ZOO TlTHERLEY : 8-PHENYL-1 : 3-BEKZOXAZINE-4-ONE.XXII. -2 -Phen yl- 1 : 3 -benzoxaxine-4-one,BY ARTHUR WALSH TITHERLEY.IN his criticism of the metoxazone theory of labile isomerism in theacylsalicylamide group, Auwers (Bey., 1907, 40, 3506) attributed thechange of N-benzoylsalicylamide (I) to 0-benzoylsalicylamide (111) (byboiling with acetic acid) to the intermediate formation of the unsatur-ated ring (11) by loss of water, which then immediately decomposedthis hypothetical ring, yielding the 0-benzoyl derivative, thus TlTHERLEY : 2-PHENYL-1 : 3-BENZOXAZINE-4-ONE. 201This view was shown to be untenable by Titherley and Hicks(Trans., 1909, 95, 908), but it appeared desirable to effect if possiblethe synthesis of this ring-compound in order that a study of itsproperties might be made in relation to the derivatives in the acyl-salicylamide group.Several attempts were made to prepare it fromphenylbenzometoxazone," C 6 H 4 < ~ ~ ~ ~ & , h , by eliminating two hydro-gen atoms, and from 0- and N-benzoylsalicylamides by the action ofdehydrating agents, like phosphoric oxide and zinc chloride, butwithout success.Eventually the compound was synthesised from phenyl salicylateand phenylbenzamidine by loss of phenol and aniline, thus :c6H4<O--C 'O*p Ph + PhOH + PhNH,,and i t was later found that the same ring-compound could be obtainedwhen special precautions were observed, by the action of hydrogenchloride on either N- or 0-benzoylsalicylamide at temperaturesbetween l l O o and 140O.These independent syntheses, which arediscussed later, and the properties of the compound place its consti-tution beyond doubt, and i t is, therefore, 2-phenyl-1 : 3-benzoxazine-4-one.The compound has the same melting point, 1 0 6 O , as that of itsisomeride, benzoylsalicylnitrile, CN*C6H4* OBz, from which, however, itdiffers greatly. It is markedly additive in properties, and its moststriking addition is with water. With pure water or dilute alkalis itis almost unaffected, but in presence of hydrogen ions it rapidly addsone molecule of water, a quantitative yield of N-benzoylsalicylamidebeing produced. This simple circumstance finally disposes of AUWBL'S~Shypothetical contention which involves direcb hydrolytic rupture atthe double linking, but is intelligible on the author's theory that theintermediate hydroxy-derivative (IV) is too unstable to exist, asalready frequently indicated, and at once rearranges to N-benzoyl-salicylamide, thus.:The still alternative explanation that hydrolytic rupture takes place* The correct name of this compound is 2-phenyldihydro-1 : 3-benzosazine-4-onebut in view of its use in previous papers, '' phenylbenzometoxazone " is retained inthe present one202 TI’I’HERLEY : 8-PHENYL-1 : 3-BENZOXAZiNE-4-OKE.between the phenolic oxygen atom and the carbon atom to which thephenyl group is attached, is discounted by the comparatively greatstability of the unsaturated ring-compound with alkali, which wouldfavour such a rupture, and the only possible mechanism appears to bethe above, which is identical with that occurring in the conversionof nitriles into amides through the agency of acids.The above mechanism throws some light upon the decomposition bywater in presence of hydrogen ions of certain allied compounds inwhich the double linked C:N pair appews in the molecule. Auwersobserved (Ber., 1904, 37, 2249) that the comparatively stable unsatur-ated ring (V) yields a hydrobromide which in aqueous solution addswater and passes into the salt of o-aminobenzyl acetate (VII), whilstJ.F. Thorpe has shown that all true imino-compounds (VIlI) are at oncedecomposed by dilute acids, giving ketonic derivatives (X). In thesecases a similar mechanism is evidently at work, in which (1) thehydroxyl group attaches itself to carbon arid the hydrogen atom tonitrogen, and (2) the resulting unstable compcund (VI or IX) changeseither by wandering of a hydrogen atom or by loss of ammonia intoa stable derivative :(VIII.) (IX.) (X.That such intermediate hydroxy-compounds are first formed in thecases observed by Auwers and Thorpe appears to follow from thecomplete analogy of these reactions to that of 2-phenyl-1 : 3-benz-oxazine-4-oneY where the production of the intermediate hgdroxy-compound must be admitted on the grounds already stated andin view of’ the author’s previous observations.Contrasting the intermediate hydroxy-derivatives in the author’scase (XI) with that of Auwers (XII) :CO*TH NH*QMe*OH(XI. 1 (XII.)%H4<()-CPh.()H C6H4<CH2.0 2it is clear that when the hydrogen atom of the hydroxyl groupwanders, (XI) yields an N-acyl derivative, whilst (XII) yields an0-acyl derivative. Theoretically, each could yield an 0- or N-acylderivative according to whether the hydrogen atom wandered tonitrogen or to the phenolic oxygen atom, and the final result probablydepends on the conditions which favour one or other of thesealternatives. Thus, in Auwers’s case it is natural that the hydroTITHERLEY : 2-PHEKYL-1 : 3-BENZOXAZINE-4-ONE. 203bromic acid would favour the production of the 0-acyl derivative,which is a base and yielded the salt, rather than the neutral N-acylderivative, which is an amide. I n the author's case neither the0- nor N-benzoylsalicylamide is a base, and the rearrangement yieldsBimply the stable form (N-benzoyl).At t,he same time it must benoted that, although neither of the isomerides is a base in the ordinarysense, 0-benzoylsalicylamide as a primary amide must be more basicthan its isorneride, which is secondary. Experiments which have beenmade on the two isomerides show that the former is more basic, andyields a definite hydrochloride when treated in presence of benzenewith hydrogen chloride. It might therefore be anticipated thathydrogen chloride should effect the rearrangement of N-benzoyl-salicylamide into its isomeride [through the hydroxy-form (Xl)].Such a change has already been partly effected by continued boilingwith pure acetic acid (McConnan and Titherley, Trans., 1906, 89,1331), a fact which supports the supposition that the intermediatehydroxy-form may open in either of the two alternative waysaccording to conditions (acetic acid favouring the production of themore basic 0-benzoyl derivative).Experiments on the action ofhydrogen chloride on N-benzoylsalicylamide in benzene showed thatno change took place in the cold, but that in presence of boilingxylene slow and incomplete rearrangement to 0-benzoylsalicylamideoccurred. As the experiments proceeded, however, it became apparentthat, besides rearrangement, chemical elimination of water was takingplace at about 130'. By using ethylene dibromide or anisole assolvents and distilling off the water as produced, the latter action wasaccelerated, and 3-phenyl-1 : 3-benzoxazine-4-one was formed in largequantities.The latter was even more easily obtained by the action ofhydrogen chloride on 0-benzoylsalicylamide in boiling toluene orxylene if the water formed was continually removed. The hydrogenchloride virtually behaves as a catalyst, since relatively smallquantities were found to be suf€icient to effect the change. It isprobable that the loss of water is not due to direct dehydration,but to the intermediate formation of the chloro-derivative (XIII),which immediately loses hydrogen chloride and apparently is incapableof existence 204 TITHERLEY : 2-PHENYL-1 : 3-BENZOXAZINE-4-ONE.When 2-phenyl-1 : 3-benzoxazine-4-one (11) in benzene solution istreated with dry hydrogen chloride, a white, crystalline compoundcontaining chlorine is precipitated, which on exposure to atmosphericmoisture rapidly passes into N- benzoylsalicylamide, but on treatmentwith aniline yields a yellow, crystalline solid, melting at 106", whichwas found to have the constitution OH*C6H4*CO*N:CPh*NKPh(ealicylphenylbenzamidine). From this behaviour it was supposedthat the ring-compound had added hydrogen chloride, giving thechloro-compound (XIII), or, by rearrangement, the compoundOH*C,H,*CO*N:CCIPh, either of which might be expected to behaveas above with water and aniline (see below).But from the fact thatthe hydrogen chloride additive product in presence of benzeneregenerates the original ring-compound (1.1) with alkali, and from itsother properties, it must be concluded that the product is merely thesalt, C,H,<Co*8'Hc1, which with water would a t once give N-benzoyl-0-GPhsalicylamide (as explained above) ; whilst its curious behaviour withaniline subsequently became clear when it was found that the freering-compound itself readily unites with aniline, producing the aboveamidine derivative according to the following scheme :(XIV.)CO-IjTHc6H4<OH CPh:NPhco-fiC6H4<OH CPh-NHPh'(XVI.)The reaction is reversible, and the anilino-derivative (salicylphenyl-benzamidine) is a tautomeric substance.Its tautornerism comprisesthe three forms (XIV), (XV), and (XVI), of which the last tworepresent ordinary amidine tautomerism, and it is evident that one ofthese two formuls must be given to the substance itself, melting at106O, because of its phenolic properties.It loses aniline, however,below looo, and passes quantitatively into 2-phenyl-1 : 3-benzoxazine-4-oneY provided that the aniline is removed, and the easy reversibilityof the above process would seem to imply that the compound isbordering on the verge of ring-formation (XIV). That the phenolichydroxyl group and the group *CPh:NPh are in intimate stericassociation is confirmed by the difficulty with which the substancTITHERLEY 2-PHENYL-1 : 3-RENZOXAZINE-4-ONE. 205dissolves in dilute aqueous sodium hydroxide, and at first it wassupposed that the ring formula (XIV) must be adopted for thesubstance. This view, however, is inconsistent with the brightyellow colour of the substance, for whilst phenylbenzometoxazone and2-phenyl- 1 : 3-benzoxazine-4-one are colourless, the alkali and ammoniaderivatives of N-acylsalicylsmides are bright yellow, and the phenolicnature of the anilino-derivative is confirmed by the strong ferricchloride reaction it gives.In the synthesis of 2-phenyl-1 : 3-benzoxazine-4-0110 by heatingphenyl salicylate with phenylbenzamidine, t h i s anilino-compound isthe first product of the reaction, and it is owing to its reversibledecomposition, as above, that the unsaturated ring compound (IT)is obtained, and for the same reason the yield of the latter iscomparatively small, thus :(11.1A considerable quantity of diphenylbenzamidine (XVII) is alsoproduced as a, result of a secondary reaction, which also appears to bereversible, between the aniline and salicylphenylbenzamidine, thus :CO-YH PhNHz CO*NH, NPhC6H4<OH CPh:NPh C6H4<OH CPheNHPh.(XVII.)EXPERIMENTAL.2-Phenyl-l : 3-benxoxnxine-4-one.(a) Pmparation, from PIkn,yZ SaZicylate :-(a) 21.4 Grams of phenylsalicylate and 19.6 grams of phenylbenzamidine were heated at 110' forsix and a-half hours, during which aniline and phenol were producedin quantity.On cooling, the resulting viscid, yellow syrup was stirredwith 150 C.C. of water and 100 C.C. of 10 per cent. sodium hydroxideto remove phenol and salicylphenylbenzamidine, and the yellowalkaline solution was decanted off. The remaining syrup was washedwith water, dissolved in 200 C.C. of pure benzene, and the benzenesolution washed with alkali and water.In order to remove anilineand diphenylbenzamidine, the benzene solution was shaken for a fewminutes with abont 300 C.C. of 5 per cent. sulphuric acid, the upperbenzene solution washed with water, and $hen as rapidly as possibl206 TITHERLEY : 2-PHENYL-1 : 3-BENZOXAZlXE-4-ONE.shaken with dilute alkali, to remove the N-benzoylsalicylamide formedduring washing with acid and to neutralise the remaining traces ofacid. (During the former operation some losl; was entailed, because ofthe great instability of the ring compound* in presence of diluteacid.)The alkali extract was found to contain besides phenol, aconsiderable quantity of ssllicylphenylbenzamidine, which was pre-cipitated on adding acid and then decomposed, giving N-.benzoylsalicyl-amide. The acid extract contained aniline and diphenylbenz.lmidine.The latter was isolated (4 grams), and found to be identical withWallach's compound (Armalen, 1877, 184, 83).The clear benzene solution WAS washed twice with water, dried, andevaporated a t about 40".Massive crystals remained, together with asyrupy portion, which crystallised after several hours. After mashing.with a little ether and draining on porous porcelain, the crude product(12.5 grams) was crystallised from 700 C.C. of light petroleum, thesolution being allowed to cool very slowly to avoid deposition of syrup.White or transparent, colourless leaflets or plates (m. p. 104";10.5 grams) separated, which on recrystallisation from light petroleumwere obtained in the pure state and then molted sharply a t106-107°:0.2193 gave 0.6024 CO, and 0.0759 H,O.0.1688 ,, 9.3 C.C.N, (moist) at 20° and 757 mm. N=6*28.C1,H,O,N requires C: = 75.34 ; H = 4.03 ; N = 6.27 per cent.2- Plienyl-1 : 3-6enxoxaxine.4-one is very soluble in chloroform oracetone, readily so in benzene, methyl alcohol, or ethyl acetate, andrather less so in ethyl alcohol, ether, or pyridine. It dissolves in about'70 parts of light petroleum (b. p. 90-120"). It may be recoveredunchanged after solution in these solvents, even pyridino. Itssolutions give no coloration with alcoholic ferric chloride. I n glacialacetic acid it readily dissolves, but owing to the presence of rnoistur2the solution soon deposits N-benzoylsalicylamide.Behuvwu?* with Water.-Ths cornpound is insoluble in wLtter, but onboiling an oil is produced which dissolves appreciably.The clear hotaqueous solution if rapidly cooled becomes turbid, and after a shorttime fine needles of the unaltered compound separate, but not withoutproduction of some N-benzoylsalicylamide.was studied by diluting an alcoholic solution largely, but insufficiently,to induce separation. The clear solution remained clear for aboutthree hours, after which it became faintly turbid, owing to separatior~+i If ether is used instead of benzene as a solvent, the whole of the ring-compoundis decomposed during the washing with acid.C = 74.92 ; R = 3.85.0.3480, by Kjeldahl's method, required 16.0 c,c. N/lO-HGl. N = 6.3'3.Its bchaviour witTITHERLEY : 2-I'HENTL-1 : 3-BENZOXAZINE-4-ONE. 207of the highly insoluble N-benzoylsalicylamide, which, however, wasnot completely precipitated in twenty-four hours.Whether purewater free from carbonic acid would behave like ordinary distilledwater has not been tried. Traces of acids jinorg$nic and organic)enormously accelerate the action of the water, and it was shownroughly that the velocity depends on the concentration of the acidadded. If 20c.c. of a 0.05. per cent. solution are treated with 0.05 C.C.of 10 per cent. hydrochloric acid, the decomposition is complete inabout seventy seconds. Accurate velocity measurements have notbeen made, but the quantitative nature of the conversion was easilyshown by decomposing a known weight in dilute alcohol.Onacidifying the clear solution with a few drops of hydrochloric acid,after a few seconds a voluminous, microcrystalline precipitate *separated, which was collected and dried at 100' :0.2 2 30 gave 0.23 9 8 N-benzoylsalicylsmide.The purity of the N-benzoylsalicylamide was shonn by the sharpmelting point (206') before recrystallising.Behaviour with Acids.-The cyclic compound has only a weaklybasic character, and with aqueous acids is decomposed withoutpreviously dissolving. It does not yield a picrate in benzene oralcoholic s o h tion. Dry hydrogen chloride in benzene solutionprecipitates its hydrochloride, which, however, could not be isolated ina pure condition for analysis, owing to the great ease with which itis decomposed by atmospheric moisture.Behuviour with Alkalis.-With cold aqueous sodium hydroxide andammonia, the compound is comparatively stable and only very slowlychanged to the salt of N- benzoylsalicylamide, but sodium hydroxideimmediately decomposes it in alcoholic solution.Dry ammoniaproduces a yellow colour with an alcoholic solution of the compound,and probably an additive compound like that obtained with aniline isformed ; this is under investigation.Behaviour with Phosphorus Pentachloride-The compound, on treat-ment with phosphorus pentachloride in presence of chloroform,instantly gives a bright lemon-yellow, crystalline solid,C,4H,0NCI,,POC13,identical with th :t obtained by Titherley and Hicks (Zoc. cit.) by theaction of phosphorus pentachloride on phenylbenzometoxazone a thigher temperatures.Beduction.-The reduction of 2-phenyl-l : 3-benzoxazine-d-one* The reaction affords a very delicate test for 2-phenyl-1 : 3-benzoxazine-4-one.On dissolving mintite amourit in a drop of alcohol and treating with about 1 C.C. ofwater and a trace of any dilute acid, a precipitate appears suddenly after a fewseconds.C,,H,02N requires 0.2410 N-benzoylsalicylamide208 TITHERLEY : 2-PHENYL-1 : 3-BENZOXAZINE-4-ONE.cannot be effected by sodium amalgam, in alkaline solution, or byacid reducing agents, owing to ring rupture, With aluminiumamalgam, however, the reduction proceeds rapidly.A dilute alcoholic solution of the substance was treated withan excess of aluminium amalgam, and water added gradually.Whenthe reduction was complete, the alcoholic solution mas filtered andallowed to evaporate. The resulting solid was washed with dilutealkali, dried, and extracted with boiling alcohol. A white, sparinglysoluble solid remained behind, which has not yet been identified, and thealcoholic solution deposited needles consisting of impure phenylbenzo-metoxazone. By carefully recrystallising from alcohol, the latter wasobtained pure in fine needles melting at 169", and was easilyidentified. The melting point after mixing with a specimen ofphenylbenzometoxazone synthesised from salicylamide and benz-aldehyde mas 168-169'. The yield was about 30 per cent., and asimilar reduction performed in an ethereal solution containing5 per cent.of alcohol gave a yield of 40 per cent.(b) Preparation from O-BenxoyEsaZic~Zamide.-A mixture of 3 gramsof 0-benzoylsalicylamide and 30 C.C. of xylene contained in a distillingflask was heated in a bath to 145O, and when the solution becameclear a stream of dry hydrogen chloride was passed in for a fewminutes. By gradually raising the temperature of the bath to 155O,the xylene was allowed to distil over drop by drop, carrying with itthe water formed in the reaction. The hydrogen chloride actedcatalytically, and only a relatively small quantity was necessary, butin order to make good that lost by distillation, a few bubbles of thegas were passed into the liquid at intervals of three minutes, Therate of distillation was regulated so that the majority of the xylenehad passed over in an hour.About 0.2 gram of water passed overwith it, and an oil remained in the flask which partly crystallised oncooling and consisted of a mixture of 2-phenyl-1 : 3-benzoxazine-4-0110and its hydrochloride and unchanged 0-benzoylsalicylamide, togetherwith a little xylene. When cold it was treated with 20 C.C. of benzeneand well shaken with an excess of 3 per cent. sodium hydroxide.A considerable quantity of lemon-yellow solid (the sodium derivativeof N-benzoylsalicylamide produced by rearrangement of O-benzoyl-salicylamide) separated, which was mostly removed by repeatedlywashing the benzene extract with water. The benzene extract wasthen dried, filtered, and evaporated at 50°, when large, colourlessplates were left, together with an oil which crystallised on cooling.The solid was practically pure 2-phenyl-1 : 3-benzoxazine-4-one, andweighed 1.8 grams, or 65 per cent.of the theoretical yield. Onrecrystallising from about 100 C.C. of light petroleum and cooling slowlyTITHEIiLEY : 2-PHENYL-1 : 3-BENZOXAZINE-4-0EE. 209it separated in small, colourless, glistening plates, which meltedsharply at 106*, and a mixture with a specimen obtained by theamidine method of synthesis also melted at 106':0.2128 gave 0.6003 CO, and 0,0755 H,O. C=76.8; H=3.9.0.2362 ,, 13.7 C.C. N, (moist) at 23" and 755 mm. N = 6-49.C,,H90,N requires C = 75.34 ; H = 4.03 ; N = 6.27 per cent.The properties of the substance agreed in every particular withthose of the compound obtained by the amidine method.I n a similar preparation to the above, using toluene instead ofxylene as solvent, the compound was also obtained, but the yield waslower (36 per cent.of the theoretical).(c) Prepccration from ~ - ~ e n ~ o y ~ s c c ~ ~ c ~ ~ ~ r n ~ ~ e . - A similar method tothe above was employed, using anisole as a solvent at 130'. Abouttwo and a-half hours mere required for the completion of the change,and a yield of 60 per cent. of 2-phenyl-1 : 3-benzoxazine-4-one (m. p.106-107°) mas obtained. Ethylene dibromide mas also used withsuccess as a solvent, but xylene mas unsatisfactory, owing to the smallsolubility of ni-benzoglsalicylamide. During the action of the hydrogenchloride, some of the N-benzoylsalicylamide is rearranged t o O-benzoyl-salicylamide, and this was isolated in smal1:quantity in one experiment,when xylene was used as solvent, as needles melting with rearrange-ment a t 144".0.1812 ,, 0.50'70 CO, ,, 0.0635 H,O.C=76*39 ; H=3'89.Saliey Zpheny Zbenxamidine,OH*C,H 4* COON : CPh*NH P= OH* C6H,* CO *NH CPh: NPh,was obtained by the sction of aniline on 2-phenyl-1 : 3-benzoxazine-4-one or its hydrochloride by the following methods.One gram of 8-phenyl-1 : 3-benzoxazine-4-one and 0.5 ,gram ofaniline were warmed t o 50°, and the resulting yellow oil, which didnot crystallise on cooling, was dissolved in 50 C.C. of boiling lightpetroleum and the clear yellow solution allowed to cool slowly. Ayellow syrup was deposited, which, after being kept for twelve hoursin contact 'with the mother liquor, changed to a mass of beautifultransparent, yellow needles (1.3 grams), and the mother liquor yieldeda further quantity (0.1 gram), the total yield being 93 per cent.ofthe theoretical.When instead of light petroleum as a solvent, ether (15 c.c.) wasused, the pure compound separated after several hours in clusters ofneedles (0.9 gram) without any syrup being firsti deposited, and afurther quantity (0.2 gram) was obtained from the mother liquor.The substance was also prepared in the pure state by treating abenzene suspension of 2-phenyl-1 : 3-benzoxazine-4-one hydrochloridewith aniline. The identity of the yellow needles obtained by theVOL. XCVlI. 210 TITHERLEY : 2-PHENYL-1 : 3-BENZOXAZINE-4-0NE.different methods was shown by a comparison of melting point(1 05-10ri0) and other properties :0.2750 gave 19.8 C.C.N, (moist) at 17" and 772 mm.C,oH,,0,N2 requires N = 8.86 per cent.Salicylphenylbe~~xamidine is somewhat sparingly soluble in ether,moderately so in alcohol, and readily soluble in chloroform or benzene.Its solutions give a reddish colour with alcoholic or ethereal ferricchloride. Its solutions on evaporation at the laboratory temperatureyield syrups which crystallise completely only after several days, andit is probable that two or more isomeric forms in equilibrium areproduced by labile change when the yellow needles are dissolved inany solvent. If heated in a light petroleum solution for any length oftime, as in crystallising large quantities, the substance more or lessdissociates into aniline and 2-phenyl-1 : 3-benzoxazine-4-one, whichseparates to some extent on cooling in colourless needles or tuftsalongside the yellow needles of the unchanged substance. Prolongedboiling leads to decomposition of a portion, owing to the action of theaniline (see below) forming diphenylbenzamidine.Salicylphenylbenzamidine possesses weakly basic and weaklyphenolic characters; it is soluble in, but at once decomposed by,hydrochloric acid, yielding N-benzoylsalicylamide. I n a condition offine powder it is only very slowly soluble in dilute sodium hydroxide ;a yellow solution of the sodium salt results, from which, however, itis apparently impossible to recover the original substance ; on acidifica-tion with hydrochloric acid, a yellow precipitate is formed, solublein excess of acid to a nearly clear solution, from which dV-benzoyl-salicylamide is almost instantly precipitated. When salicylphenyl-benzamidine is heated a t 9 5 O , it dissociates, and eventually melts to aclear yellow liquid with an odour of aniline ; if the latter is allowedto evaporate slowly in an open shallow vessel, the yellow colourbecomes less conspicuous, and at the end of some hours a colourlessliquid results, which finally sets to a mass of colourless plates, con-sisting of 2-phenyl-1 : 3-benzoxazine-4-one, 'the yield of which wasalmost theoretical. When heated with aniline at 100' for severalhours, salicylphenylbenzamidine is partly decomposed into salicyl-amide and diphenylbenzamidine, the latter obtained in a yield of35 per cent.N=S.50.ORGANIC LABORATORY,[JNIVERSITY OF LIVERPOOL