What is EPMA

About EMPA (electron microprobe analysis)

Electron microprobe analysis (EMPA), also called electron probe microanalysis (EPMA), is an analytical technique that is used to establish the composition of small areas on specimens. EMPA is one of several particle-beam techniques. A beam of accelerated electrons is focused on the surface of a specimen using a series of electromagnetic lenses, and these energetic electrons produce characteristic X-rays within a small volume (typically 1 to 9 cubic microns) of the specimen. The characteristic X-rays are detected at particular wavelengths, and their intensities are measured to determine concentrations. All elements (except H, He, and Li) can be detected because each element has a specific set of X-rays that it emits. This analytical technique has a high spatial resolution and sensitivity, and individual analyses are reasonably short, requiring only a minute or two in most cases. Additionally, the electron microprobe can function like a scanning electron microscope (SEM) and obtain highly magnified secondary- and backscattered-electron images of a sample.

EDS (Energy-dispersive spectrometry)

Our electron microprobe is outfitted with a new SDD energy-dispersive spectrometry (EDS) system, which we use for rapid identification of phases and elemental abundances. Unlike wavelength-dispersive spectrometry (WDS), the EDS system does not "tune in" specific X-rays. Instead, a solid-state detector collects and counts all of the emitted X-rays at once, and it divides the energy spectrum into different "channels" or ranges. We can collect a spectrum of X-ray energies from 0 to 10 keV and display it on the computer screen where peaks will show up on the spectrum, corresponding to energies of elements present in a sample.

The EDS system can be used for quantitative analysis (one simply counts the X-rays received in the channels that correspond with a peak of interest), phase identification, rapid particle analysis etc. For many combinations of elements, however, the EDS system is less desirable than WDS because of its limited spectral resolution (overlapping peaks) and limited sensitivity. Still, if one wants to determine quickly what elements are present in a sample, EDS is the way to go!

WDS (Wavelength-dispersive spectrometry)

Our microprobe also has a wavelength-dispersive spectrometry (WDS) system; it is outfitted with four WDS spectrometers and eight different crystal types. Wavelength-dispersive spectrometers are "tuned" to the characteristic X-ray of interest for analysis. The "tuning" is done by scattering of X-rays from a crystal positioned between a sample and the detector. By changing the angle of incidence of the X-rays, the crystal will constructively diffract X-rays of specific wavelengths. As a result, both the crystal and the detector move to accommodate the different incident angles. Different analyzer crystals are used to cover the entire X-ray spectrum: lithium fluoride (LIF), pentaerythritol (PET), thallium acid pthalate (TAP), and artificial layered dispersive element (LDE) crystals are the most commonly used.

WDS analysis results in a spectral resolution and sensitivity an order of magnitude better than is possible with EDS analysis; the detection limits of WDS ordinarily varies between 300 and 3 parts per million (ppm). Also, in comparison to EDS, WDS offers more accurate quantitative analyses, particularly for light elements, and better resolution of overlapping X-rays peaks for improved element identification and quantification.

Full integration of EDS and WDS

Through full integration of EDS and WDS on our machine, we can combine the productivity of the EDS with the analytical sensitivity of the WDS. Both signals can be combined within our Probe for EPMA software as well as in our Thermo Noran EDS system.

SXES (Soft X-ray Emission Spectrometry)

A unique feature of the UMN system is a novel wavelength-dispersive soft X-ray emission spectrometer (SXES) as the fifth spectrometer, the first in an academic institution in North America.

The SXES spectrometer uses X-ray focusing optics, variable line-spacing gratings and a solid-state X-ray CCD detector to collect a low energy X-ray spectrum. It provides higher spectral resolution and higher sensitivity for soft X-rays (<1 keV) relative to traditional WDS, yielding information on electronic structure and bonding [1-2]. The grating and CCD detector are rigidly fixed, with no moving parts, resulting in spectra that are extremely reproducible.

A characteristic feature of the SXES is its parallel detection of the X-ray signal. This allows it to be used like a conventional energy dispersive spectrometer (EDS) and generate a spectrum without the need to move the grating or the detector (like a traditional WDS scan). Hyper-spectral data is collected simultaneously, like an SDD EDS, with orders of magnitude better energy resolution. This parallel collection of the spectra data allows for the acquisition of spectra maps, from which element maps and chemical state maps can be extracted.

The SXES spectrometer at the University of Minnesota is configured with one high-resolution (JS300N) and one broad energy range (JS2000) grating with a combined spectral range covering nominally the X-ray energy range of 85 eV through >2.5keV. This allows for the analysis of ultra-light elements with very high peak to background ratios (the spectrometer is capable of trace analysis of light elements such as C, N and B in steel at the 10-100ppm range), but also a wider element list using both first order and high order reflections including the first order L-lines of the transition elements. For example, it is possible to study iron oxidation states from the peak shapes of Fe-L spectra.

Electron images and X-ray mapping

Our microprobe is also capable of different types of imaging: secondary electron (SE) images, backscattered electron (BSE) images, and characteristic X-ray maps. Images in each of these imaging modes can be acquired, processed, and stored digitally and be displayed in either grayscale or false color.

For SE imaging, the electron microprobe functions as an SEM, providing topographic information about a sample. The spatial resolution for SE imaging is approximately 100 to 200 nm, depending on the accelerating voltage, beam current, and other operating conditions. Common applications include studies of grain morphology, precipitates on a mineral surface, microfossils (for instance, ancient diatoms), and other materials too small for visible light microscopy.

BSE images show atomic number differences within a sample. A certain fraction of the electrons in the beam are scattered "backward" out of the sample as a result of interactions with its nuclei, and these "backscattered" electrons can be used to form images. The number of electrons are backscattered increases with increasing mean atomic number of the material. In a BSE image, the brighter the area, the heavier the mean atomic mass of that material.

In X-ray maps (also called element maps), the element distributions are shown in maps that represent concentrations as colors, very much like a weather radar map. The spatial resolution of X-ray maps is approximately 1 micron. Such mapping can delineate sub-micron particles or can be carried out across surfaces up to 90mm on a side. Our lab also offers fully quantitative (background- and interference-corrected, fully calibrated) element mapping either through our EDS system or ProbeImage software.

Detailed discussion

This overview of the principles behind electron microprobe analysis starts with the electrons that surround a nucleus. Niels Bohr first proposed that electrons are restricted to particular orbits around an atomic nucleus and that all other orbits are not permitted. The allowed orbits are quantized, meaning that electrons in these orbits have particular amounts of energy. Bohr also theorized that electrons can instantaneously "jump" from one energy level to another. Energy must be added to an electron for it to jump to a higher energy level, and energy must be released for it to jump down to a lower one. Such transitions involve the absorption and emission of "packets" of electromagnetic radiation known as photons. The energies of the photons depend on the energy difference between the levels involved in an electron transition. When an electron jumps to a lower energy level, something must happen to its excess energy; it cannot retain all of the energy it had in the higher level. As a result, it emits the excess energy in the form of a photon.

Orbital electrons are arranged in energy levels, or shells, that can be filled. Thanks to Barkla and Sadler, who started their characteristic X-ray nomenclature with K and L, the levels are labeled K, L, M, N, and so on, starting from the innermost one. The K-shell is the lowest energy level, and its electrons are most tightly bound. The "spacing" of these energy levels differs for each element. The levels of, for instance, one zinc atom are identical to those for all other zinc atoms, but the levels differ from those of all other elements. Additionally, when an electron jumps from, for example, the L-shell to the K-shell in a specific element, it emits a photon with the same energy, which is equal to the difference between the two shells. Because the wavelength of a photon is inversely proportional to its energy, such a transition will always yield a photon with the same wavelength.

Since the "spacing" of energy levels differs for every element, the transitions that electrons can undergo are different for each element. An element has a unique set of allowed transitions that produce photons with characteristic energies and wavelengths. When such transitions involve inner electron shells, the emitted photon falls into the X-ray range of the electromagnetic spectrum. This constitutes the basis of electron microprobe analysis: characteristic X-rays are identified using their unique energies or wavelengths to ascertain the composition of the sample. A beam of high-energy electrons that bombards the sample enables such electron transitions.

Electron microprobes contain an electron optical column, which produces the electron beam and controls its diameter when focused on a sample. At the top is an electron "gun" comprised of a tungsten wire bent into a v-shape and heated with an electric current to about 2700 K, which frees electrons from the apex of the wire. Since electrons are negatively charged, they are accelerated by an electrical potential, between 5 and 30 kV, toward a sample. As the electrons are accelerated, a pair of electromagnetic lenses focuses the electrons like a convergent lens focuses light. One lens restricts the number of electrons that pass down the column (the beam current) while another electromagnetic lens focuses the beam on the sample and controls its diameter. These lenses and a set of apertures can focus the beam to a diameter of 0.1 microns or less.

Consequently, electron microprobe analysis is considered a spot analytical technique, which means compositional information is collected from only a small volume, not the entire sample. The beam electrons interact with a volume usually between one and nine cubic micrometers (1E-18 to 9E-18 cubic meters). This volume is known as the interaction volume of the electrons. The small interaction volume of EMPA permits a researcher to collect highly localized compositional data and to examine specimens too small to be studied with other analytical techniques. In addition, it allows for the determination of the chemical variability over the surface of a sample. Consequently, EMPA is well-suited to study specimens composed of mixed phases that one wishes to resolve and analyze in situ, leaving the contextual relationships of the phases unaltered and visible.

Within an interaction volume, numerous electron-specimen interactions occur, such as X-ray production. Characteristic X-rays, which are produced when electrons "fall" from an outer energy level to one of the inner ones, possess wavelengths and energies specific to the elements from which they are emitted. Because inner levels are ordinarily filled, an electron must be removed in order to create a vacancy. When a sample is bombarded by an electron beam, a beam electron can knock an atom's orbital electron from its shell. This process is called inner-shell ionization since the atom is left with a positive charge. It remains ionized for a short period, around 1E-14 second, before one of the outer-shell electrons jumps down to fill this vacancy. Since an electron that falls into the vacated position must lose some of its energy, a characteristic X-ray is emitted. Therefore, on exposure to a high-energy beam of electrons, every element -- except for H, He, and Li -- emits a distinctive set of characteristic X-rays that can, in turn, be detected by the spectrometers.

Secondary electrons are a result of the inner-shell ionization, the same process that produces characteristic X-rays. A secondary electron is the electron liberated from its energy level by a beam electron. It is a former orbital electron that, once freed, leaves a vacancy into which an electron from a higher energy level falls as it radiates a characteristic X-ray. These electrons have low energies, so only those created within a few nanometers of the sample surface can escape. Therefore, secondary electrons are very sensitive to surface topography and can be utilized to acquire images of a sample similar to those collected by SEMs. In fact, electron microprobes are very similar to SEMs; each can be used as simple versions of the other.

Unlike secondary electrons, backscattered electrons are not produced in a sample. They are beam electrons that have been scattered back toward the surface of the sample. When backscattered electrons re-emerge from the surface, they are collected by detectors. These electrons have energies greater than secondary electrons, so they are less sensitive to topography. Instead, the backscattered electrons are influenced by the atomic numbers of the elements in the interaction volume. In heavier elements, many electrons are backscattered as a result of a single deflection, and the electrons retain much of their original energies. In lighter elements, a backscattered electron is more likely to suffer small deflections and lose more energy before it re-emerges. This effect is used to produce images, called backscattered electron images, that show some compositional information: the images exhibit bright areas where the atomic number is high and dark areas where the mean atomic number is low. The compositional variations apparent in backscattered electron images indicate differences in mean atomic number; elements cannot be identified without characteristic X-rays.

X-rays have characteristics of both particles and waves and can be described, and therefore detected, in terms of their energies or wavelengths. An electron microprobe is equipped with an energy-dispersive (ED) spectrometer, which electronically sorts and measures X-rays with respect to their energies. Electron microprobes also have several wavelength-dispersive (WD) spectrometers, which use diffraction to sort X-rays by their wavelengths. When white light passes through a prism or diffraction grating, it divides into its constituent colors, each of which has its own wavelength. The same phenomenon occurs in WD spectrometers; the X-rays are dispersed with respect to their wavelengths by a crystal. In a particular arrangement of the sample, the crystal, and the detector, the atomic lattice of the crystal reflects just one wavelength of the incoming X-rays toward the detector. Consequently, a WD spectrometer is "tuned" to a single wavelength at a time, which means it can better resolve X-rays and obtain more accurate measurements.

An ED spectrometer works best for simple qualitative analyses in many cases because it can rapidly record the full spectrum of interest. Within seconds, the X-ray spectrum collected by an ED spectrometer can reveal the major elements in a specimen and their relative concentrations, although the error in these measurements is rather large. Further, close X-ray peaks usually have ambiguous identifications, requiring use of a WD spectrometer, and ED spectrometers are not sensitive enough to reveal X-ray peaks from minor and trace elements. Peaks are more distinct from the background in a WDS spectrum, resulting in a better accuracy and minimum detection limits that are at least ten times lower. In addition, due to the higher resolution of WD spectrometers, there is little ambiguity in peak identification. Hence, WD spectrometers are better adapted to quantitative analysis than ED spectrometers.

Quantitative analysis is essentially a comparative method. It entails the measurement of the characteristic X-rays from a sample and a set of standards analyzed under the same conditions, and correction factors for various effects are calculated by the computer. The accuracy depends largely on the similarity of the standards and the specimen. For quantitative analysis, accuracy approaching ±1 percent (relative) is attainable for major elements. It is usually worse for trace and light elements or when significant differences exist between the compositions of the standard and the sample. The precision depends on counting statistics, particularly the number of X-ray counts from the standard and sample, and the reproducibility of the WD spectrometer mechanisms. The minimum obtainable precision is about 0.5 percent, although it is higher for elements at trace concentrations.

The detection limits differ for each element and are affected by the overall composition of a sample and the analytical conditions. For most elements, the detection limits for WD spectrometers is between 30 and 300 parts per million (ppm). It must be noted, however, that these detection limits for EMPA are misleading, particularly when compared to a bulk analytical technique. The microprobe can detect an amount of material a hundred thousand times smaller than that which can be detected with neutron activation analysis, a prevalent bulk analytical technique. As a result, EMPA is well-suited to the analysis of heterogeneous specimens.