Image-Based Particle Size Analysis: Principles, Standards, Methods, and Applications

Image-based particle size analysis is a high-precision particle characterization technique that combines optical or electron microscopy with digital image processing to measure both the size and the shape of individual particles. Governed by international standards such as ISO 13322-1 and ISO 13322-2, the method delivers true number-based particle size distributions together with quantitative morphological parameters that laser diffraction alone cannot provide. This article explains the working principles, sample requirements, testing procedures, and typical applications of static and dynamic image analysis in modern materials testing laboratories.

Technical Overview

Image-based particle size analysis, often referred to simply as image analysis, is a direct particle characterization method in which magnified images of individual particles are captured, digitally processed, and evaluated to determine the particle size distribution and particle shape. Unlike ensemble techniques such as laser diffraction, which infer size from a scattering pattern and rely on optical model assumptions, image analysis measures each particle individually and provides visual evidence of the measured population. This makes it one of the few techniques capable of simultaneously quantifying size and morphology.

Working Principle

Particles are dispersed on a measurement surface (static mode) or flowed through a measurement cell (dynamic mode), illuminated, and imaged with an optical microscope, an electron microscope, or a high-speed camera. Each digital image is converted into a binary representation through thresholding and segmentation. Image analysis software then extracts geometric features from every particle silhouette and converts them into statistically valid size distributions and shape descriptors.

Static Versus Dynamic Image Analysis

Static image analysis (ISO 13322-1) records particles at rest, typically on a microscope slide or substrate, and is well suited to very fine particles, irregular shapes, and applications requiring electron microscopy resolution. Dynamic image analysis (ISO 13322-2) records particles in motion as they pass a pulsed light source and camera, enabling tens of thousands of particles to be measured within minutes and covering a broad size range, typically from about 1 micrometer up to several millimeters.

Equivalent Diameters and Shape Factors

Because projected particle images are two-dimensional, size is expressed through equivalent diameters rather than a single physical dimension. Common descriptors include the circle-equivalent diameter (the diameter of a circle with the same projected area), maximum and minimum Feret diameters, and the Martin diameter. Shape is quantified through morphological parameters such as circularity, aspect ratio, elongation, convexity, and solidity, which together reveal whether particles are spherical, angular, fibrous, porous, or agglomerated.

Testing Samples

Image-based analysis is applicable to a wide variety of particulate systems in dry or liquid form. Proper sampling and dispersion are critical, because the technique measures a comparatively small number of particles and is therefore sensitive to non-representative specimens.

Dry Powders and Granules

Free-flowing powders such as metal powders, ceramic granules, abrasive grains, refractory raw materials, food powders, and chemical intermediates can be dispersed dry onto a substrate or measured with a dry air dispersion unit. Dry dispersion must be energetic enough to break soft agglomerates yet gentle enough to avoid fracturing the primary particles.

Suspensions, Slurries, and Emulsions

Particles suspended in water, alcohol, or process liquids, including pigment dispersions, battery electrode slurries, drilling mud solids, and droplet emulsions, are typically measured wet. A suitable carrier liquid and, where necessary, a compatible surfactant or dispersant are selected to keep particles separated and prevent settling during measurement.

Fibers and Highly Non-Spherical Particles

Needle-like crystals, glass fibers, carbon materials, and flakes are ideal candidates for image analysis because their shape cannot be adequately described by a single equivalent spherical diameter. Length, width, and aspect ratio distributions are reported directly from individual particle measurements.

Specialty Engineering Materials

Metal powders for additive manufacturing, catalyst carriers, toner, pharmaceutical powders, proppants, and battery materials are frequently characterized by image analysis to control morphology-related performance attributes such as packing density, flowability, packing behavior, and reactivity.

Testing Items

Typical reported results from an image-based particle size analysis include:

• Number-based particle size distribution, convertible to volume- or area-based distributions
• Characteristic percentile diameters such as D10, D50, and D90
• Circle-equivalent, Feret, and Martin diameter distributions
• Shape parameters: circularity, aspect ratio, elongation, convexity, solidity, and roundness
• Particle counts and number concentration within defined size classes
• Agglomerate, satellite particle, and oversized particle identification
• Foreign particle and contamination screening
• Quantitative image documentation for traceability and reporting

Testing Methods

Testing follows the general workflow of sampling, dispersion, image acquisition, image processing, and statistical evaluation, in accordance with ISO 13322-1, ISO 13322-2, ISO 9276 for result representation, and related national standards such as GB/T 21649.1.

Sample Preparation and Dispersion

A representative sample is split using a rotating divider or similar technique. Particles are then dispersed either dry, with controlled air pressure, or wet, in an ultrasonically assisted dispersion cell. The particle concentration is adjusted so that particle overlap is minimized and individual silhouettes remain distinguishable, since overlapping particles bias results toward larger sizes.

Image Acquisition

In static mode, an automated microscope stage scans the specimen while a digital camera captures thousands of fields of view. In dynamic mode, particles suspended in a liquid or air stream pass through a flow cell positioned between a strobed light source and a high-speed camera, freezing each particle motion in the image. Magnification, illumination, exposure, and focus are optimized for the expected size range, and the optical system is calibrated using certified reference materials such as stage micrometers or standard latex spheres.

Image Processing and Particle Recognition

Captured images are converted to binary form through gray-level thresholding. Segmentation algorithms separate touching particles, edge detection defines particle boundaries, and watershed routines resolve aggregates where appropriate. Particles that intersect the image frame edge are either excluded or corrected according to the applicable standard to avoid measurement bias.

Statistical Evaluation

Individual particle measurements are accumulated into histograms and cumulative distributions. Sufficient particle numbers, typically several thousand to tens of thousands depending on the width of the distribution and the required statistical confidence, are measured so that the reported D-values are stable and repeatable. Accredited third-party laboratories report results with defined resolution classes and measurement uncertainties consistent with ISO 9276.

Quality Control

Method validation includes calibration verification, repeatability testing on replicate specimens, and, where available, inter-laboratory comparison or certified reference materials. Blanks are run for liquid samples to confirm that background contamination does not influence the results.

Applications

Image-based particle size analysis supports quality control, research and development, and failure analysis across many industries:

• Additive manufacturing: morphology, sphericity, and satellite detection in metal and polymer powders
• Pharmaceuticals: particle shape and size of active ingredients and excipients for flow, compaction, and dissolution behavior
• Battery manufacturing: graphite, cathode, and anode material grading and agglomerate control
• Abrasives and cutting tools: grit sizing and angularity classification of abrasive grains
• Ceramics and powder metallurgy: raw powder qualification before pressing and sintering
• Paints, coatings, and inks: pigment and filler particle shape and oversize detection
• Food science: milk powder, sugar, spice, and ingredient particle characterization
• Geology and environmental science: sediment, soil, and airborne dust classification
• Electronics and toner: toner shape factor control and contamination analysis

FAQ

What is the difference between image analysis and laser diffraction?

Laser diffraction is an ensemble method that rapidly measures large sample quantities but reports results based on an assumed equivalence model, historically volume-based spherical equivalence. Image analysis measures each particle individually, provides true number-based distributions, and additionally quantifies particle shape. The two methods are complementary: laser diffraction offers throughput, while image analysis offers morphological detail and direct visual verification.

What size range can be measured?

Dynamic image analysis typically covers approximately 1 micrometer to several millimeters. Static optical image analysis covers roughly 1 micrometer upward, and static electron-microscope-based image analysis extends down to the nanometer range, with practical lower limits near tens of nanometers.

How many particles must be counted?

Because the technique is number-based, statistical reliability depends on particle count. For narrow distributions, several thousand particles may suffice, while broad or multimodal distributions require tens of thousands of particles to stabilize the D10 and D90 values. The required count is determined by the width of the distribution and the target confidence interval.

Which standards apply?

ISO 13322-1 governs static image analysis, ISO 13322-2 governs dynamic image analysis, and ISO 9276 defines the mathematical representation of results. National standards such as GB/T 21649.1 and related ASTM and pharmacopoeial methods provide additional sector-specific guidance.

Can image analysis detect agglomerates?

Yes. Because every particle image is available for inspection, agglomerates, satellites on spherical powders, broken particles, and contaminants can be identified, counted, and classified by shape criteria, which is a key advantage over indirect sizing techniques.

How should samples be submitted?

Samples should be representative, obtained by proper sample splitting, and supplied in a quantity of typically a few grams for powders or about 50 to 100 milliliters for liquid suspensions, together with information on the expected size range, material safety data, and whether wet or dry dispersion is preferred.