8 Practical Lessons from a Particle Analysis Webinar

The answer is yes — and much more.

But what this question really signals is not a request for another number. It reflects a broader expectation: SEM-based particle analysis should integrate naturally with how materials are already described and compared.

Size distributions are a baseline requirement. The real value emerges when size can be connected to:

• Shape
• Mineralogy
• Lithotype
• Inclusion type
• Surface structure

That is when particle analysis moves from reporting numbers to understanding materials.

• Does segmentation depend on B/C?
• Can different B/C settings be saved per sample?
• What happens in heterogeneous samples?
• Does automatic B/C cause problems?

These are not software questions — they are data quality questions. Brightness and contrast define the grayscale dynamic range used for:

• Particle detection
• Segmentation
• X-ray point placement
• Classification thresholds

If brightness and contrast are unstable, segmentation becomes unstable.
If segmentation is unstable, statistics lose meaning.

The recurring theme was clear: reproducibility matters more than automation speed.

Channeling is often treated as a problem:

• It can affect BSE-based classification
• It can introduce contrast unrelated to chemistry

But users also recognize that:

• Channeling can help differentiate grains
• It can assist in grain size or texture analysis

This highlights a recurring truth in SEM-based particle analysis: many “artifacts” are only artifacts if you are not trying to measure them.

The key is knowing when to suppress an effect — and when to use it deliberately.

Real samples are messy:

• Particles overlap
• Precipitates grow on other particles
• Agglomerates dominate statistics

Handling this requires:

• Robust geometrical separation
• Careful control of histogram dynamic range
• Sometimes, multiple imaging signals (SE + BSE, or BSE + CL)
• In difficult cases, machine-learning-based de-agglomeration

The takeaway is simple: segmentation strategy must match sample reality — not idealized assumptions.

• Does the system adjust B/C between samples?
• Can it maintain stability within a sample?
• Can settings be changed mid-run?

These questions all point to the same operational reality: heterogeneous samples are the norm, not the exception.

Automation that assumes uniform contrast, particle density, or composition will eventually fail. Robust workflows acknowledge variability and actively manage it.

• Custom holders
• Saving settings per sample
• Reporting formats
• Stop criteria
• Resume-after-pause functionality

Individually, these topics sound mundane. Collectively, they define whether a system works at scale.

High-volume particle analysis lives or dies on:

• How fast jobs can be set up
• How reliably they run unattended
• How easy it is to audit and report results

These are not edge cases — they are the core of industrial particle analysis.

This points to a clear trend:

• Particle analysis is no longer isolated
• Users want to align morphology, chemistry, and mechanical properties
• Spatial correlation across techniques is becoming more important

SEM-based particle analysis increasingly acts as a spatial reference framework for other measurements.

The honest answer remains: it depends.

What matters is that:

• Statistics are monitored during acquisition
• Users can see when distributions stabilize
• Stop criteria are flexible (particles, frames, time)

Good particle analysis workflows help users know when to stop — not guess.