216 SPENCER : XXXVI1.-The Action of Bromine on Sodium and Silver A xides. By DOUGLAS ARTHUR SPENCER. Part I . Bromoamimide. DURING the course of experiments suggested by Professor H. B. Baker aiming a t the preparation of triatomic nitrogen bromine vapour diluted with nitrogen was passed over sodium azide. The colour of the gaseous mixture faded considerably but was not completely discharged however long the bromine remained in contact with the azide and the gas leaving the apparatus had a pungent but sickly smell reminiscent of hydrazoic acid and dilute bromine vapour. An aqueous solution of the gas was yellow gave a blood-red coloration with ferric chloride smelled of hypobromous and hydr-azoic acids and slowly evolved nitrogen on standing. Preliminary attempts to freeze out any compound formed resulted in violent explosions which occasionally detonated the sodium azide.The dilute gas mixture is itself extremely sensitive to shock or rise of temperature the explosion being accompanied by a flash of livid blue light whilst the glass parts of the apparatus are reduced to powder. The analysis was therefore performed indirectly as follows : Pure carbon dioxide generated by the action of boiled-out hydro-chloric acid on calcite was washed with sodium bicarbonate and dried by sulphuric acid and phosphorus pentoxide and was used to carry bromine vapour (derived from the liquid a t 5") over a large surface of sodium azide contained in a glass tube 175 cm. long and 0.5 cm. in diameter coiled into a spiral and kept a t 0" THE ACTION OF BROMINE ON SODIUM AXD SILVER AZIDES.217 The gas current was adjusted to carry about 0.05 g. of bromine over the azide per hour. The pale yellow gas so obtained was passed throcgh a hard glass tube 10 em. long packed with glass fragments and heated a t the far end by a small bunsen flame. By this means, the compound was decomposed without explosion and the gas acquired the much darker red colour characteristic of bromine vapour. The products of decomposition were passed over silver leaf into a potash nitrometer the carbon dioxide stream being stopped when the volume of gas in the latter had remained constant for 4 hour. The volume of this gas which was pure nitrogen having the normal density was measured over water and reduced t o that a t S.1'.P.The excess of silver leaf was dissolved in nitric acid and the silver bromide determined gravimetrically. The results are tabulated below. TABLE I. Expt. NO. 1 2 3 4 3 fi 7 Duration in hours. t 2 Bromine Bromine C.C. of N2 aken (g.). recovered (g.). obtained. 0.1684 0.1247 37.07 0-3938 0.2128 74.10 - 0.1549 57.60 0.1 !)3 5 0*09447 36.00 0.1 701 0.0802 25.80 O*BDOti 0.1324 50.00 0.3833 0.1706 GI*& IS3 Rr,. X. 1.41 1.26 1.13 1.11 1.14 1.13 1.15 Excluding the results of experiments 1 and 2 in which the bromine vapour was in contact with the azide for a comparatively short period the mean value is N3 Brl.13. Whilst pointing t o the presence of brornoaxoimide N,Br the analyses show that there is about 8% more bromine in the gas than is required by the simple formula.Since the quantity of bromine recovered agrees with that required by the equation Na" + Br = NaBr + N,Br it was a t first thought that this excess was due to a deficiency in the volume of nitrogen obtained, but the density viscosity and chemical behaviour of the nitrogen were normal and therefore it is improbable that any polymeric modification was present even if it could have survived the heating. A second possibility was that the reaction was reversible or incom-plete but alterations in the temperature of tlhe reaction tube, and of the time of contact of the bromine with the sodium azide, did not materially affect the final analysis. The formation of some other nitrogen bromide e.g. NBr, was a third possibility.By using a fine capillary tube as connecting link between the reaction tube and condensing vessels any explosion could be localised, and by immersing these vessels in freezing mixtures kept in un-silvered Dewar flasks standing in large beakers of water rendered less dangerous. It then became possible to freeze the substanc 218 SPENCER : out and to fractionate it by passing a stream of nitrogen over the surface whilst allowing the temperature to rise. The analysis of these fractions was carried out as follows The vapour derived from each fraction was absorbed in standardised solutions of carbonate-free caustic soda containing hydrogen peroxide. With this mixture the compound forms sodium azide and sodium bromide the hydrogen peroxide reducing the sodium hypobromite first formed : N3Br + 2NaOH = NaBrO + NaN3 + H,O.NaBrO + H,O = NaBr + H,O + 0,. By titrating the excess of sodium hydroxide with N/10-sulphuric acid and phenolphthalein the amount of alkali required to combine with the N3 and Br radicals was found. A known excess of silver nitrate was then added silver bromide and azide being precipitated. The mixture was boiled with nitric acid until all the silver azide had been decomposed and the hydrazoic acid driven off. The cooled solution was titrated with ammonium thiocyanate giving a measure of the silver nitrate required to precipitate the bromide present.* The results are summarised in Table 11. TABLE 11. Expt. Fraction. 1 1st 2nd 2 1st 2nd 3 1st 2nd 3rd a. 91.80 84.70 117-9 58.55 29.04 52.17 90-51 b .48-30 43.70 58-06 32.71 14.73 27.19 50.67 t . -loo + 5 - 18 + 5 - 20 - 10 + 5 N Rr,. 1.06 1.07 0.97 1.26 1.036 1.09 1.18 X. a = C.C. of N/lO-NaOH neutralised = N3 + Br; b = C.C. of NIlO-AgNO required to neutralise Br ; t = temperature below * The above procedure provides a method for the estimation of soluble azides in presence of halides which has decided advantages over the gravi-metric determination (Dennis and Isham J . Amer. Chem. SOC. 1907 29, 18). Owing to the number of operations involved and the appreciable solubility of silver azide at normal temperatures the latter method is tedious and liable t o error the results being usually low. On the other hand the volumetric method-titration with silver nitrate followed by treatment with nitric acid and thiocyanate-is both rapid and trustworthy if the solution is first rendered barely acid by the addition of sodium acetate and acetic acid.If potassium arsenate is used as indicator in the first titration, the fhal acid solution is colourless. The colour of silver arsenate is however, not so intense as that of the chromate and quite accurate results may be obtained with the latter. Owing to the solution being slightly acid with acetic acid a small amount of dichromate is formed and the solution is of a yellow-orange colour rather than the usual bright yellow and the use of a comparison basin is advisable THE ACTION OF BROMIXE ON SODIUM AND SILVER AZIDES. 219 which the fraction was obtained.C.C. of NIlO-NaOH equivalent to N3 = a - b. The first fractions (with a mean value of N3 Brl.02) were dark orange-red liquids which solidified at about -45" to dark red solids. The later fractions were ruby red but even these were distinctly lighter and apparently much more mobile than liquid bromine. I f the high bromine content is due to the presence of free bromine, it should be possible to remove this by treating the liquid with sodium azide. A first fraction obtained at -15" was distilled on to sodium azide kept a t -25". The colour of the liquid was not altered even after 3 hours' contact (Pound N Br 04). Also when gas derived directly from the apparatus and of approximate composition N3Br1.13 (see Table I) was condensed on to sodium azide and after 3 hours the liquid was allowed to vaporise the proportions were N3 BrlSl6.The excess of bromine therefore is not present in the free state and the most probable explanation is that it is in the form of nitrogen tribromide NBr,. When the gas was passed into water the solution contained in addition to hydrazoic and hypobromous acids traces of ammonium salts. This cannot however be taken as a proof of the presence of NBr3, for Hantzsch states that iodoazoimide (triazo iodide) gives rise t o a small quantity of ammonia on hydrolysis (compare L. Spiegel, '' Der Stickstoff," 1903 pp. 35 36). Hantzsch (Ber. 1900 33 522) by the interaction of silver azide (but not potassium azide) and iodine in ether solution a t 0" and evaporation of the ether obtained red crystals of the highly unstable iodoazoimide N31 (which he states is probably colourless when pure).A freshly prepared aqueous solution is neutral towards litmus and starch but hydrolyses fairly rapidly to hydrazoic and hypoiodous acids. I n non-aqueous solvents the compound slowly decomposes to iodine and nitrogen. When either sodium or silver azide is treated with bromine dis-solved in ether benzene or ligroin bromoazoimide is formed, but the method is not a convenient one. The bromine attacks the solvent to a certain extent and owing t o the great solubility and volatility of bromoazoimide it is impossible to separate it from the solvent. Moreover the presence of traces of water results in the immediate hydrolysis of the compound to hydrazoic and hypobromous acids.This is most striking in the case of silver azide-bromine mixtures as the admission of a drop of water results in a vigorous evolution of nitrogen due to the rapid reaction of the hypobromous acid a t once produced with the silver azide. Properties of Bromoaxoimide.-The purest sample of the bromide obtained in these experiments was a mobile very volatile orange 220 SPENCER : red liquid which changed to a dark red solid a t about -45". The pungent vapour has toxicological properties similar to hydrazoic acid causing giddiness headache and a slackening of the muscles when inhaled. Traces of the vapour irritate the eyes and cause a slight difficulty in breathing due apparently to congestion of the nasal mucous membrane. Solid liquid and vapour are as sensitive to shock as iodoazo-imide the explosion (often apparently spontaneous) being accom-panied by a flash of livid blue light.Some idea of the instability of the compound even a t -200" may be gathered from the fact that of twenty-four attempts a t freezing the compound and deter-mining its melting point only six were completed without explosion. Fortunately in the majority of cases the sphere of action was limited to a radius of about 3 feet. Within this radius all glass apparatus was reduced to powder ; beyond it a reinforced glass screen proved a sufficient protection. The liquid explodes in contact with phos-phorus arsenic sodium and silver foil but the vapour when diluted with nitrogen and passed over silver leaf or sodium gives a film of the corresponding azide and bromide.The liquid is apparently miscible in all proportions with ether, but is less soluble in benzene or ligroin. These solutions are stable for a few hours in the dark but when concentrated are liable to explode on shaking and on standing gradually decompose giving nitrogen and bromine the latter attacking the solvent. When passed into water bromoazoimide hydrolyses instantaneously, giving a mixture of hydrazoic and hypobromous acids and the solution on standing evolves nitrogen by the interaction of these acids. When bromoazoimide is passed into potassium iodide solution iodine is liberated equivalent to the hypobromous acid produced and potassium azide is obtained : N,Br + 2IZI = KN + KBr + I,. This experiment was performed by substituting a potassium iodide absorption vessel for the heated glass decomposition tube in the apparatus described on p.216 and it is possibly significant that 6 C.C. of nitrogen collected in the nitrometer. Some of this nitrogen may have been due to a slight decomposition of the bromoazoimide into its elements before it reached the absorption vessel but in view of the fact that no free bromine was detected when the compound was distilled on to sodium azide (p. 219), this does not seem probable. It is thought therefore that the nitrogen may have been derived from NBr present THE ACTION O F BROMINE ON SODIUM AND SILVER AZIDES. 221 Part I I . Since bromoazoimide is instantly hydrolysed by water the reaction in aqueous solution between bromine and sodium azicle should yield hydrazoic and hypobromous acids and therefore should differ from that between iodine and sodium or potassium azide which occurs only in presence of sulphur compounds yielding nitrogen (Raschig Chenz.Ztg. 1908 32 1203; Browne J . Amer. Chem. Xoc. 1922 44 2106). When sodium azide solution mas added to N/lO-bromine water, the colour of the mixture faded a t once to a pale straw-yellon- and nitrogen was evolved a t a rate depending on the concentration, temperature and proportions of the solutions. Approximately F I G . 1. 94% of the expected volume of nitrogen was evolved in 15 hours from concentrated solutions a t the ordinary temperature. The gas was contaminated with oxygen hydrazoic acid and hypo-bromous acid. The course of the reaction with various proportions and conceii-trations of the reactants can conveniently be followed by plotting the fall in iodine value of the solution against time.Aliquot portions of the eEervescing solution removed a t fixed intervals were added to potassium iodide and the iodine liberated by the hypobromous acid was titrated with X/lO-sodium arsenite (thiosulphate is unsuitable as it causes an instantaneous liberation of all the available nitrogen. In the figure typical curves obtained with it'/lO-solutions a t 15" are plotted. Hypobromous acid reacts somen7hat slowly x-ith hydrazoic acid, Rrowne Zoc. c i f . ) 222 SPENCER : but rapidly with sodium azide solutions (the latter are always alkn-line by hydrolysis and sodium hypobromite is dissociated to a greater extent than hypobromous acid).Curves 1 and 2 correspond therefore to the slow fall in concentration of hypobromous acid with hydrazoic acid and curves 3 and 4 to the fall in the presence of an excess of sodium azide. An abrupt change in the slope of the curves occurs when one uses more than one equivalent of sodium azide to two of bromine. NaN + Br + H,O = NaBr + HN + HBrO . (1) HBrO + 2HN3 = HBr + H,O + 3N2 - (2) Since one equivalent of hypobromous acid can decompose two equivalents of hydrazoic acid there is in the solution sufficient hypobromous acid to decompose a further equivalent of sodium azide :-HBrO + 2NaN = NaBr + NaOH + 3N . - (3) Equivalent proportions of sodium azide solution and bromine After 17 hours gas evolution had ceased and water were mixed.the solution was colourless and exactly neutral :-NaN + Br + H20 = NaBr + HBrO + HN . a (1) NaN + HBrO + HN = NaBr + H20 + 3N2 (4) . One equivalent of sodium azide was mixed with two equivalents of bromine water. The effervescence was much slower resembling that obtained when mixtures of hydrazoic and hypobromous acids react in presence of sodium bromide. NaN3 + Brz + H,O = NaBr + HBrO + HN (1) HN + &HBrO = $HBr + *H2O + l*Nz . - (2) . The solution should therefore contain equal parts of hydro-bromic and hypobromous acids amounting to one equivalent a t the conclusion of the experiment. The nitrogen evolved would be expected to carry away the greater part of these volatile acids and yet after 17 hours the solution still contained a little more than a quarter of an equivalent of acid one-half of which consisted of hypobromous acid the other half being presumably hydrobromic acid since there was only the slightest trace of azoimide.The Reaction between Bromine Water and Silver Axide.-Silver azide precipitated from a solution of sodium azide slightly acidified with nitric acid was washed until free from soluble silver salts and treated with freshly-prepared bromine water. A vigorou THE ACTION OF BROMINE ON SODIUM AND SILVER AZIDES. 223 reaction took place and silver bromide was obtained (in daylight, the yellow precipitate a t once commenced to turn slate-blue unless excess of bromine was present). The evolved gas contained traces of azoimide together with about 1% of oxygen but no nitrogen oxides and after drying with lime and phosphorus pentoxide had the density of ordinary nitrogen.About 920/o of the nitrogen expected from the equation 2AgN3 + Br = 2AgBr + 3N was obtained within 10 minutes of the mixing butl the presence of hydrazoic acid and oxygen suggests that the side reaction AgN + Br = AgBr + N,Br probably occurs in a manner analogous t o the formation of iodoazoimide (Hantzsch Zoc. cit.). The bromoazoimide was instantly hydrolysed by the water the oxygen being derived by decomposition of the hypobromous acid thus produced (compare Fleury C'ompt. rend. 1920 171 037). The formation of hydr-azoic acid in this manner accounts for the deficiency in nitrogen, as azoimide is only slowly attacked by hypobromous acid; the latter moreover reacts very rapidly with silver azide.Summary. In t h e absence of water bromine reacts with sodium and silver azides to give the highly unstable bronioazoimide : XN + Br = XBr + N,Br. This compound (m. p. about -45") whilst resembling iodoazo-imide in its general properties differs in its greater volatility and immediate decomposition by water. Bromine water reacts instantly with sodium azide solutions to give a mixture of hydrazoic and hypobromous acids which then interact to produce nitrogen. When the sodium azide is present in larger quantities than are required by the equation NaN + Br + H,O = NaBr + HN + HBrO the nitrogen evolution is more rapid owing to the interaction of the hypobromous acid with the excess of sodium azide and it is for this reason that two equivalents of bromine are able to decompose two equivalents of sodium azide.The reaction between silver azide and bromine water differs from that with iodine solutions in that it is better represented by the equation 2AgN + Br = 2AgBr -1 3N,. The only evidence for the momentary existence of bromoazo-imide in aqueous solution is the formation of a certain amount of azoimide with conscquent loss of free nitrogen. The density or viscosity of all nitrogen samples was determined 224 MILLER AND SMILES : but no indication of the existence of the polymeride N3 was obtained. My thanks are due to Professor H. B. Baker at whose suggestion, IMPERIAL COLLEGIZ OF SCIENCE AND TECHNOLOGY, and under whose supervision the work has been carried out.LONDON S.W. 7. [Received November 15th 1924. 216 SPENCER : XXXVI1.-The Action of Bromine on Sodium and Silver A xides. By DOUGLAS ARTHUR SPENCER. Part I . Bromoamimide. DURING the course of experiments suggested by Professor H. B. Baker aiming a t the preparation of triatomic nitrogen bromine vapour diluted with nitrogen was passed over sodium azide. The colour of the gaseous mixture faded considerably but was not completely discharged however long the bromine remained in contact with the azide and the gas leaving the apparatus had a pungent but sickly smell reminiscent of hydrazoic acid and dilute bromine vapour. An aqueous solution of the gas was yellow gave a blood-red coloration with ferric chloride smelled of hypobromous and hydr-azoic acids and slowly evolved nitrogen on standing.Preliminary attempts to freeze out any compound formed resulted in violent explosions which occasionally detonated the sodium azide. The dilute gas mixture is itself extremely sensitive to shock or rise of temperature the explosion being accompanied by a flash of livid blue light whilst the glass parts of the apparatus are reduced to powder. The analysis was therefore performed indirectly as follows : Pure carbon dioxide generated by the action of boiled-out hydro-chloric acid on calcite was washed with sodium bicarbonate and dried by sulphuric acid and phosphorus pentoxide and was used to carry bromine vapour (derived from the liquid a t 5") over a large surface of sodium azide contained in a glass tube 175 cm.long and 0.5 cm. in diameter coiled into a spiral and kept a t 0" THE ACTION OF BROMINE ON SODIUM AXD SILVER AZIDES. 217 The gas current was adjusted to carry about 0.05 g. of bromine over the azide per hour. The pale yellow gas so obtained was passed throcgh a hard glass tube 10 em. long packed with glass fragments and heated a t the far end by a small bunsen flame. By this means, the compound was decomposed without explosion and the gas acquired the much darker red colour characteristic of bromine vapour. The products of decomposition were passed over silver leaf into a potash nitrometer the carbon dioxide stream being stopped when the volume of gas in the latter had remained constant for 4 hour. The volume of this gas which was pure nitrogen having the normal density was measured over water and reduced t o that a t S.1'.P.The excess of silver leaf was dissolved in nitric acid and the silver bromide determined gravimetrically. The results are tabulated below. TABLE I. Expt. NO. 1 2 3 4 3 fi 7 Duration in hours. t 2 Bromine Bromine C.C. of N2 aken (g.). recovered (g.). obtained. 0.1684 0.1247 37.07 0-3938 0.2128 74.10 - 0.1549 57.60 0.1 !)3 5 0*09447 36.00 0.1 701 0.0802 25.80 O*BDOti 0.1324 50.00 0.3833 0.1706 GI*& IS3 Rr,. X. 1.41 1.26 1.13 1.11 1.14 1.13 1.15 Excluding the results of experiments 1 and 2 in which the bromine vapour was in contact with the azide for a comparatively short period the mean value is N3 Brl.13. Whilst pointing t o the presence of brornoaxoimide N,Br the analyses show that there is about 8% more bromine in the gas than is required by the simple formula.Since the quantity of bromine recovered agrees with that required by the equation Na" + Br = NaBr + N,Br it was a t first thought that this excess was due to a deficiency in the volume of nitrogen obtained, but the density viscosity and chemical behaviour of the nitrogen were normal and therefore it is improbable that any polymeric modification was present even if it could have survived the heating. A second possibility was that the reaction was reversible or incom-plete but alterations in the temperature of tlhe reaction tube, and of the time of contact of the bromine with the sodium azide, did not materially affect the final analysis.The formation of some other nitrogen bromide e.g. NBr, was a third possibility. By using a fine capillary tube as connecting link between the reaction tube and condensing vessels any explosion could be localised, and by immersing these vessels in freezing mixtures kept in un-silvered Dewar flasks standing in large beakers of water rendered less dangerous. It then became possible to freeze the substanc 218 SPENCER : out and to fractionate it by passing a stream of nitrogen over the surface whilst allowing the temperature to rise. The analysis of these fractions was carried out as follows The vapour derived from each fraction was absorbed in standardised solutions of carbonate-free caustic soda containing hydrogen peroxide. With this mixture the compound forms sodium azide and sodium bromide the hydrogen peroxide reducing the sodium hypobromite first formed : N3Br + 2NaOH = NaBrO + NaN3 + H,O.NaBrO + H,O = NaBr + H,O + 0,. By titrating the excess of sodium hydroxide with N/10-sulphuric acid and phenolphthalein the amount of alkali required to combine with the N3 and Br radicals was found. A known excess of silver nitrate was then added silver bromide and azide being precipitated. The mixture was boiled with nitric acid until all the silver azide had been decomposed and the hydrazoic acid driven off. The cooled solution was titrated with ammonium thiocyanate giving a measure of the silver nitrate required to precipitate the bromide present.* The results are summarised in Table 11. TABLE 11.Expt. Fraction. 1 1st 2nd 2 1st 2nd 3 1st 2nd 3rd a. 91.80 84.70 117-9 58.55 29.04 52.17 90-51 b . 48-30 43.70 58-06 32.71 14.73 27.19 50.67 t . -loo + 5 - 18 + 5 - 20 - 10 + 5 N Rr,. 1.06 1.07 0.97 1.26 1.036 1.09 1.18 X. a = C.C. of N/lO-NaOH neutralised = N3 + Br; b = C.C. of NIlO-AgNO required to neutralise Br ; t = temperature below * The above procedure provides a method for the estimation of soluble azides in presence of halides which has decided advantages over the gravi-metric determination (Dennis and Isham J . Amer. Chem. SOC. 1907 29, 18). Owing to the number of operations involved and the appreciable solubility of silver azide at normal temperatures the latter method is tedious and liable t o error the results being usually low.On the other hand the volumetric method-titration with silver nitrate followed by treatment with nitric acid and thiocyanate-is both rapid and trustworthy if the solution is first rendered barely acid by the addition of sodium acetate and acetic acid. If potassium arsenate is used as indicator in the first titration, the fhal acid solution is colourless. The colour of silver arsenate is however, not so intense as that of the chromate and quite accurate results may be obtained with the latter. Owing to the solution being slightly acid with acetic acid a small amount of dichromate is formed and the solution is of a yellow-orange colour rather than the usual bright yellow and the use of a comparison basin is advisable THE ACTION OF BROMIXE ON SODIUM AND SILVER AZIDES.219 which the fraction was obtained. C.C. of NIlO-NaOH equivalent to N3 = a - b. The first fractions (with a mean value of N3 Brl.02) were dark orange-red liquids which solidified at about -45" to dark red solids. The later fractions were ruby red but even these were distinctly lighter and apparently much more mobile than liquid bromine. I f the high bromine content is due to the presence of free bromine, it should be possible to remove this by treating the liquid with sodium azide. A first fraction obtained at -15" was distilled on to sodium azide kept a t -25". The colour of the liquid was not altered even after 3 hours' contact (Pound N Br 04). Also when gas derived directly from the apparatus and of approximate composition N3Br1.13 (see Table I) was condensed on to sodium azide and after 3 hours the liquid was allowed to vaporise the proportions were N3 BrlSl6.The excess of bromine therefore is not present in the free state and the most probable explanation is that it is in the form of nitrogen tribromide NBr,. When the gas was passed into water the solution contained in addition to hydrazoic and hypobromous acids traces of ammonium salts. This cannot however be taken as a proof of the presence of NBr3, for Hantzsch states that iodoazoimide (triazo iodide) gives rise t o a small quantity of ammonia on hydrolysis (compare L. Spiegel, '' Der Stickstoff," 1903 pp. 35 36). Hantzsch (Ber. 1900 33 522) by the interaction of silver azide (but not potassium azide) and iodine in ether solution a t 0" and evaporation of the ether obtained red crystals of the highly unstable iodoazoimide N31 (which he states is probably colourless when pure).A freshly prepared aqueous solution is neutral towards litmus and starch but hydrolyses fairly rapidly to hydrazoic and hypoiodous acids. I n non-aqueous solvents the compound slowly decomposes to iodine and nitrogen. When either sodium or silver azide is treated with bromine dis-solved in ether benzene or ligroin bromoazoimide is formed, but the method is not a convenient one. The bromine attacks the solvent to a certain extent and owing t o the great solubility and volatility of bromoazoimide it is impossible to separate it from the solvent. Moreover the presence of traces of water results in the immediate hydrolysis of the compound to hydrazoic and hypobromous acids.This is most striking in the case of silver azide-bromine mixtures as the admission of a drop of water results in a vigorous evolution of nitrogen due to the rapid reaction of the hypobromous acid a t once produced with the silver azide. Properties of Bromoaxoimide.-The purest sample of the bromide obtained in these experiments was a mobile very volatile orange 220 SPENCER : red liquid which changed to a dark red solid a t about -45". The pungent vapour has toxicological properties similar to hydrazoic acid causing giddiness headache and a slackening of the muscles when inhaled. Traces of the vapour irritate the eyes and cause a slight difficulty in breathing due apparently to congestion of the nasal mucous membrane.Solid liquid and vapour are as sensitive to shock as iodoazo-imide the explosion (often apparently spontaneous) being accom-panied by a flash of livid blue light. Some idea of the instability of the compound even a t -200" may be gathered from the fact that of twenty-four attempts a t freezing the compound and deter-mining its melting point only six were completed without explosion. Fortunately in the majority of cases the sphere of action was limited to a radius of about 3 feet. Within this radius all glass apparatus was reduced to powder ; beyond it a reinforced glass screen proved a sufficient protection. The liquid explodes in contact with phos-phorus arsenic sodium and silver foil but the vapour when diluted with nitrogen and passed over silver leaf or sodium gives a film of the corresponding azide and bromide.The liquid is apparently miscible in all proportions with ether, but is less soluble in benzene or ligroin. These solutions are stable for a few hours in the dark but when concentrated are liable to explode on shaking and on standing gradually decompose giving nitrogen and bromine the latter attacking the solvent. When passed into water bromoazoimide hydrolyses instantaneously, giving a mixture of hydrazoic and hypobromous acids and the solution on standing evolves nitrogen by the interaction of these acids. When bromoazoimide is passed into potassium iodide solution iodine is liberated equivalent to the hypobromous acid produced and potassium azide is obtained : N,Br + 2IZI = KN + KBr + I,.This experiment was performed by substituting a potassium iodide absorption vessel for the heated glass decomposition tube in the apparatus described on p. 216 and it is possibly significant that 6 C.C. of nitrogen collected in the nitrometer. Some of this nitrogen may have been due to a slight decomposition of the bromoazoimide into its elements before it reached the absorption vessel but in view of the fact that no free bromine was detected when the compound was distilled on to sodium azide (p. 219), this does not seem probable. It is thought therefore that the nitrogen may have been derived from NBr present THE ACTION O F BROMINE ON SODIUM AND SILVER AZIDES. 221 Part I I . Since bromoazoimide is instantly hydrolysed by water the reaction in aqueous solution between bromine and sodium azicle should yield hydrazoic and hypobromous acids and therefore should differ from that between iodine and sodium or potassium azide which occurs only in presence of sulphur compounds yielding nitrogen (Raschig Chenz.Ztg. 1908 32 1203; Browne J . Amer. Chem. Xoc. 1922 44 2106). When sodium azide solution mas added to N/lO-bromine water, the colour of the mixture faded a t once to a pale straw-yellon- and nitrogen was evolved a t a rate depending on the concentration, temperature and proportions of the solutions. Approximately F I G . 1. 94% of the expected volume of nitrogen was evolved in 15 hours from concentrated solutions a t the ordinary temperature. The gas was contaminated with oxygen hydrazoic acid and hypo-bromous acid.The course of the reaction with various proportions and conceii-trations of the reactants can conveniently be followed by plotting the fall in iodine value of the solution against time. Aliquot portions of the eEervescing solution removed a t fixed intervals were added to potassium iodide and the iodine liberated by the hypobromous acid was titrated with X/lO-sodium arsenite (thiosulphate is unsuitable as it causes an instantaneous liberation of all the available nitrogen. In the figure typical curves obtained with it'/lO-solutions a t 15" are plotted. Hypobromous acid reacts somen7hat slowly x-ith hydrazoic acid, Rrowne Zoc. c i f . ) 222 SPENCER : but rapidly with sodium azide solutions (the latter are always alkn-line by hydrolysis and sodium hypobromite is dissociated to a greater extent than hypobromous acid).Curves 1 and 2 correspond therefore to the slow fall in concentration of hypobromous acid with hydrazoic acid and curves 3 and 4 to the fall in the presence of an excess of sodium azide. An abrupt change in the slope of the curves occurs when one uses more than one equivalent of sodium azide to two of bromine. NaN + Br + H,O = NaBr + HN + HBrO . (1) HBrO + 2HN3 = HBr + H,O + 3N2 - (2) Since one equivalent of hypobromous acid can decompose two equivalents of hydrazoic acid there is in the solution sufficient hypobromous acid to decompose a further equivalent of sodium azide :-HBrO + 2NaN = NaBr + NaOH + 3N . - (3) Equivalent proportions of sodium azide solution and bromine After 17 hours gas evolution had ceased and water were mixed.the solution was colourless and exactly neutral :-NaN + Br + H20 = NaBr + HBrO + HN . a (1) NaN + HBrO + HN = NaBr + H20 + 3N2 (4) . One equivalent of sodium azide was mixed with two equivalents of bromine water. The effervescence was much slower resembling that obtained when mixtures of hydrazoic and hypobromous acids react in presence of sodium bromide. NaN3 + Brz + H,O = NaBr + HBrO + HN (1) HN + &HBrO = $HBr + *H2O + l*Nz . - (2) . The solution should therefore contain equal parts of hydro-bromic and hypobromous acids amounting to one equivalent a t the conclusion of the experiment. The nitrogen evolved would be expected to carry away the greater part of these volatile acids and yet after 17 hours the solution still contained a little more than a quarter of an equivalent of acid one-half of which consisted of hypobromous acid the other half being presumably hydrobromic acid since there was only the slightest trace of azoimide.The Reaction between Bromine Water and Silver Axide.-Silver azide precipitated from a solution of sodium azide slightly acidified with nitric acid was washed until free from soluble silver salts and treated with freshly-prepared bromine water. A vigorou THE ACTION OF BROMINE ON SODIUM AND SILVER AZIDES. 223 reaction took place and silver bromide was obtained (in daylight, the yellow precipitate a t once commenced to turn slate-blue unless excess of bromine was present).The evolved gas contained traces of azoimide together with about 1% of oxygen but no nitrogen oxides and after drying with lime and phosphorus pentoxide had the density of ordinary nitrogen. About 920/o of the nitrogen expected from the equation 2AgN3 + Br = 2AgBr + 3N was obtained within 10 minutes of the mixing butl the presence of hydrazoic acid and oxygen suggests that the side reaction AgN + Br = AgBr + N,Br probably occurs in a manner analogous t o the formation of iodoazoimide (Hantzsch Zoc. cit.). The bromoazoimide was instantly hydrolysed by the water the oxygen being derived by decomposition of the hypobromous acid thus produced (compare Fleury C'ompt. rend. 1920 171 037). The formation of hydr-azoic acid in this manner accounts for the deficiency in nitrogen, as azoimide is only slowly attacked by hypobromous acid; the latter moreover reacts very rapidly with silver azide.Summary. In t h e absence of water bromine reacts with sodium and silver azides to give the highly unstable bronioazoimide : XN + Br = XBr + N,Br. This compound (m. p. about -45") whilst resembling iodoazo-imide in its general properties differs in its greater volatility and immediate decomposition by water. Bromine water reacts instantly with sodium azide solutions to give a mixture of hydrazoic and hypobromous acids which then interact to produce nitrogen. When the sodium azide is present in larger quantities than are required by the equation NaN + Br + H,O = NaBr + HN + HBrO the nitrogen evolution is more rapid owing to the interaction of the hypobromous acid with the excess of sodium azide and it is for this reason that two equivalents of bromine are able to decompose two equivalents of sodium azide. The reaction between silver azide and bromine water differs from that with iodine solutions in that it is better represented by the equation 2AgN + Br = 2AgBr -1 3N,. The only evidence for the momentary existence of bromoazo-imide in aqueous solution is the formation of a certain amount of azoimide with conscquent loss of free nitrogen. The density or viscosity of all nitrogen samples was determined 224 MILLER AND SMILES : but no indication of the existence of the polymeride N3 was obtained. My thanks are due to Professor H. B. Baker at whose suggestion, IMPERIAL COLLEGIZ OF SCIENCE AND TECHNOLOGY, and under whose supervision the work has been carried out. LONDON S.W. 7. [Received November 15th 1924.