146.5° 147.1° 120.6° 120.6° 109.5° Amyl acetate 137.5° and amyl iodide Of the 16 pairs of liquids investigated, 5 gave mixtures having well-defined minimum boiling-points, while 2 showed no relative elevation or depression of the boiling-point. The chemical constitution of the constituents exercises a greater influence in the formation of mixtures with minimum boilingpoints than the close proximity of the boiling-points of the constituents. One constituent remaining the same, or with constituents closely related, mixtures with substances of similar chemical constitution yield similar boiling-point curves. In the next paper the writer hopes to present the results obtained with propyl and isobutyl compounds. UNIVERSITY OF MAINE. [CONTRIBUTION from the ChEMICAL LABORATORY OF THe University OF MICHIGAN.] THE VOLUMETRIC ESTIMATION OF ALUMINA, AND FREE AND COMBINED SULPHURIC ACID IN ALUMS. BY ALFRED H. WHITE. Received January 25. 1902. N judging the quality of an alum, among the important determinations are those which give the amount of soluble alumina and the amount of sulphuric acid in combination with it or existing as free acid. The alumina may be satisfactorily estimated gravimetrically, but the method is tedious. A gravimetric estimation of the sulphuric acid gives not only that combined with aluminum plus that present as free sulphuric acid, but also that present as sodium sulphate, etc., and the amount of alkalies must be known before the amount of sulphuric acid combined with aluminum can be determined. The determination of free sulphuric acid in alums by volumetric means has been repeatedly attempted. The hydrolysis of aluminum sulphate prevents direct titration with an alkali, since as fast as the free acid present is neutralized, more is formed by hydrol ysis, so that the solution will not remain neutral for an appreciable length of time until nearly all the acid originally combined with the aluminum has been neutralized, and most of the aluminum hydroxide precipitated. This cannot be made into a practicable quantitative method for the estimation of free and combined acid, as the formation of aluminate with alkaline reaction begins before all the hydroxide has been precipitated. Further if phenolphthalein is used as indicator the precipitated hydroxide obstinately holds, by adsorption, the pink color even when the solution is no longer alkaline so that the method, though perhaps giving in the hands of works chemists in constant practice results which are fairly concordant, is, because of these various errors, not to be considered a practicable analytical method. The following paper results from an attempt to work out a practicable method. A standard solution of barium hydroxide was used instead of caustic soda to avoid any trouble caused by carbon dioxide in the caustic affecting the phenolphthalein used as indicator. It was found later that the barium hydroxide possessed other advantages. To prevent the adsorption of pink color by the precipitated aluminum hydroxide obscuring the end reaction, an addition of neutral potassium sodium tartrate (Rochelle salt) was made to hold the alumina in solution. The modification proved a good one. Duplicate results checked closely. There was no precipitation of alumina during the titration, nor even of barium sulphate. The solution remained perfectly clear and colorless until the end reaction, which was sharp. After standing a short time the solution became opalescent, and then milky, but no precipitate settled out. This marked retardation of the precipitation of barium sulphate was unexpected, and the conditions under which it occurs are undergoing further investigation. To determine that the results thus obtained were accurate, a solution of aluminum sulphate was made by precipitating aluminum hydroxide from a solution of aluminum chloride with ammonia, washing the precipitate thoroughly and then dissolving in standard sulphuric acid which was afterward diluted to a liter (Solution A). The solution thus obtained was almost fifthnormal, and the amount of acid was slightly in excess of the amount theoretically necessary to form the normal aluminum sul phate. A duplicate solution of acid alone, carefully standardized gravimetrically, was used to check the results. The method of procedure was as follows: To 25 cc. of the fifthnormal alum solution was added 50 cc. of 10 per cent. neutral potassium sodium tartrate, and the mixture was titrated cold with barium hydroxide solution, using phenolphthalein as indicator. This amount of tartrate was sufficient to prevent anything more than a slight opalescence in the solution before the end reaction. Duplicate titrations required 25.85 and 25.80 cc. of fifth-normal barium hydroxide. To determine the effect of the amount of tartrate, another titration was made, using 100 cc. of tartrate instead of 50. The amount of barium hydroxide solution used was 25.85 as before. The check solution of sulphuric acid required 25.65 cc. barium hydroxide, no difference being apparent whether no tartrate, 50 cc., or 100 cc. of tartrate were present. The solution containing the aluminum sulphate therefore requires slightly more barium hydroxide than that containing the sulphuric acid alone. A possible explanation is that the ionization of the barium hydroxide is lessened to such an extent by the aluminum tartrate complex that it requires an appreciably large excess of barium hydroxide to bring about the end reaction. If, however, the barium hydroxide is standardized against a solution of aluminum sulphate made from precipitated alumina and sulphuric acid as described, the results will be constant and give accurate results when extended to other alums. The addition of sodium sulphate in amount equivalent to the amount of sulphuric acid combined with the aluminum does not affect the result; the addition of standard sulphuric acid increases the amount of barium hydroxide used by the theoretical amount. The addition of neutral tartrate and titration with barium hydroxide, therefore, affords an accurate method of determining the total sulphuric acid combined with the aluminum plus the excess of free acid, irrespective of the amount of sulphuric acid combined with alkalies. As it is well known that the hydroxy-organic acids in general have the power of preventing the precipitation of alumina, salts of other acids than tartaric were tried, among them neutral sodium citrate. Sodium citrate prevented the precipitation of alumina, retarded the precipitation of barium sulphate, and perfectly sharp end reaction, as did the tartrate, but the amount of barium hydroxide used was only a little over twothirds of that required when titrating in the presence of the tartrate. Solution A, with neutral sodium citrate added, required only 17.95 cc. barium hydroxide instead of 25.84 when tartrate was used. Experiment showed that the addition of sodium sulphate did not affect the result and that an addition of standard sulphuric acid caused the theoretically calculated increase in the barium hydroxide used. It seemed that the result obtained when titrating in the presence of citrate could be due only to the formation of a complex aluminum ion and that this might furnish the basis of a method for the estimation of the aluminum. If we assume that the barium hydroxide used when titrating in presence of tartrate, represents free acid plus acid combined with alumina, while the barium hydroxide used when titrating in presence of citrate represents free acid, plus two-thirds of acid combined with alumina, the difference represents one-third of the sulphuric acid combined with the alumina, or one-third the alumina. In the above instance, 25.84 - 17.95 7.89 cc. fifthnormal barium hydroxide, and calculating the alumina on the assumption that this is equivalent to one-third of it we find that we get 0.0805 gram of alumina as compared with 0.0831 gram obtained gravimetrically. The volumetric result is too low. It seemed entirely possible that partial hydrolysis of the alum in the citrate solution might cause more barium hydroxide to be used than called for by the above supposition and that this might account for the low result. Accordingly, among other variations, the solution of aluminum sulphate was evaporated to dryness on the water-bath and dissolved in 50 cc. of 10 per cent. sodium citrate, and titrated. There were required only 17.58 cc. barium hydroxide instead of 17.95 cc., and the alumina calculated from this is 0.0842 gram instead of 0.0831 gram obtained gravimetrically. Using a saturated solution of citrate to redissolve the alum did not give an appreciably different result. Thus, the first method gives resecond gives slightly high re = sults considerably low while the sults. Some further experiments upon the influence of concentration and time follow, made upon a C. P. aluminum sulphate in a solution of 30 grams per liter (Solution B). 3 25 335 25 355 25 50 50 21.76 Titrated at once. 50 21.67 4 25 100 50 21.70 Water added; stood fifteen minutes; then citrate added, and titrated. Water added; stood fifteen minutes; then citrate added, and titrated. Water added; stood fifteen min utes; then citrate added, and titrated. Evaporated to dryness and redissolved in 50 cc. 10 per cent. citrate; stood ten minutes before titrating; required 21.62 cc. barium hydroxide. 0.750 gram (25 cc.) of solid salt, dissolved in 50 cc. 10 per cent. citrate and titrated at once, required 21.90 cc. barium hydroxide. The above series of experiments shows that the addition of varying amounts of water and variation of time within short intervals makes but slight difference in the result. The hydrolysis is apparently a slow one. In twenty-four hours, however, equilibrium is practically complete, as is shown by another series of experiments on a commercial aluminum sulphate dissolved to a strength of 30 grams per liter (Solution C). Nos. 3 and 5 show that if sufficient time is given to allow equilibrium to be established, the results are practically the same |