234 SUDBOROUGH AND ROBERTS : XXIX. -Diortho-substituted Benxoic Acids. Part V. Formation of Xults from Dio1.tho-substituted Benxoic Acids agd Orgnnic Bases. By JOHN JOSEPH SUDBOROUGR and WILLIAM ROBERTS. IN a previous communication on this subject (Lloyd and Sudborough, Trans., 1899, 75,580), it has been shown that stereochemical influences only slightly affect the reactions between substituted benzoic acids and organic bases, these influences being completely masked by the effect of the comparative strengths of the acids and bases. Since the publication of this earlier paper, E m i l Fischer (Bey., 1900, 33, 345, 1967) has drawn attention to the fact that diortho-substituted tertiary bases are incapable of combining with an alkyl iodide to yield quatern- ary ammonium compounds.As the earlier experiments were not carried out with bases of this type, we have tried to prepare salts from diortho-substituted alkylated arylamines of the type of dimethyl-DIORTHO-SUBSTITUTED BENZOIC ACIDS. PART v. 235 mesidine and various substituted benzoic acids, including those with two ortho-substituents. We have also attempted to combine the same acids with diethyl-a- and P-naphthylamines. s-Trinitrobenzoic acid forms salts with all the bases examined ; 2 : 4 : 6-tribromo-3-aminobenzoic acid also forms salts with a number of the bases, but not many have been obtained in a crystalline form. It would thus appear that the introduction of ortho-substituents into the acid and basic molecules does not prevent the formation of saltp, provided that the acid is very strong and the base not too weak.In addition, we have attempted to settle a point raised in the earlier paper (ibid., p. 583), namely, the behaviour of various organic bases towards a weak diortho-substituted acid of the type of s-trimethyl- benzoic acid. The acid forms a well-defined, crystalline salt with a strong base such as benzplamine, and also dissolves readily in 33 per cent. aqueous trimethylamine solution, but all attempts to combine it with tertiary bases such as dimethylmesidine, dimethyl-$-cumidine, bromodimethyl-nz-xylidine, and diethy laniline have given negative results; in each case, the unaltered acid crystallised out from the different solvents, even when two or three times the theoretical amount of base was employed. It is clear that the loss of the function of salt formation is not due to the presence of ortho-substitu- ents in the basic molecules, because a simple base like diethylaniline is incapable of forming a salt with this acid.I n order t o determine whether it is due to ortho-substituents in the acid molecule, we have endeavoured t o prepare salts from the same bases and mono-substituted benzoiu acids containing substituent methyl, bromo-, and nitro-groups. The fact that no salts are formed from either nz-toluic acid and dimethylaniline, or p-toluic acid and dimethylaniline or +-cumidine indicate that the presence of ortho-substituents in the acid molecule is not necessary in order to inhibit salt formation. In addition t o the normal salts formed by the unionof one molecule of acid with one of the base, we have obtained a number of acid salts formed by the union of two molecules of acid with one of the base, ag, for example, the acid salts derived from $-cumidhe and o-toluic acid, dimethyl-q-cumidine and fimitrobenzoic acid, $-cumidine and fit-nitro- benzoic acid, dimethyl-$-cumidine and tribromoaminobenzoic acid, and dimethyl-$-cumidhe and o-nitrobonzoic acid.These compounds are similar to the acid potassium and ammonium salts derived from sub- stituted benzoic acids, which have been recently described by Farmer (Trans,, 1903, 83, 1440). According to this investigator, the probable constitution of these acid salts is represented by the formula R-co>O<g, and the molecule contains a quadrivalent oxygen R-CO.0 atm. With such a formula, it is quite possible t h a t inhibition236 SUDBOROUGH AND ROBERTS : through stereochemical influences should occur.If ortho-substituents retard or prevent the formation of additive compounds involving the carbonyl residue of the substituted benzoic acid (compare Trans., 1899, '75, 581), it is quite possible that ortho-substituents would also prevent the conversion of the bivalent oxygen atom of the hydroxyl group into a quadrivalent oxygen. We might therefore expect that diortho-substituted benzoic acids mould not give rise to acid salts. Although we have not particularly investigated this question, we have met with only one such salt derived from dimethyl-q-cumidine and tribromoamiaobenxoic acid. The salts were usually analysed by determining the amount of acid present according to the method previously described (Zoc.cit., 584). We have found that the majority of the salts can also be analysed by titrating their alcoholic solutions with standard baryta when phenol- phthalein is used as indicator. The presence of the organic base does not interfere with the titration, and good results are obtained. For example, 20 C.C. of N/20 oxalic acid required 15.4 C.C. of baryta solution, and 20 C.C. of N/20 oxalic acid mixed with 6 drops of dimethyl- +ximidine also required 15.4 C.C. of the same alkali solution. EXPERIMENTAL. A. Prepavation of Bases. C,H,Me,*NMe, [Me, : NMe, = 2 : 4 : 5 : 1 3, was obtained by a method somewhat similar to that employed by Noelting in the preparation of dimethyl-m-xylidine (Bey., 1891, 24, 563); 20 grams of +-cumidbe melting at 6S0, 70 grams of methyl iodide, 47 grams of sodium carbonate, and 500 grams of water were heated in a reflux apparatus until all the iodide had disappeared; caustic potash solution was added and the base extracted with ether. The ethereal solution, when dry, gave 17 grams of a colourless oil distilling a t 219q which was characterised by the formation of its pkatinichloride.Dimetbyl-+-cumidine, 0.3808 gave 0.1008 Pt = 26.4'7. (C,H2Me,*NMe,),,H1PtCI, requires 26.47 per cent. Dimethylmesidine [Me, : NMe, = 2 : 4 : 6 : 11, obtained in a similar manner from mesidioe, was freed from secondary base by heating in a sealed tube with methyl iodide and dry magnesia a t 100' (Fischer, Ber., 1900, 33, 1968). Bromo-m-xylidine [Me, : NH, : Br = 2 : 4 : 1 : 61, melting at 46--47O, was obtained by Fischer's method (ibid., 1971) and transformed into the tertiary base by the method employed in the case of mesidine.Diethyl a- and /3-nsphthylamines were prepared by Morgan'sDIORTHO-SUBSTITUTED BENZOIC ACIDS. PART V. 237 process (Trans., 1900, 77, 823). OF the acids employed, the 2 : 4 : 6-tribromo-3-aminobenzoic acid was prepared by the method already described (Trans., 1899, 75, 589), the s-trinitrobenzoic acid was kindly presented by the Chemische-Fabrik Griesheim, and the 8-trimethylbenzoic acid was prepared from mesitylglyoxylic acid (Meyer, AnnaZen, 1888, 246, 139). Van Scherperizeil (Bee. traw. chim., 1900, 19, 380) recommends heating the glyoxylic acid with concentrated sulphuric acid until no more carbon monoxide is evolved, but we find that in order to obtain a good yield the temperature should be carefully regulated, the best yields being obtained when the temperature is kept about 40"; above 50°, carbon dioxide begins to be evolved, and the trimethylbenzoic acid itself is destroyed.I n one experiment, we obtained 2.05 grams of trimethylbenzoic acid from 2.5 grams of the glyoxylic acid (m. p. 152'). The magenta coloration produced when glyoxy lic acid is dissolved in concentrated sulphuric acid is extremely characteristic. 33. S a l t s f r o m s - T r i n i t p . o b e n , x o i c A c i d . (~02),C6H,~C02*~HMe20C6H,Me3, slowly separated in the form of snow-white, small needles on mixing alcoholio solutions of the acid and base ; it crystallises from benzene in small needles, dissolves in hot water, alcohol, or chloroform, but is insoluble in carbon disulphide or light petroleum.Like the salts of trinitrobenzoic acid already described, it has no definite melting point ; on heating, it melts and decomposes, yielding a dark red oil ; the temperature at which this occurs varies considerably with the rate of heating. When heated fairly rapidly, it begins to turn dark at 1 1 6 O and melts at 120°. DirnethyZ-$-cumidine 2 : 4 : 6-T~initrobenxoate.--The salt, 0,7179 gave 0.438 trinitrobenzoic acid = 61.0. C,8H200,N, requires 6 1-2 per cent. Dimethylmeoidine Trinitrobenxoate, (NO,),C,H,* CO, NHMe,*C,H,Me,. -The salt, obtained in the form of small needles when alcoholic solu- tions of the two components were mixed, begins to darken at 105O, and is completely decomposed a t 116-117'; it dissolves in the ordinary solvents on heating, but is decomposed by some of these, more especially water or alcohol.When boiled for a short time with alcohol, it yields crystals melting a t 122"; these proved to be s-trinitro- benzene, as with diethyl-P-aaphthylamine they furnished purplish- black needles melting a t 125-1 16". 0.8867 gave 0.5392 trinitrobenzoic acid = 60% VOL. LXXXV. R C,,H,008N4 requires 61.2 per cent.238 SUDBOROITGH AND ROBERTS : Bromo-m-aylidine ~rinil~~~berf~~oate,(NO,),~C~H,*CO,~NH,*C,H,Me~-This salt, which was prepared by mixing alcoholic solutions of the acid and base, slowly separated in yellow, prismatic crystals which begin t o change colour at 1 30°, are completely decomposed at 150° ; it is soluble in warm alcohol. 0.7154 gave 073986 trinitrobenzoic acid = 55.6.C15H1308N4Br requires 56.2 per cent. When the alcoholic solution was boiled for a short time, it acquired a red colour, and small, red needles melting at 104-105° were obtained on cooling. This substance, which is the additive compound, C6H2Me,Br*NH,,C,H,( NO,) 3, was a1 so obtained by mixing alcoholic solutions of bromo-m-xylidine and trinitrobenzene. 0.3824 gave 0.1955 trinitrobenzene = 51.1. It dissolves readily in ether, benzene, chloroform, and warm alcohol, Bromodimethyl-m-xylidine T&nitrobenzoccte, C,,H,,O,N,Br requires 51.5 per cent. (NO,),-C,H,* C02*NHMe,* C,H,Me,Br. -The salt, which was obtained from alcoholic solutions, begins to change colour a t 1 0 5 O , and is completely decomposed at 10s'.0.6506 gave 0.3435 trinitrobenzoic acid = 52.8. C,7H,p0,N,Br requires 53.0 per cent. D i e t ~ y l - a - . n ~ ~ h t r y Z ~ ~ i n ~ s-TrinitroEerfLzoate.-This salt separated in the form of flat, slightly yellow plates when warm alcoholic solu- tions of the acid and base were mixed. The mother liquor, when kept for several days, deposited bright red needles melting at 95O. These consist of diethyl-a-naphthylamine-trinitrobenzene (Trans., 1903, 83, 1338), and if the alcoholic solution of the salt is boiled, it is completely converted into this compound. The salt decomposes at 117-118° and dissolves in alcohol, benzene, or chloroform. 0.5420 gave 0,3024 trinitrobenzoic acid = 55.S. C,lH,,O,N, requires 56.3 per cent.D iethyLP-naphth y lamine s- Frinitrobmzoate. -T h i s salt, which was also obtained from alcoholic solutions of the acid and base, begins to darken at 1 2 5 O and is completely decomposed at 132'; it dissolves in hot water, alcohol, benzene, and chloroform, and, like its isomeride, shows a tendency to lose carbon dioxide and to form diethyl-p-naphthyl- amine-trinitrobenzene melting at 116' (Trans., 1903, 83, 1340). This transformation is practically quantitative when an alcoholic solution of the salt is boiled for a short time.DIORTHO-SUBSTITUTED BENZOIC ACIDS. PART v. 239 0.4736 gave 0.2641 trinitrobenzoic acid = 65.7. C,,H,,O,N, requires 56.3 per cent. The changes which these salts undergo when their alcoholic solutions are boiled, illustrate remarkably well the difference in stability between the additive compounds of s-trinitrobenzene with aniline derivatives and those produced with naphthylamine derivatives. Dimethyl-+-cumidine and dimethylmesidine, when boiled in alcoholic solution, yield the free base and s-trinitrobenzene, whereas diethyl-a- or -P-naphthylamine trinitrobenzoate yields the corresponding coloured additive compound of the base with s-trinitrobenzene.I n order t o determine whether dimethyl-$-cumidine forms a definite compound with s-trinitrobenzene, we have made a number of experi- ments with different solutions, but in all cases, even when excess oE the base was employed, crystals of unaltered s-trinitrobenoene were deposited. A coloured compound, obtained by crystallising s-trinitro- benzene from the base and allowing the crystals to dry in a desiccator containing a little of the free base, separated in small, bright red needles with no definite melting point, as when warmed they readily give up the base, leaving a colourless residue.When exposed to the air, the substance readily leaves a colourless residue of s-trinitrobenzene. 0.5 gave 0.2831 trinitrobenzene (m. p. 122') = 56.62. C,7H2,06N, requires 56.6 per cent. An additive compound of trinitrobenzene and ybcumidine, obtained by mixing together alcoholic solutions of the constituents, crystallises in long, slender, purple-brown needles melting at llS0, and is much more stable than the compound of aniline with trinitrobenzene. NE 15.9. 0.3036 gave 41.3 C.C. moist nitrogen at 17" and 765 mm.C,,HI,O,N, requires 16.09 per cent. C. S a l t s f3-om 2 : 4 ; 6-T~1~ibromo-3-aminobenxoic A c i d . Biinethylrnesidine 2 ; 4 : 6-Trib~omo-3 -aminobenxoate, NH,*C,HBr,* C02-N HMe, C, €&Me,. -The salt was obtained by dissolving the acid in ether, adding a slight excess of dimethylmesidine, and allowing the greater part of the ether to evaporate; it crystallised from water in small needles, softening at 147" and melting at 163-165'. 0-8002 gave 0,5677 tribromonminobenzoic acid = 70 9. C,8H210,N2Br, requires 69.6 per cent. The salt readily dissolves in alcohol, and also t o a less extent in ether, benzene, chloroform, or hot water, but is very sparingly soluble in light petroleum. R 2240 SUDBOROUGH AND ROBERTS : DimethyE-11/-cumidilze Hydrogen ~ribromoaminobentzoate, 2(NH,*CBHBr,*C02H),C,,H,Me3*NMe2. -When the acid and base were mixed in moleuulap proportion, the mixture did not set t o a solid mass, but remained pasty, and when the mixture was crystallised from benzene, in which it is only moderately soluble, nodular aggregates of small, colourless needles melting at 145.5-1 46' were obtained.When titrated in alcoholic solution with standard baryta, using phenol- phthalein as indicator, 0.4062 required 17-85 C.C. of N/20 Ba(OH),. The amount required for the normal and acid salts, NH,*C,HBr,* CO,H, C,H,Me,- NMe, and 2(NH,*C,HBr,*C0,H),C,H,Me3*NMe,, are 15 -48 and 17.84 C.C. respectively. When decomposed with dilute hydrochloric acid and extracted with ether, 0.6024 gave 0.505 tribromoaminobenzoic acid = 83.8 per cent.NH2*C6HBr3* C02H,C6H2Me,*NMe2 and 2( NH20C,HBr,*C02H),C6H,Me,oNMe2 require 69.6 and 82.1 per cent. respectively. The same compound, obtained in the form of glistening, well- developed prisms by dissolving molecular quantities of the acid and base in alcohol, adding water, and allowing the solvent to evaporate slowly, melted a t 146". 0.3 required 13.24 C.C. of N/20 Ba(OH), for neutralisation ; 2 mols. acid + 1 mol. base requires 13.17 C.C. The melting point was not affected by crystallisation from benzene. We have attempted t o prepare the normal salt (1 mol. acid + 1 mol. base) by dissolving a mixture of the acid with twice its weight of base in hot benzene and allowing the solution to crystallise and also by dissolving the acid (1 mol.) in dry ether, and adding an ethereal solution of the base (4 mols.), but the same acid salt melting at 1 4 6 O was obtained in both cases.Attempts to obtain salts with diethyl-a- and -P-naphthylamines led to the formation of gummy, uncrystallisable masses. D. SaEts f r o m 2 : 4 :6-TrimethyZbe~nxoic Acid. Benxpkamine 2 : 4 : 6-t?.in.zet?~yEbenxoate is obtained by separately dissolving equivalent amounts of the acid and base in light petroleum and mixing the solutions; it crystallises from benzene in small, feathery, colourless needles melting at 1 6B0. 0,3052 gave 0.1868 trimethylbenzoic acid = 61.2. Salts could not be obtained with this acid and any of the following C17H2,02N requires 60.5 per cent.DIORTHO-SUBSTITUTED BENZOIC ACIDS. PART v. 24 1 bases : dimethylmesidine, dimethyl-+-cumidine, bromodimethyl-m- xylidine, diethylaniline.E. S a l t s of t h e T h r e e Toluic Acids. BenzyZamine o-toluate, obtained from light petroleum solutions of the acid and base, crystallises from benzene in small prisms melting at 146'. It dissolves readily in water, and to a moderate extent in light petroleum. 0.4'716 gave 0.2668 o-toluic acid = 56.5. C,H,Me*CO,*NH,*CH,Pb requires 55-9 per cent. $-Cumidine hydrogen o-toluate, 2 (C,H,Me*CO,H), C,H,Me,* NH,, pro- duced by mixing together ethereal or light petroleum solutions of the acid and base, crystallised from water in long, snow-white, silky needles, softening somewhat at 80' and melting at 82.5'. The same compound is obtained when an excess of the baee is employed. 0.5 gave 0.3315 o-toluic acid (m.p. 105) = 66.3. One mol. acid + 1 mol. base requires 50.18. Two $ 9 ,? ? ) ,, 66.8 per cent. 0.70 gram, when dissolved in sodium carbonate and extracted with ether, gave 0.236 gram of solid base =33*7. The acid salt requires 33.2 per cent. 0.25 gram of the salt, when dissolved in ethyl alcohol, required 24.8 C.C. of 21'/20 baryta solution for complete neutralisation, using phenolphtbalein as indicator. Salts of o-toluic acid with dimethylaniline, dimethyl-$-cumidhe and dimethylmesidine could not be obtained. Definite salts could not be obtained either from m-toluic acid and q-cumidine or dimethylaniline, or from p-toluic acid and the same bases. The acid salt requires 24.6 C.C. F. XaZts fi*ona t h e T h r e e N i t r o b e n x o i c A c i d s .J/-Cumidine o-Nitrobenxoate,-This salt; was obtained in the form of silky needles when ethereal or benzene solutions of the constituents were mixed; when crystallised from water, it formed long, silky, prismatic needles melting a t 133-1 34'. 0.4137 gave 0.2260 o-nitrobenzoic acid = 54.6. 0.3 required 20.1 C.C. of N/20 baryta solution for neutralisation. C16H180,N, requires 55.3 per cent. The normal salt requires 19.9.242 DIORTHO-SUBSTITUTED BENZOIC ACIDS. PART V. Dimethyl-+-cumidine hydvogen o-ruitrobenxoccte, 2( NO,. C,H,*CO,H),C,H,Me,NMe,, obtained by mixing equal weights of the acid and base in benzene solution and crystallising from this solvent, forms small, colourless crystals melting at 1245. 0.50 required 40.6 C.C. N/20 baryta solution. The acid salt requires 40.2 C.C.t,!d?umidine m-nitro6enzoate was obtained in . the form of long, colourless, feathery needles melting at 129fj-l3O0, when the acid was mixed with twice its weight of +-cumidhe and crystallised from benzene. 0.3 required 20.05 C.C. N/20 alkali. The normal salt requires 19.9 C.C. 0.5 requires 33.4 C.C. The normal salt requires 33.16. The ucid salt, 2(N0,~C6H,*C0,H),C6H,Me,*NH,, was obtained when a mixture of the acid (2 mols,) and base (1 mol.) was dis- solved in benzene and allowed to cool; it crystallises from water in pale yellowish needles melting at 144O." 0.40 required 34.3 C.C. of N/20 alkali. The acid salt requires 34.12 C.C. A mixture of the two salts was obtained when equal weights of the acid and base were crystallised together from benzene or water.The normal salt, when boiled with water, is partially transformed into the acid salt. BirnethyE~-cumidine hydrogen m-niti*obenxoate, obtained when mole- cular quantities of the acid and base are mixed in ethereal solutions, crystallises from benzene in small, colourless prisms melting at 119 * 5 O , dissolves readily in alcohol, ether, or chloroform, but is only sparingly soluble in carbon disulphide or light petroleum, and crystallises from hot water in colourless, flat plates. 0.3566 gave 0.2362 m-nitrobenzoic acid = 66.2. 0.255 required 20.77 C.C. of ~ i / 2 0 alkali. The acid salt The acid salt requires requires 67.2 per cent. 20.52 C.C. q-Cumidine p-nitrobemoate, obtained by mixing ethereal solutions of the acid and base, separates from water in small, feathery crystals melting at 160O. * The salt mentioned previously (Trans., 1899, 75, 595) as melting a t 140-141" must have been the acid salt, but the specimen actually analysed must have been the normal salt melting a t 129*5-130".STUDIES ON THE ELECTROLYTIC OXIDATION OF PHEKOLS. 243 0 3246 gave 0.1872 p-nitrobenzoic acid 1: 57.6. No definite salts could be obtained from dimethylaniline and the ~ 0 2 ~ C , H , ~ C 0 2 ~ ~ H , ~ c 6 ~ , M e 3 requires 55.3 per cent. three nitrobenzoic acids. G. S a l t s f y o m the Three Bromobefixoic Acids. +!I -Cumidine o ~ bromo benzoate, C6H,Br* CO;NH,. C6H,Me3, slowly separates in the form of fine needles from an ethereal solution of the acid and base; i t dissolves readily in benzene, alcohol, or hot water, and crystallises from either light petroleum or hot water i n colourless needles melting a t 106-106*5°. 0.555 gave 0.32863 o-bromobenzoic acid = 59.2. The normal salt requires 59.8 per cent. Ob5O required 29-87 C.C. of A720 alkali. The normal salt requires 29.76 C.C. $-Cumidine m -6romobenxoute crystallises from benzene in long, felted needles, melts at 98*5", and is readily soluble in chlorolorm, alcohol, warm ether, or benzene, 0.50 required 29.7 C.C. of N/20 alkali, The normal salt require8 29.76 C.C. No definite salts could be obtained from dimethylaniline and the three bromobenzoic acids, and an ethereal solution of t,b-cumidine and the pbromo-acid slowly deposited crystals of the free acid melt- ing at 251O. CHEMICAL LABORATORIES, UNIVERSITY COLLEGE OF WALES, ABERPSTWYTH.