General Summary of EPMA Technique
Analytical Capabilities
Sample Requirements
WUSTL JEOL JXA-8200 Instrument Summary
Software
Microanalysis Standards
Imaging Capabilties
Precision and Accuracy
Microprobe Use Policy
Writeup Summary
Instructions and Manuals (links)
References

General Summary

An electron microprobe is an electron microscope designed for the non-destructive x-ray microanalysis and imaging of solid materials. It is capable of high spatial resolution and relatively high analytical sensitivity. The EPSC Microanalysis Facility has a JEOL JXA-8200 that can acquire digital secondary-electron and backscattered-electron images, digital x-ray maps, and digital cathodoluminescence images. It is equipped with 5 wavelength-dispersive spectrometers and a silicon-drift energy-dispersive spectrometer. Most of the periodic table can in principle be analyzed (Boron through Uranium), subject to several important considerations. The analytical sensitivity ranges from a low of a few parts per million for optimum cases, to a typical detection limit of several hundred ppm, but can be as high as several weight percent for problem elements. The volume sampled is typically a few cubic microns, corresponding to a weight of a few picograms. Samples should be prepared as flat, polished solid mounts up to 1.25 inch in diameter, and must be clean and stable in a 10^-5 torr vacuum environment. After preparation, samples are coated with a 22 nm layer of carbon using a high-vacuum carbon evaporator. The quality of analyses performed depends on the quality of sample preparation, character of the sample material, and availability of appropriate primary and secondary calibration standards for the desired elements. A precision approaching 0.X% relative and accuracy as good as 1-2% can be obtained with this instrument.

The technique of electron-probe microanalysis (EPMA) is a comparative method. Under electron-beam bombardment, elastic and inelastic electron scattering occur in a volume that is typically 1 micron in size. Inelastic scattering produces both characteristic and continuum X-rays in the analytical volume. The emitted X-ray intensities of a given element is measured in a sample and compared to the emitted intensity from a standard of known composition. This raw intensity ratio is called the k-ratio. In practice we take the peak intensity P for a given element, correct it for background B, and calculate the k-ratio from:

k = (P - B)^sample / (P - B) ^standard

This is the uncorrected concentration of an element in a sample. For example, a count rate of 500 counts/sec for Si Kα in a sample vs. 1000 counts/sec in a standard (Si metal, for example) indicates approximately 50 weight percent Si is present in the sample. However, it is necessary to correct for several processes in the primary scattering volume of both the sample and standard.

In the analytical volume, electrons lose energy in a stepwise fashion, are scattered, and may also backscattered out of the sample. These processes of electron energy loss and backscattering form Z, the atomic number correction. X-rays that are generated in the analytical volume are absorbed by all elements to a differing degree and in all propogation directions. The generated X-rays must traverse through the sample to the detectors, where the emitted X-ray intensity is actually measured. The absorption along the path to the X-ray detectors and the length of that path requires a correction from measured (i.e., emitted) intensity to obtain the initial (i.e., generated) X-ray intensity. This correction for X-ray absorption along the X-ray takeoff path is called A the absorption correction. Additionally, X-rays generated by a given element, with an energy close to and higher than the critical excitation energy of another element, may fluoresce the X-rays of the other element and accentuate the X-ray intensity of the other element relative to that produced by electron excitation alone. The additional contribution of fluorescence by characteristic X-rays is called F the fluorescence correction.

The parameters for Z, A, and F are typically used in a mulitplicative fashion to perform the "ZAF correction". For microprobe analysis of an olivine (nominally Mg_xFe_1-xSiO₄) which contains O, Mg, Si, and Fe, it is necessary to correct for the effects of electron retardation and scattering, X-ray absorption, and X-ray fluorescence of all other elements on the element of interest. The emitted X-ray intensities are measured for the Kα lines of Mg, Si, and Fe on an olivine sample, and the k-ratio for each element is calculated relative to the standard used. For an olivine analysis we would use the standards synthetic forsterite for Mg and Si, and synthetic fayalite for Fe, so the emitted X-ray intensities for Mg and Si would be measured on the forsterite standard, and the emitted intensity for Fe on the fayalite standard. The k-ratios for each element are obtained from these measurements and the concentration of each element in the olivine sample is obtained from:

C = k * ZAF

That is, the concentration of Mg in the olivine sample is obtained from the k-ratio for Mg and is corrected using the factor ZAF. It is important to understand that the ZAF factors are a function of composition, so that the conversion from k to C is performed by iterative calculation. This involves essentially an nxn matrix inversion and includes many other calculations not discussed here. The analytical approach is very powerful because a minimal set of calibraion standards can be used for quantitative analysis of a given element over essentially the entire concentration range from trace element to pure element using the algorithms for the ZAF factors.

This laboratory is involved in the development and application of quantitative electron microprobe analysis to a wide array of problems and maintaines an active program in basic research in the development and refinement of microprobe analysis techniques.

Analytical Capabilities

Elements Analyzed by EPMA

• Our microprobe is nominally capable of analyzing a wide range of elements, most of the periodic table, from Boron through Uranium, at concentrations ranging from ~0.01 to 100 percent by weight. This wide range of concentration range is enabled through use of Φ(ρz) and ZAF correction algorithms, attention to matrix chemistry, and careful setup of the analytical technique. The sensitivity depends on the element under consideration and the matrix of other elements present. The technique is usually applied to inorganic materials such as metals, oxides, minerals, glasses, semiconductors, and ceramics. Organic materials may be analyzed but must be made stable in vacuum, freeze-dried, fixated with OsO₄ or similar fixative, and treated as beam-sensitive materials.

Volume Sampled by EPMA

• The volume excited by the electron beam is nominally about 1 cubic micron, corresponding to a sample size of a few picograms. The primary advantage of microprobe analysis is that the analysis is non-destructive, although some material must be consumed in making a polished mount, and that a very small amount of material is sampled for a measurement. It is the ideal technique for analyzing crystals that are zoned, or samples that are made up of intimately-mixed phases where one needs to spatially resolve the phases being studied.

Types of Sample Geometries Analyzed by EPMA

• Bulk Materials. The default sample configuration for EPMA is a so-called bulk sample, meaning a flat, micropolished surface that is infinitely thick to keV electrons, corresponding to materials with a thickness greater than ~ 25 microns. Polished thin sections and bulk-cast epoxy mounts are typically used. Analysis of bulk materials that have been polished and carbon-coated is the default scenario for EPMA; the software correction and virtually all instructions pertain to that application. The other sample geometries require different sample preparation and further involvement from the user in order to obtain corrected results.
• Thin-film and Stratified Materials. EPMA is a very senstive surface analysis technique with the capability to detect a thin surface layer down to a monolayer in thickness. The technique was calibrated by measurement of progressively buried layers which were used to develop the Φ&rho(z) method, so the analysis of stratified materials is a fundamental capability of the technique. The penetration depth of scattered electrons is a function of the accelerating voltage used and the composition of the material via the atomic number Z properties. Stratified materials are analyzed in plan view (not in cross section) in order to apply the Φ&rho(z) method to layered structures. The accelerating potential may be adjusted in order to adequately sample the layered structure of interest, and analysis at more than one accelerating voltage may be used to verify the correction layer composition and thickness. The GMR thin film software is used to process data and calculates the individual layer compositions and their thicknesses as part of the analysis, where the layer sequence and element inventory is a model proposed and refined by the user.
• Particles. For particulate materials with a grain size larger than approximately 10 microns, samples should be cast in epoxy and polished for convention bulk analysis. For particles that are smaller than ~ 5 microns, particle analysis technqiues are used for EPMA measurement. The sample is dispersed on a pyrolitic carbon planchet and analyzed using a rastered electron beam. The particle shape and maximum diameter are recorded and used for EPMA particle correction. The Armstrong-Buseck Φ&rho(z) algorithm is used to correct for the processes of electron sidescatter and differential X-ray absorption relative to a bulk standard reference. Particle analysis is thus conducted like bulk EPMA with the additional requirements of sample preparation, measurement procedure, and correction of results requiring a particle model. This capability is now built in to the Probe for EPMA software on our instrument and is routine to the extent that particles are analyzed serially by the operator. Note that quantitative particle analysis as applied here is distinct from the automated feature analysis routines that many SEM and EPMA instruments are capable of (ours included) whereby a digital image is used to define a "particle" and metrologic measurements are obtained from the digital image (i.e., perimeter, projected dimensions, etc.) and fully quantified particle analysis is not performed.

Specialized Capabilities at the WU EPMA Laboratory

We have significant expertise in EPMA and a number of capabilities that have been developed in concert with the hardware and Probe for EPMA software used on our microprobe:

• Analysis of beam-sensitive materials. By selection of analysis parameters and using the Time-dependent Intensity (TDI) correction, measurements are performed sequentially and the X-ray intensity at t0 is calculated and used for correction. Many materials exhibit sensitivity to the electron beam and this capability is fundamental to quantitative analysis. This capability has been applied to the analysis of alkali glasses, hydrous basalt glasses, fluorine-bearing compounds, and carbonates.
• Multiple-spectrometer Measurement. Analysis of elements with multiple WDS spectrometers can be routinely performed in order to obtain high precision and improved sensitivity at both major and trace element concentrations in materials. With proper background selection, multiple-spectrometer measurement, low level correction for I vs. C, and intelligent use of trace element secondary standards, we are able to perform accurate EPMA down to concentrations at the ppm level for materials that do not present other atypical analysis challenges.
• Interference Correction. The quantitative analysis by EPMA may require the correction for X-ray peak overlaps that contribute an intensity at the peak position of an element of interest. The interference correction is used to remove this contributed intensity in a quantitative and iterated fashion in order to obtain the best possible accuracy in the EPMA of complex materials.
• Comparison with Secondary EPMA Standards. Quantitative EPMA runs in our laboratory typically include secondary EPMA standards for the purpose of verifying the primary calibration and inspection of the accuracy of analysis at lower concentration levels. It is important to verify the accuracy over the concentration range of both primary and secondary standards and this procedure is used in PFE as part of our quality control methods.
• Long-term Quantitative Analysis Quality Control. Each microprobe analytical run includes secondary standards (i.e., Kakanui Hornblende and other standards). The measurements on these standards can be data mined and form a part of ongoing evaluation of EPMA accuracy and precision in our and other labs. In particular it is observed that long term X-ray intensity performance for WDS (sealed counters) surpasses that observed for EDS detector systems.
• High speed X-ray Mapping. High speed stage X-ray mapping is performed using our e2v SDD EDS coupled with 5 WDS during stage mapping runs.
• Quantitative stage mapping. The Probe Image program is used to acquire WDS stage maps which are background corrected using the mean atomic number (MAN) empirical background method, and full Φ(ρz) correction at each pixel in the map. Results are presented and processed using Golden Software Surfer and other utility programs.

Sample Requirements

Samples must be flat (planar on a micron scale), polished to 0.25 micron final polish, and either electrically conductive or made so by evaporative coating with carbon (using our high-vacuum carbon evaporator). The sample must also be stable in a high vacuum environment. The sample must not contain residual oil, uncured epoxy, etc. that will outgas and contaminate the microprobe.

For the preparation of polished samples, several commercial thin-section companies can prepare both 1" round microprobe and standard 27 mm x 46 mm thin sections and polish them to 0.25 micron final polish. See the important information regarding sample size constraints listed below. We do not have sample preparation facilities in the department for general sample polishing. Please refer to the following links for commercial laboratories and suppliers:

Commercial Laboratories
These laboratories are known to make good polished sections:
Spectrum Petrographics, Inc.
Burnham Petrographics, LLC.

How to Polish Materials and Purchasing Sources for Polishing Consumables
These companies have excellent technical information concerning sample polishing for a wide range of materials:
Buehler, Inc.
Struers, Inc.

Important Information About Sample Size Constraints

Samples for EPMA are subject to several size requirements in order to fit in the sample holders for our instrument. There are important size limitations due to the use of top-referencing holders which have a lip that extends 1 mm onto the upper surface of the glass slide. The sample cannot extend to this outer region on the upper surface of the slide. Also, no epoxy can be on the outer edge of the glass slide due to tight clearance into the top-referencing holder.

We can analyze the sample configurations listed below.

• Prefered: 1" round microprobe thin section. A round sample billet is mounted on a 1" round glass slide. Sample limitations: The 1" round glass slide cannot be larger than 1" in diameter (i.e., any epoxy that spilled over onto the edge must be removed). Additionally, the sample billet (rock slice) must be less than 0.9 inches or 23 mm in diameter with minimum 1 mm clear area along the top perimeter of the round glass slide to clear the top-referencing holder. The sample billet can be up to about 2 mm thick although standard thickness for a polished mount is 30 microns. In our standard 9-hole block, we have 6 open 1" round positions for samples, with the total number of samples divided up into groups of 6 for analysis.

• Standard rectangular petrographic thin section. Note: Only standard petrographic glass slides measuring 27 mm wide by 46 mm long will fit in the slotted thin section holder. The sample billet/rock slice itself must be less than 25 mm wide by 44 mm long with at least 1 mm clear area along the perimeter of the glass slide on the upper surface, so that it will fit into the groove of the sample holder. The slide cannot have epoxy on the edge or top of the glass along this 1 mm perimeter and the thickness of the glass slide alone cannot exceed 0.053 inches or 1.35 mm. (Sorry to be picky but the tolerances on the slotted holder are tight). We can put 4 slides plus standards into the instrument, or use a second sample block and put 6 slides into the instrument at a time.
• If you have rectangular petrographic slides that do not have the tolerances listed above (either narrow slides measuring 25 mm by 46 mm, or slides that are too thick), we can use a bottom referencing holder and can put 2 slides into the instrument at a time. This sample mounting is not preferred: If the billet is wedged on the glass slide then there may be sample tilting issues which affects the x-ray absorption path length, and we may not be able to obtain focused x-ray maps for the same reason. For these samples the total thickness may be an issue as the microprobe can only accept a thin section with less than ~2 mm total sample thickness.

• "Bulk" mounts. We can analyze round mounts that utilize samples cast in epoxy (i.e., eraldite) or conductive mounting media. The prefered mount is a 1" round as we can nominally handle 6 mounts in our 9-hole sample block. We can also analyze 1.25" and 1.5" rounds using the 150 metallurgical sample block. These sample sizes are subject to the same requirement of having at least 1 mm clear at the perimeter for the top-referencing holder.
• Other sample configurations. We do have a so-called universal sample holder that can conceivably be used for samples that are much larger in size. There are orientation and referencing issues with large samples that will require further discussion.

WU JXA-8200 Instrument Summary

The JEOL JXA-8200 electron microprobe is a fully-automated, customized instrument. This instrument has digital imaging capability and can acquire digital secondary-electron, backscattered-electron (in compositional and topographic modes), and cathodoluminescence images. Digital x-ray intensity maps can be obtained using either beam scanning or stage mapping modes. X-ray focus is maintained during automated runs by means of a Carnegie autofocus module.

The microprobe has 5 wavelength-dispersive spectrometers, of which one is a four-crystal spectrometer with light-element layered dispersive crystals, four are two-crystal spectrometers equipped with two analyzing crystals and have extended range capability, and the remaining spectrometer is a special H-type two-crystal spectrometer designed for the measurement of trace elements via low intensity x-rays. The crystal inventory and general list of elements that they can be used for is listed below:

Spectrometer Crystal Inventory on the WU JEOL JXA-8200 microprobe

• Spectrometer 1: PETJ TAP LDEB LDE1 (P-10 flow counter, spectrometer range 71-258 mm)
• Spectrometer 2: TAP LDE2 (P-10 flow counter, spectrometer range 62-258 mm)
• Spectrometer 3: PETJ LIF (sealed counter, spectrometer range 62-258 mm)
• Spectrometer 4: LIF PETJ (sealed counter, spectrometer range 62-258 mm)
• Spectrometer 5: PETH LIFH (sealed counter, spectrometer range 88-238 mm)

The P-10 flow counter uses an 90% Argon - 10% Methane mixture that is optimum for low energy X-ray detection, and the sealed counters use Xenon for detection of higher energy X-ray photons.

Here is a list of the elements that each crystal can be used to detect. Note that in practice element analytical lists are set up in advance for most routine materials, and judgement based on experience is used to select a particular x-ray line and crystal for a given application.

This analyzing crystal inventory nominally permits detection and analysis of elements B through Pu, at concentrations typically above a few hundred ppm but in selected cases as low as a few ppm. This instrument is equipped with oil-free vacuum pumps (a dry scroll roughing pump, a turbomolecular pump, and an ion pump on the electron gun). It routinely achieves ~10^-6 Torr vacuum at the sample, and typically exhibits a low carbon contamination rate. These capabilities make the instrument desirable for light-element and/or low energy x-ray analysis. This instrument has very good electronic and analytical stability, vacuum cleanliness, and capability to perform analyses with high precision and accuracy as demonstrated by a history of applications.

The microprobe has a JEOL (e2v / Gresham) silicon-drift energy-dispersive spectrometer. This spectrometer is typically used for identification of samples during routine analysis, high-speed x-ray mapping, and monitoring of sample charging by inspection of the Duane-Hunt high voltage cutoff. It can also be used to analyze elements B - U either simultaneously or in conjunction with the wavelength spectrometers.

Software Summary

Our JEOL JXA-8200 has two software operating systems. The JEOL software system runs under Solaris on a Sun Ultra 45 and allows for completely integrated control of all spectrometers, imaging, and monitoring of system hardware. This instrument also runs the Probe for EPMA (PFE) operating system, which was developed for this instrument by John Donovan, Brian Gaynor, and Paul Carpenter.

Information regarding PFE can be found here and at the Probe Software web site. It is an excellent microprobe operating system!

PFE and related programs are essentially advanced front-end programs that store their data in Microsoft Access database files. All standard compositions are maintained in a standard database and read in by PFE during correction. Probe for EPMA performs all spectrometer and stage automation activities and handles all operations during an automated microprobe run. The PFE system enables sophisticated quantitative analysis that is not part of available turnkey systems.

Microanalysis Standards

We have a relatively large set of microanalysis standards that are used for primary calibration, secondary analysis, and specialized applications. However, we typically use a small set of primary and secondary standards for routine microanalysis projects. Our microanalysis standard suite is comprised of pure element, pure oxide, end-member stoichiometric compounds, synthetic and natural silicates, and synthetic compounds and glasses. The complete list of standards and their compositions is contained in the standard.mdb file used by the Probe for EPMA software.

We routinely use 3 standard mounts. The index maps are located in the lab notebook. These are:

• Standard S1 -- Natural and synthetic silicates, oxides, and glasses
• Standard S2 -- Natural and synthetic silicates, oxides, and glasses
• Taylor mount -- Natural and synthetic metals, alloys, oxides, silicates, and other materials

There are additionally mounts which contain the synthetic REE phosphates made by Boatner, and REE glasses made by Drake and Weill, and those of Amlie, and those from the University of Edinborough.

A significant number of the Smithsonian Microbeam Standards (SMS) are mounted in S1 and S2. Information regarding the SMS suite can be found at the Smithsonian web site, namely, the bulk chemistry analyses and homogeneity values, references for reporting the use of these standards, and information about inclusions or more recent analytical data on the standards:
http://mineralsciences.si.edu/facilities/standards.htm

The Taylor mount has only printed documentation, see the standard database and the printed copy.

Imaging Capabilities

Digital imaging (secondary-electron, backscattered-electron, cathodoluminescence, x-ray, and electron-channeling pattern) may be performed on the microprobe using the JEOL hardware and software. Both wavelength-dispersive and energy-dispersive digital x-ray maps as well as secondary-electron, backscattered-electron, or CL digital images can be simultaneously acquired. Two basic x-ray mapping modes are available. Beam scanning is used for x-ray maps typically obtained using the silicon-drift energy-dispersive spectrometer, which does not exhibit x-ray defocus problems, and has a relatively high x-ray throughput that rivals the WDS spectrometers. Stage scanning is used for x-ray maps using the both the silicon-drift detector and the wavelength-dispersive spectrometers, where the stage is scanned with a fixed electron beam. X-ray intensity maps can be acquired on elements down to low concentration, dependent only on the time available for measurement, and are usually run during overnight sessions. A guide-net mapping mode allows for the acquisition of mosaic-style maps and can be used to map materials that are topographically non-uniform.

Particle analysis capabilities have been added to the Probe for EPMA software.

Digital images are acquired using the JEOL OS-9 computer and the automated mapping software, and can be saved in Jpeg or TIFF format. These images can be processed further using a number of image processing programs.

Precision and Accuracy

The precision of measurements on the electron microprobe is a function of x-ray counting statistics, which depend on the total number of x-ray counts collected on both the standard used for calibration, and also on the counts collected on the sample. The minimum precision attainable on the instrument is in the vicinity of 0.5% relative, as determined by replicate measurements on wafer standards involving extensive spectrometer movement. Spectrometer mechanical reproducibility is considered to be the limiting factor in precise measurements on our instrument. Therefore, at low total counts collected, counting statistics errors dominate, and at high total counts collected, instrumental reproducibility dominates. Precision also depends on the chemical homogeneity of both the standard used for calibration, and also that of the sample.

The accuracy of measurements on the electron microprobe depends on accurate knowledge of the composition of the primary calibration standard, and the "correctness" of the algorithm used to convert from x-ray intensity to concentration units (i.e. the ZAF or F(rz) procedure). A global accuracy statement cannot be made. However, the accuracy is typically better than 5%, but may be worse for elements subject to peak interferences, or where there is a large compositional difference between the standard and sample and a large correction factor is observed (i.e. x-ray absorption, for example).

Maintenance, Calibration and ISO 9000 Issues

The Microanalysis Facility JEOL JXA-8200 microprobe is continuously maintained under ongoing contract by JEOL engineers. High priority is given to WDS spectrometer alignment, spectrometer reproducability, detector counting system deadtime calibration and correction. Specific procedures are followed that exceed the specifications of the instrument manufacturer, in terms of spectrometer calibration and reporoducibility testing. Records have continuously been kept since acceptance of the instrument. From discussions with other operators of microprobes and field engineers, and through exchange of calibration data, it is clear that our instrument is maintained to a high standard not achieved at a number of other facilities.

Microprobe Use Policy

Time availability

Time is available on a first-come, first serve basis. Calibration, repair, or other maintenance routines have highest priority, followed by EPSC users, followed by all other users.

Training and Probe Checkout

Microprobe users need to be trained prior to instrument use. Please read the manuals before your checkout session (hard copy available for checkout). Please sit in with a fellow user for one session to familiarize yourself with procedures prior to checkout.

Sample Preparation

Your samples should be completely polished, coated, and mounted (if applicable and reasonable) before the beginning of your session.

Coordination with Lab Manager

If you need a checkout or you just need help from Paul Carpenter, please coordinate schedules before you sign up. Paul may not be available if working on another project or on travel.

Writeup Summary

Here is a summary of the instrument hardware, software, standards, and correction scheme used for typical analytical runs. Your run may entail the use of non-standard analytical protocols.

Analyses are performed on a JEOL JXA-8200 electron microprobe equipped with 5 wavelength-dispersive spectrometers, and a JEOL (e2v / Gresham) silicon-drift energy-dispersive spectrometer. Analyses are acquired using either the Probe for EPMA or JEOL analysis software, and x-ray correction is performed using the CITZAF correction software (see references for a document reference). Typical operating conditions are 15 KV accelerating potential and 25 nA probe current, but conditions appropriate for analysis of special materials warrant other values. Standards used in the facility range from pure elements and oxides to simple or complex silicates and glasses recognized throughout the analytical community. A wide range of standards appropriate to specific analytical problems may also be used.

Instructions and Manuals (links)

The instructions, manuals, and other documentation can be found using the following links.

JEOL JXA-8200 Manuals

The hardware and software manuals for the JEOL JXA-8200 electron microprobe are located in the lab in print form only.

Probe for EPMA Manuals

The Probe for EPMA online manual and pdf versions of the PFE Getting Started and Advanced Users Manuals are located here.

Return to Microprobe Lab

Page last updated on: by Paul Carpenter