A duplex stainless steel grain size statistical experiment quantitatively characterizes the grain size of both the ferrite and austenite phases, delivering the reliable microstructural data required for material qualification, welding procedure validation, and heat treatment verification. This guide explains sample preparation, metallographic etching, measurement techniques, and statistical analysis in accordance with ASTM E112, ASTM E1382, ISO 643, and GB/T 6394.
Technical Overview
Duplex stainless steels (DSS) are dual-phase alloys engineered to contain approximately 50% ferrite (α) and 50% austenite (γ) by volume. This balanced microstructure combines the high strength and chloride stress-corrosion cracking resistance of ferrite with the toughness and formability of austenite, which is why grades such as UNS S31803, UNS S32205, UNS S32750, and UNS S32760 are widely used in aggressive service environments.
Because the properties of duplex stainless steel depend strongly on microstructural scale, grain size must be treated as a statistical variable rather than a single value. A grain size statistical experiment measures a population of ferrite and austenite grains, determines the mean grain size for each phase, and characterizes the distribution, dispersion, and uniformity of that population. The results are used to verify solution-annealing quality, assess weld heat-affected zone (HAZ) behavior, and correlate microstructure with strength, impact toughness, and corrosion performance.
- ASTM E112 – Standard Test Methods for Determining Average Grain Size (comparison, planimetric, and intercept procedures)
- ASTM E1382 – Determining Average Grain Size Using Semiautomatic and Automatic Image Analysis
- ASTM E562 – Determining Volume Fraction by Systematic Manual Point Count
- ASTM E1245 – Determining the Inclusion or Second-Phase Content of Metals by Automatic Image Analysis
- ISO 643 – Steels: Micrographic Determination of the Apparent Grain Size
- GB/T 6394 – Determination of Grain Size for Metals (Chinese national standard)
- GB/T 1954 – Measurement of Ferrite Content in Stainless Steel (related phase-balance verification)
Testing Samples
Plates, Sheets, and Strips
Flat products are sampled in both longitudinal (L) and transverse (T) orientations because hot rolling and solution annealing can produce directionally elongated grain structures. Test blocks are typically taken at quarter-width positions, with the observed plane chosen according to the governing product specification.
Pipes and Tubes
Seamless and welded tubulars require sections cut from both the body and, where relevant, the weld region. For longitudinally welded pipe, specimens include weld metal, fusion line, HAZ, and base metal so that grain coarsening or undesirable phase transformation near the weld can be statistically compared with the parent material.
Forgings, Castings, and Bars
Heavy-section components exhibit grain size gradients between surface and center. Sampling therefore includes locations at the surface, mid-radius, and core, and large forgings may require multiple orientations to reveal anisotropy introduced by deformation and heat treatment.
Welded Joints and Heat-Affected Zones
Weld specimens are sectioned perpendicular to the welding direction. Narrow HAZ bands are mapped with sequential fields so that the grain size distribution across the thermal cycle can be reconstructed, supporting welding procedure qualification and repair assessment.
Specimen Preparation Requirements
- Sectioning with abundant coolant to avoid deformation or thermal alteration of the microstructure
- Mounting in resin for edge retention and safe handling
- Grinding through successive abrasive grades followed by fine polishing to a scratch-free, artifact-free surface
- Etching with a reagent that clearly discriminates ferrite from austenite
Testing Items
The core deliverables of a duplex stainless steel grain size statistical experiment are listed below, and each is reported separately for the ferrite and austenite phases whenever the two-phase structure can be resolved.
- Mean grain size of the ferrite phase, expressed as mean intercept length and ASTM grain size number G
- Mean grain size of the austenite phase, expressed as mean intercept length and ASTM grain size number G
- Grain size distribution (frequency histograms, typically log-normal in character)
- Standard deviation, coefficient of variation, and 95% confidence interval of the mean grain size
- Minimum, maximum, and selected percentile grain dimensions for outlier screening
- Grain aspect ratio and directional measurements for elongated or anisotropic structures
- Volume fraction (phase balance) of ferrite and austenite reported alongside grain size
- Grain count, number of fields measured, and uniformity assessment across the evaluated area
- Micrographs of representative fields and calibrated measurement overlays as objective evidence
Testing Methods
Etching for Phase Discrimination
Successful grain size statistics in duplex stainless steel depend on selective etching. Common reagents and practice include:
- Kalling's No. 2 reagent (copper chloride in hydrochloric acid and ethanol): attacks ferrite, leaving austenite bright, and is widely used for general phase distinction
- ELECTROLYTIC etching in 20%–40% sodium hydroxide or 10% potassium hydroxide solution at low voltage: tints the ferrite while austenite remains bright, providing high contrast for automatic image analysis
- Beraha-type tint etchants: deposit interference films that color the two phases differently, useful for both manual and digital measurement
- Glyceregia: a general-purpose etchant for revealing overall structure prior to phase-specific etching
Comparison Method
The specimen microstructure is compared against standard grain size charts at a defined magnification. The comparison method is fast and inexpensive but provides only an approximate G number and limited statistical information, so it is generally used as a screening step in duplex work where separate phase ratings are needed.
Planimetric Method
The number of grains within a known area is counted, with grains intersecting the boundary counted according to the fractional rules of ASTM E112 and ISO 643. The grain count per unit area is converted into a grain size number and mean grain area. When applied phase-selectively, this method yields independent planimetric statistics for ferrite and austenite.
Intercept (Lineal) Method
A grid of test lines is superimposed on the microstructure, and the number of grain boundaries or mean intercept lengths is measured along multiple directions. The intercept method is the preferred statistical basis for duplex stainless steel because it handles two-phase structures well, supports directional measurements on elongated grains, and connects directly to the ASTM G number, which is defined through the relation N = 2^(G−1), where N is the number of grains per unit area at the standardized 100× reference condition.
Automatic Image Analysis
Digitized micrographs are segmented using gray-level or color thresholds, after which software identifies individual grains, measures equivalent diameters and areas, and compiles full distributions per phase in conformance with ASTM E1382. Phase volume fractions are simultaneously quantified by point counting or image analysis per ASTM E562 and ASTM E1245. Calibration with a certified stage micrometer, unbiased field selection along a systematic traverse, and exclusion of etching artifacts are essential for defensible results.
Statistical Analysis
Raw measurements are processed to produce statistically meaningful conclusions:
- Sufficient population sampling: typically several hundred grain intercepts or at least several hundred grains per phase so the reported G number is stable to within approximately 0.25 units
- Reporting of mean, standard deviation, and coefficient of variation for each phase
- Confidence intervals on the mean grain size based on the number of measurements and observed scatter
- Log-normal distribution fitting, which usually describes grain populations more accurately than a normal distribution
- Comparison tests (t-test or analysis of variance) when evaluating process changes, batch differences, or HAZ versus base metal
- Separate reporting of longitudinal and transverse results for directionally worked material
- Documentation of magnification, etchant, fields measured, operator, and instrument calibration for traceability
Applications
Grain size statistics for duplex stainless steel support decisions across the full product lifecycle:
- Oil and gas: qualification of seamless and welded line pipe, subsea components, and umbilical tubing exposed to sour and chloride environments
- Chemical and petrochemical processing: microstructural verification of reactors, heat exchangers, and columns handling chloride-bearing media
- Desalination and water treatment: acceptance testing of evaporator shells, tubing, and high-brine piping
- Marine and offshore structures: confirmation that forming and welding have not coarsened grains or unbalanced the phase structure
- Pressure vessels and piping: heat treatment validation for solution-annealed components to ensure uniform fine grains and correct ferrite/austenite balance
- Welding procedure qualification: statistical comparison of HAZ and base metal grain size in support of procedure qualification records
- Failure analysis and research: linking intergranular corrosion, pitting resistance, and impact toughness to measured microstructural scale, including studies of super duplex grades and additively manufactured material
FAQ
Why must ferrite and austenite be measured separately?
The two phases form, grow, and respond to thermomechanical processing differently. Ferrite grains typically coarsen more readily at high temperature, while austenite morphology depends on nitrogen content and cooling history. A single blended average would mask phase-specific information that controls strength partitioning, toughness, and corrosion behavior.
Which standard should be selected for the experiment?
ASTM E112 or ISO 643 covers manual comparison, planimetric, and intercept methods; ASTM E1382 governs semiautomatic and automatic image analysis; GB/T 6394 applies when Chinese national standards are specified. The choice depends on the purchasing specification, the required accuracy, and whether distribution-level statistics are needed.
How many grains must be counted for a reliable result?
As a practical rule, at least several hundred grains or intercepts per phase should be measured across a sufficient number of unbiased fields so that the reported mean grain size and G number remain stable within the target uncertainty. Higher scatter or mixed structures requires larger populations.
Is grain size related to the phase balance?
They are related but distinct. Phase balance (roughly 50/50 ferrite/austenite) reflects chemical composition and heat treatment, while grain size reflects the scale of each phase. Both must be verified, and grain size statistics are commonly reported together with ferrite number or volume fraction measurements.
What grain size is typically specified for duplex stainless steel?
Correctly solution-annealed duplex plate and pipe usually exhibit a fine, uniform structure, and product specifications or engineering design codes commonly require an average grain size no coarser than approximately ASTM 5 or 6, with additional limits on local coarsening near welds. Acceptance limits should always be taken from the applicable purchase specification.
Can coarse grains be corrected?
Yes, in most cases. Re-solution annealing at the recommended temperature followed by rapid water quenching can refine and re-equilibrate the two-phase structure, after which the grain size statistical experiment is repeated to confirm compliance before the material enters service.