Scanning Electron Microscopy (SEM) with EDXRF

A primary electron beam at typical SEM energies (5 to 30 keV) penetrates a few micrometres into the sample and scatters through a teardrop-shaped interaction volume whose depth depends on beam energy and sample density. Several distinct signals come out, and several different detectors read them.

Secondary electrons are the workhorse of topographic imaging. The primary beam ionises atoms in the top few nanometres and ejects loosely bound outer-shell electrons at energies below 50 eV. The escape depth is 5 to 10 nm. The Everhart-Thornley detector sits at a side angle, biased with a positive grid to attract the low-energy SEs. SE yield depends on the angle at which the beam meets the surface, so the image reflects topography directly, producing the strong three-dimensional appearance characteristic of SEM micrographs.

Back-scattered electrons are different. A fraction of the primary beam bounces elastically off the heavier nuclei and exits with energies close to the beam energy. BSE yield scales with atomic number Z roughly as Z^0.7. A lead region (Z = 82) gives a far brighter BSE signal than a carbon region (Z = 6). The detector is an annular ring under the pole piece. The image carries atomic-number contrast rather than topography, which makes BSE the first thing a GSR analyst switches on: scan the swab stub at low magnification in BSE and the high-Z Pb-Sb-Ba particles light up as bright dots against the dark carbon-tape background.

Characteristic X-rays are the third signal. When the primary beam ionises an atom and a higher-shell electron fills the vacancy, the energy difference is released as a characteristic X-ray photon. The energy is set by atomic number through Moseley's law, the same physics that drives benchtop XRF. The EDS detector reads the X-rays as a function of energy and produces an elemental spectrum from wherever the beam is parked. Nanometre-scale beam positioning combined with elemental analysis at the same spot is what SEM-EDS gives that no other instrument does.

The column generates and shapes the primary beam. At the top sits the electron gun. The cheapest is a tungsten thermionic filament at about 2700 K; cheap, durable for about 100 hours, and is what most state SFSL labs run. The next step up is a lanthanum hexaboride (LaB6) cathode, ten times the brightness for ten times the cost. The top tier is the field-emission gun (FEG), cold-cathode or Schottky, with brightness 1000 times higher than tungsten and a spot below 1 nm. FEG-SEMs sit at IISc Bangalore, the SAIF at IIT Bombay and IIT Madras, and the CDFD Hyderabad bio-imaging facility. CFSL routine SEMs are typically LaB6 or upgraded tungsten.

Below the gun, two or three condenser lenses and an objective form the focused spot. The lenses are electromagnetic, not glass. Between condenser and objective sit the scan coils, electromagnetic deflectors that rake the beam across the sample in a raster synchronised with the display. There is no "lens that changes magnification" on a SEM; increase the magnification and the scan coils rake a smaller patch.

The sample stage is a five-axis precision mechanical stage with X, Y, Z, tilt and rotation, typically eucentric so tilting does not throw the feature out of focus. The chamber accepts samples up to about 100 mm in diameter and 50 mm in height for a routine benchtop SEM.

The detectors live in or around the chamber. The Everhart-Thornley SE detector sits at a side angle, the annular BSE detector under the pole piece, the EDS detector at a side port. Wavelength-dispersive spectrometers (WDS) use a crystal monochromator for higher-resolution elemental analysis; resolution at the Mn K-alpha drops from about 125 eV on EDS to 5 to 20 eV on WDS, with detection limit dropping from 0.1 weight percent to 0.001 weight percent. CFSL labs typically run EDS only; IIT Bombay SAIF and IISc have WDS-capable microprobes.

The column and chamber run under vacuum at 10^-5 to 10^-7 torr, held by a turbomolecular pump. Electrons would scatter off air molecules over millimetre distances at atmospheric pressure, and the filament would oxidise in minutes. Environmental SEM (ESEM) variants run the chamber at higher pressure (up to a few torr of water vapour) with a differential pumping aperture, which permits imaging of hydrated and uncoated samples at the cost of resolution and EDS performance.

A SEM cannot image a sample that charges up under the beam. Conductive samples (metals, alloys, conductive carbon, semiconductors) go straight on a stub with double-sided carbon tape and into the chamber. A spent cartridge case, a bullet jacket fragment, a brass shrapnel piece, a tool-mark replica in conductive epoxy, all go in as-is. The CFSL Pune ballistics workflow for a recovered bullet is exactly this: mount, clean with compressed nitrogen, and image the surface striations at low magnification.

Non-conductive samples (polymers, fibres, hair, paint, biological tissue, glass, paper) build up a static charge that deflects beam electrons, distorts the image, and can damage the sample. The fix is sputter coating: a small benchtop chamber that deposits a 5 to 15 nm layer of gold, gold-palladium or carbon onto the surface from a sputter target. CFSL Hyderabad runs a Quorum SC7620 as standard prep for paint cross-sections and fibre samples.

The coating choice is set by what you plan to measure. Gold gives the cleanest SE image but adds an Au M-alpha line at 2.12 keV that obscures phosphorus and sulphur in EDS. Au-Pd is similar with finer grain. Carbon is the right choice when EDS quantitation matters: the C K-alpha at 0.28 keV sits well below most analytical lines. For a GSR particle search where Pb, Sb and Ba are the question, carbon coating is the CFSL Pune default.

Biological samples need an additional step. Wet tissue, skin, botanical fragments, and hair from sexual-assault evidence kits cannot go into vacuum directly because the water would boil out and destroy the structure. The standard prep is chemical fixation (2.5 percent glutaraldehyde in phosphate buffer), graded ethanol dehydration, critical-point drying (CPD) in liquid CO2, and sputter coating. AIIMS Delhi and the National Centre for Cell Science (NCCS) Pune run CPD as routine for forensic-pathology samples.

Gunshot residue particle analysis is the highest-value SEM-EDS application in forensic ballistics casework. The classical inorganic GSR signature is the simultaneous presence of lead, antimony and barium in a single spheroidal particle of 0.5 to 10 micrometre diameter, with the elemental composition coming from the priming compound (lead styphnate, antimony sulphide, barium nitrate) and the bullet core. The reason the morphology matters is that the same three elements can appear together in environmental contamination (battery-shop dust, ceramic glaze flakes, plumbing solder). What contamination does not produce is the spheroidal melt-droplet shape that comes from the few-thousand-Kelvin instantaneous muzzle temperature and the rapid quench in air.

Collection happens at the scene or the lock-up. A trained crime-scene officer presses a carbon-tape stub against the back of the suspect's dominant hand, the web between thumb and forefinger, the face, the hair around the temples, and sometimes the inside of the cuff. Each stub is sealed in a polypropylene container under chain of custody and transported to CFSL Pune or the regional FSL with a GSR-capable SEM.

The acquisition runs at 20 to 25 keV, which gives clean excitation of the Pb L-alpha at 10.55 keV, the Sb L-alpha at 3.60 keV (or K-alpha at 26.36 keV depending on protocol), and the Ba L-alpha at 4.47 keV. The microscope runs in BSE mode at low magnification to scan the stub for high-Z particles. Modern instruments run an automated GSR workflow under Oxford Instruments AZtec Feature or the older INCA Feature: the BSE image is segmented, every particle above a brightness threshold is identified, the beam parks on each in turn, and the EDS spectrum is recorded. The full automated run takes 6 to 12 hours and typically scans the stub overnight.

The morning after, a ballistics analyst reviews the hit list. Each candidate is classified by ASTM E1588. A particle containing Pb, Sb and Ba with the right spheroidal morphology and size range is "characteristic of GSR from a lead-styphnate-primed cartridge". A particle with a partial signature (Pb+Sb without Ba, or Ba alone with right morphology) is "consistent with" or "indicative of" GSR but not characteristic. A particle that fails the morphology criteria, even with the elemental signature, is rejected as environmental contamination.

Lead-free primers (Sintox and similar, introduced to reduce indoor-range lead exposure) use titanium, zinc, strontium and gadolinium. The CFSL Pune protocol reports both the classical Pb-Sb-Ba and the lead-free Ti-Zn-Sr-Gd panels; a positive call can come from either.

Paint comparison is the second large-volume SEM-EDS application. A chip from a hit-and-run scene is embedded in epoxy, microtomed perpendicular to the layer stack, mounted, carbon-coated, and imaged in BSE at low magnification to see the layer interfaces. BSE contrast distinguishes the higher-Z primer (with titanium dioxide, zinc, calcium) from the lower-Z organic top coat. An EDS line scan across the cross-section returns the elemental composition of each layer. The CFSL Hyderabad paint section runs the suspect-vehicle paint against the scene chip layer-by-layer; a match on layer count, thickness, composition and binder (by FTIR) is what the court accepts as a positive paint comparison.

Fibre identification uses SEM for cross-section morphology (synthetic fibres have characteristic extrusion shapes: trilobal, multi-lobal, hollow, dog-bone) and EDS for delustrant and dye-related elements. Polyester contains titanium from the TiO2 delustrant; acrylic contains copper or nickel from cationic dyes. CFSL Hyderabad runs fibre SEM-EDS for sexual-assault evidence and for hit-and-run debris on victim clothing.

Glass particle analysis combines SEM-EDS for major-element composition (silicon, calcium, sodium, magnesium, aluminium) with LA-ICP-MS for the trace-element fingerprint. SEM-EDS alone classifies glass into soda-lime container, soda-lime float, borosilicate laboratory or tempered automotive glass, and the fracture-surface morphology indicates impact direction. The CFSL Hyderabad glass section runs this routinely for vehicle window glass and bottle-fragment cases.

Explosive residue particles use SEM-EDS for the elemental signature of inorganic explosives and primers. Ammonium nitrate prills from ANFO show spherical morphology in BSE. Lead azide and lead styphnate primer residues show the high-Z Pb signature. RDX, HMX and PETN are organic and give weak EDS signals but their crystalline morphology is distinctive. CFSL Pune and the NSG ballistic-explosives laboratory at Manesar both run SEM-EDS for post-blast debris, paired with FTIR and XRD for molecular and crystalline-phase confirmation.

Questioned-document work uses SEM-EDS for ink particle composition (iron-gall inks contain Fe, Cu, S; inkjet pigment inks contain Ti, Cr, S) and for paper filler analysis. Counterfeit currency analysis at CFSL Chandigarh runs SEM-EDS on the security thread and the intaglio printing against genuine Reserve Bank of India reference samples. Counterfeit pharmaceutical work uses SEM-EDS for API particle distribution, excipient morphology and coating uniformity; a substandard tablet with a non-uniform API coating fails the SEM-EDS distribution check even when the chromatographic assay passes.

The long tail includes tool-mark topography (cut-wire striations, file marks on padlocks, screwdriver marks on forced doors), hallmarking verification (Au, Ag, Pt ratios), botanical identification, hair cuticle analysis and soil heavy-mineral grains. The FSL Sector 14 Madhuban runs SEM-EDS as a shared multi-discipline resource, and the IIT Madras and IIT Bombay SAIFs provide open-access SEM-EDS to investigators without in-house capacity.