EBSD · SEM imaging

SEM imaging basics: where the SE and BSE signals come from

Start with the first picture: drop the beam voltage from 20 kV to 5 kV and watch the cloud of electron paths shrink.

How deep does the electron beam go into the sample?

Side view (cross-section). Each line is one simulated electron. Red: leaves the top surface again (backscattered electron, BSE). Blue: stops inside.

Sample material

Beam voltage20 kV

Electrons simulated600

New random run

Electron range (Kanaya-Okayama)-

X-ray range-

Typical BSE depth (median)-

SE escape depthabout 5 nm

BSE fraction, this run-

BSE fraction, Reuter fit-

Elastic mean free path at entry-

Paths drawn-

Try it: keep 20 kV and switch iron to carbon. The range grows from about 1.6 µm to about 4.4 µm and far fewer paths turn red.

What is the difference between SE and BSE images?

Top view, 48 × 24 µm. Matrix is aluminium. Top left: flat patch of a second phase. Top right: raised particle (aluminium). Bottom left: 45° slope, flat top, then a vertical edge. Bottom right: a hole 4 µm deep. The ETD sits to the right.

Detector

Second phase

Beam voltage10 kV

Working distance10 mm

Show feature names

BSE coefficient η, Al-

η, second phase-

Phase contrast, this detector-

Bright edge band width-

Signal collected (model)-

Noise on flat matrix-

Try it: with copper, switch from SE (ETD) to BSE. The copper patch goes from about 29% brighter to about 49% brighter, and the particle loses most of its bright rim.

What limits SEM resolution and image noise?

128 × 128 pixel patch of the scan. Disks: 5, 10, 20, 40, 80, 120 nm wide, each brighter than the background by the contrast C. A disk wider than its cell fills the cell.

Probe size vs current for this source and voltage. The dot is your setting. The grey line is the pixel size.

Electron source

Beam voltage15 kV

Probe current-

Dwell time per pixel-

Pixels across

Field of view-

Feature contrast C-

Collected signal per beam electron ε-

Probe diameter-

Pixel size-

Signal counts per pixel-

SNR = C√N (Rose needs 5)-

Current needed for this C-

Frame time-

Try it: at 1 nA and 10 µs the disks pass the Rose test (SNR about 8). Drop the current to 100 pA: SNR falls to about 2.5 and the disks sink into the noise.

Why does an insulator charge in the SEM, and which voltage stops it?

Electrons out per electron in, vs beam energy. Above the line y = 1 more electrons leave than arrive (surface goes slightly positive). Below it, charge builds up negative.

Insulator

Beam voltage-

Grounded carbon coat (10 nm)

SE yield δ-

BSE yield η-

Total yield σ = δ + η-

Surface charge-

E2, pulsed beam-

E2C, steady beam-

Try it: pick PMMA and slide from 10 kV down. The surface turns from negative to positive at 1.8 keV, where the curve crosses 1.

Which detector or attachment do I need?

Signal colours

secondary electrons
backscattered (BSE)
X-rays
light (CL)
transmitted electrons
ions
current (EBIC)
gas, radicals
chosen part

Not to scale. The side section is one cut through both columns; parts off this cut appear only in the top view and the close-ups. Tap any part. On a phone, swipe the side section and the top view sideways. Grey parts are not chosen.

What do you want to know?

Attachment

Show names on the drawing

Try it: pick "Crystal orientation and grains". EBSD gets badge 1 and the sample in the side section tilts to 70°; TKD gets badge 2. Then tap the EDS port in the top view: a purple X-ray arrow runs to it and the panel gives the 35° take-off angle.

All SEM attachments at a glance. Numbers are typical for a field emission SEM; your instrument will differ.
Attachment	Signal	What it measures	Resolution	Sample needs	Use it when
ETD (SE)	SE, some BSE	Surface shape, edges	1 to 3 nm (FEG, 15 to 30 kV)	Conductive or coated; any shape	General imaging, rough samples, long working distance
In-lens SE	SE1 and SE2	Fine surface detail, coatings	About 1 nm at 15 kV, 1 to 2 nm at 1 kV	Small, flat-ish, short working distance	Low kV, high resolution, nanoparticles
BSE (annular)	BSE	Mean atomic number, phases, channelling	10 to 100 nm (grows with kV)	Flat and polished for Z contrast	Which phase is where; ECCI
AsB / ESB	Energy or angle filtered BSE	Z contrast at low kV; grain contrast	A few nm at 1 to 3 kV	Short working distance, low tilt	Nanoscale phases, thin coatings, beam sensitive samples
EDS (SDD)	X-rays	Elements Z 5 and up; amounts to about 1 to 5% relative	About 1 µm at 20 kV in steel	Flat, polished, conductive for quant	Quick composition, maps, spots
WDS	X-rays, one wavelength at a time	Light elements, trace to about 100 ppm, overlaps	Same volume as EDS	Flat, polished, exact working distance	Overlapping peaks, B C N O, trace
EBSD	BSE Kikuchi pattern	Orientation, phase, grains, texture, strain	20 to 50 nm in metals at 20 kV; depth 10 to 40 nm	Flat, polished, no damage layer; 70° tilt	Grains, boundaries, phases, texture
TKD	Transmitted Kikuchi pattern	Orientation in nanograins	About 10 nm	Thin foil (about 50 to 150 nm)	Grains under 100 nm, heavy deformation
CL	Light (UV to IR)	Band gap, defects, dopants, zoning	0.1 to several µm	Must emit light; semiconductors, minerals	Defects in semiconductors, growth zoning
STEM-in-SEM	Transmitted electrons	Internal structure of a thin foil	About 1 nm at 30 kV	Thin foil on a TEM grid	Nanoparticles, lamella checks
EBIC	Current through a junction	Electrically active defects, junction depth, diffusion length	0.1 to a few µm	Semiconductor with a junction and two contacts	Solar cells, dark dislocations, device junctions
Plasma cleaner	None (a cleaner)	Removes hydrocarbon contamination	Whole chamber	Not for easily oxidised surfaces	Black boxes appear, low kV work, carbon EDS
FIB column	Ions in; SE and sputtered atoms out	Cross-sections, lamellae, 3D slices	About 5 nm imaging; damage layer about 20 nm at 30 kV in Si	Tilt to 52° so ions hit square on	Site specific cross-section, TEM lamella
FIB-SIMS	Sputtered ions, sorted by mass	Isotopes, trace, Li, H, dopants	About 50 nm lateral; tens of nm depth	Destructive; any solid	ppm sensitivity, isotopes, light elements
GIS	Gas in	Deposits Pt, W, C; etches with XeF2	Deposit line about 100 nm wide	Needle 100 to 200 µm above the surface	Protective cap before milling, welds, marks
Micromanipulator	None (a needle)	Lifts out and moves lamellae; probes	Steps of tens of nm	Needs GIS to weld	In-situ lift-out, electrical probing
In-situ stage	None (a stage)	Heating, tension, compression, cooling	Set by the detector you use	Small, must fit under the pole piece	Watching grains, cracks, phase changes live

What should I remember about SEM signals?

Voltage sets depth.The electron range grows about as E^1.67. At 20 kV it is about 1.6 µm in iron but 4.4 µm in carbon. Heavy, dense samples keep the beam near the surface.

SE = shape.Secondary electrons (SE) have a few eV and escape only from the top few nm. Tilted faces and edges give more of them, so SE images show topography.

BSE = atomic number.Backscattered electrons (BSE) keep most of the beam energy. Their fraction η rises from about 0.06 for carbon to about 0.49 for gold, so heavier phases look brighter.

Noise sets the limit.To see a contrast C you need about (5/C)² collected electrons per pixel (Rose criterion). More current means a bigger probe, so small features need a bright source.

More detail: equations, assumptions and limits of the models

Electron paths (widget 1). Single-scattering Monte Carlo as in Joy (1995). Elastic scattering: screened Rutherford cross-section σ = 5.21e-21 (Z/E)² · 4π/[α(1+α)] · [(E+511)/(E+1024)]² cm² with screening α = 3.4e-3 Z^0.67 / E (E in keV). Step length is drawn from the mean free path λ = A/(N_A ρ σ). Angle: cos θ = 1 - 2αR/(1+α-R), R random. Energy loss between steps: Joy-Luo stopping power dE/ds = -78500 (ρZ/AE) ln(1.166(E + kJ)/J) keV/cm, with J = (9.76Z + 58.5 Z^-0.19) eV and k = 0.731 + 0.0688 log10 Z. Paths stop at 0.5 keV or at 5% of the beam energy, whichever is higher. The "Reuter fit" readout is the measured reference. At 20 kV the simulated BSE fraction lands close to it for light and mid elements (carbon about 0.06, aluminium about 0.16, iron about 0.30, against 0.06, 0.15 and 0.28 measured) but runs high for gold (about 0.55 against 0.49). At 5 kV every element runs a few hundredths high (gold about 0.58). Screened Rutherford scattering is known to overestimate large-angle scattering for heavy atoms at these energies; full Mott cross-sections fix this, and the page does not use them. The mean free path and the depths shown come from the same cross-section, so for gold they carry the same error.
Range. Kanaya-Okayama: R (µm) = 0.0276 A E^1.67 / (Z^0.889 ρ), E in keV, ρ in g/cm³. X-ray range uses E^1.67 - Ec^1.67, with Ec the edge energy of the line shown (K line if the beam energy is at least 1.5 × Ec, else L or M).
SE escape depth. About 5 λ_SE with mean escape depth λ_SE about 1 nm for metals (Seiler 1983). It is several times larger in insulators. The mark is not to scale except at low voltage.
BSE coefficient. Reuter (1972): η = -0.0254 + 0.016 Z - 1.86e-4 Z² + 8.3e-7 Z³, for normal incidence, and roughly constant from 5 to 30 kV. Below about 5 kV the spread between light and heavy elements shrinks; widget 2 does not model that. Tilt: η(θ) = (1 + cos θ)^(-9/√Z) (Arnal et al. 1969), scaled to match Reuter at θ = 0.
SE image (widget 2). SE yield δ ∝ 1/cos θ (capped at 6) times (1 + βη) with β = 2.5 for SE made by leaving BSE (SE2; Reimer gives values of about 2 to 3). The ETD also collects SE made where BSE hit the chamber (SE3, taken as 1.5η), some direct BSE, and is shadowed on faces turned away from it. The in-lens detector gets SE1 and SE2 only and loses signal fast as the working distance grows. Edge brightness: height minus its average over a disk 0.4 × the range wide. Collection efficiencies are simple made-up curves chosen for the right trend, not measured values. BSE collection: annular detector with 2 mm inner and 10 mm outer radius, cosine emission.
Probe (widget 3). d² = d_G² + d_s² + d_c² + d_d², with d_G = (2/π)√(I/β)/α (from the brightness β), d_s = 0.5 C_s α³, d_c = C_c α ΔE/E, d_d = 1.22 λ/α. The page picks the α that gives the smallest d. C_s = C_c = 2 mm. Brightness at 20 kV: tungsten 1e5, LaB6 1e6, Schottky 1e8 A cm⁻² sr⁻¹, scaled in proportion to voltage. Energy spread: 1.5, 1.0 and 0.5 eV. These are typical textbook values; real columns differ.
Noise. Signal counts per pixel N = ε I τ / e. Shot noise is √N. Rose: a feature of contrast C is seen when C√N ≥ 5, so the threshold current is I = 25 e / (ε C² τ). Detector noise and the scan generator are ignored.
Drawings (widget 5). Not to scale. Near the sample, sizes are enlarged: the working distance is drawn about 10 times too big so the parts can be read. Port angles in the top view are typical; makers differ, so check your own chamber. In-column detector names differ by maker: Zeiss calls them InLens (SE) and EsB (BSE); Thermo uses TLD and ICD, and the names change by model. The FIB sits at 52° (Thermo), 54° (Zeiss) or 55° (Tescan) from the electron column. WDS take-off is 40° on microprobes and often about 35° on SEM-mounted units. EDS take-off is drawn at 35°.
Charging (widget 4). Universal SE yield curve δ/δm = 1.28 (E/Em)^-0.67 [1 - exp(-1.614 (E/Em)^1.67)] (Lin and Joy 2005) with Em = 0.4 keV for all four materials. δm is set so that δ + η = 1 at the pulsed-beam E2 measured by Rau et al. (2008). η is Reuter's fit at the mass-averaged Z. Under a steady beam the real balance point E2C is lower (Rau et al.), shown as a second mark.

Questions people ask about SEM imaging

What is the difference between secondary and backscattered electrons?

Secondary electrons (SE) are knocked out of the sample atoms and have less than 50 eV, so only those made in the top few nm escape. Backscattered electrons (BSE) are beam electrons that bounce back out, keeping most of their energy. SE mostly show surface shape; BSE mostly show average atomic number.

What accelerating voltage should I use in the SEM?

Use low voltage (1 to 5 kV) for surface detail and for insulators, because the beam stays shallow. Use 15 to 20 kV for BSE atomic number contrast, EBSD and EDS, where you need signal and enough energy to excite X-ray lines. The electron range in widget 1 tells you how deep each choice looks.

Why are edges bright in SEM images?

Near an edge the electrons spreading inside the sample reach a second free surface, so more secondary electrons escape. Tilted faces also give more SE, about as 1/cos of the tilt. The bright band is about as wide as the region the spreading electrons reach, so it grows with voltage.

What is the difference between an in-lens and an Everhart-Thornley detector?

The Everhart-Thornley detector (ETD) sits at the side of the chamber. It collects SE from all around, including SE made by BSE hitting the chamber walls, and it gives shadows on faces turned away from it. An in-lens detector sits inside the column. It sees mostly SE from the beam spot, so images are sharper and more surface sensitive, but it needs a short working distance.

How do I stop an insulating sample from charging?

Coat it with a thin grounded layer (carbon or metal), or image near the E2 voltage where as many electrons leave as arrive, or use a low-vacuum mode. Lower current and faster scanning also help. For many ceramics and polymers E2 under a steady beam is only 1 to 3 kV.

Why is my SEM image so noisy at high magnification?

Each pixel gets only the electrons that land on it during the dwell time. To see a contrast of 10% you need about 2500 collected electrons per pixel. At high magnification people often lower the current to shrink the probe, which lowers the counts. Fix it with longer dwell, frame averaging, or a brighter source.

What sets the resolution of an SEM?

The probe diameter, the pixel size, and for SE2 and BSE images the spread of electrons in the sample. The probe diameter grows with current and depends on source brightness and lens errors. Signal-to-noise then decides whether a small, low-contrast feature can actually be seen.

EDS or WDS: which should I use?

EDS first. It gives the whole spectrum in seconds and is fine for most elements above about 0.1 wt%. Switch to WDS when peaks overlap (for example S K with Mo L or Pb M), when you need numbers for boron, carbon, nitrogen or oxygen, or when you need trace levels near 100 ppm. WDS is slow, needs a polished sample at the exact working distance, and needs tens of nanoamps.

EBSD or ECCI for dislocations?

ECCI. It uses the BSE detector with the grain tilted to a Bragg condition and shows single dislocations in a bulk sample, down to about 100 nm below the surface. EBSD does not see single dislocations; its KAM and GND maps give a density of geometrically necessary dislocations from small orientation changes, and the value depends on the step size. Use EBSD to find the grain and its orientation, then ECCI to see the dislocations in it.

What does a plasma cleaner remove?

Hydrocarbon contamination: the thin oily film from pump oil, fingerprints, tape and air that the beam cross-links into dark rectangles. Oxygen radicals in the plasma turn it into CO, CO2 and water, which the pumps remove. It does not remove oxide layers, dust or scratches, and it can oxidise some metals and etch some polymers.

Is FIB-SIMS the same as SIMS?

No. FIB-SIMS is a mass spectrometer bolted onto a FIB-SEM. It sorts the ions the gallium beam knocks out, so you get isotopes, lithium, hydrogen and trace elements at parts per million with about 50 nm resolution, on the same spot you just imaged. A standalone SIMS (NanoSIMS, ToF-SIMS, magnetic sector SIMS) is its own instrument with better sensitivity and mass resolution. Both are destructive and both need standards for numbers.

Why does EBSD use a 70° tilt and 20 kV?

EBSD patterns come from BSE that leave the top tens of nm. Tilting to 70° makes more of them leave toward the detector, and 15 to 20 kV gives enough signal and sharp bands. The interaction volume at that voltage sets how fine a grain you can map, which is why TKD on thin foils is used for nanoscale grains.

Where does this lead next?

References

Show the 22 references

Goldstein, J.I., Newbury, D.E., Michael, J.R., Ritchie, N.W.M., Scott, J.H.J., Joy, D.C. Scanning Electron Microscopy and X-Ray Microanalysis, 4th ed. Springer, New York, 2018. Ch. 2 (backscattered electrons), 3 (secondary electrons), 4 (X-rays), 16 (EDS), 21 (trace analysis, EDS and WDS), 28 (cathodoluminescence), 29 (EBSD and ECCI), 30 (FIB). doi:10.1007/978-1-4939-6676-9
Reimer, L. Scanning Electron Microscopy: Physics of Image Formation and Microanalysis, 2nd ed. Springer Series in Optical Sciences 45, Springer, Berlin, 1998. doi:10.1007/978-3-540-38967-5
Joy, D.C., Luo, S. An empirical stopping power relationship for low-energy electrons. Scanning 11 (1989) 176-180. doi:10.1002/sca.4950110404
Joy, D.C. Monte Carlo Modeling for Electron Microscopy and Microanalysis. Oxford University Press, New York, 1995.
Kanaya, K., Okayama, S. Penetration and energy-loss theory of electrons in solid targets. J. Phys. D: Appl. Phys. 5 (1972) 43-58. doi:10.1088/0022-3727/5/1/308
Reuter, W. The ionization function and its application to the electron probe analysis of thin films. In: Shinoda, G., Kohra, K., Ichinokawa, T. (eds.), Proc. 6th Int. Conf. on X-ray Optics and Microanalysis, University of Tokyo Press, 1972, p. 121.
Arnal, F., Verdier, P., Vincensini, P.-D. Coefficient de rétrodiffusion dans le cas d'électrons monocinétiques arrivant sur la cible sous une incidence oblique. C. R. Acad. Sci. Paris 268 (1969) 1526.
Seiler, H. Secondary electron emission in the scanning electron microscope. J. Appl. Phys. 54 (1983) R1-R18. doi:10.1063/1.332840
Rose, A. Television pickup tubes and the problem of vision. Advances in Electronics 1 (1948) 131-166.
Lin, Y., Joy, D.C. A new examination of secondary electron yield data. Surf. Interface Anal. 37 (2005) 895-900. doi:10.1002/sia.2107
Rau, E.I., Fakhfakh, S., Andrianov, M.V., Evstafeva, E.N., Jbara, O., Rondot, S., Mouze, D. Second crossover energy of insulating materials using stationary electron beam under normal incidence. Nucl. Instrum. Methods Phys. Res. B 266 (2008) 719-729. doi:10.1016/j.nimb.2007.12.093
Schwartz, A.J., Kumar, M., Adams, B.L., Field, D.P. (eds.) Electron Backscatter Diffraction in Materials Science, 2nd ed. Springer, New York, 2009. doi:10.1007/978-0-387-88136-2
Humphreys, F.J. Characterisation of fine-scale microstructures by electron backscatter diffraction (EBSD). Scripta Mater. 51 (2004) 771-776. doi:10.1016/j.scriptamat.2004.05.016
Keller, R.R., Geiss, R.H. Transmission EBSD from 10 nm domains in a scanning electron microscope. J. Microsc. 245 (2012) 245-251. doi:10.1111/j.1365-2818.2011.03566.x
Zaefferer, S., Elhami, N.-N. Theory and application of electron channelling contrast imaging under controlled diffraction conditions. Acta Mater. 75 (2014) 20-50. doi:10.1016/j.actamat.2014.04.018
Yacobi, B.G., Holt, D.B. Cathodoluminescence Microscopy of Inorganic Solids. Plenum Press, New York, 1990. doi:10.1007/978-1-4757-9595-0
Edwards, P.R., Martin, R.W. Cathodoluminescence nano-characterization of semiconductors. Semicond. Sci. Technol. 26 (2011) 064005. doi:10.1088/0268-1242/26/6/064005
Leamy, H.J. Charge collection scanning electron microscopy. J. Appl. Phys. 53 (1982) R51-R80. doi:10.1063/1.331667
Vickerman, J.C., Briggs, D. (eds.) ToF-SIMS: Materials Analysis by Mass Spectrometry, 2nd ed. IM Publications and SurfaceSpectra, Chichester, 2013.
Wirtz, T., Philipp, P., Audinot, J.-N., Dowsett, D., Eswara, S. High-resolution high-sensitivity elemental imaging by secondary ion mass spectrometry: from traditional 2D and 3D imaging to correlative microscopy. Nanotechnology 26 (2015) 434001. doi:10.1088/0957-4484/26/43/434001
Vladár, A.E., Postek, M.T., Vane, R. Active monitoring and control of electron-beam-induced contamination. Proc. SPIE 4344 (2001) 835-843. doi:10.1117/12.436794
Mayer, J., Giannuzzi, L.A., Kamino, T., Michael, J. TEM sample preparation and FIB-induced damage. MRS Bull. 32 (2007) 400-407. doi:10.1557/mrs2007.63

Cite this page: Tripathy, Manisha. “SEM Imaging Basics: Interaction Volume, SE vs BSE Contrast.” untethered atom, 2026, https://untetheredatom.com/ebsd/sem-imaging-basics.

BibTeX

@misc{tripathy2026semimaging,
  author = {Tripathy, Manisha},
  title  = {SEM Imaging Basics: Interaction Volume, SE vs BSE Contrast},
  year   = {2026},
  howpublished = {\url{https://untetheredatom.com/ebsd/sem-imaging-basics}},
  note   = {Interactive web tool}
}