An energy dispersive spectrometer on a scanning electron microscope reads the characteristic X-rays the beam excites, which turns an image into an elemental measurement at the point the beam sits. It is the most common analytical addition to an electron microscope and the most commonly over-read, because the numbers it prints look like a chemical analysis and are not one unless several conditions are met.
eds microscopy, and what the detector adds to the column
The electron beam excites inner shell transitions, and the detector collects the X-rays and sorts them by energy, so a spectrum is accumulated for whichever pixel or area the beam addresses. The analysis volume is a micrometre or so across and deep, which is much larger than the imaging resolution, so a feature smaller than that is measured together with its surroundings. Accelerating voltage decides which lines are excited at all, and a voltage chosen for a pretty image often excludes the line an element needs.
What standardless quantification means
Most reported compositions come from a standardless calculation: the software fits the spectrum, applies physical corrections for atomic number, absorption and fluorescence, and normalises the result to one hundred per cent. Because it normalises, an element that was missed still leaves a total of one hundred, which is why a result can look complete and be wrong. Quantification against measured standards of similar matrix, on a flat polished specimen at a known geometry, is what turns the numbers into an analysis with an uncertainty.
Spot, line and map
A spot spectrum is the fastest and most quantitative because the counts accumulate at one place. A line scan shows how composition changes across an interface and is the honest way to report a diffusion or a coating profile. A map assigns colour by element and is the most persuasive and least quantitative output, because dwell time per pixel is short and the map's brightness reflects counts rather than concentration. A report that shows a map should also show the spectra the conclusions rest on.
A sem microscope for sale, and what the analysis needs from it
Buying a used instrument for elemental work asks more than imaging does. The chamber needs the port and the working distance the detector is specified at, the stage needs to hold the specimen at the right geometry repeatably, and the detector itself has a window and, on older units, a liquid nitrogen dewar that a silicon drift detector has replaced. Ask what the detector's resolution is in electronvolts at manganese, whether the software licence transfers, and whether standards and a calibration routine come with it.
A sem for sale listing, and the questions before the price
The questions are the same for any electron microscope and sharpen when analysis is involved: the source type and its hours, vacuum and pump history, which detectors are included, the state of the stage and its motors, and who will decommission, move and reinstall it. Then the room, because vibration, stray magnetic field and cooling water decide whether the instrument performs at all. An instrument nobody will service, or whose software runs only on an unsupported operating system, is a parts donor rather than a purchase.
Questions people ask about sem eds
Is an EDS result a chemical analysis?
Only when it is quantified against standards of similar matrix on a flat polished specimen. A standardless result normalised to one hundred per cent is a screening measurement.
Why is a small particle's composition wrong?
The analysis volume is about a micrometre across, so anything smaller is measured together with the material around and beneath it.
Map or spectrum?
Maps show where elements are; spectra support what the concentrations are. A conclusion about composition should cite spectra.