Atomic emission lines

Most of the lines we see in LIBS spectra are atomic emission lines of particular wavelengths that are result of the energy levels of the electron shell of atom in question. In hot plasma valence electrons jump on higher energy levels and when they return back to a lower energy atomic orbital that's when a light emission occurs. There are also molecular emission lines for some specific molecules that we can see in LIBS spectrum.

There is no fully comprehensive list of all atomic emission lines. Some of the best atomic line datasets are are:

• NIST ASD Atomic Spectral Database at https://www.nist.gov/pml/atomic-spectra-database. At least 293550 lines according to my own extraction of their database. NIST has a LIBS specific NIST LIBS interface which is often a nice starting point for LIBS analysis.

• Kurucz

Robert L. Kurucz passed away in March 2025https://www.keefefuneralhome.com/memorials/robert-l-kurucz/5559934/. He had dedicated much of high life to understand stars and in doing that computed models and spectra and datasets of billions of lines. We are very thankful for his work. linelists at http://kurucz.harvard.edu/linelists.html. 2.31million "good" lines and 850million predicted lines. They are freely available for download.

• VALD Vienna Atomic Line Database at https://vald.astro.uu.se/. According to Pakhomov et al.

Pakhomov, Y., Piskunov, N. and Ryabchikova, T. (2017) ‘VALD3: current developments’. arXiv. Available at: https://doi.org/10.48550/arXiv.1710.10854. in 2015 VALD offered 1.2 million "good" lines with accurate wavelengths and 250 million predicted, but it's been updated since and includes Kurucz datasets so the total number of lines presently(in 2025) in VALD is certainly more. VALD can be used after a quick email registration which is open for everyone.

So we have millions or hundreds of millions of known emission lines, depending if we consider the predicted lines as known. These huge numbers make it clear how atomic transitions and emissions is not an easy problem to comprehensively describe.

Only a tiny fraction of these lines will be significant in our analysis, but to find out which ones we want to use the most complete datasets.

The VALD and Kurucz datasets have separate collections for hyperfine splitting of atomic lines. Pakhomov, Yu.V., Ryabchikova, T.A. and Piskunov, N.E. (2019) ‘Hyperfine Splitting in the VALD Database of Spectral-line Parameters’, Astronomy Reports, 63(12), pp. 1010–1021. Available at: .

Atomic line datasets processing

A request from VALD gives us a table of atomic lines, for example (Nickel atomic lines):

For most purposes we only need know the element, ionization, wavenumber, "log gf" and J values so we only take those columns and get the following kind of table:

Nickel atomic lines:

The relative intensities of atomic lines(at same ion level and plasma parameters) is given by 10^loggf. In our calculations we normally want the g values as well though so this isn't maybe enough? Need to test if these simpler ratios would work for our purposes.

Kurucz and NIST ASD dataset have the same values(formatted differently) but also the g/J degeneracy values so those are our preference at the moment.

What is "log gf"?

It is calculated log_10 g*f,

where:

• $g$ is the statistical weight of the lower energy level, and $f$ is the oscillator strength of the transition.

from this value together with plasma parameters we can calculate the relative intensities of the atomic lines.

Why atoms have the lines they do?

• Electrons are standing waves bound to an atom. The different energy levels are different possible harmonics of these standing waves around an atom. Giving more energy to the electron cloud will tune up the electron to a higher frequency and thus change the music.

• But the conductor of the orchestra is cheap and only wants to hear the lowest possible frequencies and all soloing electrons playing their own tune will eventually fall back in to their designated place where they fill every harmonic one by one from the lowest energy upwards.