Energy-dispersive X-ray spectroscopy (EDS or EDX) is an analytical technique that determines the elemental composition of a material by measuring the energies and intensities of the X-rays it emits following excitation. It is widely used with electron microscopes to connect microscopic features with chemical composition. Each chemical element produces a characteristic set of X-ray lines, allowing elements to be identified and, with suitable corrections, their concentrations estimated. EDS generally provides elemental rather than molecular information: it does not by itself establish crystal structure, identify chemical bonds, or uniquely determine a compound.
Physical principles
Characteristic X-ray emission begins when an incident particle or photon removes an inner-shell electron from an atom. An electron from a higher-energy shell can then fill the vacancy. The energy difference may be emitted as an X-ray photon:
Here, the energies refer to the initial and final electronic states of the atom. Because inner-shell binding energies depend strongly on atomic number, the emitted photon energies provide a fingerprint of the element.
Transitions ending in the K shell produce K-series lines; those ending in the L and M shells produce L- and M-series lines. Labels such as Kα and Kβ distinguish groups of transitions involving different starting subshells or shells. An element may therefore contribute several peaks to a spectrum.
Not every inner-shell vacancy produces an X-ray. A competing process, Auger emission, transfers the available energy to another electron, which escapes from the atom. The probability of X-ray emission, called the fluorescence yield, depends on the element and the shell involved.
In electron-excited EDS, the incident beam must have sufficient energy to ionize the shell of interest. Electrons also generate a continuous X-ray background through bremsstrahlung, or braking radiation, as they are deflected and decelerated in the sample. The measured spectrum consequently contains characteristic peaks superimposed on a continuum.
Excitation methods and related techniques
EDS describes how X-rays are measured, rather than a single way of exciting the sample.
Electron-beam excitation
In a scanning electron microscope (SEM), a focused electron beam scans a specimen, and an EDS detector collects X-rays from the irradiated region. This arrangement supports point analysis, line scans, and elemental maps.
EDS can also be integrated into a transmission electron microscope (TEM), including scanning transmission operation. Thin specimens reduce the volume in which electrons interact with matter, allowing chemical analysis of much smaller features than is normally possible in bulk specimens.
X-ray excitation
An incident X-ray beam can also produce characteristic emission through X-ray fluorescence (XRF). When the emitted photons are measured directly by their energies, the method is energy-dispersive X-ray fluorescence, or EDXRF. Laboratory, portable, and synchrotron-based instruments use this approach.
Electron-excited EDS and EDXRF share detection principles, but their excitation physics, sampling volumes, backgrounds, and quantitative corrections differ. Performance figures for one configuration should not be assumed to apply to the other.
Energy dispersion versus wavelength dispersion
In energy-dispersive detection, individual photons are sorted electronically according to their measured energies. In wavelength-dispersive X-ray spectroscopy (WDS), an analyzing crystal separates X-rays according to wavelength through Bragg’s law.
EDS records a broad energy range simultaneously and is well suited to rapid surveys and mapping. WDS generally offers finer spectral resolution and can resolve overlaps that are difficult for conventional EDS. The two techniques are often complementary.
Detectors and instrumentation
Most contemporary electron-microscope EDS systems use a silicon drift detector (SDD). Earlier systems commonly used lithium-drifted silicon detectors, generally operated with liquid-nitrogen cooling. SDD systems typically use thermoelectric cooling and can process high count rates.
When an X-ray is absorbed in the detector’s active silicon, it creates electron–hole pairs. The collected charge is approximately proportional to the energy deposited:
where is the number of pairs and is the mean energy needed to create one pair, approximately in silicon under typical detector operating conditions. Electronic processing converts each charge pulse into an energy estimate and assigns it to a channel in the spectrum.
Important detector characteristics include:
- Energy resolution: the width of a measured spectral line, conventionally expressed as its full width at half maximum. Many conventional SDD systems achieve roughly 125–140 eV at the manganese Kα line near 5.9 keV under specified operating conditions.
- Collection solid angle: the fraction of emitted X-rays intercepted by the detector, governed by detector area, distance, and geometry.
- Detection efficiency: the probability that an incoming photon produces a usable signal. Detector windows, surface layers, and active thickness make this strongly energy-dependent.
- Count-rate capability: the rate at which events can be processed without unacceptable losses or distortion.
- Dead time: the time during which electronics are unavailable to process additional events.
Longer pulse-processing times can improve energy resolution but usually reduce count-rate capability. Detector specifications therefore depend on the conditions under which they are measured.
Spectra, identification, and artifacts
An EDS spectrum plots counts against photon energy. Element identification compares measured peak positions and patterns with known characteristic lines. Reliable assignments consider multiple lines where available, rather than relying on one apparent peak.
Several effects complicate interpretation:
Peak overlap. Lines from different elements can lie closer together than the detector can resolve. Sulfur Kα and molybdenum Lα are a familiar example. Software fitting can separate some overlapping contributions, but an apparently good fit does not guarantee a unique identification.
Escape peaks. A silicon characteristic X-ray generated inside the detector may escape, leaving less energy in the detector than the incident photon carried. The recorded event then appears approximately 1.74 keV below the original photon energy.
Pulse pile-up. Photons arriving too close together may be recorded as a single event with approximately their combined energy. Such events can create false peaks or distort existing ones.
Background and absorption. Weak peaks may be obscured by the continuum, while low-energy photons may be absorbed in the specimen, detector window, surface contamination, or intervening material.
Extraneous emission. A coating, specimen holder, mounting medium, support film, or nearby instrument component may contribute X-rays. A detected element is not necessarily located in the particular feature visible in the image.
Energy calibration establishes the relationship between channel number and photon energy. Quantitative analysis additionally requires checks on detector response, spectral processing, and measurement stability.
Quantitative analysis
Peak intensity is not simply proportional to elemental concentration in all specimens. The measured signal depends on ionization probability, electron transport, X-ray absorption, fluorescence, detector efficiency, and specimen geometry.
For bulk electron-excited analysis, a common approach compares the background-subtracted intensity from a specimen with that from a reference standard measured under comparable conditions:
The resulting intensity ratio is converted into composition using matrix corrections. Common approaches include:
- ZAF corrections, accounting for atomic-number effects, absorption, and secondary fluorescence.
- methods, describing X-ray generation as a function of mass depth and correcting for its subsequent modification within the specimen.
Standardless methods use modeled intensities and stored sensitivity information instead of directly measured standards for each analysis. They are useful for rapid compositional estimates, but their accuracy depends on the validity of the underlying assumptions. Normalizing reported concentrations to 100% can conceal missing elements or systematic errors.
Thin-specimen TEM analysis often uses different procedures. The Cliff–Lorimer method relates elemental concentration ratios to X-ray intensity ratios through sensitivity factors, provided that absorption and other relevant effects are sufficiently small or separately corrected.
Results may be reported as mass fractions or atomic fractions; these are not interchangeable. Oxygen or other elements inferred from assumed stoichiometry should also be distinguished from elements measured directly.
Counting statistics contribute to measurement uncertainty. For an ideal independent count with total , the standard deviation is approximately . Weak-peak measurements incur additional uncertainty from background subtraction, peak overlap, and corrections. Longer acquisition improves counting precision but does not remove systematic errors.
Spatial resolution and elemental mapping
The spatial resolution of EDS is determined by the region producing the detected X-rays, not merely by the electron-beam diameter.
In bulk specimens, electrons scatter and lose energy beneath the surface, producing an interaction volume whose dimensions depend on beam energy, composition, and density. At conventional SEM operating voltages, the X-ray generation region may extend from hundreds of nanometres to several micrometres, depending strongly on the material and conditions. A nanometre-scale electron probe therefore does not imply nanometre-scale EDS resolution.
Lower beam energies can reduce the interaction volume, but they also restrict which shells can be excited and may force analysis of lower-energy lines with stronger absorption or more difficult overlaps. Thin TEM specimens can support nanometre-scale or, under specialized conditions, atomic-scale analysis; beam spreading, specimen thickness, channeling, signal strength, and damage remain important.
An elemental map associates an X-ray signal with each image pixel. Modern spectrum imaging stores an entire spectrum at each position, enabling later background subtraction and peak fitting. A map formed simply by summing counts within an energy window can contain contributions from overlapping lines and continuum background.
Map brightness is not automatically a direct measure of concentration. Topography, thickness, absorption, and acquisition settings can all alter contrast.
Capabilities and limitations
EDS is valued for rapid, multielement analysis with little additional specimen preparation beyond that needed for the host instrument. It can investigate major and minor constituents, locate inclusions, and relate composition to microstructure.
Its main limitations include:
- Very light elements: conventional EDS does not detect hydrogen or helium. Detection of lithium requires specialized instrumentation; other light elements are strongly affected by window transmission, absorption, and contamination.
- Trace sensitivity: detection limits are element- and matrix-dependent. Favorable conventional SEM measurements may reach approximately 0.1 mass percent, but this is not a universal specification; overlap, short acquisition, or unfavorable geometry can make limits substantially worse.
- Limited chemical-state information: conventional EDS usually cannot determine oxidation state or distinguish compounds with the same elemental composition. Small chemical shifts are generally below its resolving power.
- Specimen assumptions: rough, porous, layered, or heterogeneous materials may violate the assumptions used in bulk quantification.
- Beam effects: charging, contamination, heating, material loss, or irradiation-induced change can alter the specimen during acquisition.
EDS does not determine crystal structure. Materials with identical elemental composition but different structures require complementary methods such as X-ray diffraction or electron diffraction.
Development and applications
Energy-dispersive analysis expanded with the development of semiconductor X-ray detectors and electronic pulse-height analysis in the second half of the twentieth century. Its integration with electron microscopy made elemental microanalysis widely accessible. Silicon drift detectors subsequently increased practical collection rates and supported faster spectrum imaging and multichannel detector arrangements.
In materials science and manufacturing, EDS is used to examine alloys, ceramics, electronic devices, coatings, inclusions, corrosion products, and failure surfaces. In geology, it supports mineral identification and the study of compositional variation within rocks. It is also used in particle analysis, forensic examination, archaeology, and cultural heritage research, where elemental information can help characterize pigments, glass, metals, and alteration products.
In biological and other beam-sensitive specimens, preparation and irradiation can redistribute or remove constituents. Consequently, interpretation depends not only on the spectrum but also on specimen history, acquisition conditions, and complementary structural or chemical evidence.