{"id":16714,"date":"2026-04-23T07:02:48","date_gmt":"2026-04-23T07:02:48","guid":{"rendered":"https:\/\/www.jet-mills.com\/?p=16714"},"modified":"2026-04-23T07:04:33","modified_gmt":"2026-04-23T07:04:33","slug":"dry-powder-sampling-seven-details-that-determine-whether-your-psd-data-reflects-reality","status":"publish","type":"post","link":"https:\/\/www.jet-mills.com\/fr\/dry-powder-sampling-seven-details-that-determine-whether-your-psd-data-reflects-reality\/","title":{"rendered":"\u00c9chantillonnage de poudres s\u00e8ches\u00a0: sept d\u00e9tails qui d\u00e9terminent si vos donn\u00e9es PSD refl\u00e8tent la r\u00e9alit\u00e9"},"content":{"rendered":"

R\u00e9partition granulom\u00e9trique<\/a> data is only as good as the sample it came from. In dry powder processing \u2014 fraisage au jet<\/a>, ball milling, impact milling, or air classification \u2014 the measurement instrument is rarely the problem. The sample is. A poorly drawn sample from a segregated batch, a surface scoop from a container where coarse particles have floated to the top, or a sample contaminated by moisture during handling can make a stable, well-controlled process look like it is out of specification.<\/p>\n\n\n\n

The practical consequence is real: engineers adjust process parameters based on bad data. A false reading showing the product is too coarse triggers a classifier speed increase or a grinding time extension that was not needed. A false reading showing the product is too fine causes the opposite. Either way, unnecessary process changes consume energy, waste time, and introduce actual variability into a process that may have been perfectly stable.<\/p>\n\n\n\n

This article covers why dry powder is inherently difficult to sample representatively. It also explains specific protocols that produce samples that actually reflect the batch.<\/p>\n\n\n\n

\n
\"\"<\/figure>\n\n\n\n
\"\"<\/figure>\n<\/figure>\n\n\n\n

Why Dry Powder Sampling Goes Wrong<\/mark><\/h2>\n\n\n\n

Particle Size Segregation<\/mark><\/h3>\n\n\n\n

Dry powder segregates \u2014 coarse and fine particles separate spatially \u2014 whenever the powder experiences gravity, vibration, or airflow. The mechanism differs by context: in a static container, fines percolate downward between coarser particles under gravity, concentrating at the bottom while coarser material rises to the top. In a pneumatically conveyed stream, fines are carried preferentially by airflow into dead zones and low-velocity regions. This is while coarser particles report more directly to the primary collection point. In a vibrating container (such as during transport), fines migrate downward and coarse particles rise \u2014 the Brazil nut effect.<\/p>\n\n\n\n

The practical consequence for sampling is that no single position in a batch represents the full PSD of the batch. A sample taken from the top surface of a container will be enriched in coarse particles; a sample taken from the bottom will be enriched in fines. Neither is wrong \u2014 both are real measurements \u2014 but neither represents the overall batch PSD that will be experienced by the downstream process.<\/p>\n\n\n\n

Surface Area, Static Charge, and Agglomeration<\/mark><\/h3>\n\n\n\n

Fine dry powders \u2014 particularly those with D50 below 10 microns from fraisage au jet<\/a> or fine impact milling \u2014 have high specific surface area and accumulate electrostatic charge. These two properties cause particles to adhere to container walls, sampling tool surfaces, and to each other. When a sample is scooped from a container, the finest particles adhere to the scoop surface and are not proportionally transferred to the sample container. The sample is therefore coarser than the bulk material.<\/p>\n\n\n\n

Agglomeration adds a different error. Van der Waals forces and electrostatic attraction between fine particles form soft agglomerates \u2014 clusters of particles that behave as a single larger particle during laser diffraction measurement if the dispersion conditions are not adequate to break them. An agglomerated sample from a well-milled product will report a bimodal or broadened PSD with an apparent coarse tail that does not reflect the true primary particle size distribution.<\/p>\n\n\n\n

Dry Powder Characteristic<\/strong><\/td>Sampling Consequence<\/strong><\/td>Error Direction<\/strong><\/td><\/tr>
Wide PSD \u2014 coarse and fine together<\/td>Gravity segregation in container; coarse and fine separate spatially<\/td>Single-point sample misrepresents the batch<\/td><\/tr>
High specific surface area and static charge<\/td>Particles adhere to sampling tools; fines lost to walls<\/td>Sample is coarser than bulk \u2014 apparent under-grinding<\/td><\/tr>
Soft agglomerates (static, van der Waals)<\/td>Agglomerates report as coarse particles in laser diffraction if not dispersed<\/td>Apparent coarse tail \u2014 false reading<\/td><\/tr>
Hygroscopic surface (pharmaceuticals, food)<\/td>Moisture absorption during sampling exposure causes agglomeration<\/td>D50 inflates; true fine distribution masked<\/td><\/tr>
Low bulk density, easy fluidisation<\/td>Airflow during handling redistributes particles<\/td>Unpredictable size fractionation by air currents<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Seven Sampling Details That Make a Measurable Difference<\/mark><\/h2>\n\n\n\n
\"Poudre<\/figure>\n\n\n\n

1. Use Multi-Point Sampling, Never Surface Sampling<\/mark><\/h3>\n\n\n\n

A single surface sample from a bag, bin, or product container is the most common sampling error in dry powder processing. The surface is systematically enriched in coarse particles from gravitational segregation. The only representative sample from a static container is one that includes material from multiple depths \u2014 top, middle, and bottom \u2014 combined in proportion to the container volume at each level.<\/p>\n\n\n\n

For a bag or small container, insert a sampling thief (a hollow tube with a side opening) vertically to the bottom, then extract a core sample. Combine cores from at least three positions across the container. For a pile on a conveyor or floor, take samples from the top, the middle layers, and the base using a tube sampler rather than a surface scoop.<\/p>\n\n\n\n

2. Sample Across the Full Width of a Moving Stream<\/mark><\/h3>\n\n\n\n

The most representative sample from a production line is a cross-cut sample taken across the full width of the powder stream as it falls from a conveyor, classifier outlet, or mill discharge. This is the gold standard method \u2014 a sample taken this way captures the full PSD as it exits the process, before any post-process segregation occurs. The practical implementation is a traversing sample cutter that passes repeatedly across the full stream width and collects a time-averaged sample.<\/p>\n\n\n\n

In-line sampling from a pipe or duct \u2014 using a side-entry probe \u2014 is less reliable because particles in different size fractions have different velocities and trajectories in the flowing stream. A probe at a fixed point in a pipe preferentially captures particles with certain aerodynamic characteristics rather than a representative cross-section.<\/p>\n\n\n\n

3. Use a Rotary Sample Divider for Sub-Sampling<\/mark><\/h3>\n\n\n\n

Once a bulk sample has been collected from multiple positions, it typically needs to be reduced to the quantity required for instrument analysis (a few grams to tens of grams for most laser diffraction instruments). Manually scooping a sub-sample from the bulk introduces another segregation step \u2014 fine particles in the bulk will be unevenly distributed, and a random scoop will not capture them proportionally.<\/p>\n\n\n\n

A rotary sample divider (riffle splitter) consistently divides a powder into equal fractions by distributing the falling stream across multiple collection chutes. The sub-sample from a rotary divider is statistically equivalent to the bulk. Manual sub-sampling from a scoop is not, regardless of how carefully it is done.<\/p>\n\n\n\n

4. Check for Hard vs. Soft Agglomerates Before Choosing Dispersion Method<\/mark><\/h3>\n\n\n\n

Agglomerates in dry powder fall into two categories that require different handling. Soft agglomerates \u2014 held together by electrostatic attraction or weak van der Waals forces \u2014 can be broken by the pneumatic dispersion unit of a laser diffraction instrument at standard pressure (1-3 bar). The instrument’s report should be checked for consistency across multiple injection pressures: if the PSD becomes significantly finer as dispersion pressure increases from 1 to 3 bar, soft agglomerates are present and the measurement should be taken at the higher pressure.<\/p>\n\n\n\n

Hard agglomerates \u2014 held by sintering, chemical bonding, or strong mechanical interlocking \u2014 cannot be broken by standard pneumatic dispersion. These require ultrasonic treatment in a liquid medium before dry measurement, or wet laser diffraction with ultrasonic assistance. Attempting to measure a hard-agglomerate powder by dry dispersion alone will consistently give a falsely coarse PSD.<\/p>\n\n\n\n

5. Seal Samples Immediately After Collection<\/mark><\/h3>\n\n\n\n

Many dry powders \u2014 pharmaceutical excipients, food-grade minerals, hygroscopic fine chemicals \u2014 absorb moisture rapidly from ambient air. Even a few minutes of open exposure in a humid environment is sufficient to cause measurable surface moisture that promotes agglomeration and inflates the measured D50. The sample container should be sealed within seconds of collection, not minutes.<\/p>\n\n\n\n

Use dry, pre-cleaned containers. If the production environment is humid (above 60% relative humidity), consider transferring samples in a portable desiccator or using nitrogen-flushed sealed containers for the transfer step. These precautions are routine in pharmaceutical applications and are worth adopting for any hygroscopic fine powder regardless of industry.<\/p>\n\n\n\n

6. Mix Before Sub-Sampling, But Gently<\/mark><\/h3>\n\n\n\n

Even a multi-point bulk sample collected correctly will have some inhomogeneity from local variations. Mixing the bulk sample before sub-sampling reduces this variation. However, the mixing method matters: vigorous agitation of a fine dry powder can generate electrostatic charge that promotes agglomeration, and tumbling in an open container exposes the surface to ambient humidity.<\/p>\n\n\n\n

The recommended approach is gentle rolling or inversion of a sealed container \u2014 20-30 roll cycles is sufficient for most powders. Avoid high-energy mixing methods (vortex mixer, shaker) that generate charge or heat in the sample.<\/p>\n\n\n\n

7. Record the Sampling Protocol as Part of the PSD Report<\/mark><\/h3>\n\n\n\n

In regulated industries (pharmaceutical, food, electronics), the sampling method is as important as the measurement result for product release decisions. A PSD report without a documented sampling protocol cannot be meaningfully audited or disputed. If a customer challenges the result, there is no way to assess whether the measurement or the sampling was the source of the discrepancy.<\/p>\n\n\n\n

Even in non-regulated industries, recording the sampling protocol enables comparison of results across production runs, shifts, and facilities. If PSD data varies unexpectedly, a documented protocol allows the investigation to determine whether the sampling changed rather than the process.<\/p>\n\n\n\n

Sampling Protocol Checklist for Dry Milled Powder<\/mark><\/h3>\n\n\n\n

\u2022 Collection points: Minimum 3 positions (top, middle, bottom) from static containers; full cross-cut from moving streams
\u2022 Tools: Sampling thief or tube sampler for containers; traversing cutter for streams. No surface scoops.
\u2022 Sub-sampling: Rotary riffle splitter for reducing bulk to analysis quantity. No manual scooping.
\u2022 Agglomerate check: Test PSD at 1 bar and 3 bar dispersion pressure \u2014 if D50 changes by more than 10%, soft agglomerates are present; use 3 bar
\u2022 Sealing: Seal sample container within 30 seconds of collection. Use dry, pre-cleaned container.
\u2022 Mixing: 20-30 gentle rolls in sealed container before sub-sampling
Documentation: Record sampling position, time, operator, container type, and dispersion pressure used<\/p>\n\n\n\n

How Mill and Classifier Performance Affects Sampling Reliability<\/mark><\/h2>\n\n\n\n
\"talc-jet-milling\"<\/figure>\n\n\n\n

Some PSD variation attributed to sampling is actually real process variation from the milling circuit. A milling system that produces consistent, narrow PSD batch after batch requires fewer samples, less frequent sampling, and produces samples that are genuinely representative because the spatial PSD variation within the batch is small.<\/p>\n\n\n\n

A mill or classifier running with unstable feed rate, worn classifier wheel, or fluctuating air pressure will produce various batches. No sampling protocol can make a variable process look consistent. When sampling shows unexpected variation, the first diagnostic question should be whether the variation is in the milling process or in the sampling technique. This requires comparing samples taken at the mill outlet against samples taken from the finished product container. If the mill outlet samples show consistent PSD but the finished product samples show wide variation, the source is post-process segregation during storage or transport, not the milling process.<\/p>\n\n\n\n

Questions About Sampling from an EPIC Powder Milling Line?<\/mark><\/h3>\n\n\n\n

Poudre \u00c9PIQUE<\/a> Machinery’s application engineers can advise on sampling protocol design for specific materials and mill types. It includes recommended sampling points, sampling frequency, and dispersion settings for laser diffraction analysis. We can also help you assess whether PSD variation is coming from the milling process or from sampling.<\/p>\n\n\n\n

Contactez-nous<\/a> with your material, your milling target, and the PSD variation. We will give you a specific recommendation.<\/p>\n\n\n\n

Ask a Technical Question: www.jet-mills.com\/contact-us<\/a>
Explore Our Jet Mill Range: www.jet-mills.com<\/a><\/p>\n\n\n\n

\n