SEM/EDS Sample Preparation Guide

A scanning electron microscope (SEM) scans the sample under vacuum with a focused electron beam. The sample must therefore satisfy four basic conditions: it must be vacuum-compatible (free of water, solvent or volatile components), conductive enough to dissipate the charge deposited by the beam, mechanically secure on its holder, and unable to contaminate the chamber or the detectors. Failure on any of these points degrades image quality and the accuracy of EDS data immediately.

Variable-pressure (VP) SEM neutralises charge build-up by admitting a low pressure of gas into the chamber, which allows polymers, ceramics and biological specimens to be examined without a coating. The trade-off is a modest loss of resolution and EDS count rate, so for high-magnification imaging or precise elemental analysis a conductive coating remains the most dependable route. Which path to take should be decided according to the goal of the analysis, before the sample is submitted.

Solid pieces are mounted on aluminium stubs with double-sided conductive carbon tape or with silver or carbon paint. Tape is fast and clean, but on thick or irregular pieces poor contact leads to charging; a thin bridge of conductive paint from the edge of the sample to the stub resolves this. The sample must not be placed under vacuum until the solvent of the paint has fully evaporated. For porous or fragile pieces, embedding in epoxy followed by polishing gives both a flat surface and a secure mount.

Powders are the most frequently mis-prepared category. A thick layer heaped onto carbon tape produces overlapping particles and drifting clusters in the image, while in EDS the carbon of the tape contaminates the signal. The correct approach is to sprinkle the powder very sparsely and then remove loose particles with compressed air or a gentle tap. For nanopowders, ultrasonic dispersion in ethanol or isopropanol followed by drying a single drop on a silicon wafer or polished aluminium makes it possible to image individual particles.

Non-conductive or poorly conductive samples examined in high-vacuum mode are coated by sputtering with a metal or carbon film a few nanometres thick. Gold and gold-palladium give fine-grained films with high secondary-electron yield and are the standard choice for morphology imaging. Typical thicknesses lie between 5 and 15 nm; thicker films bury fine surface detail, and at high magnification the grain structure of the coating itself becomes visible.

If EDS analysis is planned, the choice changes. The gold L and M lines overlap with several elements, notably phosphorus, sulphur and zirconium, and distort quantitative results. Carbon coating is preferred in that case: its signal stays in the low-energy region of the spectrum and does not interfere with metallic elements. Carbon coating is, however, unsuitable when the carbon content of the sample itself must be measured; variable-pressure mode or a very thin chromium or iridium film is then considered.

The sample must be completely dry before coating, and after coating there must be a conductive path from the sample to the stub. Coated samples should be stored in closed boxes away from humidity. If the same sample will later be examined by other methods such as Raman or XRD, the possible influence of the coating on those analyses should be assessed in advance.

Reliable quantitative analysis by energy-dispersive X-ray spectroscopy (EDS) requires a flat, polished surface. On rough surfaces X-rays are absorbed by protrusions of the sample itself before reaching the detector, and light elements are systematically under-reported. Metallographic sectioning, grinding and polishing with diamond suspension is the standard route; polishing residues such as silica or alumina must then be removed thoroughly, otherwise silicon and aluminium peaks that do not belong to the sample appear in the spectrum.

The electron beam generates an "interaction volume" below the surface that extends from roughly one to several micrometres depending on accelerating voltage and material density. At 20 kV in a light matrix this volume reaches several micrometres, so the EDS spectrum of a coating thinner than one micrometre inevitably contains signal from the substrate. For thin films and small particles, lowering the accelerating voltage to 5–10 kV shrinks the analysed volume, at the cost of limiting the excitation of high-energy K lines.

On irregular fracture surfaces or powders, EDS should be regarded as semi-quantitative only. For such samples, elemental maps and point analyses are best used to show relative differences between regions rather than absolute percentages, and the report should state this limitation explicitly. Where a true composition is required, a polished cross-section or a pressed and polished powder pellet is the appropriate route.

Charging shows up as bright patches, streaks and image drift; it is usually caused by an inadequate coating, a broken conductive path between sample and stub or an accelerating voltage that is too high. Outgassing manifests as a vacuum that fails to reach its target and an image that deteriorates over time; moist samples, incompletely cured epoxy and greasy surfaces are the main sources. Drying samples overnight in a vacuum desiccator or a low-temperature oven before analysis prevents most of these problems.

Contamination is the third major issue. Surfaces touched with bare hands show sodium, chlorine and carbon peaks, poorly cleaned polished surfaces show silicon and aluminium, and regions left under the beam for long periods accumulate carbon. Samples should always be handled with clean gloves and tweezers and transported in closed containers, and overview images and EDS measurements should be completed before any area is scanned at high magnification for an extended time.

A good SEM/EDS report starts with a good request: the origin of the sample, its expected composition, whether it is conductive, any moisture or volatile content and the actual question to be answered (morphology, thickness, composition, cause of failure). These details determine the preparation route and avoid unnecessary coating or examination in the wrong mode. NAKTEM plans sample preparation together with the analysis and documents the coating, accelerating voltage and working distance used in the report.