THE IIY DROBROMIDES OF UNDECYLENIC ACID. 1191 CXXVI.-Y%c fIycld)~-omicles of Undecylenic Acid. By JAMES WALKER and JOHN s. LUMSDEN. BRUNNER (Bey-., 1886, 18, 2226), by the addition of hydrogen bromide to undecylenic acid, CI,H,,02, obtained a bromundecylic acid, C,,H,,O,Br, which fused at 35’. No details of the mode of preparation or purification are given. The same acid is mentioned by Nordlinger (Ber., 1890, 23, 2357), but again no details are given. Nardlinger also, by the addition of hydrogen bromide to the methyl and ethyl esters of undecylenic acid, prepared the corresponding esters of a bromundecylic acid, As it has been proved with practical certainty that ufidecylenic acid has the formula CH,:CH*[CH,],*CO,H, compounds obtained from it by the addition of hydrogen bromide must have one or other of the formula : CH,Br*CH,*[CH,],*CO,H and CH,*C€€Br*[CH,~,*CO2H. Nordlinger, on the strength of the rule that the halogen atom in such additions generally takes up the position more remote from the carboxyl group, assumes that Brunner’s acid has the formula CH,Br*[CH,]9*C0,H and that the esters which he himself prepared are the esters of this acid.Since these esters have been made the starting point for determining the constitution of various substances, i t is plainly of importance to know if Nordlinger’s assumptions are justified, CH,Br*[CH,],*CO,H is not Brunner’s acid, which has in all probability the other con- stitution, C?H,*CHBr*[CH,],*CO,H. Both these acids are simul- taneously formed by the addition of hydrogen bromide to undecylenic acid, and no doubt the esters of both are produced by the addition of hydrogen bromide to undecylenic esters, a circumstance which would explain the divergent results obtained by different observers in syntheses involving their use.In the present paper, it is shown that the acid1192 WALKER AND LUMSDEN: Addition of Hydrogen Bromide to Undecplenic Acid. The hydrogen bromide which we used in the following experiments was in all cases free from bromine, and dried by means of phosphoric oxide. Addition without the use of a Solvent.-The following is a typical experiment. Seven grams of undecylenic acid were saturated with hydrogen bromide at the ordinary temperature. The acid fused and became slightly warm. Absorption ceased when 35 grams of the gas had been taken up.Dry air was then led through the liquid product to remove the excess of hydrogen bromide. This occasioned a loss of 0.5 gram, so that 7 grams of undecylenic acid had united with 3.0 grams of hydrogen bromide, the calculated quantity being 3.1 grams. After remaining for some time in a vacuous desiccator, the liquid partially solidified, and the crystals (4 grams) were separated from the oil (6 grams) by filtration. The crude crystals were spread on porous tiles to remove the oil which still adhered to them, and were recrystallised from warm light petroleum. The substance separated in clusters of needles which melted a t 50°. The oil did not solidify at the laboratory temperature even after long-continued standing, but crystallised on cooling to Oo.The crys- tals were washed on the ice-cooled filter with cooled light petroleum, and were then spread on porous tiles. On recrystallisation from light petroleum, they melted at 359 In this experiment, Brunner's acid mas formed in about equal propor- tion with another acid of higher melting point, The proportions of the two acids vary very much in different experiments-, sometimes one, sometimes the other, preponderating. Addition in Ethereal Xolut iom-Seven grams of undecylenic acid were dissolved in anhydrous ether, and the solution saturated at the ordinary temperature with hydrogen bromide, the excess of which was after- wards removed by a current of dry air. On evaporation of the ether, the residue solidified, and was purified by spreading on a porous tile and recrystallisation from light petroleum.Five grams of Brunner's acid melting at 35' were thus obtained. Tbe presence of the ether seems to favour the formation of the isomeride of lower melting point. Addition in, Toluene Solution.-Twenty-five grams of undecylenic acid were dissolved in 20 grams of toluene, the solution cooled in ice, and saturated with hydrogen bromide. When saturation was com- plete, a solid separated, which was filtered off, thoroughly mixed with a little light pstroleum at Oo, and refiltered. After drying on a tile, its weight was found to be 16 grams and its melting point 49-50°.THE HYDROBROMIDFX OF UNDECYLENIC ACID. 1193 The filtrates on cooling to Oo deposited a further amount of solid, which was purified as already mentioned.I n all, 26.5 grams of the acid of higher melting point were obtained from the 25 grams of undecglenic acid, and this we have found to be the best method of preparing it. w-Bromundecylic Acid, CH,Br*[CH,],*C02H, This acid, prepared in toluene solution in the manner just detailed, is insoluble in water, but freely soluble in the ordinary organic sol- vents, for example, alcohol, ether, chloroform, or benzene. Whilst easily soluble in warm light petroleum, it is only sparingly SO at the ordinary temperature, and very sparingly so a t 0". Light petroleum therefore forms the most convenient solvent for its recrystallisation. I t usually separates in clusters of fine needles, which, when pure, melt sharply a t 51' without discoloration or evolution of gas. The sub- stance gave the following numbers on analysis : 0.1472, heated in a Carius tube with 0.4323 AgNO,, required 1 9 9 C.C. of decinormnl NH,CNS to precipitate the excess of AgNO, Br = 30.05.C,,H,,O,Br requires Br = 30.19 per cent. The substance may be warmed either by itself or in organic solvents to a temperature considerably above its melting point without noteworthy evolution of hydrogen bromide occurring. The aqueous solutions of its alkali salts also are compsratively stable, showing little tendency to the separation of bromide and regeneration of undecylenic acid. o-Hydroxyundecylic Acid, CH2(0H)*[CH,],*C0,H. Thirteen grams of o-bromundecylic acid were dissolved in the calcu- lated quantity of normal sodium hydroxide solution, and the resulting liquid was warmed for 12 hours a t 60-70' with excess of freshly precipi- tated silver hydroxide, the mixture being constantly agitated by means of a Witt stirrer.The action was then complete, bromine in m y form being absent from the solution, The silver compounds were removed by filtration, treated with a little warm dilute sodium hydroxide solution, and again filtered. The filtrates mere united, and acidified with nitric acid, whereupon a solid acid separated, which was filtered off, washed with water, dried on porous tiles, and crystallised from benzene. On recrystallisation from much hot water, the acid separated in clusters of very long, fine needles, and melted a t 70". From the above quantity of bromundecylic acid, 5.5 grams of the purified hydroxy-acid were obtained.0.0801 gave 0*0800 H,O and 0.1922 CO,. 0.1206 of the calcium salt gave 0.0372 CaSO,. C = 65.43 ; H = 11.10. C,,H,,O, requires C = 65.35 ; H = 10.S9 per cent. Ca= 9.07. (C,1H,0,)2Ca requires Ca = 9*05 per cent.1194 WALKER AND LUMSDEN: The acid is thus monobasic, and has the composition of a hydroxy- undecylic acid. I t is easily soluble in alcohol or ether, moderately so in benzene, separating readily from a warm solution, and sparingly in light petroleum. One hundred parts of water at 20" dissolve 0.04 part of the acid. As it is more than 20 times as soluble in boiling water, it can easily be purified by crystallisstion frdm this solvent. A solution of the sodium salt containing 1 part in 200 is freely precipitated on addition of soluble salts of barium, caloium, strontium, silver, zinc, or mercury.The barium, strontium, calcium, and silver salts are much more soluble a t the boiling point than st the ordinary temperature. The calcium salt was purified for analysis by recrys- tallisation frorn boiling water. The solution of sodium salt of tho above strength affords only a slight precipitate with a soluble mag- nesium salt. A more concentrated solution of the sodium salt, however, gave a curdy, somewhat stringy, precipitate, which dissolved slightly on heating, and was reprecipitated in a granular form on cooling the filtered solution, As this hydroxyundecylic acid is derived from the straight-ohain undecylenic acid, it also has its carbon atoms in a straight-chain. On oxidation with chromium trioxide, it is converted into a dicarboxylic acid having the same number of carbon atoms as itself.The hydroxyl which it contains is therefore part of a primary alcohol group. The acid therefore must have the constitution CII,(OH)I[~H~]!,*CO~H, and consequently the bromo-acid from which i t was derived must be CH,Br*[CH,],*CO,H. n-Nonanedicnrboxylic Acid, CO,H*[CH,],=CO,H, One gram of o-hydroxyundecylic acid was dissolved in glacial acetic acid, and to the solution were added 10 c,c. of a 1 : 10 solution of chromium trioxide in glacial acetic acid, A slight rise of temperature was observed when the solutions were mixed. The mixture was allowed to remain overnight, and on the following morning was gently warmed to complete the oxidation. The solution was then poured into cold water, when a white precipitate separated.This was filtered off, washed with water, dried on porous tiles, and recrystallised first from benzene and then from boiling water, In this way, 0.6 gram of an acid melting at 110' was obtained, Analysis yielded the following results. 0.1323 gave 0,2968 CO, and 0.1112 H,O. 0.1545 required 28.5 C.C. of N/20 NaOH for neutralisation. C=61*17; H=9.34. C,,H,,O, requires C = 61 -11 ; H = 9.26. Re- placeable H = 0-922 j calculated for C,HI,(CO,H), replaceable H = 0.926,THE HYDROBltOMIDES OF UNDECYLBNIC ACID. 11115 The substance is thus a nonanedicarboxylic acid, and its derivatiob from undecylenic acid shows i t to have the normal structure. rt-Nonanedicarboxylic acid resembles sebacic acid, and the - other higher members of the series of normal saturated dibasic acids in ap- pearance and properties.It is soluble in alcohol or ether, can be crys- tallised readily from warm benzene, and is very sparingly soluble in light petroleum. One hundred parts of water a t 20' dissolve 0,014 part of the acid. It is much more soluble in boiling water, from which i t separates on cooling in lustrous scales, indistinguishable in appearance from those obtained under similar conditions from n-decanedicarb- oxylic acid and n-brassylic acid. The melting point of the purified substance is l l O o , and falIs into place in the series of melting points of the group of normal acids to which it belongs, thus : Acid (evcii). iilelting point. Acid (odd). Succinic .................. Ad il'i c .....................Sitberic .................. Scbacic .................. Decanedicarboxylic.. .... Dadecaneclicarboxylic . , 181" 119 141 133 127 123 98" 103 107 110 113 Gliitaric:. I'iuiclic. Azelaic. Noiinnetlicarboxylic. 15 mssylic. A solution Of the sodium salt containing 0.6 grstu in 100 C.C. was precipitated in the cold by solutions of calcium, zinc, silver, or mercuric salts. Barium salts gave no immediate precipitate in the cold, but a granular precipitate formed on heating the solution t o the boiling point. Magnesium salts gave no precipitate in either hot o r cold solution. A solution containing 3 grams of the sodium salt in 100 C.C. gave a slight precipitate with magnesium nitrate in the cold, which greatly increased in bulk on heating tho solution to looo. The calcium salt is very nearly equally soluble in hot and in cold water.Byomundecylic Acid, m. p. 35". This acid, originally prepared by Brunner, is by no means so stable as the o-bromundecylic acid melting a t 51'. It melts sharply a t 35", hut the fusion is usually accompanied by the formation of minute bubbles of gas, no doubt hydrogen bromide. Traces of hydrogen bromide are lost on warming the liquid, either by itself or in solution. Analysis of a specimen recrystallised from warm light petroleum gave 29.9 per cent. of bromine instead of 30.2 per cent. required by the formula CllH2,0,Br. The acid is freely soluble in the ordinary organic solvents, but insoluble in water. It is considerably more soluble in light petroleum than the isomeric w-bromundecylic acid, and can easily be crystallised from this solvent, the crystals assuming the form of thin plates or1186 THE HYDROBROMIDES OF UNDECYLENIC ACID, oblique prisms.I n contact with an alkaline solution, i t readily loses the elements of hydrogen bromide with simultaneous formation of undecylenic acid. Thus an attempt was made to convert it into the corresponding hydroxy-acid by treatment of its sodium salt with silver hydroxide. The action was conducted at the ordinary temperature, and took a fortnight for completion, the product being then worked up as detailed under w-hydroxyundecylic acid. The resulting acid was liquid, and only partially solidified on long standing. It was found to be unsaturated, and analysis showed it to consist of impure undecylenic acid. This tendency of the bromo-acid to revert to undecylenic acid made direct proof of its constitution difficult to procure.The tendency to lose hydrogen bromide would of itself, however, indicate that the bromine is not primary. Moreover, we know the primary bromo-acid, and since the mode of formation from undecylenic acid admits of only two possible constitutions, and the possibility of stereoisomerism is excluded, it is practically certain that the bromundecylic acid melting at 35’ has the constitution CH,*CHBr*[CH,],*CO,H, in opposition to Nordlinger’s assumption. Xyndlmis of n-Brassylic Acid. Both Komppa (Bey., 1901, 34, 897) and Krafft and Seldis (Ber., 1900, 33, 3571), starting from the ester obtained by the addition of hydrogen bromide to ethy€ undecylenate, have by means of a malonic ester synthesis prepared undecanedicarboxylic acids, which both originally held to be the normal acid, for both postulated that the bromundecylic ester with which they worked was the w-derivative.The acids which they obtained, however, differed entirely from each other; that of ICrafft and Seldis melted at 113-114’, being iden- tical with ordinary brassylic acid, whilst Komppa’s acid melted a t 8 2 O and gave entirely different derivatives. In order to ascertain definitely which was the normal acid, we per- formed the synthesis with an ester prepared from pure w-bromunde- cylic acid. I n converting this acid into its ethyl ester, it is necessary to employ more alcohol than usual, for although it dissolves readily enough in pure alcohol, it is thrown out of solution when sul- phuric or hydrochloric acid is added.Five grams of the acid were heated with 150 C.C. of atlcohol and 5 C.C. of sulphuric acid on the water-bath for 6 hours. This yielded 4.5 grams of the ethyl ester, boiling at 188-190’ under 18 mm. pressure. On performing a malonic ester synthesis with this amount in the usual way, the liquid became neutral after 20 hours’ heating. The mixed esters were extracted and dried, and saponiEed by addingWALKER AND LVMSDEN : N-DECANEDICARBOXYLIC ACfD. 1107 them while hot to a boiling solution of alcoholic potash. A white precipitate of potassium salt separated in a few minutes. From this the tribasic acid was liberated by hydrochloric acid, and appeared as a pure white, bulky, flocculent precipitate which melted when the water was warmed, and solidified on cooling.The dry substance melted in bulk a t about 60°, and in a melting point tube a t 70°, with evolution of carbon dioxide, which became very brisk above 100'. After all the carbon dioxide had been driven off, the residual dibasic acid was recrystallised from boiling water. It formed pearly scales melting at 11 1-1 12'. On recrystallisation from warm benzene, i t melted a t 112-113°, and agreed in its other properties with the brassylic acid derived from erucic acid, which has therefore undoubtedly the normal structure. No doubt the esters employed by Krafft and Seldis on the one hand, and Komppa on the other, although appbrently prepared in the same way, contained different proportions of the two possible bromunde- cylic esters (compare p. 1191), so that in the one case the normal undecanedicarboxylic acid resulted from a malonic ester synthesis, and in the other case an isomeric acid. Fileti and Ponzio (abstract in Bey., 1893, 26, Ref. S11, from J. pr. Chem., [ ii], 48, 323) give the solubility of brassylic acid in 100 parts of water at 24' as 0.74. This is obviously erroneous, as even acids so low in the series as suberic and sebacic acids do not attain even approximately this degree of solubility. An experiment with our synthetic brassylic acid gave a solubility of 0.004 in 100, which is in harmony with the results obtained for the other members of the series (see following paper). UNIVERSITY COLLEGE, DUNDEE.