Spectrum Matching is an automatic function that allows the user to search through a custom built spectrum library to find similar spectra. This greatly simplifies identification of unknowns by comparing them to a group of potential candidates and reduces the complexity of finding discrepancies and similarities between spectra.
Spectrum Match is a versatile tool with a multitude of applications, including:
- Failure analysis and quality control - defects can be compared to a library of potential contaminants.
- Process control - changes in distribution and composition can quickly be identified with the combination of phase mapping and spectrum matching.
- Reverse engineering - material constituents can be matched against a library of standard components found in the material.
Whichever the application, users will now be able to find their match with the click of a button.
The matching of the active spectrum to those found in the library file is based on a chi squared goodness of fit test, comparing either concentrations or spectra directly (Figure 1). The match sensitivity is adjustable, giving the user full control over the match results, ranging from completely identical to roughly in the same ballpark. The potential matching spectra are overlaid on the original spectrum with the match percentage giving a measure of how similar or dissimilar the spectra are along with a visual representation of the differences. Figure 2 shows an example spectrum with two overlaid spectra with a match value of 81.53% for SbS and 62.64% for CaSO4. A visual inspection quickly tells us that the discrepancies between the sample and the best matching spectrum are to be found in the Si, Ti, Ba and Zr content.
Figure 2. EDS spectrum with two spectra overlaid with unknown (green)
fitting the SbS material (blue) with an 81.53% match.
Building a spectrum match library in the TEAM™ software is made easy through the included spectrum search utility. The user is able to search the full database of spectra stored on the system either through manual selection of specific project nodes or by adding filters to identify spectra with the desired characteristics. The filters include:
- Acquisition date
- Presence of given elements (with range in atomic %, weight % or net intensity)
- Sample name
- kV used
Once the spectra are selected, they are saved to a Spectrum Matching Library file, which can be shared amongst users and systems. It is even possible to add all data in the system database to a Spectrum Matching Library file in order to finally answer the nagging question of “Hmm, where did I see this before?”