Probably you could find something on that at the site of Prof. Ernst Pernicka (here on ResearchGate) - probably the best archaeometallurgist for tin worldwide.

I work with ICP OES for analyzing tin. You have two ways: 1) direct but you aggregate about 1% your standard of hicrocloric acid, because the matriz of tin standard is HCl like dilute you must mantein the matriz aprox and read directly, I obtein good results 2) Use the hydride generation, with sodium borohydride, FIAS, this method give me very good results for low concentrations, ppb. I preparing samples with aqua regia so in the matriz have HCl and use 189.927 nm with high purge. Respect to information, you can view the method in the application note from Perkin Elmer, related with FIAS.

Unfortunately my experiences in the analyses of tin in medieval statues of brass and red brass (12th century) are published in German. The analyses were performed by ICP-OES and AAS (flame). Problems as described in your question are not relevant for the analyses of digested brass and bronze.

Due to the extreme valuableness of the statues my samples were in the range of 10 milligrams. The samples were digested in 10 mL aqua regia (a.r.). I preferred to use quartz glass dishes; they allow high temperature changes without breaking to sherds, which would be unexcusable considering the valuable sample materials.

The digestions were evaporated to dryness (not totally dry ! ). 10 mL of 1-normal HNO3 was added. Considering the generally high amounts of tin in brass (one to several weight percent) and the instrumental sensitivity allows the high dilution (10 mg sample with 1 % Sn in 10 mL yields a solution with 0.1 ppm Sn). After stirring the solution thoroughly it was stored for about 24 hours in the quartz dishes to allow total solution of the digest. To avoid evaporation during this time the dishes were covered with Parafilm (double). The ICP analyses on tin followed standard procedures. Analyses by AAS (flame) must consider the relatively high solution consumption of the method. So I used the injection method, which yields ballistic signals. Via normal aspiration tubing I injected portions of 50 µL into the flame. By this you can achieve a perfect calibration curve and do lots of analyses out of 10 mL.

As a routine in my lab we did huge numbers of acceptance analyses of brass fittings and taps on tin. By this we also verified the method.

Your concerns about tin analyses are certainly correct when going to the field of samples with tin-organics (soils, sludges, environment). These materials require adequate digestion, which totally disintegrates the organic matrix.

Thank you very much for your useful comments.