Scanning Electron Microscopy in Metallurgy: SEM, EDS, and EBSD Techniques
Scanning electron microscopy (SEM) is the most versatile microstructural characterisation platform available to the practising metallurgist. Where optical microscopy is bounded by the diffraction limit of visible light at approximately 200 nm, the SEM focuses an electron beam with effective wavelengths of 10–100 pm — enabling surface imaging at 1–5 nm resolution on modern field-emission instruments. Coupled with energy-dispersive X-ray spectroscopy (EDS) for elemental mapping and electron backscatter diffraction (EBSD) for crystallographic orientation analysis, the SEM delivers compositional, morphological, and crystallographic data simultaneously, from the same specimen, in a single session.
Key Takeaways
- SEM resolution (1–5 nm) exceeds optical microscopy by approximately two orders of magnitude; depth of field is 100–300× greater, making SEM the definitive tool for fracture surface analysis.
- Secondary electrons (SE) reveal surface topography; backscattered electrons (BSE) provide atomic-number (Z) contrast for phase identification without etching.
- EDS detects characteristic X-rays from all elements simultaneously, providing semi-quantitative spot compositions, line scans, and 2D elemental maps with ~0.1–0.3 wt% detection limits.
- EBSD captures Kikuchi diffraction patterns from individual grains to map crystal orientation (IPF maps), phase distribution, grain boundary character, and local plastic strain (KAM maps).
- EBSD requires a deformation-free surface: final polish with colloidal silica (OPS, 0.04–0.06 μm) or electropolishing is mandatory; conventional mechanical polish introduces a deformed Beilby layer that degrades pattern quality.
- SEM fractography identifies failure mechanisms definitively: fatigue striations, microvoid coalescence dimples, cleavage river marks, and intergranular fracture each have distinct and diagnostic morphologies.
SEM Working Principles and Electron–Specimen Interaction
The SEM operates by focusing a fine electron probe — formed by electromagnetic condenser and objective lenses acting on electrons emitted by the gun — onto the specimen surface. The probe is swept across the surface in a two-dimensional raster by deflection coils, and signals generated at each probe position are collected synchronously to build a digital image. The entire electron optical column is maintained under high vacuum (10−5–10−7 Pa) to prevent beam scattering by residual gas molecules.
The Interaction Volume
When the primary electron beam enters the specimen, it does not stop at the surface. Electrons undergo multiple elastic and inelastic scattering events, spreading laterally and losing energy as they penetrate. The resulting interaction volume is a teardrop-shaped region whose size depends on:
- Accelerating voltage (V₀): Higher voltage = larger interaction volume and greater penetration depth. At 5 kV in steel the interaction volume is ~0.2 μm across; at 20 kV it expands to ~1–2 μm.
- Mean atomic number (Z): Heavier elements (higher Z) scatter more strongly, confining the interaction volume closer to the surface. Low-Z materials (Al, C) allow deeper penetration.
- Specimen density (ρ): Higher density reduces penetration range. The Kanaya-Okayama range formula gives a useful estimate:
Kanaya–Okayama electron range:
R_KO = (0.0276 × A × V₀^1.67) / (Z^0.889 × ρ) [μm]
Where:
A = Atomic weight [g/mol]
V₀ = Accelerating voltage [kV]
Z = Atomic number
ρ = Density [g/cm³]
Example: Iron (A=55.85, Z=26, ρ=7.87 g/cm³) at 20 kV:
R_KO = (0.0276 × 55.85 × 20^1.67) / (26^0.889 × 7.87)
≈ 1.8 μm interaction depth
Different signals exit from different depth zones within this interaction volume, each with its own spatial resolution and information content.
Signals Generated and Their Information Content
| Signal | Origin / Mechanism | Emission Depth | Information | Spatial Resolution |
|---|---|---|---|---|
| Secondary electrons (SE) | Low-energy electrons (<50 eV) ejected by inelastic scattering of primary beam | 0–5 nm (surface only) | Topography; surface morphology; edge enhancement | 1–5 nm |
| Backscattered electrons (BSE) | Primary electrons elastically scattered back with high retained energy | 5–50 nm | Atomic number (Z) contrast; phase identification; channelling contrast (crystal orientation) | 5–20 nm |
| Characteristic X-rays (EDS/WDS) | Inner-shell ionisation followed by characteristic X-ray emission | 0.1–2 μm (full interaction volume) | Elemental composition; maps; line profiles; quantitative analysis | 500 nm–2 μm |
| EBSD diffraction patterns | Bragg diffraction from crystal planes in tilted specimen | 10–50 nm (near-surface crystal) | Crystal orientation; phase identification; grain boundary character; strain mapping | 50–200 nm |
| Cathodoluminescence (CL) | Photon emission from radiative recombination of electron–hole pairs | Variable (100 nm–1 μm) | Defect mapping in semiconductors; band gap variations; rarely applied to metals | 50–500 nm |
| Absorbed current (EBIC) | Net current absorbed by specimen = beam current minus emitted electrons | Full interaction volume | P-N junction mapping; depletion regions; semiconductor device characterisation | 10–100 nm |
Table 1 — Electron–specimen interaction signals in SEM: emission depth, information content, and achievable spatial resolution. EDS resolution is interaction-volume limited and far coarser than SE imaging resolution.
Secondary Electron Imaging and Fractographic Analysis
Secondary electron imaging provides the topographic contrast and depth of field that make SEM indispensable for fracture surface examination (fractography). The depth of field at a given magnification scales inversely with numerical aperture; because the SEM uses an extremely small aperture angle (1–10 mrad), depth of field at 1000× magnification is typically 100–300 μm — 100 to 300 times greater than an optical microscope at the same magnification. Rough fracture surfaces, fatigue crack surfaces, and corrosion pits are therefore fully in focus across the entire image field.
Diagnostic Fractographic Features
Parallel ridges perpendicular to crack propagation direction. Each striation represents one loading cycle. Striation spacing equals the crack growth increment per cycle, da/dN. Used to back-calculate loading history from fracture surface.
Hemispherical depressions formed by nucleation, growth, and coalescence of voids at inclusions or second-phase particles. Equiaxed dimples indicate tensile (mode I) fracture; elongated dimples indicate shear. Confirms ductile fracture mechanism.
Flat, reflective crystallographic planes with characteristic river marks (steps between adjacent cleavage planes) converging toward the crack origin. Diagnostic of brittle transgranular fracture in BCC metals (ferritic steel, α-Ti) at low temperature or high strain rate.
Smooth, polyhedral grain surfaces exposed without dimples or striations. Indicates grain boundary embrittlement: temper embrittlement (Sb, Sn, P segregation), hydrogen embrittlement, liquid metal embrittlement, or stress corrosion cracking at grain boundaries.
Combinations of the above features on a single fracture surface are common in real failures. Quantitative area fraction analysis of each mode (via image thresholding) enables objective failure mechanism assessment and comparison with test specimens.
SEM imaging of pit morphologies (crystallographic, hemispherical, lacy cover), combined with EDS confirmation of oxide composition and halide contamination, provides definitive characterisation of corrosion-initiated failures. See the pitting corrosion article for mechanism detail.
Backscattered Electron Imaging: Atomic Number and Phase Contrast
Backscattered electron (BSE) yield (ηBSE) increases monotonically with mean atomic number Z according to the empirical relationship:
BSE yield (approximate): η_BSE ≈ -0.0254 + 0.016Z - 1.86×10⁻⁴Z² + 8.3×10⁻⁷Z³ At Z=26 (Fe): η_BSE ≈ 0.28 At Z=74 (W): η_BSE ≈ 0.48 At Z=13 (Al): η_BSE ≈ 0.17 Higher η_BSE = brighter image in BSE detector
This Z-dependence makes BSE imaging a rapid, etching-free method for phase identification by contrast. In practice:
- Tool steels / hard metals: WC particles (ZW=74) appear bright against a dark Fe-Co binder matrix (ZFe=26, ZCo=27). Mo2C and TiC are readily distinguished.
- Aluminium alloys: CuAl2 precipitates (high Cu, Z=29) appear bright against the Al matrix (Z=13). Fe-rich intermetallics (Al3Fe, Al6Fe) are distinguishable from Mg2Si.
- Superalloys: γ' (Ni3Al) precipitates vs. γ matrix; TCP phases (σ, μ) containing high-Z elements (Re, W, Mo) appear distinctly brighter than the γ matrix.
- Duplex stainless steel: Ferrite and austenite have essentially the same mean Z (both Fe–Cr–Ni based), so BSE cannot distinguish them reliably. EBSD phase mapping is required.
Electron Channelling Contrast (ECC)
In addition to Z-contrast, BSE imaging with a sensitive detector reveals orientation-dependent channelling contrast: grains with different crystallographic orientations channel the incident beam differently, producing subtle brightness differences between adjacent grains with identical chemistry. This allows grain structure to be observed without etching in polished specimens, and can reveal dislocation substructure and deformation band contrast — though at lower contrast than EBSD.
Energy-Dispersive X-ray Spectroscopy (EDS): Elemental Analysis
EDS detects the characteristic X-rays emitted when the primary electron beam ejects inner-shell electrons and higher-energy electrons relax to fill the vacancies. Each element produces X-rays at defined energies (keV) characteristic of the atomic transition: Kα transitions for light and medium elements, Lα and Mα for heavy elements. A silicon drift detector (SDD) — the modern replacement for Si(Li) detectors — captures all energies simultaneously, building a complete spectrum of detected counts versus X-ray energy within seconds.
Types of EDS Measurement
Spot analysis positions the beam on a specific feature (inclusion, precipitate, grain boundary region) and acquires a spectrum for quantification. Standardless quantification using ZAF or φρZ matrix correction gives ±1–2 wt% accuracy for major elements; with calibrated standards this improves to ±0.1–0.5 wt% and enables trace element detection approaching 0.1 wt%.
Line scan steps the beam along a defined path while continuously acquiring spectra, building composition profiles across microstructural features: segregation at grain boundaries, carbide depletion zones adjacent to sensitised stainless steel welds, interdiffusion zones in dissimilar metal welds, coating–substrate interfaces.
Element mapping (X-ray mapping) acquires a full spectrum at every pixel in the image raster, producing 2D colour maps showing the distribution of each detected element simultaneously. Modern SDDs at 106 counts per second enable maps of 512×512 pixels in 5–15 minutes. EDS mapping is the fastest method to identify phase distributions, segregation patterns, and contamination sources in metallographic sections.
Limitations of EDS and When to Use WDS
EDS energy resolution is approximately 130–150 eV FWHM — sufficient to separate most elemental peaks but inadequate for light elements where overlapping peaks cause systematic errors. Critical EDS limitations include:
- Light elements (B, C, N, O): Their very low-energy X-rays (0.1–0.5 keV) are strongly absorbed by the specimen and detector window; accuracy is limited and detection requires special thin-window or windowless EDS detectors with ZAF correction for absorption.
- Peak overlaps: Ti Kβ overlaps with V Kα; Cr Kβ overlaps with Mn Kα; Ba L with Ti Kα. Deconvolution is required and introduces additional uncertainty.
- Detection limits: ~0.1–0.3 wt% for most elements. For trace element analysis (grain boundary segregation, impurity certification), wavelength-dispersive spectroscopy (WDS) in an electron probe microanalyser (EPMA) provides 20–100 ppm detection limits and significantly better accuracy.
EBSD: Crystallographic Orientation and Phase Mapping
Electron backscatter diffraction is the most powerful crystallographic technique available in the SEM. When the primary electron beam enters a crystalline specimen tilted to 70°, a proportion of electrons satisfy Bragg's law for specific crystal planes and diffract, forming pairs of parallel lines — Kikuchi bands — in the diffracted beam. These are projected onto a phosphor screen positioned in front of the tilted specimen and captured by a CCD or CMOS camera, forming the Kikuchi pattern (also called EBSP: electron backscatter pattern).
From Kikuchi Patterns to Orientation Maps
Each Kikuchi pattern is unique to the crystal orientation at the point of measurement. Automated indexing software: (1) detects Kikuchi bands using the Hough transform; (2) measures band widths and interband angles; (3) matches this geometry against a database of crystal structures (BCC, FCC, HCP, etc.) to identify the phase and determine the three Euler angles defining the crystal orientation. Modern SEMs with optimised detectors index 500–3000 patterns per second, enabling maps of millions of measurement points within a practical session time.
EBSD Map Products
| Map Type | What It Shows | Typical Application |
|---|---|---|
| Inverse pole figure (IPF) map | Grain colour coded by crystallographic direction parallel to a chosen specimen axis (e.g., ND = rolling direction normal). Each colour = one orientation. | Grain size, grain shape, texture intensity; identifying preferred orientations after forming or heat treatment |
| Phase map | Each phase (FCC, BCC, HCP, etc.) assigned a unique colour based on crystal structure identified from Kikuchi pattern geometry | Austenite/ferrite ratio in duplex SS; martensite/austenite in TRIP steel; retained austenite measurement; α/β titanium ratio |
| Grain boundary map | Boundaries coloured by misorientation angle: low-angle (sub-grain, <15°), high-angle (>15°), special CSL boundaries (Σ3 twin, Σ9, etc.) | Recovery/recrystallisation state; twin density in austenite; sensitisation-risk boundaries in SS; fatigue crack path analysis |
| Kernel average misorientation (KAM) | Average misorientation between a point and its neighbours within a kernel radius; proxy for GND density and plastic strain | Deformation mapping around welds, cracks, indentations; quantifying heterogeneous deformation in dual-phase steel; HAZ strain analysis |
| Band contrast (BC) / Image quality (IQ) | Intensity of Kikuchi pattern bands; poor in deformed or poorly prepared regions | Pattern quality check; distinguishing heavily deformed zones (low BC) from recrystallised grains (high BC) |
Table 2 — EBSD map types: data product, information content, and principal metallurgical applications.
Retained Austenite Measurement by EBSD
Quantifying retained austenite (RA) volume fraction in martensitic and bainitic steels is a critical quality control measurement — RA fraction affects dimensional stability, fatigue performance, and TRIP effect in advanced high-strength steels. EBSD phase mapping directly resolves FCC austenite from BCC/BCT martensite at the individual grain level. Advantages over X-ray diffraction (XRD) RA measurement: (1) spatial resolution allows correlation with microstructural position (e.g., RA between martensite laths vs. blocky RA at prior-austenite grain boundaries); (2) no averaging across the full specimen area; (3) simultaneous information about grain morphology and orientation. Typical agreement between EBSD and XRD RA measurements is within ±1–2% absolute for specimens above ~2% RA. See the martensite lath and plate types article for the role of RA in martensitic microstructures.
Specimen Preparation for SEM Analysis
Specimen preparation requirements scale with the technique being applied: SE/BSE imaging has the most relaxed requirements; EDS imposes intermediate demands; EBSD is the most demanding and the most sensitive to surface preparation artefacts.
Electropolishing for EBSD
For metals amenable to electropolishing — aluminium alloys, copper and copper alloys, most austenitic stainless steels, titanium alloys in perchlorate-based electrolytes — electrochemical material removal produces a deformation-free surface superior to any mechanical polishing sequence. The surface Beilby layer (an amorphous or nanocrystalline layer introduced by mechanical abrasion) degrades Kikuchi pattern quality and must be entirely removed. Standard electropolishing conditions for austenitic stainless steel: 10% perchloric acid in acetic acid at 0–5 °C, 20–30 V DC, 30–60 seconds.
FIB (Focused Ion Beam) Preparation
Focused ion beam milling uses a gallium ion beam to cut, mill, and thin specimens with nanometre precision. FIB preparation enables site-specific cross-sections through specific microstructural features — a fatigue crack tip, a coating delamination, a specific inclusion — that are inaccessible to conventional preparation. FIB-prepared lamellae (typically 80–100 nm thick, 15–20 μm wide) are extracted and transferred to TEM grids using a micromanipulator, enabling atomic-resolution TEM/STEM analysis of exactly the feature identified in SEM. The FIB is also used to produce EBSD cross-sections perpendicular to the polished surface for 3D EBSD serial sectioning.
Comparative Summary: SEM Techniques and Alternative Characterisation Methods
| Technique | Spatial Resolution | Information | Preparation | Limitation |
|---|---|---|---|---|
| SEM-SE imaging | 1–5 nm | Topography, morphology | Minimal; conductor coat if needed | Surface only; no composition |
| SEM-BSE imaging | 5–20 nm | Phase (Z contrast), orientation (channelling) | Polish; no etch needed | Cannot distinguish same-Z phases |
| EDS mapping | 500 nm–2 μm | Elemental distribution (2D, semi-quant) | Polish; flat surface | ~0.1–0.3 wt% detection limit; light elements poor |
| EBSD mapping | 50–200 nm | Crystal orientation, phase, grain boundaries, strain | Deformation-free polish (OPS/EP/FIB); 70° tilt | Surface crystal only; amorphous = no pattern |
| Optical microscopy | ~200 nm | Microstructure (etched); phase fractions | Polish + etch | Diffraction limited; no composition or orientation |
| EPMA-WDS | 500 nm–1 μm | Quantitative composition; light elements; 20–100 ppm detection | Polish; flat, carbon-coated | No orientation data; slow; dedicated instrument |
| TEM/STEM | 0.05–0.2 nm (atomic) | Atomic structure; dislocations; precipitate interfaces; atomic-column EDS | Thin foil (<100 nm) by FIB or jet polish | Very small area; destructive; complex preparation |
| Atom probe tomography (APT) | 0.1–0.3 nm (3D) | 3D atom-by-atom elemental maps; grain boundary segregation at ppm level | Sharp needle tip (50–100 nm apex) by FIB | Very small volume (~50×50×200 nm); expensive; complex |
Table 3 — Comparison of SEM-based techniques with optical microscopy, EPMA, TEM/STEM, and atom probe tomography by resolution, information content, preparation requirement, and principal limitation.
Industrial Applications of SEM in Metallurgical Practice
Failure Analysis
SEM is the primary diagnostic instrument in engineering failure analysis. A systematic fractographic examination follows the fracture surface from the macroscopic crack origin identified under low magnification to the microscopic fracture mode at higher magnification. In a well-conducted failure analysis: SE imaging establishes fracture mode and crack propagation direction; BSE imaging identifies inclusions or second-phase particles at the fracture origin; EDS spot analysis confirms inclusion chemistry (sulfide, oxide, silicate); EBSD phase mapping identifies any unexpected phase (e.g., sigma phase in stainless steel, untempered martensite in a weld HAZ). The combination provides an unambiguous failure mechanism diagnosis that correlates with material specification, processing records, and service conditions. For corrosion-related failures, EDS mapping of oxide and salt deposit chemistry at the initiation site is essential.
Weld HAZ Characterisation
The weld heat-affected zone presents steep microstructural gradients over distances of hundreds of micrometres — from coarse-grain martensite or bainite immediately adjacent to the fusion line through fine-grain, intercritical, and subcritical zones. EBSD mapping across the HAZ in a single scan resolves grain size gradients, phase distribution (residual austenite, martensite fraction), misorientation gradient (KAM as proxy for residual stress/strain), and grain boundary character. This level of spatially-resolved information is inaccessible to optical microscopy or hardness traverses alone. EBSD-derived grain size and texture data feed directly into fracture mechanics models for critical assessment of weld quality. The HAZ microstructure article provides the metallurgical context for interpreting SEM/EBSD data from weld cross-sections.
Quality Control in Advanced High-Strength Steels (AHSS)
Third-generation AHSS (medium-Mn TRIP/TWIP steels, press-hardened PHS grades, quench-and-partition Q&P steels) rely on precise retained austenite fractions and morphologies for their mechanical performance. EBSD phase mapping provides austenite fraction, austenite grain size, and austenite stability (via KAM distributions) simultaneously, enabling rapid feedback on heat treatment optimisation. EDS mapping confirms Mn and Al partitioning between austenite and ferrite/martensite, which governs austenite stability. See the bainite microstructure guide for the role of bainitic ferrite in AHSS microstructures.
Chemical Etchants for Steel Metallography: SEM and Optical Reference
| Etchant | Composition | Target Microstructure | Notes |
|---|---|---|---|
| Nital 2% | 2 ml HNO₂ in 98 ml ethanol | General steel structure; grain boundaries; ferrite/pearlite/martensite/bainite | Most widely used steel etchant; 5–30 s immersion; rinse in ethanol immediately |
| Picral 4% | 4 g picric acid in 100 ml ethanol | Cementite outlines; pearlite lamellae; distinguishes bainite from pearlite more clearly than nital | Does not reveal ferrite grain boundaries; often used after nital for bainite/pearlite distinction |
| Klemm's reagent | 50 ml saturated sodium thiosulfate + 1 g potassium metabisulfite | Tint etch: martensite (brown/yellow), bainite (blue), ferrite (white) | Colour etching; requires optical microscopy; not applicable in SEM BSE |
| Béchet–Beaujard | Aqueous picric acid + detergent (e.g., Teepol) | Prior austenite grain boundaries in hardened steel | Requires clean, flat surface; 5–60 min immersion; reveals PAGB decoration by carbide films |
| Kallings No. 2 | 5 g CuCl₂ + 40 ml HCl + 25 ml ethanol + 30 ml H₂O | Stainless steels: delta ferrite, austenite, sigma phase; nickel alloys | Electrolytic or immersion; reveals σ phase as dark particles at grain boundaries |
| Vilella's reagent | 1 g picric acid + 5 ml HCl + 100 ml ethanol | Stainless steel weld microstructures; distinguishes δ-ferrite from austenite | Colours ferrite selectively; reveals weld solidification structure and HAZ grain coarsening |
Table 4 — Common chemical etchants for steel and stainless steel metallography. Etching is used for SE/BSE imaging enhancement; EBSD specimens must not be etched.
Frequently Asked Questions
What is the resolution advantage of SEM over optical microscopy?
What is the difference between secondary electrons and backscattered electrons in SEM?
What EDS accuracy can be expected in SEM compositional analysis?
What does EBSD measure and what is a Kikuchi pattern?
What is the difference between SEM and TEM?
How does SEM help identify hydrogen embrittlement failure?
Why must SEM specimens be tilted to 70° for EBSD analysis?
What specimen preparation is required for high-quality EBSD?
What is kernel average misorientation (KAM) and how is it used in metallurgy?
Recommended References
Disclosure: MetallurgyZone participates in the Amazon Associates programme. If you purchase through these links, we may earn a small commission at no extra cost to you. This helps support free technical content on this site.
Further Reading
