The

| contain carbonates as impurities. To ascertain whether or not these precipitates were due to incomplete oxidation of SO3 to SO4(BaSO3), or to carbonate present in the Na2SO3, 5 cc. dilute H2SO, and a few drops of 1 per cent KMnO4 were added to each of the precipitates. failure of the mixtures to bleach the potassium permanganate showed that all three sulphites contained carbonates. Finally, Merck's reagent grade crystallised sodium sulphite (Na2SO3.7H2O) was tried, and was found free of carbonate. The anhydrous sulphites were found to contain about 80 per cent Na2SO3, while the crystallised sulphite contained 67 per cent Na2SO3. Separate portions of a mixture of CaCO3 and SiO2 containing increasing amounts of CO3 and 500 grms. SO3 as Na2SO3.7H2O were treated by the same procedure with the results given in Table XIII. TABLE XIII.

1. A scheme is proposed by means of which the acids of Group I. may be systematically detected and the relative amounts approximately estimated. The chief features of the method are:

(a) The preliminary precipitation of all of the acids of Group I. with a mixture of BaCl2 and CaCl2.

(b). The treatment of the precipitate with HCl to dissolve all but SO4.

(c) The oxidation by means of H2O2 of the sulphite to sulphate and the precipitation of the latter with BaCl2.

(d) Removal of Ba with (NH4)2SO4.

(e) Separation of F, AsO4, and PO4 from BO, and C4H4O6 in alkaline solution in the presence of sufficient ammonium salts.

(f) Separation of C204 and F from AsO, and PO, in an acetic acid solution containing an excess of Ca ions. (g) Reduction of AsO4 to AsO3 with H2SO3 and separation from PO4 by H2S in HCl solution.

(h) Separate tests for carbonates and chromates. (e) Separate test for the determination of the state of

oxidation of arsenic.

IN a recent paper (Am. Journ. Sci., 1917, xliv., 371) published from this laboratory an account has been given of the preparation of esters derived from the a and 8 sub. stituted ethyl alcohols and of an investigation of the effect of their constitution upon the rate of hydrolysis. It was shown by a study of a chloroethyl acetate, a-chloroethyl propionate, and a-ethoxyethyl acetate that the substitution of halogen or an alkoxyl group in the a-position of the alkyl radical of an ester accelerates the decomposition of the ester to such an extent that the reaction velocity can. not be measured. In the case of all three esters of this type that were hydrolysed acetaldehyde formed one of the hydrolysis products. It was further observed that the

substitution of hydroxyl, methoxyl, ethoxyl, chlorine, and bromine, in the B-position of the ethyl radical of ethyl acetate, produces a considerable retardation of the rate of hydrolysis. The hydroxyl and ethoxyl groups, and chlorine, produced practically the same degree of retardation; methoxyl and bromine a somewhat smaller retardation. In preparing the esters necessary for this study it was found that the 3-hydroxyethyl acetate may be obtained by refluxing equimolecular quantities of ethylene glycol and glacial acetic acid for eight hours over twice the theoretical quantity of anhydrous copper sulphate; and that B-ethoxyethyl alcohol is formed by digesting equimolecular quantities of 3-bromoethyl acetate and sodium ethylate for half an hour in alcoholic solution.

In the present article this investigation has been extended to the study of esters derived from halogen substituted propyl alcohols, and the results are here given to show the effect of this substitution in the alkyl portion of the esters upon the rate of hydrolysis.

:

The following esters were prepared and hydrolysed :Unsymmetrical 8-monochloroisopropyl acetate,3CH.O.CO.CH3.

CH3

CH2CI

Preparation of Materials.

Symmetrical 3,3'-dichloroisopropyl acetate was prepared by treating the 8,3'-dichlorohydrin of glycerol, which was in stock, with acetyl chloride in slight excess of the theoretical amount by the method of Henry (Ber., 1871, iv., 704; Bull. Sci. Acad. Roy. Belg., 1906, xlii., 261; Rec. Trav. Chim., 1907, xxvi., 89). The pure ester boiling at 201-203° was obtained by fractionation. The halogen content of the ester was also determined as a further test of the purity of the substance.

Chlorine found-I. 41.45 per cent; II. 41'92 per cent. Chlorine calculated-41'47 per cent.

Unsymmetrical 8,7-dichloropropyl acetate was prepared by combining equimolecular quantities_of_the chlorohydrin and acetyl chloride according to R. de la Acena (Comptes Rendus, 1904, cxxxix., 668). The ester purified by fractional distillation, boiled at 197-1980. The halogen content was found by analysis.

Chlorine found-I. 40:33 per cent; II. 40'59 per cent. Chlorine calculated-41'47 per cent.

8,7-dichloropropyl alcohol used in the preparation of this ester was obtained by passing dry chlorine gas into allyl alcohol at o° until the theoretical quantity of chlorine gas had been absorbed.

This method of preparation is discussed in the literature by Tollens (Ann., 1870, clvi., 164), Hübner and Müller (Ann., 1871, clix., 179), Munder and Tollens (Zeit. Angew. Chem., 1871, xiv., 252). The chlorohydrin was separated by fractionation, boiling at 182-183°. The yield from one hundred grms. of allyl alcohol varied from eighteen to twenty-two grms. The allyl alcohol used was prepared by the method of Koehler (Bull. Soc. Chim., 1913 [4], xiii., 1103).

Unsymmetrical 8,y-dibromopropyl acetate was obtained by treating the 8,y-dibromohydrin, which was in stock, with acetyl chloride in slight excess. The ester purified

by fractionation, boiled at 227-228°. Aschan (Ber., 1890, xxiii., 1827) prepared this ester by treating the dibromobydrin with acetic anhydride.

B-monochloroisopropyl alcohol was prepared from allyl chloride by the method of Oppenheim (Ann. Spl., 1868, vi., 367; Rec. Trav. Chim., 1902, xxi., 535). Two mols. of concentrated sulphuric acid were gradually added to one mol. of allyl chloride at o°, when a thick reddish brown oily substance was formed, which became dark brown in colour upon standing. The reaction seemed to be additive in character, and may be represented by the following equation :

The flask containing the reaction mixture was imbedded in snow and water, and after each addition of acid was shaken vigorously. The mixture was not permitted to warm up, because when the temperature was allowed to rise clouds of hydrochloric acid gas were evolved and charring took place, indicating decomposition. After standing twenty-four hours, the dark-coloured chloroisopropyl acid sulphate was diluted with eight to ten times its volume of water, and heated in a flask fitted with a reflux condenser on a water-bath from one to two hours. The hydrolysis may be represented by the equation

The solution of the hydrolysis products was then distilled, and the distillate collected until a temperature of 128-130° was reached, when the distillation ceased and sulphur dioxide was evolved. The distillate was then neutralised with potassium carbonate, and the oily liquid separating was extracted with ether, salting out with sodium chloride. The ether extract was then dried over freshly fused potassium carbonate. The ether was distilled off, the residue passing over at 120-130°, which, when redistilled, boiled definitely at 126-127, as described by Henry (Rec. Trav. Chim., 1903, xxii., 209).

In the account of this work done by Oppenheim (Ann. Spl., 1868, vi., 367), the details of the preparation of the chloroisopropyl acid sulphate are not given. After carrying out the distillation of the solution containing the products of hydrolysis, Oppenheim separated the 8-monochloroisopropyl alcohol from the distillate by saturating it with potassium carbonate. Much better results were obtained by the author by neutralising the solution with potassium carbonate at this point and extracting with ether as described above.

The 8-monochloroisopropyl acetate was prepared by causing equimolecular quantities of 3-monochloroisopropyl alcohol and acetyl chloride to react. The resulting ester was purified by fractionation, boiling at 149-150°, the same temperature reported by Henry (loc. cit.).

The allyl chloride used in the preparation of the unsymmetrical 8-monochloroisopropyl alcohol was prepared by the method of Tollens (Ann., 1870, clvi., 154), by treating three mols. of allyl alcohol with one mol. of phosphorus trichloride. The allyl alcohol was placed in a distilling flask connected with a Liebig's condenser, and the phos phorus trichloride was added through a dropping funnel, the flask being cooled with ice-water and shaken vigorously from time to time as the addition was carried on. The reaction may be represented by the following equation3CH2=CH.CH2OH+PC13 → CH2=

=CH.CH2CL+H3PO3. When all the phosphorus trichloride had been added the

| temperature of the water-bath was maintained at 40° until all the hydrochloric acid was evolved, then raised sufficiently to distil off the fraction of allyl chloride boiling at 45-50°. During the distillation the flask was shaken vigorously at intervals to assist in volatilising the allyl chloride and separating it from the thick and heavy phosphorous acid.

In performing this fractionation it was found that if the temperature were permitted to rise to the vicinity of the boiling point of allyl alcohol, an explosion occurred wrecking the apparatus and setting free red phosphorus, which was distributed through the distillation apparatus, while at the same time a strong garlic odour of phosphine or hydrogen phosphide was formed. De Mole (Rec. Trav. Chim., 1876, ix., 48) and Béchamp (Comptes Rendus, 1856, xlii., 227) record the same difficulty in carrying on similar work.

When the distillation was completed the distillate was washed with water; the allyl chloride forming a layer on the surface of the water solution was removed by a separatory funnel, dried over calcium chloride, and redistilled, boiling at 46-47°.

Meeting with the above-mentioned difficulty in the method of procedure just described, another treatment was adopted with satisfactory results. The procedure in this case was based upon the fact that allyl chloride is insoluble in water, while phosphorous acid is soluble. The allyl alcohol was placed in a flask fitted with a reflux condenser and the phosphorus trichloride was added by means of a dropping funnel supported by a grooved cork in the upper end of the condenser. The flask was held in a water-bath filled with ice-water, and shaken at intervals during the addition of the chloride of phosphorus. When the addition was complete the reaction mixture was refluxed for an hour, permitted to cool, and washed with a sufficient quantity of water to remove the phosphorous acid. The allyl chloride was then separated, dried, and redistilled as already described in the former procedure.

The latter method of treatment seemed to be very satis factory, and the yield of this method showed a slight increase over that of the first, due to the fact that a small amount of the allyl chloride was held back from distillation by the phosphorous acid. This was shown by dissolving up the phosphorous acid left as a residue after the distillation in the first process. The yield obtained by treating allyl alcohol with the phosphorous trichloride varied from 39'4 per cent to 37.8 per cent of the theoretical.

Hydrolysis of Esters.

Esters Derived from Halogen Substituted Propyl Alcohols-These esters were hydrolysed in decinormal hydrochloric acid by the methods already outlined (Am. Journ. Sci., 1917, xliv., 371) in connection with the hydrolysis of ester derived from the B-substituted ethyl alcohols. Hydrolysis measurements were made of the 3,3'-dichloropropyl acetate and the By-dichloropropyl acetate at 25°, 35°, and 45°, and of the 3-monochloroisopropyl acetate at 35° and 45°, and constants were calculated for these esters as may be seen in Table I., and these results are further sumarised in Table II.

The dichloro-substituted propyl acetates were found to be soluble in water only to the extent of about four cubic centimetres per litre. The 8,7-dibromopropyl acetate was much less soluble, making it impossible to dissolve enough of this ester in decinormal hydrochloric acid to make satisfactory velocity measurements. The 8-monochloroisopropyl acetate proved to be more soluble than the dichlorosubstituted propyl acetates, eight cubic centimetres of this ester dissolving in one litre of water.

In the case of the dichloro-substituted propyl acetates one cubic centimetre of each ester was dissolved in 250 cc. of decinormal bydrochloric acid previously warmed in the thermostat to the required temperature. The flask was then vigorously shaken, and as soon as the ester was completely dissolved 25 cc. of the reaction

Temperature.

Time in minutes

25°.

25°.

(67'4) (165)

25.2

65'3

75'4 184

26.6

74.8 184

165'0 27'5 83.3 27.8 84'9 201

202

27.6

66 6

165'0

27'9

83.1

200

75'5 188

203

75'0 185

75.6 186

185 186

26'4 66.7 26.8 66'8

k x 105

Averages

Average (duplicate).

Two cubic centimetres of the B-monochloroisopropyl acetate were dissolved in 250 cc of decinormal hydrochloric acid, and hydrolysis measurements were made as in case of the other esters. On titrating the portions of the reaction mixture withdrawn at the intervals recorded in Table I., with decinormal sodium hydroxide and then with silver nitrate, it was found that no halogen was liberated in the titration made at 35°. At 45° in the final titration there was a small amount of halogen acid set free, for which correction was made in calculating the

165'0 28.5 83.1 167'0 27'9 83.0 166'0

On making a comparison of the respective rates of hydrolysis of B-chloroethyl acetate and 8-chloroisopropy! acetate with those of their corresponding unsubstituted esters it is seen that the retardation in the case of the 8-chloroisopropyl acetate is much greater. This may attributed to the iso-character of this ester. It is also to be observed that the more effective retarding influences of the dichloro-substitution in the 3,8'-position may be due to the iso-structure of the ester in this case. E. W. Dean (Am. Fourn. Sci., 1914 [4], xxxvii., 331) in his study of the effect of the substitution of the acyl radical of an ester upon the rate of hydrolysis found that substituted and unsubstituted isobutyrates decompose more slowly than the corresponding normal esters. This investigation seems to furnish further evidence of the retarding influence of the iso-structure upon the rate of hydrolysis of the

ester.

The retarding effect produced by the substitution of halogen in the B-position of the alkyl radical forms a striking contrast to the accelerating influence produced by substitution in the a-position, where the decomposition of the ester is promoted to such an extent that the rate of hydrolysis is not measurable. As a result of this study it is seen that the position of the halogen with respect to the carboxyl group has an important influence upon the hydrolysis. On making a comparison of the results found in this investigation with those obtained by Drushel (Idem., 1912 [4], xxxiv., 69), in studying halogen substitution of the acyl radical of an ester, it is interesting to note that substitution in the a-position of either the alkyl or the acyl radical of an ester produces a decided accele ration of the rate of hydrolysis, but that this acceleration is much greater in the case of the alkyl substitution. degrees show very little variation. An exception is found The temperature coefficients for an increase of ten and 35°, where the coefficient seems to be 2.9 (see in the case of the B,y-dichloropropyl acetate between 25° Table II.).

Summary.

1. A convenient method of procedure is introduced for the treatment of allyl alcohol with phosphorus trichloride in the preparation of allyl chloride based upon the insolubility of allyl chloride and the solubility of the phos. phorous acid produced in the process.

heim brought about the separation of this alcohol by 2. In preparing 8-monochloroisopropyl alcohol, Oppensaturating the distillate obtained from the solution of hydrolysis products with potassium carbonate. It was found in performing this operation that a more convenient and complete separation was effected by salting out the alcohol with sodium chloride and extracting with ether.

3. Substitution of chlorine in the 8-position of the alkyl radical of the ester produces a marked retardation of the rate of hydrolysis. Substitution of two chlorine atoms in the B,y-position produces a greater and in the 6,8'-position a still greater retardation.

4. Evidence is furnished to show that the iso-structure in the molecule of an ester produces a retarding influence on the rate of hydrolysis.

The writer is indebted to Prof. W. A. Drushel for his co-operation in this work. Journal of the American Chemical Society, xli., No. 3, March, 1919.

same numerical value as in a sphere of the same volume. In the present work it was not only possible to test this conclusion experimentally, but it was also of great service in comparing the present results with those of Cameron and Ramsay, who used for this reaction glass tubes instead of spherical bulbs. Accordingly, one spherical reaction

EMANATION. I.*

THE COMBINATION OF HYDROGEN AND OXYGEN.

By S C. LIND. (Continued from p. 34).

5. Influence of the Size of the Reaction Bulb. (Law of the Inverse Square of the Diameter). INCREASING the size of the reaction bulb influences the velocity constant of the chemical reaction in two oppositely directed ways:-First, with the pressure remaining constant, the effective path of the a-particle is lengthened directly by 07 of the radius of the spherical bulb (S. C. Lind, loc. cit.), and hence, other things being equal, the velocity is directly proportional to the diameter of the bulb, provided that the diameter does not exceed the range of any of the a-particles. Second, a given amount of chemical action or volume change will produce a change of pressure dependent on the volume of the reaction bulb. The smaller the bulb the larger the change of pressure produced, in proportion to the cube of the diameter. Combination of these two opposite effects shows that the velocity constant expressed with reference to pressure, as in Equation 1, must vary inversely with the square of the diameter of the reaction bulb. This data from Table I. are summarised in Table II. so as to show that this is

actually the case to a high degree of precision for bulbs varying in diameter from 1 to 5.5 cm. and in volume over almost 200-fold.

(a) Extrapolated value. See following paper on "Recoil Atoms."

(b) Vessel cylindrical for sake of comparison with spherical ones).

This appears to establish thoroughly the nature of the law governing the influence of the size of the sphere on the velocity of the reaction. There appear to be no reasons why it should not apply to any other gaseous reaction as well as to the one under consideration. It appears to the writer to confirm his earlier theory of the average path of a particles in spherical volumes with reference to their chemical effect.

On passing to volumes other than spherical it has not been possible as yet to give a mathematical treatment of the average path. A graphical treatment for cylinders, such as will be represented by glass tubes of moderate dimensions, showed that the average path would have the

an actual volume of 6.787 cc., equal to the volume of a sphere 2.375 cm. in diameter. In Table II. the data for this cylindrical vessel have been reported together with the spheres, and it is found that the cylindrical volume furnishes a value of 83.3 for (ku/λ)/D, agreeing with the other values for true spheres. This is of considerable interest in that it enables a direct comparison between most of Cameron and Ramsay's results and the present ones. (See Section 9 for a later discussion). It may be of interest to inquire how great the diameter of the reaction bulb may become before the validity of the formula (ku/A)=(84 1/D) is affected. Evidently it still holds for the largest bulbs used in these experiments, and,

as will be seen in Section 9, it still holds within 2 per cent for Scheuer's results in a bulb of 7.18 cm. diameter at a pressure of 1580 mm. Even on considering that the stopping power of the electrolytic mixture is only about one-half that of air, and that the average path is about 0.7 of the radius, it is evident that many a particles are being completely absorbed before reaching the wall, while many others reach the end of their range where the ionisation is no longer proportional to the path travelled. These two effects tend to compensate each other, but when the diameter of Scheuer's bulb reaches 8:94 cm. at a pressure of 1680 mm. the value of kμ/A drops 6.7 per cent below the theory for uniform path, and continues to fall for larger sizes. The limit of the applicability of the formula appears to be at about 7 cm. for 1580 mm. of 2H2O2, which would be a diameter of to cm. at 760 mm., corresponding to an average path in air of about 3.7 cm.

It was the writer's earlier view (loc. cit.) that the general law would hold only over the first 2 or 3 cm. of path of the a-particle where ionisation remains constant. This would doubtless be true for a single type of a-particle; for example, from Ra C alone, but comparison of Bragg's combined ionisation curves indicates that ionisation per length of path in emanation in equilibrium would remain almost constant up to 4 cm. of air, agreeing very well with the results of the foregoing paragraph ("Studies in Radioactivity," 1912, p. 21; Phil. Mag., 1905, [VI.), x., 323).

6. Influence of Changing the Proportions of Hydrogen and Oxygen.

All reactions discussed in the foregoing sections have been carried out using electrolytic hydrogen and oxygen in exact proportions of 2 to 1 by volume. The effect of an excess of either gas can be predicted on the assumption that the change thus produced in the specific ionisation of the gas mixture will change the reaction velocity correspondingly. The specific (molecular) ionisation compared with air is, according to Bragg, 1'09 for oxygen and 0'24 for hydrogen ("Studies in Radio-activity," Macmillan, 1912, p. 65). Consequently, an initial excess of oxygen should increase the relative reaction velocity. The velocity constant calculated by Equation I should be initially higher than the normal, and should continue to rise as the mixture became relatively richer in oxygen with the progress of the reaction. With initial excess of hydrogen the case should be exactly the opposite; the velocity constant should start abnormally low and show a further fall as the mixture enriched in hydrogen.

Both cases have been experimentally investigated and the predictions found to be fully confirmed. Since the specific ionisation, however, is variable in these reactions Equation I is not strictly applicable. The development of a new equation taking into account the changing ionisation is so complicated that the simpler method has been adopted of using Equation I to show that the change