Although the system sulfur trioxide-water has been extensively studied, there is still much disagreement in the literature. In certain parts of the diagram uncertainty exists even regarding the nature of the solid phase which is in equilibrium with solution.The melting points and various other properties of hydrates of sulfur trioxide as pure compounds were studied by many authors,3 as were the behavior of various solutes, including water, in absolute sulfuric acid as a ~o l v e n t .~ Four extensive investigations5 of the phase diagram have used thermal analysis.In view of the discrepancies in reported results, a different approach, namely, the equilibrium solubility method, was used in this investigation for the determination of all portions of the diagram except those pertaining to pure compounds, eutectics and peritectics. In this method a mixture of solid and liquid phases is agitated a t constant temperature until equilibrium is established. The liquid phase is then sampled and analyzed for sulfur trioxide content. By such procedure equilibrium can be ensured, solubilities can be correlated with known and well-defined phases, and the complete solubility curves of stable and metastable solids can be obtained.
ExperimentalThree constant temperature baths were used. Between 0 and 40' the bath temperature was controlled within 0.02', between -50 and 0' within 0.2", and between -75 and -50" within 0.1'.Five-junction copper-constantan thermocouples were calibrated a t the boiling and freezing points of water, a t the sodium sulfate, mercury and carbon dioxide points. The e.m.f. of these thermocouples was measured by a Leeds and Northrup Type K potentiometer, and temperatures -~ ( I ) This study, completed in 1943, was supported in part by a grant from the Manufacturing Chemists' Association of the United States. (2) (a) Present address: Shell Development Company, Emeryville, California; (b) present address: Eastman Kodak Co., Rochester, N. (c) 0. Hulzmann and W. Biltz, 2. anorg. ellgcm. Chcm., 918, 369 (1934); (d) E. Moles and C. R. de Robles, Anal. SOC. &an. 5s. y quim., 84, 331 (1936).were calculated by the method recommended by Scott .6 All temperature measurements were made with the thermocouple immersed in the actual equilibrium mixture and at an accuracy believed to be better than 0.02'.These were equipped with liquid-seal stirrers, sulfuric acid being used All-glass equilibrium vessels were used. as sealing liquid.Sulfuric acid samples were prepared by distillation of reagent grade sulfuric acid i n an all-glass still. In all distillations the first and last quarter portions were rejected. The samples in the fuming range were prepared by distilling sulfur trioxide from 60% oleum into sulfuric acid prepared as above. All solutions were colorless, and were discarded if any discoloration appeared.Samples of a composition between that of the expected solid and liquid phases were placed in the appropriate constant temperature bath and brought to temperature. The composition of the solution was adjusted, when ne...
