The MLA uses three X-ray analysis techniques to identify mineral species: point X-ray, area X-ray and
X-ray mapping.

Point X-ray analysis

In a typical sample measurement the MLA performs one X-ray analysis (typically at least 2,000 counts) for each grey level region identified within a segmented particle. The spectrum is collected at the centre of a phase to avoid contamination from bordering phases and hence acquire the "cleanest" spectrum possible. This spectrum is linked to its corresponding particle and grain in the segmented image to generate what is known as an X-ray image. The stored spectra are then compared with a predefined list or library of standard mineral spectra to complete the identification procedure and produce a classified image. In the classified image below, four of the five delineated phases were identified as chalcopyrite and the other as bornite. The classified image, and specifically the pixel phase data contained within, is the basis for all further quantitative analysis. In addition to particle location data, the pixel data and related mineral characteristics are stored in a data base for subsequent presentation using the MLA software: DataView.

X-ray mapping

X-ray mapping imposes a grid over an entire particle image, or specific grains within the particle, and collects X-ray data at each grid point to determine the mineral identity. Mineral identification using mapping requires significantly more time (greater than one order of magnitude) than point and area X-ray analysis as many more spectra are collected to generate a comprehensive mineral map.

The three fundamental X-ray analysis methods of point and area X-ray analysis and X-ray mapping are the building blocks of all MLA measurement modes involving X-ray identification.

Mineral Standards Library

Mineral identification through X-ray analysis requires a library of mineral standards. This library is usually constructed before an automated run and involves the collection of a high quality X-ray spectra spectrum for each mineral in the sample. The building of a standards library directly from the sample ensures that measurement conditions are reflected in the standards, such as beam energy, and it also provides for an elemental deportment that better reflects the chemistry of the sample.

The spectra identification process involves an error based search for the measured spectrum in the
standards library to find the most probable fit. Statistics on the spectral pattern matching can be examined to assess the confidence of classification for specific minerals.