For the determination of sodium carbonate in commercial caustic soda, there is no satisfactory direct volumetric method, and usually a determination of carbon dioxide is resorted to. Winkler’s method yields higher results, owing to the precipitation of barium silicate, and to the precipitation of basic barium carbonate, when carried out according to the usual procedure.
Theoretically Warder’s method is suitable for mixtures containing much hydroxide and little carbonate, like commercial caustic soda, but the usual procedure is far from precise. Slight errors in the two end points affect the result seriously. If 45 cc. of a standard acid are used for the complete titration and an error of 0.04 cc. is made at each end point, the result for carbonate would be affected by 0.36 per cent when the errors are cumulative in nature.
Impurities that might occur in appreciable quantities in commercial caustic soda are sodium silicate, sodium aluminate,
iron, lime, and magnesia. Sodium silicate behaves like sodium hydroxide during the titration and offers no interference in the carbonate determination. Sodium aluminate is decomposed with the separation of aluminum hydroxide at the first end point. Aluminum hydroxide as well as ferric hydroxide, lime, and magnesia consume part of the acid for the second end point and, when present in appreciable quantities, cause a positive error in the carbonate determination.