Information Content

LIBS spectra fluctuate very wildly from pulse to pulse. Therefore statistical analysis is required to be able to deduce elemental concentrations in a sample. The information content of a spectrum and its proximity to a pure element spectrum vary vastly with the number of spectra used in averaging or the estimated floor level used for baseline correction. Using tools from information theory and fuzzy matching algorithms we aim at determining the information gain with each new spectrum and try to benchmark our matching algorithms.

Masters' thesis

You will analyse the information gain with each additional spectrum and put your algorithm to the test with a large amount of spectral data on various samples. You will evaluate the effect of various experimental deviations from the ideal case, learn how to use the NIST's LIBS API, and enhance robust matching algorithms. You should bring basic programming skills, preferably in python and C, general interest in spectroscopy, data analysis and mathematics as well as enthusiasm for hands on experiments. We offer a motivated team with backgrounds in experimental physical chemistry, physics, information theory supporting you throughout your work, as well as a well established network in academia and industry alike. Please send us a letter of motivation, a current CV portraying the current progress of your studies (must), samples of your programming (must) and your general interests (nth) to office[at]