Considering that Amir Liba (UCLA) wanted to determine sulfur in proteins, if
there is to be any hope of an accurate determination, there must be a closed
vessel digestion as the intial decomposition will lead to evolution of
volatile organo-sulfur compounds. I would certainly start with peroxide to
try to trap any free R-SH as R-SO2H (maybe in basic solution to form the
far-less volatile salts), then follow with cold, concentrated nitric and HCL.
Perhaps CEM or Milestone has a recipe ready.
The ease with which you can smell various sulfur compounds is indicative of
the calibration problem that will ensue if you do not have everything in the
same chemical species. H2S is highly volatile, NaSH is far less so; even if
it is nominally Na2S in very strong base, there is a small amount of H2S in
equilibrium, and you can smell 1 E-12 M. The range of volatilities, in my
experience with sulfur determinations, is the reason for Luis Morales' very
high RSDs. The calcium in thermal waters should be a very stable signal if
it is, indeed, interfering at the selected wavelength. Try using several
wavelengths 180.670, 181.972, and 182.562 nm (in purge), which have relative
sensitivities of 5:2.5:1 (a nice way to tell how good your purge is - echelle
users beware the blaze, though; the difference between purge and vacuum
wavelengths is about 0.061nm). Only the first has the Ca interference,
unless instrument resolution is poor -- Si at 180.74 should be resolved by
several spectral bandwidths.