Dear All,
I have not read the Carmody volume and I will ask Ty for one.  However we
developed a technique some 20 years ago in which we have considerable
confidence. There are of course some caveats.There may possibly be other
dissolved species. When dealing with solutions, we prefer to precipitate CdS
using Cd-acetate crystals because we have found that the reaction of Cd++
with dissolved sulfide is fast and quantitative when we  passed natural gas
containing traces of H2S through various metal ion solutions. Precipitation
of other metal sulfides tended to give lower yields in these gas flow tests
if only one trap were used. This could be overcome by having a second trap
and combining the two precipitates.However in the case of precipitating CdS,
there was never a hint of the bright yellow-orange colour of the precipitate
in the second trap. This of course does not mean that quantitative
precipitation will not occur with some of the other alternatives when the
appropriate salt is added to a closed container of solution....but it seemed
that what always worked quantitatively in the gas flow experiments was the
best choice for additions to solutions.
Cd and some other alternatives such as Pb are undesirable from the viewpoint
of environmental toxicity and has lead to precipitation with Zn++ in some
laboratories.
It must also be remembered that some CdSO4 will definitely precipitate. ( I
have not investigated whether sulfate precipitates with some of the other
alternatives) So our approach is to first add
Cd-Ac followed by BaCl2 crystals under anaerobic conditions. Then decant
anaerobically. The supernatant can then be oxidized with peroxide,
HNO3-Br2, or another oxidant. If BaSO4 precipitates, it probably corresponds
to one or more other S-containing ions in intermediate valence states eg
SO3=.. H2S is evolved from the CdS/CdSO4/BaSO4 mixture ( it can be wet) by
addition of HCl to a reactor/reflux column and trapped again as CdS. The CdS
is subsequently converted to Ag2S by adding ammoniacal AgNO3 solution ( by
complex ion formation,  NH4- prevents precipitation of AgCl...and of course
Ag+ should not be added to the original sample because of Ag-halide
precipitation).
The question then becomes whether the precipitate of BaSO4 with minor CdSO4
(after washing and drying) is suitable for O-isotope (as well as S-isotope)
determinations. We think that it gives reasonable answers using the
traditional graphite reduction method and should  pose no problem with the
more recent pyrolysis approach. One must always keep in mind the Cd toxicity
and proceed accordingly. The alternative is to add only BaCl2  to another
aliquot of the sample for O-isotope analyses.
Roy Krouse