1046 CHATTAWAY AND ORTON: A SERIES OF CIV.-A Series of Substitzctecl Nitrogen Chlorides and tlzei?. Relation to the Substitu.tion of Halogeit in Anilides cantl Anilines, By F. D. CHATTAWAY and K. J. P. ORTON. COMPOUNDS in which a halogen is attached to nitrogen have been little studied, although a number of isolated examples are known. Among such may be mentioned nitrogen chloride, nitrogen iodide, the chlorine derivatives of tt few aliphatic amines (Tscherniak, Ber,, 1876, 9, 143; Norton and Tscherniak, Bull. Xoc. Chim., 1878, [ii], 30, 106), of acetanilide (Bender, Ber., 1886, 19, 2272), of benzamide (ibid.), of succinimide (ibid.), and of phenylnitramine (Bamberger, Ber., 1895, 27, 376), the bromine derivatives of the aliphatic amides (Hofmann, Ber., 1882, 15, 407 and 752), and of benzanilide (Line- barger, Aww.Chem. J., 1S94, 16, 218), the iodine derivatives of succin- imide (Bunge, Annalen Su&., 1870, 7, 119), of acetamide (Seliwanow, Ber., 1893,26, 985), and of forxuanilide (Comstock and Kleeberg, Amer. Chem. J., lS90, 12, 500), pzethylic and ethylic chlorimidocarbonates (Sandmeyer, Bey., 1886, 9, 862), and the chlorimide obtained from the oxime of benzophenone (Beckmann, Ber., 1886,19, '388). Some years ago, the attention of one of us was drawn to a substance obtained by Witt (Ber., 1875, 8, 1226) by the action of hypochlorous acid on acetanilide, and this was employed in a simple process for prepar- ing metadichlorobenzene in quantity (Chattaway and Evans, Trans., 1896, 69, 848). In continuation of our work upon nitrogen iodide, werecently took up the study of this compound, and have obtained a series of substituted nitrogen chlorides which undergo remarkable intramolecular transformations and are of extreme interest from their bearing on the theory of substitution.We find that, by the action of hypochlorous acid under carefullySUBSTITUTED NITROGEN CHLORIDES. 1047 regulated conditions, compounds of this nature can be obtained from most substances containing hydrogen attached to nitrogen. I n the present paper are considered the disubstitution products of nitrogen chloride containing formyl, acetyl, or benzoyl, together with a phenyl or chlorophenyl residue. These form a well-defined group of compounds, phenyl acetyl nitrogen chloride, discovered by Bender, being one of the simplest members.They are readily obtained from the corresponding formanilide, acetanilide, or benzanilide by inter- action with hypochlorous acid. R*CO*NR’H i- HOCl= R*CO*NR’Cl + H,O. They are all stable compounds of low melting point, crystallising well in large, colourless, transparent prisms or plates, and reacting very readily with alcohol, hydrochloric acid, hydrocyanic acid, potassium iodide, and alkaline hydrogen peroxide. ( a ) With alcohol, the anilide is reformed, and aldehyde, ethylic chloride, and other substances are liberated, ethylic hypochlorite, which breaks up into aldehyde and hydrochloric acid, being probably first produced. R*CO*NR’Cl + C,H,* OH = R*CO*NR’H + C,H,* OC1. CH,* CH,* OC1= HC1+ CH,*CHO. ( b ) With strong hydrochloric acid, chlorine is set free and the anilide regenerated.I n some cases, however, under the influence of the acid, a portion of the compound undergoes an intramolecular transformation similar to that which takes place when it is heated. (c) Withd hydrocyauic acid, the anilide and cynnogen chlorido are formed . R*CO*NR’Cl + HC1= R*CO*NR‘H + GI,. R*CO*NR’Cl + HCN = CNCl + R*CO*NR’H. ( d ) With an acid solution of potassium iodide, the anilide is repro- duced and iodine liberated. R*CO*NR’Cl+ 2HI = R*CO*NR’H + HC1+ I,. (e) With alkaline hydrogen peroxide, oxygen is liberated and the anilide reformed. R*CO*NR’Cl+ H,O, = R*CO*NR’H + HC1+ 0,. When warmed with various anilines, a vigorous reaction takes place, and the anilide is regenerated, whilst a chlorine substitution product of the aniline is formed, These compounds all undergo a remarkable isomeric change.When an unsubstituted phenyl residue is attached to the nitrogen, the1048 CHATTAWAY AND ORTON : A SERIES OF chlorine atom associated with the latter is transferred to the ring in the para-position relatively t o the substituted amido-group ; a change which was first observed by Bender in the case of phenyl acetyl nitrogen chloride. When the para-position is occupied, the chlorine is transferred to the ortho-position, and when both the para- and one ortho-position are occupied, to the remaining ortho-position. The behaviour of these substituted nitrogen chlorides resembles, i n this respect, that of the substituted nitramines and sulphamates (compare Bamberger, Ber., 1894, 27,359 ; 1895,28,399 ; 1897,30,1248,2247). In these transformations of the phenyl acyl nitrogen chlorides, how- ever, we have never, so far, observed t,he simultaneous formation of ortho- and para-compounds.The latter are formed when possible to the exclusion of the former. From the existence and behaviour of these compounds, it seems certain that, in the chlorination of nnilides or anilines, the halogen first replaces a hydrogen atom of the NH- or NH2-group, the nitrogen chloride thus formecl becoming subsequently transformed into an isomeric substituted anilide or aniline. The series of changes which takes place when an anilide is chlorinated may be illustrated by the following formuh, further chlorination proceeding on similar lines with the production, although less readily, of the 2 : 4 : 6-trichloro- anilide (p.1052). R*CO*NH R*CO*N*Cl R*CO*NH R*CO*N*CI R*CO*NH The study of these compounds affords, we believe, an explanation of many well-known facts. In the chlorination of anilides and anilinea, not only is the substitution unusually easy, but the entering halogen atom always takes up a position in the nucleus para- or ortho- relatively to the NH- or NH,-group, provided these positions are unoccupied, other substituting groups which may be present having apparently no directing influence. If, as me believe, the action of chlorine or a chlorinating agent on an anilide and probably on an aniline leads to tho production of a nitrogen chloride which then undergoes isomeric change, the position taken up by the halogen in the nucleus depends chiefly on the relation of the nitrogen atom to the carbon atoms of the riug, whilst other groups present have no oppor- tunity of exerting any specific orientating influence, It is possible that in every case an addition of hypochlorous acid to the amide precedes the formation of these nitrogen chlorides, and our work on nitrogen iodide, which proves that the latter is formed fromSUBEITITUTED NITROGEN CHLORIDES.1049 ammonium hypoiodite, supports this view. We have not, however, so far observed any indication of the existence of such compounds. We have prepared substituted nitrogen bromides which undergo exactly similar intramolecular transformations, and show that bromi- nation is effected in an exactly analogous manner. Phenyl Pormyl Nityoyen Chloyide, C,H,*NCl*CHO.This substance is best prepared by adding the calculated quantity of a solution of bleaching powder t o a saturated solution of formanilide containing an excess of potassium bicarbonate, or to formanilide suspended in a solution of the bicarbonate. It separates as an oil which slowly solidifies, and calcium carbonate is at the same time formed. The bleaching powder solution is run in very slowly, the mixture being vigorously stirred by the aid of a turbine; after the addition, stirring is continued for a n hour or so to ensure complete con- version of the anilide into the chloride. The solid is then filtered off, and the phenyl formyl nitrogen chloride extracted by chloroform, This solution must be rapidly evaporated by a blast of air, as the impure nitrogen chloride very quickly changes into parachloroformanilide if the solvent is distilled off or allowed to volatilise.The reddish- crystalline mass which separates from the chloroform is purified by recrystnllisation from a mixture of chloroform and light petroleuni. It is thus obtained in long, lustrous prisms terminated by pyramids. The melting point is 47'. 0,1934 gave 0.1790 AgC1. 01 = 22.89. C,H,NOCl requires C1= 32.83 per cent. It is very soluble in chloroform, carbon bisulphide, and benzene, but only sparingly so in light petroleum. It gives all the characteristic reactions of the nitrogen chlorides with hydrochloric acid, a solution of potassium iodide, and an alkaline solution of hydrogen peroxide, whilst with alcohol, formanilide is regenerated.On keeping for a few days, it shows signs of decomposition, developing n chlorous smell and becoming coloured. I n a short time, a considerable portion is found to have changed into parachloroformanilide. This change takes place very rapidly if the substance is warmed ; great heat is developed, and the action may become violent if more than cz small quantity is employed. This transformation takes place quietly when the chloride is warmed under water at about 50'; the transparent oil appears to boil, and considerable heat is developed. The darker oil which is formed solidifies, as the water cools, to a mass of cqstals of para- chloroformanilide (m. p. 102O), which is quite pure after crystallising once from water. The yield is quantitative. VOL.LXXV, 4 A1050 CHATTAWAY AND OBTON: A SERIES OF Parachlorophenyl Fovmyl Nitrogen Chloride, C6H4Cl *NCl*CHO. This chloride is prepared and purified in a manner exactly similar t o that used in the case of the preceding substance. It crystallises well i n long, colourless prisms terminated by pyramids, melts at 95-96', and has all the general characteristics of the group. On keeping, it soon becomes slightly pink, and slowly changes at the ordinary temperature into 2 : 4-dichloroformanilide. On heating at 200-210° for a few minutes, it is completely and quietly converted into the dichloroformanilide. 0.1954 gave 0.2950 AgC1. C1= 37.33. C7H,NOCI, requires C1= 37.37 per cent. 2 : 4-Dichlorophenyl Formy1 Nitrogen Cldoride, C,H,CI,*NCI*CHO. Formanilide can be directly converted into this compound by adding a solution in acetic acid to a large excess of a solution of bleaching powder and finally heating on the water-bath. The oil thus obtained is difficult to purify, as it generally contains a little unchanged para- chlorophenyl formyl nitrogen chloride.It is preferable therefore t o start with 2 : 4-dichloroformanilide. A solution of the latter in acetic acid is added t o an excess of a solution of bleaching powder, and the mixture heated and thoroughly shaken until the oil which separates becomes transparent. The chloride solidifies on cooling, and is re- crystallised from a mixture of chloroform and light petroleum. It forms thick, colourless, glistening plates melting at 44O. 0.2018 gave 0.3846 AgC11. C1= 47-12.C7H4NOCI, requires C1= 47.44 per cent. Although this chloride develops a marked chlorous smell on keeping, decomposition takes place only to a very small extent. On heating, no sudden change occurs, but at about 200' chlorine is evolved, and the residue consists chiefly of 2 : 4-dichloroformanilide. Phen y I A cet y I Nitrogen Ch lovide ( Acet oclJoranilide), C,H,*NCI* CO*CH,. This compound was first obtained by Bender (Zoc. cit.), who prepared it by adding bleaching powder to a saturated solution of acetanilide in water acidified with acetic acid. This method is unsatisfactory, as it is very difficult to obtain the nitrogen chloride in quantity," and as the slightest excess of acetic acid causes the complete conversion of * Castor0 (Gaxxetla, 1898, 28, ii, 312) recently failed to confirm Bender's observations.SUBSTITUTED NITROQEN CHLORIDES.1051 the substance into pmachloracetanilide. It is, however, very easily and rapidly prepared by the action of bleaching powder on acetanilide suspended in excess of a solution of potassium bicarbonate. The chloride is extracted with chloroform, and recryetallised from a mixture of chloroform and light petroleum. The yield amounts to 96 per cent. of the theoretical. It crystallises in large, transparent, apparently rectangular plates, not in needles as stated by Bender, and melts at 91'. It dissolves readily in dilute acetic acid, long needles of para- chloracetanilide separating as the liquid cools. This conversion also takes place slowly on keeping, being nearly complete in 2 or 3 weeks.On treatment with hydriodic acid, a reaction occurs analogous t c that which we have observed in the case of nitrogen iodide; one atom of halogen attached to nitrogen, in this case the whole of the halogen, liberates two atoms of iodine in accordance with the equation given on p. 1047. This reaction is characteristic of the nitrogen halogen link- ing. The determination of the iodine liberated is readily effected by dissolving the substance in chloroform, and shaking with a solution of potassium iodide acidified with acetic acid. 0.3952 liberated I = 46.02 C.C. N/10 iodine. C1, as :NC1, = 20.64. C6H,*NC1* CO*CH, requires C1, as :NU, = 20.9 per cent. Parachlorophenyl Acetyl Nitvogen Chloride, C,H,Cl*NCl* CO*CH,. This substance is prepared from parachloracetanilide in the manner described above.I n crystalline form, it resembles the phenyl acetyl nitrogen chloride very closely, and possesses all the general characters of the group, The melting point is 82'. C1= 34.61. C,H7NOCl, requires C1= 34.76 per cent. 0.1942 gave 0.2718 AgC1. The estimation of chlorine attached to nitrogen gave the following 0.4426 liberated I = 42.9 C.C. N/10 iodine. C6H,C1* NCl* CO*CQ, requires C1, as :NCl, = 17.38 per cent. A t 165O, transformation into 2 : 4-dichloracetanilide takes place almost explosively. This change occurs quietly and quantitatively when the substance is heated under water. It first melts, and, when the water begins to boil, becomes converted into its isomeride with considerable evolution of heat and apparent ebullition of the oil, which afterwards solidifies, results : C1, as :NCI, = 17.18.1052 CHATTAWBY AND ORTON: A SERIES OF 2 : 4-DiclJo~ophmyl AcetyZ Nitrogen, Chlwide, C6H,Cl,* NC1* CO*CH,. This substance can be prepared either directly from acetanilide, or, better, from 2 : 4-dichloracetanilide, The solution of acetanilide in acetic acid is poured into a large excess of a solution of bleaching powder ; the precipitated anilide gradually clots together as the action proceeds, and finally, when the mixture is warmed on the water-bath, forms a, clear, yellow oil which often solidifies on cooling t o a hard mass of the impure chloride.To remove 2 : 4-dichloracetanilide, and more especially parachlorophenyl acetyl nitrogen chloride, which are probably present as impurities, the product is redissolved in acetic acid and again poured into a solution of bleaching powder, the mixture being warmed; it is then recrystallised from a mixture of chloroform and light petroleum. It is simpler to pour a solution of the pure 2 : 4-dichloracetanilide in acetic acid into excess of bleaching powder and apply heat.The oil which separates is shaken with n warm acid solution of bleaching powder t o complete the conversion of the anilide ; it solidifies when cooled, and is finally purified by recrystal- lising from a mixture of chloroform and light petroleum, It melts at 7S0. The yield is quantitative. 2 : 4-Dichloropheayl acetyl nitrogen chloride resembles the two pre- ceding nitrogen chlorides very closely in appearance, properties, and behaviour towards reagents ; it is, however, more stable, and can be kept practically unchanged for months.It is decomposed, but very gradually, by heating under boiling water. Its solution in acetic acid evolves chlorine slowly when heated, but many hours boiling are required before it is completely converted into 2 : 4-dichloracetanilide. In a sealed tube, a nearly quantitative yield of 2 : 4 : 6-trichloracetanilide is obtained after 4 hours heating with acetic acid at 145O. Heated alone, it decomposes a t 200°, evolving chlorine and forming a dark, coloured mass, consisting largely of 2 : 4-dichloracetanilide. This compound has been carefully analysed and studied, as it appears to be the main constituent of a substance obtained by Witt (Ber., 1875, 8, 1226) by the action, at 80°, of excess of bleaching powder solution on a solution of ncetanilide in acetic acid.The oil thus formed could not be solidified, and was regarded as an additive product of hypochlorous acid and 2 : 4-dichloracetanilide, C,H,Cl,*NH* CO*CH, + H001. His analyses agree only approximately with this formula; for example, the carbon is 5 per cent. too high, and great difficulty was experienced in the combustion owing to a burst of chlorine which occurs when the substance is first heated, a peculiarity also observed by us in the case of the nitrogen chloride. We estimated the carbon and hydrogen qnite easily when the substanceSUBSTITUTED NITROUEN CHLORIDES. 1053 was mixed with a long layer of lead chromate, with the following results : 0.3194 gave 0.3238 CO, and 0.0515 H,O.0.3452 0516 liberated I = 43.7 C.C. N/10 iodine. C,H7NOCI, rcquires C = 40.26 ; H = 2.54 ; N = 5.88 ; C1= 44.65 ; C1, as :NCI, = 15.02 per cent. Two determinations of the molecular weight by Raoult's method were made, using 10 grams of benzene as solvent. 0.2010 lowered the freezing point 0.42". C=40*23 ; H=2*63. ,, 1'7.2 C.C. nitrogen at 13' and 756 inm. N=5.86. 0.1606 ,, 0.3906 AgC1. C1= 44-75. C1, as :NCl, = 14.88. 11. wt. = 234.5. 0.6442 ,, 9 , ,, 1.295'. M. wt. = 243.7. C,H,NOCl, requires a molecular weight of 238.45. The behaviour of Witts's oil is identical in every respect with that of our crystalline 3 : -l--dichlorophenyl ncetyl nitrogen chloride, When prepared by his method, it frequently solidifies with the greatest difficulty, owing to the presence of impurity, probably parachloro- phenyl acetyl nitrogeii chloride, which can only be removed by the treatment described above.Phenyl Benxoyl iVitroge?b Chloride, CGH,*NC1* CO* C,H,. This substance is prepared by the general method from benzanilide and bleaching powder in the presence of potassium bicarbonate, but the reaction takes place less readily than with the formyl and acetyl compounds. It crystallises in colourless plates from a mixture of chloroform and light petzoleum, and melts at 77". 0.2014 gave 0.1228 AgC1. On heating the melted chloride to 130-130°, benzoyl chloride is given off, whilst a portion is converted into parachlorobenzanilide, The latter change is brought about quantitatively if the nitrogen chloride is heated under water for some time, C1= 15-0s.C,,HI,NOC1 requires C1= 15.33 per cent. ParachlorophewJ Benxop l Xt/*oge?a Chloride, We have not succeeded in obtaining this substance pure, as para- chlorobenzanilide is not attacked at the ordinary temperature by hypoclilorous acid. A t 70-80°, either in the presence of potassium bicarbonate or acetic acid, a reaction takes place, but at this tempera- ture the chloride becomes partly converted into 2 : 4-dichlorobenz- anilide, which in turn forms, with the hypochlorous acid, 2 : Cdichloro- VOL, LXXV. 4 B1Og!i4 RYAN : SYNTHETICAL PREPARATION OF CILUCOSIDES. phenyl benzoyl nitrogen chloride. On extracting the product with chloroform, an oil is obtained which solidifies only with great difficulty. Analysis of the recrystallised product showed that it consisted of about 30 per cent. of parachlorophenyl benzoyl nitrogen chloride, together with 70 per cent. of 2 : 4-dichlorophenyl benzoyl nitrogen chloride. 2 : 4-Dic?~Zoro~,?~e~t~Z Bemoy Z Nityogen C?L l o d e , C,jH5CI,*NC1*CO*C,H5. This compound, like the corresponding formyl and acetyl derivatives, can either be obtained directly from benzanilide or from 2 : 4-dichloro- benzanilide, by the action of bleaching powder in the presence of acetic acid a t a temperature of from 80-90". It resembles other members of the group in appearance and properties, and melts at 86". On heating at 150--300°, benzoyl chloride is evolved and a tarry mass is left from which 2 : 4-dichlorobenzanilide can be isolated. 0.2096 gave 0.3993 AgCl. 01= 35.3. C1,H,NOCl, requires C1= 35.44 per cent, We have obtained similar compounds from ninny other su bstitutecl anilides, from secondary amines, and from other substances in which hydrogen is attached to nitrogen, and me desire to reserve the in- vestigation of these compounds. We have also obtained substituted nitrogen bromides resembling the nitrogen chlorides very closely in properties. CHEMICAL LABORATORY, ST. BARTNOLOMEW'S IIOSPITAL, E. C,