This paper deals with the various sources of error in different steps of the procedure for determination of sulfur trioxide in gypsum samples, using hydrochloric acid as solvent for gypsum and barium chloride to precipitate barium sulfate. The paper suggests reduction in the amount of acid and barium chloride used in order to minimize errors. Thus, it justifies the need for the revision of Section 12 of ASTM, Chemical Analysis of Gypsum and Gypsum Products (C 471).

ASTM C 471 recommends use of 50 mL of 1 + 5 hydrochloric acid (sp gr 1.19) as solvent for a 0.5-g gypsum sample and 20 mL of 10% barium chloride as precipitant. This results in too high a concentration of acid and chloride ions in reaction solution before and after precipitation, justifiable only for dissolution of impure gypsum samples and formation of coarse barium sulfate crystals to give rapid filtering. The method is silent about the final volume to be made of reaction solution, which determines the acidity before precipitation. It is also silent about the use of and the type of filtering crucibles.

The greatest source of error is the occlusion of chloride ions caused by both hydrochloric acid and barium chloride in addition to adsorption of H⁺, coprecipitation of barium chloride and Ca⁺⁺ ions, and increased solubility of barium sulfate in acidic solution.

Therefore, the following suggestions have been made (1) reduce the initial amount of hydrochloric acid used as solvent to 25 mL of 1 + 5 hydrochloric acid for the 0.5-g test sample, (2) use 20 mL of 6% barium chloride solution (preferably old), (3) make the volume of the reaction solution up to 400 to 500 mL before precipitation of barium sulfate, and (4) use filtering crucibles. These changes will increase the accuracy of the sulfate determination by minimizing error caused by high acid and chloride concentration and by maintaining quick dissolution and filtering. Analyses performed with these changes have been found to be consistent and accurate.