Particle size testing by image analysis is a direct optical measurement technique that determines particle size distribution and particle shape by capturing, digitizing, and evaluating images of individual particles. Governed by internationally recognized standards such as ISO 13322-1, ISO 13322-2, and GB/T 21649, this method delivers both quantitative size data and morphological information that indirect techniques cannot provide. This article explains the technical principles, measurable parameters, sample types, testing procedures, and typical industrial applications of image analysis based particle size testing.
Technical Overview
Particle size testing by image analysis is a static or dynamic optical characterization technique in which individual particles are captured as digital images and subsequently measured using computer-based image processing algorithms. Unlike ensemble techniques such as laser diffraction, which infer particle size from scattered light patterns, image analysis measures each particle directly, producing a number-based size distribution together with detailed shape descriptors.
The technique is standardized in ISO 13322-1 (particle size analysis by image analysis methods, Part 1: static image analysis methods) and ISO 13322-2 (Part 2: dynamic image analysis methods). Corresponding national standards include GB/T 21649.1 and GB/T 21649.2. Static image analysis typically uses an optical or electron microscope to examine particles deposited on a substrate, covering a measuring range of approximately 1 micrometer to several millimeters depending on the optics used. Dynamic image analysis records particles suspended in a flowing stream using a high-speed camera and stroboscopic illumination, generally covering approximately 1 micrometer to 30 millimeters.
Because every particle is individually imaged and counted, the method provides exceptional sensitivity to oversized particles, agglomerates, and shape anomalies. Results are expressed as number-based distributions by default and can be mathematically converted to volume- or mass-based distributions for comparison with other techniques. Accredited third-party laboratories routinely apply image analysis as a complementary or referee method for verifying results obtained by laser diffraction, dynamic light scattering, or sieve analysis.
Testing Samples
Metal Powders
Spherical metal powders for additive manufacturing (3D printing), powder metallurgy, and thermal spraying are frequently characterized by image analysis because shape factors such as sphericity and circularity directly influence flowability, packing density, and laser absorption during building processes.
Pharmaceutical Powders
Active pharmaceutical ingredients (APIs), excipients, and micronized powders are tested to verify particle size and shape consistency, which affect dissolution rate, bioavailability, content uniformity, and tableting behavior.
Battery Materials
Cathode and anode materials, including lithium compounds, graphite, and silicon-based particles, are measured to confirm particle size distribution and morphology, which influence electrode coating quality, energy density, and cycle life.
Ceramic and Abrasive Powders
Ceramic precursors, silicon carbide, alumina, and diamond abrasives are analyzed because grit morphology and size distribution determine grinding efficiency, surface finish quality, and tool wear behavior.
Suspensions, Emulsions, and Slurries
Liquid-dispersed systems such as emulsions, pigment suspensions, and drilling fluids can be tested using dynamic image analysis, provided the particles are properly diluted and dispersed to avoid coincident imaging.
Fibers, Granules, and Irregular Particles
Elongated fibers, needle-like crystals, granules, pellets, and highly irregular particles are ideally suited to image analysis, since single equivalent diameters reported by other techniques cannot adequately describe their morphology.
Testing Items
Image analysis particle size testing reports both size distribution data and morphological parameters, including but not limited to the following:
- Particle size distribution: D10, D50, D90, D97, mean diameter, and full distribution histograms (number-based and volume-converted)
- Circle-equivalent diameter (CED): diameter of a circle having the same projected area as the particle
- Feret diameter: maximum Feret, minimum Feret, and mean Feret diameters measured between parallel tangents
- Martin diameter: chord length dividing the projected area into two equal halves
- Aspect ratio: ratio of maximum to minimum Feret diameter, indicating particle elongation
- Circularity (form factor): 4πA/P², describing how closely a particle resembles a perfect circle
- Sphericity and roundness: shape indices relevant to powder flowability and packing behavior
- Convexity and solidity: parameters revealing surface roughness, concavities, and agglomeration
- Particle count, number concentration, and oversized (foreign) particle detection
- Agglomerate identification and friability assessment
Testing Methods
Sample Preparation
Representative sampling is performed using techniques such as cone-and-quarter splitting or rotary riffling to ensure statistical validity. For static analysis, a small quantity of powder is dispersed on a microscope slide, either dry or in a suitable mounting medium, and spread to obtain a monolayer of separated particles. For dynamic analysis, the sample is dispersed in a carrier liquid with appropriate dispersing agents, or transported in a dry air stream, and diluted until particles pass the measurement zone individually. Correct dispersion is critical: excessive energy may fracture friable particles, while insufficient energy leaves agglomerates intact and biases results toward larger sizes.
Static Image Analysis Procedure
The dispersed sample is imaged using a calibrated optical microscope (or electron microscope for sub-micrometer particles) at a magnification selected so that the smallest particles span a sufficient number of pixels for accurate measurement. Automated motorized stages capture numerous fields of view to build a statistically representative particle population, typically several thousand to tens of thousands of particles. Digital images are processed through grayscale adjustment, thresholding or edge detection, particle boundary recognition, and exclusion of overlapping or out-of-focus particles. Each recognized particle projection is then measured to yield equivalent diameters and shape descriptors, and results are compiled into distribution statistics.
Dynamic Image Analysis Procedure
In dynamic image analysis, the dispersed suspension or dry particle stream flows through a measurement cell positioned between a pulsed light source and a high-resolution digital camera. Short exposure flashes freeze particle motion, producing sharp silhouettes of moving particles. The instrument software segments each frame, measures individual particles in real time, and continuously accumulates data until the target particle count and statistical stability are achieved. Modern systems employ two cameras with different magnifications to widen the measurable size range in a single run.
Calibration and Quality Control
Systems are calibrated using certified reference materials such as NIST-traceable latex spheres or stage micrometers. Focus verification, pixel calibration, threshold validation, and repeat measurements are conducted in accordance with ISO 13322 requirements. Measurement uncertainty is minimized by controlling particle count, verifying that the smallest particles exceed the recommended minimum pixel diameter, and confirming the absence of coincident particles or image saturation.
Data Interpretation and Reporting
Test reports typically include number-based and transformed volume-based distributions, percentile diameters (D10, D50, D90), distribution width indicators, representative particle images, and population statistics for shape parameters. Because image analysis is a number-based technique, a small number of large particles can significantly shift volume-based results, and this difference must be considered when comparing data with laser diffraction or sieving results.
Applications
Image analysis particle size testing serves a wide range of industries and technical purposes:
- Additive manufacturing: verifying powder sphericity, satellite content, and size distribution for laser and electron beam melting processes
- Pharmaceuticals: supporting formulation development, generic drug equivalence studies, and regulatory submissions requiring particle morphology data
- Battery and new energy materials: controlling cathode and anode particle morphology for consistent electrode performance
- Abrasives and grinding tools: confirming grit size grading and angularity to meet product specifications
- Toners and printing materials: ensuring particle roundness and size uniformity for transfer efficiency and image quality
- Food science: characterizing powders, emulsions, and crystalline ingredients affecting texture and stability
- Ceramics and advanced materials: optimizing sintering behavior through raw powder morphology control
- Geology and environmental science: quantifying sand, sediment, and airborne particulate shape and size distributions
- Quality control and failure analysis: detecting oversized particles, contaminants, and agglomerates in incoming materials and finished products
FAQ
What is the difference between image analysis and laser diffraction?
Laser diffraction is an ensemble method that infers size distribution from an angular light scattering pattern and reports a volume-based distribution, whereas image analysis measures each particle individually and reports a number-based distribution plus shape parameters. Image analysis is the preferred method when morphology, oversize tails, or agglomeration must be characterized, while laser diffraction offers faster measurement for routine volume-based distributions. The two methods often yield slightly different results due to different equivalent diameter definitions, and cross-method comparison should be performed with this in mind.
What particle size range can image analysis measure?
Static image analysis with optical microscopy typically covers approximately 1 micrometer to several millimeters, extendable downward to nanometer scales when scanning electron microscopy is used. Dynamic image analysis systems generally cover approximately 1 micrometer to 30 millimeters, with dual-camera configurations broadening the range in a single measurement. The practical lower limit is governed by optical resolution and the requirement that each particle span a minimum number of pixels.
How many particles must be measured for a valid result?
Statistical reliability improves with particle count. For broad distributions, several thousand particles are commonly measured, and for tightly classified powders tens of thousands may be required. ISO 13322-1 emphasizes that the measured particle count must be sufficient to ensure the confidence interval of the reported percentiles meets the intended precision, particularly for coarse tails of the distribution.
When should static or dynamic image analysis be chosen?
Static image analysis is well suited to fine powders, precise morphology studies, and small sample quantities, offering high magnification and excellent shape detail but limited throughput. Dynamic image analysis is preferred for larger sample volumes, suspensions, faster routine quality control, and coarser particles, as it continuously analyzes a flowing stream and measures far larger particle populations in a shorter time.
Can image analysis results be compared with sieve or laser diffraction data?
Yes, with appropriate caution. Conversion between number-based and volume-based distributions is possible mathematically, and equivalent diameters can be aligned with sieve aperture sizes for granular materials. However, because each technique responds to different particle properties, small systematic differences are expected, and accredited third-party laboratories typically establish correlation data for specific materials before making cross-method acceptance decisions.
How are agglomerates handled during testing?
Dispersion conditions are optimized during method development to separate loosely bound agglomerates without fracturing primary particles. Image analysis uniquely allows agglomerates to be visually identified and quantified through shape descriptors such as convexity, and representative images can be included in the report to document the dispersion state of the sample.