I) Optimize EDS measurements.

II) Consideration of peak overlap.

Peak overlap is a common problem in EDS measurement since isolated peaks are rarely found in routine analysis of complex materials. However, if overlapping peaks have at least some energy regions without energy overlap, the interference can then be dealt with “overlap factors” [1] by the following procedure:
          i) Collect a spectrum from a standard containing a single element with its X-ray peaks only;
          ii) Obtain the net window integral for a proper peak of this element. The first choice of the X-ray peak is its primary peaks (see the excel file). However, if the primary peaks will fully overlap with the peaks of the other elements in the unknown complex materials, then another proper peak should be used.
          iii) Repeat i) and ii) for other elements with pure elements or simple compounds;
          iv) Determine the relative fraction of the window integrals picked up in the energy windows in the spectra, for all the elements, obtained from the pure materials;
          v) Construct a complete matrix of overlap factors for all the elements in the unknown complex materials;
          vi) Record a spectrum or a map from the unknown materials;
          vii) Evaluate the composition of the unknown materials by using a set of simultaneous equations based on the obtained overlap factors. In more efficient practice, this step can be done by creating scripts based on Digital Micrograph  (DM) if the EDS data is recorded on or can be extracted to DM interface.

The overlap factor method avoids any detailed knowledge of peak shape, which cannot accurately be simulated because it depends on sample stage configuration, beam current, beam size, pixel time, etc. However, this method requires a considerable amount of experimental work to establish the overlap factors. Acquiring a comprehensive library of experimental data for all elements involves a lot of time and expense. Fortunately, only the data of limited elements is needed for a specific job, for instance, in a specific research group or company. If X-ray detectors from the same manufacturing process and microscopes of the same model from the same microscope manufacturer have similar characteristics, then the same profile library can in principle be used at the first level of qualification and even quantification. However, if the X-ray detector is changed, or the pulse processing time constant is changed, or the resolution, linearity, sample position, or calibration changes with count rate (or pixel time) or degrades over time [1], then overlap factors will have to be re-evaluated for highly accurate analysis, especially when very complex materials are studied.

The overlap factor method is very good for dedicated analysis involving a fixed set of known elements. However, it normally presents very poor results if the overlaps are severe.

III) Avoid artifacts.

Precise correction for elemental quantification extracted from EDS maps in crystalline specimens is a difficult task because it is complicated by the existence of electron diffraction, and channeling and blocking effects (refer to page3767); it would require:
         i) Measurement of intensity in the diffraction plane;
         ii) Knowledge of the crystal structure;
         iii) Knowledge of the orientation of the crystals;
         iv) Measurement of the specimen thickness.

[1] Peter J. Statham, Limitations to Accuracy in Extracting Characteristic Line Intensities From X-Ray Spectra, J. Res. Natl. Inst. Stand. Technol. 107, 531–546 (2002).