Characterization Method Finder

Look at a polished and etched section to see grain size, phase distribution, inclusions and layer thickness. The first step in examining a microstructure.

Principle Metals are viewed in reflected light. An etchant attacks grain boundaries and phases at different rates, which turns them into light and dark contrast.

Used for Microstructure on a section Grain size, orientation and texture Thickness of films and layers Precipitates Fracture surface Internal defects Surface shape and roughness Phase identification and crystal structure

Typical figures

Lateral resolution

About 0.2 µm (limit set by the wavelength of light)

Depth resolution / information depth

Image of the polished plane (no depth information)

Detection limit

— (no elemental analysis)

Analyzed depth / volume

Surface of the polished section

Elements

None (phases are inferred from color, shape and etching response)

Quantitative accuracy

Area fractions and grain size can be quantified by image analysis

Sample and effort

Size

Powders and small pieces can be mounted in resin; large parts must be sectioned

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Cut or polish The sectioned and polished part is destroyed

Preparation

Moderate Cut → mount in resin → polish → etch. Half a day to a day, depending on material and etchant

Time

Short Minutes to tens of minutes once the sample is ready

Cost

Low Usually cheap on a shared university instrument

Pitfalls

• Polishing scratches and smeared layers are highlighted by etching and can be mistaken for microstructure. Etching does not remove earlier damage
• Resolution is about 0.2 µm, so fine precipitates and the inside of martensite laths are not resolved (go to SEM or TEM)
• A section is a 2-D cut, so grain shape and size look different depending on the cutting direction. For rolled material, examine several sections in defined directions

Often combined with SEM · EBSD · HV/HR/HBW

Standards ASTM E3 (guide to preparing metallographic specimens) / ASTM E407 (microetching metals and alloys) / JIS G 0551 / ASTM E112 (determining grain size) / ISO 3887 (determining the depth of decarburization of steels)

Microstructure Atlas Ferrite + pearlite (hypoeutectoid steel) Lath martensite (quenched low-carbon steel) Grain size (grain size number and mean grain size)Calculators Iron–carbon diagramFormulas Grain size number (ASTM E112) Hall–Petch relation

Terms Pearlite Martensite Hall–Petch relationship

Sources: MyScope SEM; Metkon (ASTM E407); ISO 3887

Image surface topography (SE) and differences in mean atomic number (BSE) with far better resolution and depth of field than an optical microscope. The workhorse for fracture surfaces and fine microstructure.

Principle A finely focused electron beam is scanned over the surface, and the number of secondary electrons (SE) and backscattered electrons (BSE) coming out sets the brightness of each pixel.

Used for Surface shape and roughness Fracture surface Microstructure on a section Thickness of films and layers Precipitates Internal defects Grain size, orientation and texture

Typical figures

Lateral resolution

About 1–5 nm (around 1 nm with a field-emission gun, about 5 nm on a typical instrument)

Depth resolution / information depth

SE from the top ~15 nm; BSE from deeper (depends on accelerating voltage)

Detection limit

— (use EDS for composition)

Analyzed depth / volume

SE: the very surface; BSE: the upper part of the electron interaction volume

Elements

Not from the image (BSE brightness is only a guide to mean atomic number)

Quantitative accuracy

Calibrate the magnification with a standard before measuring sizes

Sample and effort

Size

Whatever fits the chamber (instrument dependent); large parts must be cut

Conductivity

Insulators need a ~10 nm conductive coating or low-vacuum imaging

Vacuum

Normally high vacuum; wet samples must be dried or imaged in low vacuum

Destructive?

Non-destructive for surfaces; sections must be cut and polished

Preparation

Minimal Little for surfaces; polishing for sections; coating for insulators

Time

Short Tens of minutes to a few hours per sample

Cost

Low Usually cheap on a shared university instrument

Pitfalls

• Reading BSE contrast as composition. Topography and crystal orientation (channeling) also change the brightness, so check on a flat polished surface together with EDS
• Mistaking charging artifacts (bright, smeared areas) on insulators, or the grain of a coating, for microstructure
• Trusting the on-screen scale as is. Calibrate the magnification with a standard before measuring sizes

Often combined with EDS (EDX) · EBSD · FIB

Microstructure Atlas Precipitates (on grain boundaries and inside grains)Failure Analysis Fatigue Hydrogen embrittlement Ductile overload fracture Brittle fracture

Terms SEM (scanning electron microscope) Fatigue striations

Sources: MyScope SEM; EAG technique chart

Map the orientation of every grain to obtain grain size, texture, grain-boundary misorientation, phase distribution (by crystal structure) and a guide to local strain.

Principle In the SEM the sample is tilted to 70°, the diffraction pattern of backscattered electrons (Kikuchi pattern) is recorded, and the crystal orientation is solved point by point.

Used for Grain size, orientation and texture Phase identification and crystal structure Microstructure on a section

Typical figures

Lateral resolution

20–80 nm (material dependent). Grains smaller than 50 nm are difficult (TKD on a thin foil reaches 2–20 nm)

Depth resolution / information depth

Top ~10 nm of the surface

Detection limit

— (no elemental analysis)

Analyzed depth / volume

Top ~10 nm layer

Elements

None (combine with simultaneous EDS to separate phases)

Quantitative accuracy

Orientation precision about 0.2° (about 0.001° with cross-correlation analysis for strain)

Sample and effort

Size

Whatever fits the SEM chamber

Conductivity

A conductive coating weakens the patterns, so low-vacuum imaging is preferred for insulators

Vacuum

Normally vacuum (low-vacuum instruments exist)

Destructive?

Cut or polish Needs a flat polished surface, so that part is cut and polished

Preparation

Extensive Mechanical polishing followed by a strain-free finish (chemical or electrolytic polishing). Laborious

Time

Moderate Tens of minutes to hours per map (fast detectors exceed 3000 points per second)

Cost

Moderate Medium (fewer SEMs carry EBSD)

Pitfalls

• Polishing damage left at the surface kills the patterns and leaves many points unindexed. Mechanical polishing alone is often not enough
• Grain size depends on the misorientation threshold used to define a boundary and on the step size. State them when comparing
• The mapped area is small (hundreds of µm to mm). Check the texture of the whole sheet with XRD pole figures

Often combined with SEM · XRD (sin²ψ, pole figures) · OM

Microstructure Atlas Grain size (grain size number and mean grain size)Formulas Grain size number (ASTM E112) Hall–Petch relation

Terms Recrystallization Hall–Petch relationship

Sources: DTU Nanolab EBSD; SERC (Carleton College); EAG technique chart; ASM EDFAS (EBSD detector)

Image dislocations, precipitates and interfaces down to atomic columns using electrons transmitted through a specimen thinner than 100 nm. With electron diffraction and EDS/EELS, structure and composition come from the same spot.

Principle Bright- and dark-field contrast comes from differences in diffracted intensity, and HRTEM forms lattice images by interference of transmitted and diffracted waves. STEM-HAADF collects electrons scattered to high angles, so heavier atoms look brighter (Z contrast).

Used for Precipitates Microstructure on a section Phase identification and crystal structure Thickness of films and layers Local composition and element maps Internal defects Grain size, orientation and texture Chemical state (oxidation state, bonding)

Typical figures

Lateral resolution

0.1–0.3 nm (TEM point resolution 0.14 nm, 0.1 nm in an aberration-corrected STEM; about 0.2–0.25 nm without correction)

Depth resolution / information depth

Projection through the whole specimen thickness (<100 nm)

Detection limit

EDS about 0.5 wt% (0.01–0.1 wt% under good conditions)

Analyzed depth / volume

A region of µm size or less in a foil thinner than 100 nm

Elements

EDS from B upwards (poor sensitivity to light elements); EELS down to Li

Quantitative accuracy

EDS is semi-quantitative to a few percent

Sample and effort

Size

Disks about 3 mm across, particles on a grid, or FIB lamellae

Conductivity

Not needed (thinning is what matters)

Vacuum

Required (dry, beam-tolerant specimens)

Destructive?

Cut or polish Destroys the part that is thinned

Preparation

Extensive Thin by electropolishing, ion milling, FIB or ultramicrotomy. Needs skill and time

Time

Long Days to weeks including preparation

Cost

High (often outsourced)

Pitfalls

• The volume examined is tiny (<100 nm thick, µm across). One field may not represent the sample, so survey with SEM first and then choose the site
• Preparation damage mistaken for microstructure. FIB lamellae carry a Ga-damaged layer on both faces; milling at 30 kV alone leaves about 20 nm per side in Si
• Bright- and dark-field contrast mainly reflects diffraction. Check whether it is compositional with HAADF (brightness set by atomic number) or EDS

Often combined with FIB · SAED/CBED · APT

Microstructure Atlas Precipitates (on grain boundaries and inside grains)Formulas Magnitude of the Burgers vector Structure factor and diffracted intensity

Terms Dislocation Burgers vector Precipitation hardening

Sources: Univ. Duisburg-Essen ICAN; Microscopy Australia TechFi; EAG technique chart; MyScope EDS

Mill a chosen site with µm precision to expose a cross-section or to make TEM lamellae and APT needles. Combined with an SEM, repeated slice-and-image gives 3-D microstructure.

Principle Ga ions focused below 10 nm sputter atoms from the surface. The milled face is imaged with the electron beam (SEM).

Used for Thickness of films and layers Microstructure on a section Internal defects

Typical figures

Lateral resolution

Beam below 10 nm; imaging at SEM resolution (about 1 nm)

Depth resolution / information depth

Serial sectioning gives 3-D data on regions 100 nm–10 µm in size

Detection limit

— (combine with EDS etc.)

Analyzed depth / volume

A local region of µm to tens of µm

Elements

—

Quantitative accuracy

—

Sample and effort

Size

Whatever fits the SEM chamber

Conductivity

Insulators charge and should be coated

Vacuum

Required

Destructive?

Cut or polish Destroys the milled spot (locally)

Preparation

Moderate Decide the site first; deposit a protective layer before milling

Time

Moderate Hours per cross-section; half a day to a day per TEM lamella

Cost

Moderate Medium to high (instrument time is expensive)

Pitfalls

• The milled face carries a Ga-damaged layer (about 20 nm per side in Si at 30 kV). Finish TEM lamellae at low voltage (3–7 nm at 2 kV)
• The accessible area is small, so locate the feature with SEM or optical microscopy before milling

Often combined with SEM · TEM/STEM · APT

Sources: Microscopy Australia TechFi; Univ. Twente NanoLab; EAG technique chart

Trace the surface with a sharp tip and build a height map at nm scale. Measures roughness, steps and particle heights without vacuum, on insulators too.

Principle A laser reads the deflection of the cantilever carrying the tip, and the height is followed so that the tip–surface force stays constant.

Used for Surface shape and roughness Thickness of films and layers

Typical figures

Lateral resolution

2–150 nm (set by the tip radius)

Depth resolution / information depth

Height resolution 0.01 nm

Detection limit

—

Analyzed depth / volume

Surface only (scan up to about 90 µm square and 5 µm in height)

Elements

None

Quantitative accuracy

Height values depend on tip shape and scan conditions

Sample and effort

Size

Small pieces to wafers (some stages take 300 mm)

Conductivity

Not needed (insulators are fine)

Vacuum

Not needed

Destructive?

Non-destructive

Preparation

Minimal Almost none (keep the surface clean)

Time

Short Minutes to tens of minutes per site

Cost

Low Usually cheap on a shared university instrument

Pitfalls

• The tip shape is convolved into the image: a blunt tip makes narrow trenches look shallower and narrower and bumps look wider
• The field of view is small (up to about 90 µm). Survey with an optical microscope and pick a representative site
• Steps above a few µm and very rough surfaces are a problem (use a laser confocal microscope or white-light interferometer)

Often combined with Laser confocal microscopy and white-light interferometry · SEM · Nanoindentation

Sources: EAG AFM; EAG technique chart; Penn State AFM

Measure the 3-D shape of a surface optically and without contact, and turn roughness, steps, wear-scar depth and indent shape into numbers.

Principle A laser confocal microscope picks up only in-focus light to find the height; a white-light interferometer finds the height where the interference fringes are strongest.

Used for Surface shape and roughness Thickness of films and layers Fracture surface

Typical figures

Lateral resolution

About 0.3–0.5 µm (lateral resolution is diffraction-limited, 0.61λ/NA: about 0.26 µm at 405 nm and NA 0.95; 0.5 µm in a white-light interferometer example)

Depth resolution / information depth

White-light interferometry 0.1 nm (phase mode) to 6 nm (scanning mode)

Detection limit

—

Analyzed depth / volume

Surface only (one field is mm-sized, stitched up to several cm)

Elements

None

Quantitative accuracy

Data drop out or go wrong on steep slopes and under transparent films

Sample and effort

Size

Small pieces to parts tens of cm across (depends on the stage)

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Non-destructive

Preparation

Minimal Almost none

Time

Short Minutes per site

Cost

Low Usually cheap on a shared university instrument

Pitfalls

• On steep or rough surfaces light does not return, leaving gaps or noise in the data
• Under a transparent film, light returns from both the film surface and the interface, distorting the shape. Film thickness can be measured only for transparent films of known refractive index
• Roughness values depend on the evaluation length or area and on the filter (cut-off). Use the same conditions when comparing

Often combined with AFM · OM · HV/HR/HBW

Sources: Evident laser microscope principles; EAG optical profilometry

Trace the surface with a diamond stylus to measure step heights (film thickness) and surface roughness, over long distances in one scan (up to 150 mm on one instrument).

Principle A stylus pressed on with a small force is moved across the surface, and its vertical motion is recorded as a surface profile.

Used for Surface shape and roughness Thickness of films and layers

Typical figures

Lateral resolution

Set by the stylus tip radius (2 µm for a standard stylus)

Depth resolution / information depth

Heights down to the ångström level (instrument example)

Detection limit

—

Analyzed depth / volume

The traced line

Elements

—

Quantitative accuracy

Valleys narrower than the stylus and steep slopes are rounded off by the stylus shape

Sample and effort

Size

Small pieces to plates on the stage; portable units measure on large parts (scan lengths up to 150 mm on one instrument)

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Non-destructive The stylus touches the surface (the force can be set low)

Preparation

Minimal Almost none (film thickness is measured at a step)

Time

Short

Cost

Low Cheap

Pitfalls

• Features finer than the stylus tip (for example 2 µm) are not resolved; valleys come out shallow and peaks broad
• Only one line is measured, so check variation by measuring several lines
• Film thickness is measured as a step, so part of the film has to be removed to make one

Often combined with Laser confocal microscopy and white-light interferometry · AFM · Spectroscopic ellipsometry

Sources: Rochester profilometer

Shine X-rays at grazing angles on a thin film on a flat substrate and use the change of reflectivity with angle to obtain film thickness, density and interface width (roughness). For single and multilayer films about 1 nm to 1 µm thick.

Principle Reflectivity is measured from near the critical angle of total reflection upward, and thickness, density and interface width are found by fitting a curve calculated from a layer model.

Used for Thickness of films and layers Surface shape and roughness

Typical figures

Lateral resolution

— (average over the X-ray footprint, which must be laterally uniform)

Depth resolution / information depth

Films about 1 nm to 1 µm thick

Detection limit

—

Analyzed depth / volume

X-ray footprint × film

Elements

Elements are not measured directly (no chemical information)

Quantitative accuracy

A layer model (number and order of layers) is assumed for fitting, so a different model gives different values

Sample and effort

Size

Films on flat substrates (wafers, plates)

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Non-destructive (the film must not change during the measurement)

Preparation

Minimal Almost none (keep the surface clean)

Time

Moderate Measurement and model fitting take time

Cost

Moderate

Pitfalls

• A substrate that is not flat, or a film that is laterally non-uniform within the footprint, cannot be analyzed
• Surface contamination or reacted layers act as extra layers. Keep the surface clean before measuring
• XRR alone gives no elements or chemical states. Combine with XPS or similar for composition

Often combined with Spectroscopic ellipsometry · XRD · XPS (ESCA)

Standards ISO 16413 (thickness, density and interface width of thin films by X-ray reflectometry)

Sources: ISO 16413; Diamond Light Source XRR

Measure the change in polarization of reflected light (Ψ and Δ) at many wavelengths to obtain thin-film thickness and optical constants such as refractive index, without contact or damage. Sensitive to sub-monolayer changes, and up to tens of µm for transparent films.

Principle Polarized light is reflected at an oblique angle and the change in polarization is measured. The data do not give thickness directly; a layer model with optical-constant dispersion formulas is fitted to them.

Used for Thickness of films and layers

Typical figures

Lateral resolution

— (average over the illuminated spot)

Depth resolution / information depth

Sensitive to sub-monolayer changes; about 0.5 nm to 20 µm for transparent films (instrument example)

Detection limit

—

Analyzed depth / volume

Illuminated spot × film

Elements

Elements are not measured (optical constants are)

Quantitative accuracy

A layer model and dispersion formulas are assumed for fitting, so a different model gives different values

Sample and effort

Size

Flat, reflective samples (wafers, glass plates, etc.)

Conductivity

Not needed (dielectric, organic, semiconductor and metal films)

Vacuum

Not needed

Destructive?

Non-destructive (no contact)

Preparation

Minimal Clean and dry the surface

Time

Short Each point is quick, but model fitting can take time

Cost

Moderate

Pitfalls

• Thickness comes from fitting a model to Ψ and Δ, so wrong assumptions about the layers or optical constants give wrong values
• Surfaces that are not flat or do not reflect light cannot be measured
• Compare with XRR or cross-sections to check that the model is reasonable

Often combined with XRR · Stylus profilometer (roughness and step height) · SEM

Sources: Vanderbilt ellipsometer; J.C. Bose Univ. ellipsometer

Record X-ray radiographs while rotating the sample to see internal pores, cracks, inclusions and fiber orientation in 3-D without cutting.

Principle The X-ray attenuation at each point (set by density and atomic number) is reconstructed by computation from projections taken in many directions.

Used for Internal defects Microstructure on a section Thickness of films and layers Density and porosity

Typical figures

Lateral resolution

About 1/1000–1/2000 of the sample width (finer for smaller samples: µm for mm-sized samples; instruments reaching 0.1 µm exist)

Depth resolution / information depth

Same as lateral (3-D voxels)

Detection limit

— (small density differences are hard to see)

Analyzed depth / volume

The whole sample (mm to tens of cm, as far as X-rays penetrate)

Elements

None (attenuation contrast only)

Quantitative accuracy

Porosity and defect volume depend on how the image threshold is set

Sample and effort

Size

From a pellet of powder to large parts (resolution drops as size grows)

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Non-destructive

Preparation

Minimal Almost none

Time

Moderate Hours to a day for scanning and reconstruction

Cost

Moderate Medium (instruments are less common)

Pitfalls

• Resolution is set by sample size. A large part scanned whole will not show fine cracks; cut out a smaller piece if needed
• Artifacts such as beam hardening (bright rim, dark center) mistaken for density differences
• Closed cracks and phases of similar density are hard to distinguish

Often combined with RT · UT · SEM

Sources: SERC (Carleton College); EAG technique chart

Find out quickly which elements are present, and roughly how much, at a point, line or area hit by the electron beam in an SEM or TEM. Good for element maps.

Principle The beam knocks out inner-shell electrons, and a semiconductor detector sorts and counts the characteristic X-rays by energy.

Used for Local composition and element maps Average composition Precipitates Thickness of films and layers Composition versus depth

Typical figures

Lateral resolution

1 to a few µm in the SEM (set by the interaction volume); a few nm in a thin TEM foil

Depth resolution / information depth

1 to a few µm in the SEM

Detection limit

0.1–0.5 wt% (0.02–0.05 wt% reported under good conditions)

Analyzed depth / volume

µm-sized volume

Elements

Be–U (Be–F can be detected but are hard to quantify; H, He and Li cannot be measured)

Quantitative accuracy

A few percent relative for major elements on a flat polished surface against standards; standardless results carry larger errors and are semi-quantitative

Sample and effort

Size

Whatever fits the SEM chamber

Conductivity

Insulators need a carbon (or other) coating or low vacuum (semi-quantitative in low vacuum)

Vacuum

Normally vacuum

Destructive?

Non-destructive Surfaces can be measured as they are (quantification needs a polished surface; depth or layer thickness needs a cut section)

Preparation

Moderate Almost none for qualitative work; a flat polished surface for quantification

Time

Short Seconds to minutes per point; tens of minutes to hours per map

Cost

Low Cheap; usually fitted to the SEM

Pitfalls

• Trusting quantified C, N and O. Light-element X-rays are strongly absorbed and their peaks change shape between compounds. C cannot be measured on carbon-coated samples, and O is often calculated rather than measured
• Peak overlaps (S, Mo and Pb near 2.3 keV; Ti and Ba; Ti and V) lead to reporting elements that are not there
• Spot analysis of particles or layers smaller than 1–2 µm also picks up the surroundings and the substrate
• Rough fracture surfaces and loose powders do not give correct quantitative values (quantify on flat polished surfaces)

Often combined with WDS/EPMA · SEM · XPS (ESCA)

Standards ISO 22309 (quantitative EDS analysis of elements from Na upwards)

Terms EDS (energy-dispersive X-ray spectroscopy)

Sources: MyScope EDS; Newbury & Ritchie 2013; Microscopy Australia TechFi; Oxford Instruments (EDS vs WDS)

With about ten times the spectral resolution of EDS, quantify trace elements down to ~100 ppm and elements with overlapping peaks accurately against standards. Also used for line scans such as carbon profiles across case-hardened layers.

Principle Characteristic X-rays are Bragg-reflected by an analyzing crystal and counted wavelength by wavelength; each spectrometer measures one wavelength at a time.

Used for Local composition and element maps Average composition Trace elements and impurities Precipitates Thickness of films and layers Composition versus depth

Typical figures

Lateral resolution

1 to a few µm (set by the interaction volume)

Depth resolution / information depth

1 to a few µm

Detection limit

100–300 ppm (a few ppm under favourable conditions)

Analyzed depth / volume

µm-sized volume

Elements

Be (B on some instruments) to U. H, He and Li cannot be measured; errors grow for Be–O

Quantitative accuracy

0.5–1.5% relative for major elements (>1 wt%); standards are essential

Sample and effort

Size

Flat samples from a few hundred µm to about 50 mm (one facility's example)

Conductivity

Insulators are carbon coated

Vacuum

Required

Destructive?

Cut or polish Needs a flat polished surface

Preparation

Moderate Polish to a flat mirror finish and carbon coat

Time

Moderate Minutes per point; longer with many elements

Cost

Moderate Medium (fewer instruments)

Pitfalls

• Without standards there is no accuracy. Calibrate with standards of similar composition
• Oxidation states such as Fe²⁺ versus Fe³⁺ cannot be distinguished
• Spectrometers scan one wavelength at a time, so many elements take long. Screen the elements with EDS first

Often combined with EDS (EDX) · XRF · XPS (ESCA)

Standards ISO 3887 (determining the depth of decarburization of steels)

Calculators Diffusion and surface-treatment simulatorFormulas Error-function solutions (carburising, diffusion couple)

Terms Carburizing

Sources: SERC (Carleton College); Univ. Arizona EPMA; Oxford Instruments (EDS vs WDS); ISO 3887

Determine elements from Na to U nondestructively and quickly from the fluorescent X-rays excited by an X-ray beam. Used for alloy sorting, plating composition and thickness, and powders or liquids.

Principle Primary X-rays eject inner-shell electrons, and the element-specific fluorescent X-rays emitted as the atoms relax are sorted by energy (ED) or wavelength (WD).

Used for Average composition Trace elements and impurities Thickness of films and layers Local composition and element maps

Typical figures

Lateral resolution

Usually an average over a mm-sized area (tens of µm with micro-XRF)

Depth resolution / information depth

A few µm to several mm (strongly dependent on the X-ray line and the matrix)

Detection limit

Around ppm (instrument and element dependent; worse for light elements)

Analyzed depth / volume

Irradiated area × information depth

Elements

Usually Na–U. WDXRF with multilayer analyzers also reaches Be–O (including C, N and O), but sensitivity and accuracy drop for elements lighter than Na (H and Li cannot be measured; handheld units cannot measure elements lighter than Mg)

Quantitative accuracy

Accurate when calibrated with standards of similar composition; standardless fundamental-parameter results are convenient but can carry large errors in some matrices

Sample and effort

Size

Powders and liquids to large parts (large parts with a handheld unit on site)

Conductivity

Not needed

Vacuum

Not needed (vacuum or helium improves light-element sensitivity)

Destructive?

Non-destructive

Preparation

Minimal Measured as is; fused beads or pressed pellets for better accuracy

Time

Short Minutes

Cost

Low Cheap

Pitfalls

• Reading the composition of plating, oxide, paint or dirt as that of the base metal. The information depth ranges from a few µm to several mm depending on the element
• Matrix effects change the intensities, so values are off unless calibrated with standards of similar composition
• Light elements (lighter than Na) lose sensitivity and accuracy even with WDXRF and are not measured by ED or handheld units. Use combustion analysis for C in steel

Often combined with ICP-OES / ICP-MS · EDS (EDX) · GD-OES

Standards ISO 3497 (coating thickness by X-ray spectrometry)

Sources: SERC (Carleton College); HORIBA XRF; Evident handheld XRF; EAG technique chart; Rigaku multilayer analyzers

Dissolve the sample in acid and feed the solution into a hot plasma to quantify about 70 elements: down to ppm in the solid with ICP-OES and ppb with ICP-MS. The standard route to average composition and trace impurities.

Principle A hot argon plasma excites and ionizes the atoms; elements are sorted by emission wavelength (OES) or ion mass (MS).

Used for Average composition Trace elements and impurities

Typical figures

Lateral resolution

— (average of all the dissolved material)

Depth resolution / information depth

—

Detection limit

About 1 ppm in the solid for ICP-OES and 10 ppb for ICP-MS (solution limits multiplied by the dissolution and dilution factor; interferences make them an order of magnitude worse or more)

Analyzed depth / volume

All the dissolved material (one facility asks for at least 1 g of solid)

Elements

About 70 elements. ICP-OES cannot measure C, N, O, halogens, noble gases or H (one facility's list)

Quantitative accuracy

Uncertainty of 5% or less (20% or less near the limit) in one facility's example

Sample and effort

Size

Powder, chips or small pieces (cut from large parts)

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Destroys the sample Consumed by dissolution

Preparation

Extensive Dissolve completely in acid (HF or fusion for refractory materials). Laborious

Time

Moderate Hours to a day to dissolve; weeks to a month if outsourced

Cost

Moderate Medium (often outsourced)

Pitfalls

• Losing elements through undissolved residue or volatilization. Complete dissolution is assumed
• Spectral overlaps (OES) and polyatomic ions (MS) produce false or inflated results
• Taking the solution detection limit as the limit in the solid (multiply by the dilution factor)
• No information on spatial distribution (segregation, surface layers)

Often combined with XRF · Combustion and inert gas fusion analysis (C, S, O, N, H) · GD-OES

Sources: PSI ICP-OES; EAG ICP; EAG technique chart

Quantify C and S (combustion) and O, N and H (inert gas fusion) in metals quickly, down to ppm levels. Used for carbon in steel, oxygen and hydrogen in titanium, hydrogen in weld metal and so on.

Principle Combustion burns the sample in oxygen and measures the CO₂ and SO₂ by infrared absorption. Inert gas fusion melts the sample in a graphite crucible and measures the CO (from O), N₂ and H₂ by infrared absorption or thermal conductivity.

Used for Average composition Trace elements and impurities

Typical figures

Lateral resolution

— (average of the whole sample)

Depth resolution / information depth

—

Detection limit

From about 0.1 ppm (element and instrument dependent)

Analyzed depth / volume

The whole sample burned or fused

Elements

C, S, O, N and H only

Quantitative accuracy

Calibrated with reference materials

Sample and effort

Size

Chips or small pieces (cut from large parts)

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Destroys the sample Consumed by burning or fusion

Preparation

Moderate Cut to a set size and remove surface contamination

Time

Short Minutes per sample

Cost

Moderate Medium (fewer instruments, but analysis is fast)

Pitfalls

• Diffusible hydrogen escapes quickly even at room temperature. In weld metal the value falls the longer the delay (about 20% between 20 s and 300 s before quenching in liquid nitrogen in one study), so keep samples cold
• Total hydrogen (fusion) and diffusible hydrogen released on heating are different quantities. State which was measured
• No spatial information (segregation, decarburized or carburized layers). Use EPMA, GD-OES or hardness profiles on a section

Often combined with Spark OES · ICP-OES / ICP-MS · GD-OES

Standards ASTM E1019 (C, S, N and O in steel, iron, Ni and Co alloys) / ASTM E1447 (H in reactive metals such as Ti and Zr) / ISO 3690 (diffusible hydrogen in weld metal)

Failure Analysis Hydrogen embrittlement

Sources: ASTM E1019; ASTM E1447; Klett et al. 2020; EAG technique chart

Fire a spark between a metal sample and an electrode and use the emitted light to determine many elements at once, including C, N, P, S and B. Used for composition control of steels, cast irons and nonferrous metals.

Principle A spark discharge in argon vaporizes and excites the sample surface, and the intensities of element-specific emission lines are measured. Lines of C, N, P and S lie in the vacuum ultraviolet, which air absorbs, so the spectrometer is evacuated.

Used for Average composition Trace elements and impurities

Typical figures

Lateral resolution

— (average over the spark spot)

Depth resolution / information depth

—

Detection limit

Below 10 ppm for some elements (instrument examples: below 10 ppm for B and Pb, below 1 ppm for Zn)

Analyzed depth / volume

Spark spot × surface layer

Elements

Many elements in metals, including C, N, P, S and B. N needs a capable instrument; use gas analysis for O and H (some instruments handle O in copper and H in titanium alloys)

Quantitative accuracy

Calibrated with standards of similar composition (prepared the same way)

Sample and effort

Size

Metal pieces that can be ground flat and cover the spark stand opening

Conductivity

Conductive metals only (the sample is an electrode)

Vacuum

Not needed (the sample is in argon; the spectrometer is evacuated)

Destructive?

Cut or polish Leaves a spark mark; the surface is ground

Preparation

Moderate Grind the surface flat and keep it clean; spark on sound metal without pores or pits

Time

Short Each measurement is quick

Cost

Low Cheap

Pitfalls

• Sparking on pores or pits changes the discharge and gives wrong values. Grind and measure sound metal
• Touching or contaminating the ground surface changes the result. Keep it clean after grinding
• If standards are not prepared the same way as samples, the calibration does not match

Often combined with Combustion and inert gas fusion analysis (C, S, O, N, H) · XRF · ICP-OES / ICP-MS

Standards ASTM E415 (spark atomic emission analysis of carbon and low-alloy steel)

Sources: ASTM E415; Hitachi High-Tech OES

Sputter the surface with a glow discharge while recording the emitted light to obtain composition depth profiles from nm to about 100 µm (plating, nitrided layers, oxides, paint) for almost all elements including H, in a short time.

Principle An argon glow discharge with the sample as cathode sputters surface atoms; their emission in the plasma is dispersed and recorded element by element against time (= depth).

Used for Composition versus depth Thickness of films and layers Average composition Trace elements and impurities

Typical figures

Lateral resolution

— (average over a circle a few mm across; no lateral mapping)

Depth resolution / information depth

About 10% of the depth (for layers thicker than 10 nm)

Detection limit

1–100 ppm (element and matrix dependent)

Analyzed depth / volume

Sputtered circle × depth (nm to about 100 µm; over 150 µm reported)

Elements

Almost all elements including H, C, N and O

Quantitative accuracy

Calibrated with standards (matrix-independent calibration is possible)

Sample and effort

Size

Pieces or sheets with a flat area larger than 4 mm across

Conductivity

Not needed (RF discharge for insulators)

Vacuum

No high vacuum, but the sample face is exposed to low-pressure argon

Destructive?

Cut or polish Leaves a sputter crater

Preparation

Minimal None if there is a flat surface

Time

Short Minutes per measurement

Cost

Moderate Medium (fewer instruments)

Pitfalls

• Depth resolution is about 10% of the depth, so deeper interfaces look broader. Do not mistake the interface width for a diffusion width
• Laterally everything is averaged: islands, local precipitates or non-uniform layers are not seen (laterally uniform layers are assumed)
• Converting time to depth needs the sputter rate (material dependent) and a measured crater depth

Often combined with XPS (ESCA) · SIMS · ICP-OES / ICP-MS

Standards ISO 3887 (determining the depth of decarburization of steels)

Calculators Diffusion and surface-treatment simulatorFormulas Error-function solutions (carburising, diffusion couple) Diffusion distance and time–temperature equivalence

Terms Nitriding Carburizing

Sources: IFW Dresden GD-OES; HORIBA GD-OES; ISO 3887

Heat the sample at a constant rate and measure the hydrogen released at each temperature. Used to find how much hydrogen a steel contains and where (in which traps) it is held.

Principle The sample is heated and the released hydrogen is measured, either with a mass spectrometer in vacuum (TDS) or carried by argon to a gas chromatograph (TDA). Changing the heating rate shifts the desorption peak temperatures, and the shift gives the binding energy of each trap.

Used for Average composition

Typical figures

Lateral resolution

— (whole sample)

Depth resolution / information depth

—

Detection limit

About 1 wt ppm (about 50 atomic ppm) in steel with one instrument

Analyzed depth / volume

The whole sample

Elements

Hydrogen

Quantitative accuracy

Hydrogen held in different ways is separated by the desorption curve (peaks) versus temperature

Sample and effort

Size

Small pieces (cut out)

Conductivity

Not needed

Vacuum

Depends on the setup (TDS in vacuum; gas-chromatograph TDA in a furnace with flowing argon)

Destructive?

Cut or polish Cut out; heating drives the hydrogen out

Preparation

Extensive Keep hydrogen-charged samples in liquid nitrogen until measurement so the hydrogen stays in

Time

Moderate (longer with slow heating)

Cost

Moderate (few instruments)

Pitfalls

• Water adsorbed on the surface also gives a hydrogen peak (a peak near 350 °C after about a minute in laboratory air was reported). Decide how samples are handled
• Diffusible hydrogen escapes even at room temperature. Keep samples cold between charging and measurement
• In thin cut samples, surface working and condition shift the peak temperatures (cutting reduced the measured hydrogen in one report)

Often combined with Combustion and inert gas fusion analysis (C, S, O, N, H) · SIMS · SEM

Failure Analysis Hydrogen embrittlement

Sources: Tapia-Bastidas et al. (TDS); AIST HYDROGENIUS (TDS); Enomoto & Hirakami 2015 (ISIJ Int.)

Identify the elements within a few nm of the surface and their chemical state (metal or oxide, bonding partner). The standard tool for passive films, contamination and plating adhesion failures.

Principle The kinetic energy of photoelectrons ejected by X-rays gives their binding energy, which identifies the element and shifts slightly with chemical state (chemical shift).

Used for Chemical state (oxidation state, bonding) Composition versus depth Thickness of films and layers Organics, polymers and contamination

Typical figures

Lateral resolution

10 µm to 2 mm (instrument dependent; smallest analysis areas of about 30–100 µm reported)

Depth resolution / information depth

Top 1–10 nm

Detection limit

About 0.1 at% (0.1–1 at%)

Analyzed depth / volume

Tens of µm to mm across × a few nm

Elements

Li–U (H and He cannot be measured)

Quantitative accuracy

Mostly semi-quantitative with sensitivity factors. In sputter depth profiles, sputtering can change the chemical state

Sample and effort

Size

Small pieces that fit the instrument (some take wafers)

Conductivity

Not needed (insulators are fine)

Vacuum

Required (ultra-high vacuum compatible)

Destructive?

Non-destructive for surface analysis; depth profiling sputters

Preparation

Minimal Almost none (do not touch or contaminate the surface)

Time

Moderate Hours per sample

Cost

Moderate Medium (fewer instruments)

Pitfalls

• It sees only the top few nm, so adventitious carbon and native oxide from air are easily mistaken for the composition of the material
• Ion sputtering during depth profiling changes chemical states (for example by reduction). Do not over-trust states deep in the profile
• Elements below about 0.1 at% are not seen

Often combined with AES · SIMS · GD-OES

Terms XPS (X-ray photoelectron spectroscopy) Passivity

Sources: EAG XPS; VUB SURF XPS

Analyze the elements in the top few nm at a chosen spot with an electron beam focused to 10–20 nm. Used for tiny particles, grain-boundary segregation and depth profiles of thin films.

Principle Elements are identified by the energy of Auger electrons emitted by atoms excited by the beam. Auger electrons can escape only from the top few nm.

Used for Composition versus depth Local composition and element maps Precipitates

Typical figures

Lateral resolution

10 nm and above (beam can be focused to 10–20 nm)

Depth resolution / information depth

Top 3–10 nm; 2–20 nm in depth profiles

Detection limit

0.1–5 at% (element and instrument dependent)

Analyzed depth / volume

A ~10 nm spot × a few nm

Elements

All except H and He

Quantitative accuracy

Semi-quantitative with sensitivity factors (better against samples of known composition)

Sample and effort

Size

Small pieces that fit the instrument (some take 300 mm wafers)

Conductivity

Insulators charge and are difficult

Vacuum

Required

Destructive?

Non-destructive for surface analysis; depth profiling sputters

Preparation

Minimal Almost none

Time

Moderate Hours per sample

Cost

Moderate Medium (fewer instruments)

Pitfalls

• Insulators charge and are hard to analyze (use XPS for chemical states of insulating surfaces)
• Little chemical-state information (use XPS if you need the state)
• Elements below 0.1–1 at% are not seen

Often combined with XPS (ESCA) · SEM · SIMS

Sources: EAG AES; KIT KNMFi AES

Sputter the surface with ions and mass-analyze the secondary ions to obtain depth profiles of dopants and impurities below ppm levels (10¹⁰–10¹⁶ atoms/cm³).

Principle A primary ion beam (O₂⁺, Cs⁺ etc.) erodes the sample and the ejected ions are counted isotope by isotope in a mass spectrometer.

Used for Composition versus depth Trace elements and impurities Local composition and element maps

Typical figures

Lateral resolution

Depth profiles over areas of 10 µm and above; about 1 µm in imaging

Depth resolution / information depth

From about 0.5 nm (a few Å)

Detection limit

10¹⁰–10¹⁶ atoms/cm³ (ppm and below)

Analyzed depth / volume

µm to hundreds of µm across × depth (a few Å to tens of µm)

Elements

H–U including isotopes (adsorbed water and hydrocarbons give a hydrogen background, suppressed by ultra-high vacuum and cryogenic trapping)

Quantitative accuracy

Concentrations require standards (for example ion-implanted samples)

Sample and effort

Size

Small pieces that fit the instrument

Conductivity

Insulators charge up, so charge compensation with an electron beam is needed (instrument dependent)

Vacuum

Required (vacuum-compatible solids)

Destructive?

Cut or polish Leaves a crater (destructive)

Preparation

Minimal Almost none

Time

Moderate Hours per sample

Cost

High (often outsourced)

Pitfalls

• Without standards there are no concentrations. Calibrate with standards in the same matrix (ion-implanted samples etc.)
• Destructive: the crater remains and the spot cannot be reused
• No chemical bonding information (use XPS for bonding)

Often combined with XPS (ESCA) · APT · GD-OES

Formulas Diffusion from a fixed surface amount (Gaussian) Error-function solutions (carburising, diffusion couple)

Terms SIMS (secondary ion mass spectrometry)

Sources: EAG SIMS; CAMECA SIMS on insulators; CAMECA hydrogen by SIMS

Field-evaporate atoms one by one from a needle-shaped specimen and reconstruct in 3-D which atom was where. Used for the composition of nm-sized precipitates, clusters and grain-boundary segregation.

Principle A needle cooled to 25–80 K is held at high voltage, and voltage or laser pulses ionize surface atoms; time of flight gives the mass and the detector position gives the original location.

Used for Precipitates Local composition and element maps Trace elements and impurities

Typical figures

Lateral resolution

0.5 nm lateral and 0.3 nm in depth in metals (about 1 nm in semiconductors and insulators)

Depth resolution / information depth

0.3 nm

Detection limit

About 10 ppm

Analyzed depth / volume

About 50 nm square × 100–500 nm deep

Elements

H–U (residual hydrogen in the analysis chamber gives a background, so deuterium is sometimes used to tell them apart)

Quantitative accuracy

Accuracy drops for elements with overlapping mass peaks

Sample and effort

Size

A needle with apex radius below 100 nm (made by FIB etc.)

Conductivity

Voltage pulsing for conductors, laser pulsing for semiconductors and insulators

Vacuum

Required (cryogenic, ultra-high vacuum)

Destructive?

Cut or polish The part made into a needle is destroyed

Preparation

Extensive Needles made by FIB etc. (hours each). Needs skill

Time

Long Days to weeks including preparation

Cost

High (few instruments)

Pitfalls

• The analyzed volume is tiny (50 nm square × a few hundred nm). Choose the site with TEM or SEM before making the needle
• Needles can fracture in the high field before the run is finished (material dependent)
• Position and composition accuracy degrade at interfaces and in complex geometries

Often combined with TEM/STEM · FIB · SIMS

Microstructure Atlas Precipitates (on grain boundaries and inside grains)

Terms Precipitation hardening

Sources: EAG APT; Khanchandani et al. (H in APT)

Identify crystalline phases and measure lattice parameters, crystallite size, strain and crystallinity on powders, plates and thin films as they are, in a short time.

Principle The positions and intensities of the diffracted beams that satisfy the Bragg condition (2d sinθ = λ) are matched against database (ICDD) patterns.

Used for Phase identification and crystal structure Grain size, orientation and texture Precipitates

Typical figures

Lateral resolution

Usually an average over 0.1 to several mm (20–25 µm with micro-diffraction)

Depth resolution / information depth

The 1/e attenuation depth of Cu Kα X-rays is about 4 µm in iron, 70 µm in Al and 3 mm in polyethylene (calculated from NIST attenuation coefficients); a few nm at grazing incidence

Detection limit

A minor phase from about 1%

Analyzed depth / volume

Irradiated area × X-ray penetration depth (varies greatly with material)

Elements

Elements are not measured directly (seen as crystalline phases)

Quantitative accuracy

Phase fractions are approximate (phases below about 1% are missed)

Sample and effort

Size

Sub-milligram powders to plates and parts that fit the stage

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Non-destructive

Preparation

Minimal Almost none (pack powders flat)

Time

Short Minutes to an hour per sample

Cost

Low Cheap (common in universities)

Pitfalls

• Confusing Scherrer "crystallite size" with grain or particle size. One particle can contain many crystallites. Above about 100 nm the estimate is unreliable, and instrumental broadening must be corrected and strain broadening separated
• Amorphous phases give no peaks and cannot be identified
• Phases below about 1% are missed
• Powder packing or rolling texture changes intensity ratios so that patterns do not match the database

Often combined with XRD (sin²ψ, pole figures) · SAED/CBED · TEM/STEM

Calculators XRD analysis workbench XRD analysis workbenchFormulas Bragg’s law Interplanar spacing (cubic) Scherrer equation Williamson–Hall plot Vegard’s law

Terms XRD (X-ray diffraction) Bragg's law Scherrer equation Extinction rules

Sources: EAG XRD; Rigaku crystallite size; NIST X-ray attenuation

Measure near-surface residual stress from the shift of a diffraction peak (lattice spacing) while tilting the sample, and texture (pole figures) from the diffracted intensity as the orientation is changed.

Principle In the sin²ψ method the spacing d is measured at several tilts ψ; stress follows from the slope of d against sin²ψ and the elastic constants of the measured plane.

Used for Residual stress Grain size, orientation and texture

Typical figures

Lateral resolution

Average over the irradiated area (0.1 to several mm)

Depth resolution / information depth

Average over a few to tens of µm from the surface (5–50 µm in most materials; Cr Kα is used for steel). Deeper levels are measured after removing layers by electropolishing

Detection limit

—

Analyzed depth / volume

Irradiated area × a few to tens of µm

Elements

—

Quantitative accuracy

Coarse grains mean few diffracting grains and poorer precision (10–100 µm grains are fine)

Sample and effort

Size

Pieces to parts that fit the stage (portable instruments for large structures)

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Non-destructive (layers are removed by electropolishing for depth profiles)

Preparation

Minimal Measure the as-received surface; do not grind or polish

Time

Moderate Tens of minutes to hours per point

Cost

Moderate Medium

Pitfalls

• Grinding, machining or wire brushing adds surface stress. Measure as received, and remove layers only by electropolishing or chemical attack
• It gives the stress within a few to tens of µm of the surface, not in the interior (use neutron diffraction for the interior)
• Texture or coarse grains make d versus sin²ψ non-linear; taking a slope regardless gives wrong values
• Use the elastic constants of the measured plane (X-ray elastic constants), not bulk values

Often combined with ND · XRD · EBSD

Standards ASTM E2860 (residual stress by XRD in bearing steels) / ASTM E81 (preparing quantitative pole figures)

Failure Analysis Fatigue Stress corrosion crackingCalculators XRD analysis workbenchFormulas Peak shift and lattice strain Stress, strain and Hooke’s law Stoney equation

Sources: NPL GPG 52; EAG XRD

Determine crystal structure, orientation and orientation relationships from a selected region of nm to a few hundred nm inside the TEM. Identifies the phase of a single precipitate.

Principle The diffraction pattern of electrons transmitted through the specimen is recorded. Selected-area diffraction (SAED) chooses the region with an aperture; convergent-beam diffraction (CBED) takes the pattern from a finely focused probe.

Used for Phase identification and crystal structure Precipitates Grain size, orientation and texture

Typical figures

Lateral resolution

SAED about 250 nm (set by the aperture); CBED a few nm

Depth resolution / information depth

The whole specimen thickness (<100 nm)

Detection limit

—

Analyzed depth / volume

A small region of a thin foil

Elements

— (use EDS or EELS for composition)

Quantitative accuracy

Lattice spacings are usually accurate to 1–3% (0.1% with careful calibration), poorer than XRD

Sample and effort

Size

TEM foils

Conductivity

Not needed (thinning is what matters)

Vacuum

Required

Destructive?

Cut or polish Destroys the part made into a foil

Preparation

Extensive As for TEM (make a foil)

Time

Long Days to weeks including preparation

Cost

High (often outsourced)

Pitfalls

• Spacings are good to about 1–3%. Use XRD for precise lattice parameters
• Errors in camera length (which changes with specimen height) shift d. Calibrate with a standard
• The selected region is small, so phase fractions of the whole sample are unknown

Often combined with TEM/STEM · XRD · EBSD

Formulas Structure factor and diffracted intensity Interplanar spacing (cubic)

Terms Extinction rules

Sources: Univ. Duisburg-Essen ICAN; Czigány & Kis 2023

Because neutrons penetrate deep into metals, map residual stress inside components and study crystal structures containing light elements such as H and Li, and magnetic structures.

Principle Neutrons are scattered by nuclei. Scattering strength does not follow atomic number, so light elements that X-rays barely see, and isotopes of the same element, can be distinguished.

Used for Residual stress Phase identification and crystal structure Grain size, orientation and texture

Typical figures

Lateral resolution

Gauge volume about 1–10 mm³

Depth resolution / information depth

Same (mm scale)

Detection limit

—

Analyzed depth / volume

Total path length up to about 60 mm in steel and 100 mm in Al

Elements

Light elements (H) and isotopes are easy to distinguish

Quantitative accuracy

Stress accuracy about 10 MPa in Al and 30 MPa in steel in one example

Sample and effort

Size

Small pieces to large parts

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Non-destructive (a stress-free reference may be cut out)

Preparation

Minimal Prepare a stress-free reference (d₀)

Time

Long Months from beam-time proposal; days of measurement

Cost

High (reactor or spallation facilities)

Pitfalls

• A stress-free reference (d₀) is required, and how it is prepared changes the result a lot
• Only possible at large facilities (reactors, accelerators), with long lead times from proposal to measurement
• The near-surface region (tens of µm) is hard to resolve with mm-sized gauge volumes (use XRD for the surface)

Often combined with XRD (sin²ψ, pole figures) · CT

Standards ISO 21432 (residual stress by neutron diffraction)

Failure Analysis Fatigue

Sources: NIST NCNR; Veqter neutron diffraction

Measure the temperatures and heats of melting, crystallization, glass transition, precipitation and phase transformations while heating or cooling at a constant rate. Used for glass transitions and crystallinity of polymers and transformation temperatures of alloys.

Principle DTA records the temperature difference between sample and reference, DSC the difference in heat flow. After calibration, DSC peak areas give heats.

Used for Transition temperatures and reactions Organics, polymers and contamination Precipitates

Typical figures

Lateral resolution

— (whole sample)

Depth resolution / information depth

—

Detection limit

DSC is more sensitive than DTA, so weak events such as glass transitions are easier to see

Analyzed depth / volume

All of a small sample

Elements

—

Quantitative accuracy

Quantitative heats with DSC; DTA suits high temperatures (above 800 °C) and mainly gives temperatures

Sample and effort

Size

Small pieces or powder that fit the pan (cut from larger items)

Conductivity

Not needed

Vacuum

Not needed (usually under flowing inert gas)

Destructive?

Destroys the sample A small piece is cut and altered by heating

Preparation

Moderate Cut a small piece and seal it in a pan

Time

Short One to a few hours per run

Cost

Low Cheap (common in universities)

Pitfalls

• Transition temperatures shift with heating rate. Compare samples at the same rate, and go slower for accurate values
• Decomposition or evaporation overlapping a transition spoils the heat (use only while the sample mass is constant)
• Solid-state transformations in metals can release little heat and be missed (combine with dilatometry)

Often combined with TGA · Dilatometry · DMA

Calculators Iron–carbon diagram

Terms Glass transition temperature Crystallinity Precipitation hardening

Sources: EAG DSC; Bhadeshia (Cambridge); Labcompare (DSC vs DTA)

Record the change in sample mass on heating to find moisture and volatiles, decomposition temperatures, filler content and weight gain from oxidation.

Principle The sample sits on a balance in a furnace and its mass is logged in inert or oxidizing gas.

Used for Decomposition, moisture and filler content (mass change on heating) Transition temperatures and reactions Organics, polymers and contamination

Typical figures

Lateral resolution

— (whole sample)

Depth resolution / information depth

—

Detection limit

Mass changes below a microgram (instrument example; samples of about 1–150 mg)

Analyzed depth / volume

The whole sample (up to about 1 g in one example)

Elements

—

Quantitative accuracy

Weighing precision ±0.01% in one example

Sample and effort

Size

Small pieces or powder that fit the pan

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Destroys the sample A small piece is cut and altered by heating

Preparation

Moderate Cut a small piece and place it in the pan

Time

Short One to a few hours per run

Cost

Low Cheap

Pitfalls

• It does not say what was lost. Identify evolved gases with TGA-MS or TGA-FTIR
• Results depend on the atmosphere (inert or air); weight gain by oxidation and loss by decomposition can overlap

Often combined with DSC/DTA · FTIR

Formulas Oxidation kinetics (parabolic and linear)

Terms Parabolic rate law

Sources: EAG TGA; UW–Madison WCNT TGA

Record the change in length while changing temperature to obtain the thermal expansion coefficient, transformation temperatures of steels (Ac₁, Ac₃, Ms etc.) and sintering shrinkage.

Principle The extension of a rod specimen is transmitted by a push rod to a displacement sensor outside the furnace. Phase transformations change the volume, so the curve bends.

Used for Transition temperatures and reactions Thermal expansion and conduction

Typical figures

Lateral resolution

— (whole sample)

Depth resolution / information depth

—

Detection limit

—

Analyzed depth / volume

The whole rod specimen

Elements

—

Quantitative accuracy

Transformation temperatures and fractions are fixed by combining the curve with metallography

Sample and effort

Size

Rods mm to cm long (machined)

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Cut or polish A specimen is machined out

Preparation

Moderate Machine a rod of set shape

Time

Moderate A few hours per run (depends on the thermal cycle)

Cost

Moderate Medium

Pitfalls

• Transformation temperatures change with heating and cooling rates (and differ between continuous cooling and isothermal holding). State the conditions when comparing
• Do not identify phases from kinks in the curve alone. Confirm with metallography
• Rolled and composite materials expand differently in different directions. Record the specimen orientation

Often combined with DSC/DTA · OM · LFA

Standards ASTM E228 (linear thermal expansion with a push-rod dilatometer) / ASTM A1033 (phase transformations of hypoeutectoid carbon and low-alloy steels)

Calculators Thermal Expansion & Stress Lab Iron–carbon diagramFormulas Thermal expansion

Terms Coefficient of linear thermal expansion Martensite

Sources: ASTM A1033

Heat the front of a thin disk with a light pulse and obtain thermal diffusivity from how fast the rear face warms up; multiply by specific heat and density for thermal conductivity.

Principle An infrared detector records the rear-face temperature rise, and diffusivity is calculated from the thickness and the time to reach a set fraction of the maximum rise.

Used for Thermal expansion and conduction

Typical figures

Lateral resolution

— (through-thickness value of the whole disk)

Depth resolution / information depth

—

Detection limit

—

Analyzed depth / volume

The whole disk

Elements

—

Quantitative accuracy

Conductivity is diffusivity times specific heat and density, so it carries their errors too

Sample and effort

Size

Small pieces that can be machined into disks

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Cut or polish Machined into a disk

Preparation

Moderate Machine a disk; usually spray with graphite

Time

Moderate Minutes per temperature; hours for a temperature series

Cost

Moderate Medium

Pitfalls

• Translucent polymers and glasses, and shiny metals, need a graphite coating to be measured correctly. Too much graphite lowers the value for highly conductive Cu and Al (4% with a ~5 µm layer in one example)
• Homogeneous, isotropic, opaque samples are assumed. Analyze layered or anisotropic materials with their orientation and structure in mind
• Conductivity = diffusivity × specific heat × density, so errors in specific heat and density go straight into it

Often combined with Dilatometry · DSC/DTA

Standards ASTM E1461 (thermal diffusivity by the flash method)

Formulas Thermal diffusivity Heat conduction (Fourier’s law)

Terms Thermal diffusivity

Sources: ASTM E1461; NETZSCH AN 066

Apply a small oscillating deformation while changing temperature to measure storage modulus, loss modulus and tan δ. Detects the glass transition of polymers more sensitively than DSC.

Principle The phase lag between oscillating force and deformation separates the elastic (E′) and viscous (E″) components. Tg is taken as the onset of the E′ drop, the E″ peak or the tan δ peak.

Used for Transition temperatures and reactions

Typical figures

Lateral resolution

— (whole specimen)

Depth resolution / information depth

—

Detection limit

— (transitions hard to see by DSC are often clearer)

Analyzed depth / volume

The whole specimen

Elements

—

Quantitative accuracy

Tg depends on the definition (E′ onset < E″ peak < tan δ peak) and on frequency. Use the same definition and frequency when comparing

Sample and effort

Size

Bars or strips (held in clamps)

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Cut or polish A specimen is cut

Preparation

Moderate Cut flat, untwisted bars or strips

Time

Moderate (longer with slow heating)

Cost

Moderate

Pitfalls

• Tg differs between definitions (E′ onset, E″ peak, tan δ peak). State which was used
• Higher frequencies give higher Tg. Compare at the same frequency
• Heating thick specimens fast makes the specimen temperature lag. Heat thick specimens slowly

Often combined with DSC/DTA · TGA · Tensile testing

Standards ASTM E1640 (assignment of the glass transition temperature by DMA)

Terms Glass transition temperature Thermosetting resin

Sources: TA Instruments DMA

Pull a specimen to measure Young's modulus, the yield point or 0.2% proof stress, tensile strength, elongation and reduction of area. The basic strength values of a material.

Principle The specimen is pulled at a constant rate while force and extension (extensometer) are recorded to give the engineering stress–strain curve.

Used for Hardness, strength and ductility

Typical figures

Lateral resolution

— (whole gauge length)

Depth resolution / information depth

—

Detection limit

—

Analyzed depth / volume

The gauge section

Elements

—

Quantitative accuracy

Elongation depends on specimen shape and gauge length, so compare like specimens

Sample and effort

Size

Enough material to machine a standard specimen (small specimens exist)

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Destroys the sample Broken

Preparation

Extensive Machine a specimen of standard shape

Time

Moderate Minutes to tens of minutes per specimen (days to machine specimens)

Cost

Low Cheap (common in universities)

Pitfalls

• Confusing the (upper) yield point with the 0.2% proof stress. Use proof stress when there is no distinct yield point
• Reading the fall in engineering stress after necking as weakening of the material (the cross-section is shrinking; look at true stress)
• Elongation depends on specimen shape and gauge length. Do not compare elongations of different specimen types directly

Often combined with HV/HR/HBW · Charpy impact testing · SEM

Standards JIS Z 2241 (tensile testing of metallic materials) / ISO 6892-1 (tensile testing of metallic materials at room temperature)

Failure Analysis Ductile overload fractureFormulas Stress, strain and Hooke’s law True stress and true strain Ductility: elongation and reduction of area Strain hardening (Hollomon)

Terms 0.2% offset yield strength Young's modulus Work hardening

Sources: TWI tensile testing; Princeton MAE 324 necking

Measure hardness from the size or depth of an indent. Quick, possible on finished parts, and used for hardness profiles through heat treatments, case-hardened layers and welds.

Principle Vickers uses the diagonals of the indent of a square diamond pyramid, Brinell the diameter of a ball indent, and Rockwell the depth of a cone or ball indent.

Used for Hardness, strength and ductility Thickness of films and layers

Typical figures

Lateral resolution

Set by the indent size (for example, about 30 µm diagonal at 0.98 N in a 200 HV steel)

Depth resolution / information depth

Indent depth and its surroundings

Detection limit

—

Analyzed depth / volume

A small volume around the indent

Elements

—

Quantitative accuracy

Conversions between scales and to tensile strength are approximate and limited to tables for specific materials

Sample and effort

Size

Small pieces to parts (Rockwell or portable testers on finished parts)

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Non-destructive Leaves a small indent (portable testers measure parts as they are). Hardness profiles through layers need a cut section

Preparation

Moderate A polished surface for micro-Vickers; a flat surface for Rockwell and Brinell

Time

Short Tens of seconds to minutes per indent

Cost

Low Cheap

Pitfalls

• Indents too close to each other or to an edge give wrong values because of work hardening around earlier indents or lack of support. Keep the spacings set by the standard
• On thin samples or thin layers the anvil or substrate shows through. Respect the minimum thickness in the standard, and use nanoindentation for thin films
• Conversions between HV, HRC and HB and to tensile strength are approximate; use only tables matching the material and heat treatment

Often combined with Tensile testing · Nanoindentation · OM

Standards JIS Z 2244-1 (Vickers hardness test) / JIS Z 2245 (Rockwell hardness test) / JIS Z 2243-1 (Brinell hardness test) / ASTM E92 (Vickers and Knoop hardness) / ASTM E110 (Rockwell and Brinell hardness by portable testers) / ASTM E140 (hardness conversion tables) / ISO 3887 (determining the depth of decarburization of steels)

Calculators Diffusion and surface-treatment simulator

Terms Work hardening Hardenability Carburizing Nitriding

Sources: Instron TechNotes 35; ASTM E140; ISO 3887; ASTM E110

Record force and depth continuously while indenting to nm–µm depths, and obtain hardness and Young's modulus of thin films, small phases and surface layers.

Principle From the load–depth curves on loading and unloading, the contact area and the initial unloading slope (contact stiffness) are found and hardness and modulus are calculated.

Used for Hardness, strength and ductility

Typical figures

Lateral resolution

Positioning below 50 nm (indents need µm spacing)

Depth resolution / information depth

Displacement noise below 0.1 nm, up to about 20 µm

Detection limit

—

Analyzed depth / volume

A small volume around the indent

Elements

—

Quantitative accuracy

Indenting shallower than 10–20% of the film thickness is the usual rule, but for a hard film on a soft substrate the substrate shows through even at 5%

Sample and effort

Size

Small pieces to wafers

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Cut or polish Leaves a small indent

Preparation

Moderate Flat, smooth surface (roughness scatters the values)

Time

Short Minutes per indent; hours for a map

Cost

Moderate Medium

Pitfalls

• Substrate properties mix into the film value. For a hard film on a soft substrate the substrate shows through even at shallow depths
• Rough surfaces and films thinner than 10 nm give semi-quantitative values only
• Indentation hardness (instrumented) and Vickers hardness are defined differently. Do not compare the numbers directly

Often combined with HV/HR/HBW · AFM · SEM

Standards ISO 14577-1 (instrumented indentation test for hardness and materials parameters)

Terms Young's modulus

Sources: EAG nanoindentation; Chen & Vlassak 2000

Break a notched specimen with one swing of a pendulum and compare toughness (whether it fails in a brittle way) by the absorbed energy. Tests at several temperatures reveal the ductile–brittle transition.

Principle The energy used to break the specimen is found from the difference between the height the pendulum is raised to and the height it swings up to afterwards.

Used for Toughness, fatigue and creep

Typical figures

Lateral resolution

— (whole specimen)

Depth resolution / information depth

—

Detection limit

—

Analyzed depth / volume

Around the notch

Elements

—

Quantitative accuracy

Values scatter even at one temperature, so several specimens are averaged. Values are for comparison with specifications and cannot be converted into fracture toughness

Sample and effort

Size

Enough material to machine a standard specimen

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Destroys the sample Broken

Preparation

Extensive Machine to standard shape, cut the notch, condition to temperature

Time

Moderate Minutes per specimen (days to machine and condition specimens)

Cost

Low Cheap

Pitfalls

• Using absorbed energy as a fracture toughness for design. Results are for comparison with each other or with a specification
• Comparing at one temperature only. In the transition region small temperature differences change the value greatly, so test at several temperatures and draw the curve
• Also record the crystalline (brittle) versus fibrous (ductile) fraction of the fracture and the lateral expansion

Often combined with Tensile testing · SEM

Standards JIS Z 2242 (Charpy impact test of metallic materials) / ISO 148-1 (Charpy pendulum impact test) / ASTM E23 (notched bar impact testing)

Failure Analysis Brittle fractureFormulas Stress intensity factor and fracture toughness

Terms Fracture toughness

Sources: TWI (impact testing)

Apply cyclic loading and count cycles to failure to obtain the S–N curve and the fatigue limit (fatigue strength). Crack growth rate (da/dN) tests are another form.

Principle Several specimens at different stress amplitudes are cycled until they fail or reach a set number of cycles (run-out).

Used for Toughness, fatigue and creep

Typical figures

Lateral resolution

— (whole specimen)

Depth resolution / information depth

—

Detection limit

—

Analyzed depth / volume

The gauge section

Elements

—

Quantitative accuracy

Lives scatter widely (log life is close to normally distributed), so test enough specimens and treat the data statistically

Sample and effort

Size

Material for several standard specimens

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Destroys the sample Broken

Preparation

Extensive Machine specimens and finish their surfaces

Time

Long 10⁷ cycles take about 12 days at 10 Hz; weeks to months overall

Cost

High (long use of the test machine)

Pitfalls

• Deriving a design curve (mean minus two standard deviations) from few data. Small samples underestimate the standard deviation, which is unconservative
• Putting run-outs (unbroken specimens) into the regression as if they had failed
• Striation spacing on the fracture surface is a guide to crack advance per cycle. Read it together with SEM fractography

Often combined with SEM · XRD (sin²ψ, pole figures) · Tensile testing

Standards ASTM E466 (force-controlled constant-amplitude axial fatigue tests)

Failure Analysis FatigueFormulas Fatigue life (Basquin, Coffin–Manson) Paris law (fatigue crack growth) Mean-stress effect (modified Goodman) Miner’s rule (linear damage)

Terms Fatigue limit Fatigue striations Paris law Miner's rule Modified Goodman diagram

Sources: TWI (fatigue statistics); NIMS (STAM feature)

Hold a constant load at high temperature and record strain against time (creep curve) and time to rupture. The basis of life assessment for high-temperature components.

Principle A specimen in a furnace carries a constant load while its extension is logged for a long time. Rupture times are organized by temperature and stress (Larson–Miller parameter etc.).

Used for Toughness, fatigue and creep

Typical figures

Lateral resolution

— (whole specimen)

Depth resolution / information depth

—

Detection limit

—

Analyzed depth / volume

The gauge section

Elements

—

Quantitative accuracy

Some programmes aim at the stress that causes rupture in 100,000 hours (about 11 years)

Sample and effort

Size

Material for several standard specimens

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Destroys the sample Broken (or deformed for a long time)

Preparation

Extensive Machine specimens

Time

Long Hundreds to tens of thousands of hours

Cost

High (long use of furnaces and machines)

Pitfalls

• Extrapolating short-term results too far. Check the assumptions of parameter methods (no change of mechanism) and state the extrapolation range
• Small temperature differences change rupture life greatly. Verify temperature control and measurement

Often combined with Tensile testing · SEM · TEM/STEM

Standards ASTM E139 (creep and creep-rupture tests of metallic materials)

Failure Analysis Creep ruptureFormulas Larson–Miller parameter Power-law creep (steady state)

Terms Creep Power-law creep Larson–Miller parameter Deformation mechanism map

Sources: NIMS (STAM feature)

Measure molecular and lattice vibrations from tiny wavelength shifts in scattered laser light to identify carbon forms (diamond, graphite, amorphous), oxides and polymers, and stress in semiconductors, at a 1 µm spot. No vacuum and nondestructive.

Principle Scattered light whose wavelength has changed by exchanging energy with molecular or lattice vibrations (Raman scattering) is dispersed and recorded.

Used for Organics, polymers and contamination Phase identification and crystal structure Chemical state (oxidation state, bonding) Residual stress

Typical figures

Lateral resolution

1 µm and above

Depth resolution / information depth

1–5 µm in confocal mode

Detection limit

1 wt% and above

Analyzed depth / volume

A spot about 1 µm across

Elements

Sees bonds and crystal types, not elements (pure metals give no peaks; oxides and rust can be measured)

Quantitative accuracy

Quantification needs standards

Sample and effort

Size

Particles to small pieces

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Non-destructive Usually non-destructive

Preparation

Minimal Almost none

Time

Short Minutes per spot

Cost

Low Cheap

Pitfalls

• Strong fluorescence swamps the much weaker Raman signal (try another excitation wavelength)
• Quantification is difficult without standards
• Each spectrum comes from a ~1 µm spot. Measure many spots on inhomogeneous samples

Often combined with FTIR · XRD · XPS (ESCA)

Terms Graphene Carbon nanotube

Sources: EAG Raman; Shimadzu rust analysis

Identify polymers, organics, contamination and coatings from infrared absorption by functional groups. With ATR, simply pressing a solid onto the crystal probes roughly the top 0.1–2 µm.

Principle Infrared light at the frequencies of molecular vibrations is absorbed; an interferometer and Fourier transform record all wavelengths at once.

Used for Organics, polymers and contamination Chemical state (oxidation state, bonding)

Typical figures

Lateral resolution

15–50 µm (FTIR microscope)

Depth resolution / information depth

About 0.1–2 µm with ATR

Detection limit

1–10 wt% for quantifying known components; 5–20% for identifying unknowns

Analyzed depth / volume

Illuminated area × penetration depth

Elements

Sees functional groups, not elements

Quantitative accuracy

Quantification needs standards and uniform thickness

Sample and effort

Size

Particles visible by eye to small pieces

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Non-destructive Usually non-destructive

Preparation

Minimal None with ATR; KBr pellets or extraction are sometimes used

Time

Short Minutes per sample

Cost

Low Cheap

Pitfalls

• Metals themselves cannot be measured (only contamination or films on them)
• Water absorbs strongly, so dry wet samples first
• Contamination films thinner than about 100 nm are hard to see (use XPS for the outermost surface)

Often combined with Raman spectroscopy · TGA · XPS (ESCA)

Terms Thermosetting resin

Sources: EAG FTIR; Innsbruck ATR-FTIR

Sweep the potential in solution and measure the current to obtain corrosion potential, corrosion rate, the passive range and pitting potential. AC impedance (EIS) follows the resistance of films and coatings and how it changes.

Principle A potentiostat sweeps the potential in a three-electrode cell (working electrode = sample, counter and reference electrodes). Corrosion current comes from Tafel extrapolation, or from polarization resistance and Tafel slopes (Stern–Geary equation).

Used for Corrosion behavior

Typical figures

Lateral resolution

— (average over the electrode area)

Depth resolution / information depth

—

Detection limit

—

Analyzed depth / volume

The surface in contact with the solution

Elements

—

Quantitative accuracy

Polarization resistance is taken from the linear range around the corrosion potential (about 10 mV or less). Scan rate and solution resistance change the value

Sample and effort

Size

Small pieces with a defined exposed area

Conductivity

A conductive sample (electrode) is required

Vacuum

Not needed

Destructive?

Cut or polish The surface is corroded

Preparation

Moderate Mask all but the measured area; polish to a consistent finish

Time

Short Tens of minutes to hours per run (including stabilization)

Cost

Low Cheap

Pitfalls

• Fast scans mix in charging current (slow scans such as 0.1 mV/s are used)
• Not correcting the potential error from solution resistance (IR drop)
• Not stating the reference electrode (SCE, Ag/AgCl, SHE), so results cannot be compared
• Results depend on test solution and temperature. Use conditions close to service

Often combined with SEM · XPS (ESCA) · EDS (EDX)

Standards JIS G 0577 (pitting potential of stainless steels) / ASTM G59 (potentiodynamic polarization resistance) / ASTM G61 (cyclic potentiodynamic polarization for localized corrosion)

Failure Analysis Stress corrosion crackingFormulas Corrosion from Faraday’s law Nernst equation

Terms Standard electrode potential Passivity Pitting corrosion Galvanic corrosion Stress corrosion cracking

Sources: Metrohm AN-COR-003

Weigh the sample dry, saturated with water and suspended in water to obtain bulk density and the fraction of pores connected to the surface (open porosity). A basic measure of how dense a sintered body is.

Principle The difference between saturated and suspended weights gives the sample volume, and the weight of absorbed water gives the open-pore volume.

Used for Density and porosity Internal defects

Typical figures

Lateral resolution

— (whole sample)

Depth resolution / information depth

—

Detection limit

—

Analyzed depth / volume

The whole sample

Elements

—

Quantitative accuracy

Pores not connected to the surface (closed pores) are not counted

Sample and effort

Size

Small pieces to parts that fit the balance

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Non-destructive (the sample is soaked in water)

Preparation

Moderate Weigh dry, weigh saturated, weigh suspended in water

Time

Short (excluding drying and soaking)

Cost

Low Cheap

Pitfalls

• Closed pores are not counted, so do not judge density from open porosity alone (use sections or X-ray CT for closed pores)
• Incomplete saturation underestimates open porosity
• Materials that react with or dissolve in water cannot be measured as they are

Often combined with CT · OM · SEM

Standards ISO 18754 (density and apparent porosity of fine ceramics)

Sources: Muğla Univ. Archimedes method; ISO 18754

Send ultrasound into a part and receive reflections from internal flaws (cracks, voids, inclusions, weld defects) to estimate their position and size. Also measures wall thickness.

Principle A probe sends ultrasound at about 1–25 MHz through a couplant (water, oil etc.), and the arrival time and height of echoes from flaws and the back wall are read.

Used for Internal defects Thickness of films and layers

Typical figures

Lateral resolution

— (higher frequency is more sensitive to small flaws)

Depth resolution / information depth

Depth from the echo time

Detection limit

—

Analyzed depth / volume

The full thickness of the part

Elements

—

Quantitative accuracy

Echo heights are compared with reference blocks to estimate flaw size

Sample and effort

Size

Small pieces to large structures (on site)

Conductivity

Not needed (sound must pass)

Vacuum

Not needed

Destructive?

Non-destructive

Preparation

Minimal Smooth the contact surface and apply couplant

Time

Short Minutes per location and up

Cost

Low Cheap

Pitfalls

• Coarse-grained materials (austenitic welds, castings) attenuate the sound and give high noise
• Without calibration on reference blocks, flaw sizes are misjudged. Needs skilled operators
• Without couplant the sound does not get in

Often combined with RT · CT · ET

Standards JIS Z 2344 (general rules for pulse-echo ultrasonic testing of metals) / JIS Z 2355-1 (ultrasonic thickness measurement: method)

Terms Heat-affected zone

Sources: Istanbul Tech. Univ. (NDT); JSNDI

Reveal surface-breaking cracks and pinholes as colored or fluorescent indications using a penetrant and developer. Works on any non-porous material: metals, ceramics, plastics.

Principle The penetrant seeps into flaws; after the surface excess is removed, the developer draws it out to form an indication.

Used for Surface cracks (nondestructively)

Typical figures

Lateral resolution

—

Depth resolution / information depth

Surface-breaking flaws only

Detection limit

—

Analyzed depth / volume

Surface

Elements

—

Quantitative accuracy

The indication is larger than the flaw (it bleeds)

Sample and effort

Size

Small pieces to large parts (on site)

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Non-destructive (liquids are applied)

Preparation

Moderate Clean off oil, paint and dirt thoroughly and dry

Time

Short Tens of minutes (including dwell and development)

Cost

Low Cheap

Pitfalls

• Flaws that do not break the surface (internal or subsurface) are not found
• Porous materials retain penetrant everywhere and cannot be tested
• Insufficient pre-cleaning leaves oil or paint plugging the flaw mouths, so flaws are missed

Often combined with MT · ET · SEM

Standards JIS Z 2343-1 (penetrant testing: general principles)

Failure Analysis Fatigue

Sources: Istanbul Tech. Univ. (NDT); JSNDI

Magnetize steel or another ferromagnetic material and collect magnetic particles where flux leaks out at surface and near-surface cracks.

Principle Flux leaks out at a flaw and the particles (fluorescent ones exist) gather there to form an indication. Sensitivity is best when the flux is perpendicular to the flaw.

Used for Surface cracks (nondestructively) Internal defects

Typical figures

Lateral resolution

—

Depth resolution / information depth

Surface and just below

Detection limit

—

Analyzed depth / volume

Near-surface

Elements

—

Quantitative accuracy

The indication length gives a guide to crack length

Sample and effort

Size

Small pieces to large parts (on site)

Conductivity

Ferromagnetic materials only (steel, cast iron, Ni, Co…)

Vacuum

Not needed

Destructive?

Non-destructive

Preparation

Moderate Remove dirt and thick coatings

Time

Short Minutes to tens of minutes

Cost

Low Cheap

Pitfalls

• It does not work on non-ferromagnetic materials (austenitic stainless steel, Al, Cu…); use penetrant or eddy current testing
• Cracks parallel to the flux are hard to see. Magnetize in two directions

Often combined with PT · ET · UT

Standards JIS Z 2320-1 (magnetic particle testing: general principles)

Failure Analysis Fatigue Hydrogen embrittlement

Sources: Istanbul Tech. Univ. (NDT); JSNDI

Induce eddy currents in a conductor with a coil to find surface and near-surface cracks, differences in material (conductivity) and plating or coating thickness, quickly and without contact. Used for 100% inspection of tubes and wire.

Principle Flaws or material changes disturb the eddy currents and change the coil impedance. Higher frequencies see only shallower depths (skin effect).

Used for Surface cracks (nondestructively) Thickness of films and layers Internal defects

Typical figures

Lateral resolution

—

Depth resolution / information depth

Penetration depends on frequency, conductivity and permeability (the standard depth of penetration is where intensity falls to 37%)

Detection limit

—

Analyzed depth / volume

Near-surface

Elements

—

Quantitative accuracy

Signal size is judged against reference standards

Sample and effort

Size

Small pieces to large parts, tubes and wire

Conductivity

Conductors only

Vacuum

Not needed

Destructive?

Non-destructive

Preparation

Minimal Almost none (surface condition affects the signal)

Time

Short Minutes (high speed on production lines)

Cost

Low Cheap

Pitfalls

• Only conductors can be tested
• Deep flaws are not seen; penetration is set by frequency, conductivity and permeability
• Flaws parallel to the coil winding or the scan direction are easily missed
• Without reference standards, signal size cannot be related to flaw size

Often combined with PT · MT · UT

Standards JIS Z 2316-1 (eddy current testing: general principles)

Terms Eddy-current loss

Sources: Istanbul Tech. Univ. (NDT); TWI (eddy current depth); JSNDI

Pass X-rays or gamma rays through a weld or casting onto film or a detector to find internal porosity, lack of fusion and cracks. Leaves a permanent record.

Principle Voids and low-density regions let more radiation through and appear darker on the image.

Used for Internal defects

Typical figures

Lateral resolution

— (as a guide, voids about 2% of the thickness are visible: about 0.5 mm in 25 mm of steel)

Depth resolution / information depth

The thickness is projected (no depth information)

Detection limit

—

Analyzed depth / volume

The whole penetrated thickness

Elements

—

Quantitative accuracy

Judged from density differences on the image (image quality checked with IQIs)

Sample and effort

Size

Small pieces to large structures (on site)

Conductivity

Not needed

Vacuum

Not needed

Destructive?

Non-destructive

Preparation

Minimal Almost none (film or detector must fit behind)

Time

Short Minutes to expose; hours including processing and evaluation

Cost

Moderate Medium (radiation control required)

Pitfalls

• Thin cracks show only when the beam is parallel to the crack plane. Shoot from several directions
• Radiation protection (controlled areas, qualified staff) is required
• The image is a projection, so the depth of a flaw is unknown (use UT or CT)

Often combined with UT · CT

Standards JIS Z 3104 (radiographic testing of steel welds)

Terms Heat-affected zone Shrinkage cavity

Sources: Istanbul Tech. Univ. (NDT)