Pretty amazing reading today – successful detection of sulfur and phosphorous in the visible region of the spectrum – in steel, at useful levels (e.g. below 0.005% for S and below 0.035% for P).  See, with LIBS both of these elements are typically detected in the deep UV, between 175 and 190 nm, which requires an argon purge (beam path, spectrometer, and sample) and a lot of patience.  

Wu and co-workers used visible lines of S (545.3 nm) and P (602.4 nm) in conjunction with a random forest (RF) regression to quantify sulfur and phosphorous in a set of 9 steel standard samples.  An echelle spectrometer with a spectral resolution of 6000 (λ/ λΔ) was used to collect high-resolution spectra.  Laser pulse energy and emission collection delay time were optimized for each of the lines using a detector gate width of 1.5 μs, and the spectral range input into the RF algorithm was varied to optimize the R2 and root mean square error of calibration.  The RF regression was compared with partial least squares regression to predict S and P content in steel.  Based on these results, further investigation of the long term stability of the RF calibration methods seem warranted.

To read more:   S. Wu et al., Quantitative analysis of nonmetal elements in steel using laser-induced breakdown spectroscopy combined with random forest, Anal. Methods, 2015, 7, 2425