration: too little material raises noise; too much can create multiple-scattering errors.
- Background and cleanliness: bubbles, contaminated cells and carryover may appear as extra size populations.
- Sample history: humidity exposure, compaction, transport and storage can alter agglomeration before measurement.
Use a dispersion study, not a single arbitrary setting
During method development, measure the sample at several controlled dispersion energies or times. If D90 decreases and then reaches a stable plateau while D50 changes little, the coarse tail was likely dominated by removable agglomerates. If the result continues to move without stabilizing, the method may be changing the sample or suffering from instability. Record the validated endpoint and use it consistently for incoming QC.
Repeatability is just as important as resolution. Run replicate preparations, not only repeat readings from the same beaker. Separate sample-preparation variation from instrument-reading variation, and use a representative sampling method for the bulk powder.
7. Why does TiO₂ particle size distribution matter in electronic ceramics?
In electronic ceramics, TiO₂ is often one reactant within a multicomponent powder system. Its distribution affects how uniformly titanium contacts the other raw materials before calcination. That can influence reaction distance, mixing efficiency, packing and the likelihood of local regions that convert differently from the bulk.
Fine fraction
A larger fine fraction can increase surface area and may accelerate some reactions, but it can also increase dispersant demand, slurry viscosity, moisture sensitivity and agglomeration tendency. Very fine particles may be difficult to meter and may respond differently to binder addition or granulation. The correct D10 is therefore process-specific.
Coarse tail and hard agglomerates
A coarse tail can indicate large particles, aggregates, agglomerates or contamination. In a ceramic route, persistent coarse units can create nonuniform contact with other reactants, survive milling, change green-body packing or leave local porosity and unreacted regions after firing. The risk is controlled by both size and strength: a 5 μm soft agglomerate that reliably breaks during milling is different from a 5 μm fused aggregate.
Distribution width and packing
A narrow distribution improves consistency and can simplify process control. A broader or bimodal distribution may sometimes improve packing because smaller particles occupy spaces between larger ones. However, a broad curve caused by uncontrolled agglomeration is not the same as an intentionally engineered bimodal powder. The specification should be linked to green density, slurry rheology, calcination response, sintered microstructure and the electrical property that matters for the component.
This is why particle-size approval should not be separated from chemical purity. Trace Na, K, Fe, Ca, Mg, Si and Al can affect electronic ceramics, while particle contact controls how the designed composition actually reacts. Review PSD together with the full impurity profile, phase, moisture, loss on ignition and batch consistency.
8. How should buyers evaluate XC-258 particle size data?
XC-258 electronic-grade titanium dioxide is a sulfate-process, high-purity rutile TiO₂ developed for electronic ceramic and functional material applications. The current public TDS lists a typical D50 of 0.54 μm, 99.6% TiO₂ content, 99.8% rutile content and controlled trace impurities.
The word typical is important. A typical value describes expected product behavior; it is not automatically a guaranteed release limit. For qualification, ask which particle-size values appear on the lot COA, which are controlled internally, and which test method and preparation conditions apply. See also our guide to reading a high-purity TiO₂ COA for electronic ceramics.
A practical XC-258 evaluation plan
- Use the TDS to screen basic suitability: phase, purity, typical D50 and impurity profile.
- Request the current specification and a representative lot COA; distinguish limits from typical values.
- Ask for the laser-diffraction method, including wet/dry route, dispersion medium, energy and reporting basis.
- When PSD is critical, review D10, D50, D90, Span and the full curve rather than D50 alone.
- Compare at least three lots with the same customer-side method and representative sampling.
- Use SEM to check morphology and persistent clusters if process behavior does not match the bulk PSD.
- Validate milling, slurry rheology, mixing, calcination, phase formation, sintering and final electrical performance.
Do not create an incoming limit from one supplier certificate without a process study. First establish the distribution window associated with acceptable manufacturing performance, then set limits that the measurement system can reproduce.
9. Particle-size checklist for purchasing and incoming QC
| Question | Why it matters | What to record |
|---|---|---|
| What does “particle size” mean here? | It may refer to primary particles, dispersed clusters, granules or an equivalent diameter. | Measurand and terminology |
| Which method generated the data? | Different methods do not produce interchangeable results. | Standard, instrument and software version |
| Is the result volume-, number- or intensity-based? | The weighting can shift percentiles substantially. | Dv, Dn or intensity basis |
| What are D10, D50, D90 and Span? | They describe center and width better than D50 alone. | Values, units and full curves |
| How was the sample dispersed? | Preparation determines whether agglomerates remain. | Wet/dry, medium, dispersant, pressure, sonication and time |
| Are coarse particles persistent? | Hard aggregates can survive production milling and create defects. | Energy study, microscopy and oversize result |
| How representative is the sample? | Fine powders segregate and small test portions may miss coarse material. | Sampling and splitting procedure |
| Is the method precise enough for the limit? | A tight specification is useless if test uncertainty is comparable to the tolerance. | Repeatability, reproducibility and control sample |
| Does PSD correlate with process performance? | A laboratory number matters only when linked to production outcomes. | Slurry, calcination, sintering and final-property data |
10. Frequently asked questions
Is D50 the average particle size?
No. D50 is a median, not necessarily an arithmetic mean. It marks the 50% point on a stated distribution basis.
Is a lower D50 always better for electronic ceramics?
No. Finer powder can increase surface area and reactivity, but it may also raise viscosity, dispersant demand, agglomeration and handling difficulty. The optimum depends on the formulation and process.
Can two laboratories obtain different D50 values for the same TiO₂?
Yes. Differences in sampling, wet or dry dispersion, dispersant, sonication, optical parameters, obscuration, instrument design and analysis settings can shift the result.
Does laser diffraction measure primary particles?
Not automatically. It measures the optical response of the particles or clusters present under the test conditions and calculates an equivalent-sphere distribution. Microscopy is needed to support a primary-particle claim.
What does a high D90 usually indicate?
It indicates a relatively coarse upper portion of the measured distribution. Possible causes include large particles, agglomerates, aggregates, contamination or insufficient dispersion; the cause must be investigated.
Is Span a specification for narrow particle size distribution?
Span is a useful normalized width descriptor, but a target must be established with one controlled method and linked to process performance. Lower is not universally better.
Can BET surface area replace particle size distribution?
No. BET measures accessible surface area. A BET-equivalent diameter requires assumptions about shape, density and porosity and cannot reveal a coarse tail or bimodal PSD.
What particle-size data should a TiO₂ supplier provide?
At minimum, the supplier should identify the method, basis and preparation conditions behind D50. When PSD is critical, request D10, D90, Span or the full curve, plus evidence of lot-to-lot control and a clear distinction between specification limits and typical values.
Technical references
- ISO 13320:2020, Particle size analysis — Laser diffraction methods.
- NIST Recommended Practice Guide: Particle Size Characterization.
- ISO nanotechnology vocabulary for primary particles, aggregates and agglomerates.
- HORIBA technical guidance on D10, D50, D90 and Span.
- Role of Particle Substructure in the Sintering of Monosized Titania, Journal of the American Ceramic Society.
Need a particle-size and impurity review for XC-258?
Share your ceramic system, process route, current PSD method and application-critical limits. ICC can provide the current XC-258 documentation and support sample evaluation for electronic ceramic and functional material development.


