1340 SIDGWICK THE FREEZING POINT OF WET BENZENE, CXL1V.-The Fyeezing Point of Wet Benzene and the Injihence of Drying Agents. By NEVIL VINCENT SIDGWICK. DURING the purification of benzene by freezing for cryoscopic purposes irregularities were observed in the freezing points of successive fractions which were found to be due to traces of water absorbed from the air. This suggested the determination of the maximum depression of the freezing point of benzene by water. I n a previous paper (T. 1915 107 675) it was stated that this depression amounts to 0*042O but later work has shown that the readiness with which benzene absorbs water from the air had been underestimated and that the original benzene was not really dry. If proper precautions are taken to dry the benzene its freezing point is lowered 0*1000 by the addition of excess of water." This affords a convenient method for determining the efficiency of various drying agents.I f the wet benzene is shaken with 8 solid dehydrating agent which is insoluble in it until a constant freezing point is obtained the ratio of the depression observed in the presence of the drying agent to that produced by pure water is a measure of the lowering of the vapour pressure of the water by the latter and hence of the efficiency of the drying agent This efficiency must be the same (at temperatures near 5 O ) for a solution in any solvent in which the drying agent is insoluble. E X P E R I M E NTAL. Two samples of benzene each about a litre were used one ( A ) was kindly given me by Dr.Hewitt and Mr. W. J. Jones; it was of English origin and had been purified by them. The other ( B ) was Kahlbaum's purest. Neither gave more than the faintest trace of the hdophenine reaction. I n the later stages irregularities of several hundredths of a degree were observed in the freezing points which disappeared when the measurements were made in the presence of potassium carbonate. When the freezing points appeared to be constant the benzene * After this work was finished a paper appeared by Richards Carver and Schumb ( J . Amer. C h ~ m . SOC. 1919,41 2024) in which two experiments are described on the depression of the freezin? point of pure benzene by water, giving th? values 0.093' and 0'096" mean 0.095". This is in satisfactory agreement with the value 0*100" adopted in the present paper.These were further purified separately by repeated freezing AND THE INFLUENCE OF DRYING AGENTS. 1341 was still further purified by the method described by Richards and Shipley ( J . Amer. Chem. SOC. 1914 36 1825) and by Richards and Barry (ibid. 1915 37 993). It mas boiled for some hours with clean sodium (previously washed with the benzene) in a flask with a bent reflux condenser and drying tube. The condenser was then inclined downwards and connected to the side-tube of a carefully dried Beckmann apparatus. A steady stream of dried air was passed into the apparatus through the sheath of the stirrer, which was expanded into a bulb in the usual way to prevent moist air from being drawn in when the stirrer was raised.It is essential that the access of moist air should be prevented; in one experiment the condenser tube was accidentally separated from the Beckmann apparatus for a minute and this caused a drop in the freezing point of several thousandths of a degree. To the condenser tube between the water- jacket and the freezing-point apparatus a rather narrow side-tube was sealed so as to point downwards when the condenser was in position for distillation. The first runnings escaped through this tube which was then closed by attaching a test-tube to it with an air-tight cork. The tube then filled up with liquid and the rest of the distillate ran straight into the freezing-point apparatus. Several readings of the freezing point having been taken two or three drops of water were added the tube was warmed with the hand and well stirred and the freezing point again observed.This gave the difference between the freezing point of pure benzene and that of benzene saturated with water (the triple point solid benzene-liquid benzene-water). Then about half a gram of dry, powdered potassium carbonate was added and thoroughly stirred, and the freezing point again taken. A smaller quantity of the salt was then added in the same way and if this raised the freez-ing point another small quantity was added; more than three additions were never required. This gives the temperature of the quadxple point solid benzene-solid potassium carbonate-benzene solution-saturated aqueous solution of potassium carbonate. These operations were performed with successive frozen-out fractions of the purified benzene and the results are given in the following table.The temperatures are referred to the arbitrary zero of the Beckmann thermometer which remained constant during the experiments; each is the mean of several concordant readings, corrected for the temperature of the emergent stem. The fractions marked A are the English benzene B the German; the index number following this shows the number of times it had been frozen out; thus A 7 is the fraction obtained by freezing out the English beqzene seven times 1342 SIDGWIUK THE JFREEZING POINT OF WET BENZENE, 111. I. 11. Wet with Dried by With potassium Fraction. sodium. wet,er. carbonaDe. I.-11. 111.-11. A 7 ............ 3.120" 3.021' 3-080' 0.099" 0.059" A 7 ............3.121 3-02 1 3.081 0.100 0.060 ............ 0.060 A S - 3-023 3.083 - ............ - 0.098 - A 9 3.119 3.021 B 6 ............ 3.115. 3.015 3.075 0.100 0.060 ............ 0.061 B 7 - 3.020 3-08 1 -B8 ............ 3.121 3.022 - 0.099 -............ 0.061 B S - 3.022 3.08 1 -Means A.... 0.099 0.060 B.... 0.100 0-061 The freezing point is thus lowered O*lOOo by saturation with water. If the molecular weight of the dissolved water is 18 its solubility at 5.4" (assuming the cryoscopic constant 51O) must be 0.035 gram per 100 grams of benzene. Groschuff (Zeitsck. EZektrochem. 1911 17 348) found that 100 grams of benzene a t 3O dissolve 0.030 gram of water and a t 23O 0.061 which would mean about 0.032 gram a t 5*4O agreeing with the observed depression for a molecular weight of 18.The value 0.24 gram in 100 grams at 22" given by Herz (Ber. 1898 81 267) is presumably erroneous. On the addition of potassium carbonate water is withdrawn from the benzene t o form a saturated solution of the salt the vapour pressure of which is lower than that of pure water; the concentration of the benzene solution diminishes and its freezing point rises by 0'061" being 0.039O lower than that of pure benzene. Other dehydrating agents were then examined in the same manner. Carefully dried and powdered specimens of sodium sulphate copper sulphate and calcium chloride were used and also powdered sodium hydroxide and pure phosphoric oxide ; this last blackens rapidly in impure benzene but remains quite colourless in the pure liquid.All these compounds differ from potassium carbonate in forming hydrates with the water and it is perhaps for this reason that some of them especially copper sulphate act more slowly and must be stirred with the liquid for some time before their full effect appears. In these experiments the benzene was not in each case distilled over sodium but a mixture of the fractions A 9 and B8 which had been so treated was saturated with water and aEter its freezing point had been measured it was dried with the saIt and the rise of freezing point so produced was observed AND THE INFLUENCE 0%' DRYING AGENTS. 1343 Drying agent. None ........................... Sodium sulphate ............ Potassium carbonate ...... Copper sulphate ............ Calcium chloride ............Sodium hydroxide.. .......... Phosphoric oxide ............ Elevation of freezing point.. 0.024" 0.061 0.089 0.091 0.098 0.100 -Depression by water + agent. o*looo 0.076 0.039 0.01 1 0.009 0.002 0-000 The removal of water by phosphoric oxide seems to be quite complete. The middle column in the above table gives a measure of the relative efficiency of the drying agents examined. If water in benzene solution is unimolecular its concentration in the solution is proportional to the pressure of its vapour and hence the depression given in the last column is directly proportional to the vapour pressure of the system drying agent-lowest hydrate (or saturated solution in the case of potassium carbonate)-vapour . The last column in the following table gives the values of the tension of aqueous vapour calculated on this hypothesis. The pre-ceding column gives the actual freezing points based on the revised value given by Richards Carver and Schumb (see foot-note on p. 1340) for pure benzene. System. pure ........................... saturated with water ... wet + sodium sulphate ... , + potassium carbonate ,) + copper sulphate ... ) + calcium chloride ... y +sodium hydroxide , + phosphoric oxide . Freezing point. 5493" 5.393 5.420 5.454 5.482 5.484 5-491 5.493 Tension ?f aqueous vapour in mm. 6.73 5.11 2.62 0.74 0.61 0.13 0.00 -The relative efficiency of these substances is of course the same for drying solutions in any other solvents in which they do not dissolve. ORGANIC CHEMISTRY LABORATORY, OXFORD. [Received August 13thy 1920.