Laboratory instrument image representing scale-aware electron microscopy analysis

SEM guide

SEM particle size statistics start with calibration and sampling.

A segmentation mask can look convincing while the distribution remains biased by scale, agglomerates, field selection or an undefined particle diameter.

Scale

Verified physical units per pixel.

n

Included particles and fields.

Shape

Explicit size descriptor.

Evidence summary

Evidence summary

Reliable SEM particle sizing requires a verified pixel scale, a stated particle-size descriptor, a segmentation rule that can be audited against the image, and fields selected independently of the desired result. Report particle count, excluded objects, sampling strategy and distribution rather than only a mean.

Key takeaways

  1. 01Calibrate from metadata or an intact scale bar before measuring pixels.
  2. 02State the size definition: equivalent circular diameter, Feret diameter, area or another descriptor.
  3. 03Treat touching particles and edge objects with an explicit rule.
  4. 04Sample multiple representative fields and report the full distribution and particle count.
Laboratory instrument image representing scale-aware electron microscopy analysis
Editorial instrument image. Particle measurements must be made from original calibrated micrographs, not from this illustration.

Verify scale before segmentation

Use original image metadata when available and cross-check the rendered scale bar. Record pixel size, magnification, working distance and any image resizing or cropping.

  • Reject screenshots with uncertain scaling.
  • Do not infer size from magnification alone.
  • Keep the original micrograph beside the analysis copy.
[1] International Organization for Standardization, 2014

Define what particle size means

Non-spherical particles do not have one diameter. Equivalent circular diameter, minimum and maximum Feret diameters, projected area and aspect ratio answer different questions.

  • Choose the descriptor before measuring.
  • Report shape metrics when morphology is anisotropic.
  • Do not mix descriptors across samples.

Make segmentation reviewable

Overlay masks and object boundaries on the SEM image. A distribution should be traceable back to individual included, split, merged and excluded objects.

  • Document threshold and cleanup rules.
  • Set an edge-object policy.
  • Review touching-particle separation.
[2] Shanthi et al., 2018

Sample fields, not favorite images

Use a planned field-selection rule across independent regions or specimens. Report the number of fields, particle count and whether the distribution changes materially by field.

Methodology and scope

This guide follows static image-analysis principles and adapts them to SEM micrographs. It focuses on traceable calibration, segmentation review and sampling rather than claiming that one thresholding algorithm fits every morphology.

Limitations

  • A two-dimensional projection does not uniquely recover a three-dimensional particle size or shape.
  • Agglomeration, charging, contrast variation and overlapping particles can bias segmentation.
  • A scale bar embedded in a resized image can be invalid if the image and bar were transformed differently.
  • Fields chosen for visual appeal are not a representative sampling design.

References

  1. [1]

    Particle size analysis — Image analysis methods — Part 1: Static image analysis methods

    International Organization for Standardization. ISO 13322-1:2014 (2014).

    Open source ↗
  2. [2]

    Automatic detection of particle size distribution by image analysis

    Shanthi et al.. Micron (2018).

    doi:10.1016/j.micron.2017.12.002 ↗

Suggested citation

Suggested citation

SciPhys Research Team. “SEM Particle Size Analysis.” SciPhys, August 5, 2026. https://www.sciphys.com/blog/sem-particle-size-analysis

Apply the workflow

Audit particle measurements on your SEM image.

Upload the original micrograph, verify its scale and keep object masks connected to the reported distribution.

Analyze an SEM image

Sign-in required · starts with an SEM upload

Review SEM capabilities

What it does

Built around scientific evidence.

Verify scale before segmentation

Use original image metadata when available and cross-check the rendered scale bar. Record pixel size, magnification, working distance and any image resizing or cropping.

  • Reject screenshots with uncertain scaling.
  • Do not infer size from magnification alone.
  • Keep the original micrograph beside the analysis copy.

Define what particle size means

Non-spherical particles do not have one diameter. Equivalent circular diameter, minimum and maximum Feret diameters, projected area and aspect ratio answer different questions.

  • Choose the descriptor before measuring.
  • Report shape metrics when morphology is anisotropic.
  • Do not mix descriptors across samples.

Make segmentation reviewable

Overlay masks and object boundaries on the SEM image. A distribution should be traceable back to individual included, split, merged and excluded objects.

  • Document threshold and cleanup rules.
  • Set an edge-object policy.
  • Review touching-particle separation.

Workflow

From raw files to research decisions.

01

Calibrate

Confirm the file, units, acquisition settings and sample context before calculation.

02

Segment

Apply a documented method while keeping parameters and intermediate evidence visible.

03

Audit

Inspect diagnostics, compare samples consistently and export the evidence with the result.

FAQ

Questions researchers ask first.

How many particles should an SEM size distribution include?+

There is no universal number. Define precision and sampling goals, use multiple representative fields and report both particle and field counts so sampling uncertainty can be evaluated.

What diameter should I report for irregular particles?+

Use a descriptor tied to the research question, such as equivalent circular diameter plus Feret diameters and aspect ratio, and state it explicitly.

Can AI segment every SEM image automatically?+

No. Contrast, topology, charging and agglomeration vary. Automated masks must remain visible and reviewable against the original image.

Continue exploring

SEM analysis methods

See scale-aware image analysis.

AFM roughness guide

Review surface metrics and preprocessing.

Editorial policy

How SciPhys separates evidence and interpretation.