1372 BAXTER AND FARGHER 1 3-BENZODIAZOLEARSINIC CXXIX.-1 3-Benxodiazolearsinic Acids and their Reduction Products. By ROBERT REGINALD BAXTER aiid ROBERT GEORGE FARGHICR. SINCE the discovery of salvarsan (Ehrlich and Bertheim B e y . 1912, 45 756) many attempts have been made to prepare derivatives or closely allied products which would render unnecessary the some-what elaborate technique involved in the clinical use of the original compound which owing to its pronounced acidity has to' be con-verted into the disodium salt in turn alkaline in reaction before use. Neosalvarsan the sodium N-methylenesulphinate of salvarsan, introduced to avoid these difficulties possesses many advantages, such as ready solubility and neutral reaction but in spite of these, there is a consensus of opinion in favour of salvarsan as its spire chaeticidal action appears to be stronger and more certain.In this and in other cases the attempts have led to the formation of acid derivatives giving neutral salts with alkalis. The authors desired 021 the other hand to obtain hydrochlorides soluble in water and sufficiently less acid in reaction than salvarsan to admi ACIDS AND THEIR. REDUCTION PRODUCTS. 1373 of their direct use and it was with this object in view that the present investigation was commenced since it was considered prob-able that the hydrochlorides of the arsenobenzenes derived from 1 3-benzodiazole (benzoglyoxaline) would fulfil these requiremenh. Further considering the properties of the glyoxaline nucleus they might be expected to exert acidic as well as basic properties and thus closely resemble salvarsan itself.Additional interest would attach to such compounds in view of their relationship to 3 4 5 3’ 4’ 5’-hexa-aminoarsenobenzene and its N-methyl derivatives (D.R.-P. 286667 286668 286854 286g55, 294276 ; E.P. 7488 and 8041 of 1913 ; U.S.P. 1081079 ; Benda Ber., 1914 47 1316; Karrer Ber. 1916 49 1448) which are stated t o possess the unexpected property of dissolving in alkali hydrogen carbonates f ormiiig carbamates with the same degree of alkalinity as that of normal blood serum and a relatively low toxicity So far as the authors are aware the beiizodiazolearsisic acids have not been examined Bertheiin (Rer. 1911 44 3092) who described 3 4-diaminophenylarsinic acid and several of its deriv-atives making no mention of them.They are however readily obtained by the well-known reaction of ortho-diamines with formic or acetic acid whereby the acyl derivative first formed suffers loss of a molecule of water if a monoacyl or of a molecule of acid if a diacyl derivative with consequent closing of the ring (compare Ladenburg Ber. 1875 8 677; Wundt Ber. 1878 11 826). The comparative stability of 3 li-cliacet?/lan~i.no~her~~lal.sinic acid, which crystallises from boiling water unchanged and does not alter save for slight decomposition when heated a t 250°/20 mm. and, generally the stability of diacyl-o-diamines (compare Bistrzycki and Ulffers Bcr. 1890 23 1876) seems t o indicate that in the absence of anhydrides the reaction proceeds with intermediate :‘ormation of a moiioacyl derivative.5 6-l)iamino-m-tol?/lars~?~ic acid ( I ) which together with 3 4-diaminophenylarsinic acid (Bertheim lw. cit.) formed the starting material for the investigation was obtained by the reduction of 6-nitro-6-amino-~~z-tolylarsinic acid (this vol. p. 989) with sodium hyposulphite. It undergoes all the condensation3 typical of an (D.R.-P. 269660; E.P. 1667 of 1914). /\ CH,NH I I N \/ () /\/ I )AsOP, NH2 / \ q N H20,AJ ’ ‘ / /\/\/ CH,(\JNH, \/ \rLN AsOs K2 () CH, (1.1 01.1 (111.) 3 ~ 1374 BAXTER AND FARGHER 2 ~-BENZODIIIZOJiWLRSINIU o-diamine giving for example with sodium nitrite in acid solution, 7-methyl-1 2 3-benzotriazole-5-arsinic acid (11) and with phen-ant h raquinone 4-m e t h y l phenarzt h raph enaaine - 2-ar sinic acid ( 111).The latter derivative shows the colour reactions charachristic of phena~thraphenazine. With boiling glacial formic acid the acids readily yielded 7-methyl-l 3-benzodiazole-5-arsinic acid (IV) and 1 3-benzodiazole-5arsinic acid (V) respectively. CH NH / A /\/\CH I \1 "HCH H208As\/-N H20,Ad I - - - h (V. 1 \/ (IV.) The action of acetic acid did not proceed quite so smoothly. 3 4-Diaminophenylarsinic acid although stated (Bertheim loc. cit.) t o be readily soluble in acetic acid is far from being so and on boiling with acetic acid acetic anhydride or a mixture of the two, gives rise to very highly coloured products which are difficult to purify. If the reaction is carried out as described in the experi-mental portion of this paper the main product is 3:4-diacetyE-aminophenylarsinic acid which on heating with water in a sealed tube a t 130° yields 2-methyl-1 3-benzodiazole-5-arsinic acid (VI).I n the case of the higher homologue the desired reaction takes place predominantly with the formation of 2 7-dimethyl-1 3-be~zo~zazole-5-arsinic acid (VII). Incidentally it may be men-tioned that it has since been f ouiid that if 3 4-diaminophenylarsinic acid dissolved in the requisite amount of sodium hydroxide (40 per cent. solution) to form the monosodium salt is treated with sufficient acetic anhydride to neutralise the sodium hydroxide and form acetic acid with the water present and then boiled the main product of the reaction is the benzodiazole derivative.(VII.) NH NH (VIII.) The reduction of the above acids with hypophosphorous acid, although it proceeds normally leads to gelatinous products difficul ACID8 AND TREIR REDUOTION PBODUOTS. 1376 to handle and consequently the reduction was carried out by means of sodium hyposulphite. This gives pale yellow arsenobenzenes, insoluble in water which were converted into hydrochlorides by precipitating their solutions in acetic acid with excess of hydro-chloric acid or by treating the suspensions of the bases in water with just sufficient hydrochloric acid t o bring about solution and precipitating as before. The hydrochlorides form pale yellow powders soluble in water but very sparingly so in the usual organic solvents. Their aqueous solutions react strongly acid to litmus, but neutral to methyl-orange.It was expected that this acidity would prove too great for pur-poses of intravenous injection and that it would be difficult to distinguish between the effect due t o the acidity and that due to the arsenic compound. This was confirmed by experiments with 5 Eil-arseno-l 3 1' 3I-benzodiazo7e (VIII) by Miss Soref of the Wellcome Physiological Research Laboratories. EXPERIMENTAL. 3 4-Dinmino~~eny7arsinic A cid. The preparation of considerable quantities of this acid was carried out according t o the directions given by Bertheim (Zoc. cit.), save that i t w'as found advisable to omit the boiling of the solution after the reduction with sodium hyposulphite as this diminished the yield and gave a more highly coloured product.1 3-Ren,zodinzo?e-5-nl.siltic Acid (V). 'Ten grams of 3 4-diaminophenylarsinic acid were boiled under a reflux condenser for six hours with 100 C.C. of glacial formic acid. After removal of the excess of formic acid by distillation water was added when 7.5 grams of crystalline material rapidly separated. To remove adhering traces of colouring matter the product was treated with charcoal in dilute alkaline solution. On making neutral to methyl-orange crystallisation commenced a t once practically the whole being recovered in a pure state. 1 3-Benzodiazole-5arsinic acid crystallises from water in which it is sparingly soluble in clusters of minute flattened prisms which are anhydrous. On heating it gradually darkens above 250° and decomposes rapidly about 297O (corr .).Found N=11-4; As=30*8. C,H,O,N,As (242.1) requires N = 11.6 ; As = 30.96 per cent 1376 DAXTER AND FARUHER 1 3-BENZODIAZOLEARSMIO 5 ; 51-Arseno-1 3 ; 1' 3t-benzodiazole (VIII). A solution of 5 grams of 1 :3-bcnzodiazole5-arsinic acid in 100 C.C. of water containing 1.2 grams of sodium hydroxide was treated with 50 grams of sodium hyposulphite and 11.5 grams of magnesium chloride dissolved in 300 C.C. of water and the mixture heated with stirring at 60° for two hours in an atmosphere of carbon dibxide by which time the precipitation of the yellow arsenobenzene was complete. After cooling the product was collected well washed with water suspended in 90 C.C. of water, dissolved by the addition of sufficient hydrochloric acid to form the dihydrochloride filtered from traces of undissolved matter and precipitated by pouring into an equal volume of concentrated hydrochloric acid.The hydrochloride was collected well washed with alcohol and ether and dried in a vacuum over sulphuric acid. The yield amounted t o 3.9 grams. 5 5t-14 rseno-l 3 1' 3/-benzodinzoZe is obtained as a bright yellow powder practically insoluble in water or the usual organic solvents, sparingly if a t all soluble in methyl alcohol containing hydro-chloric acid but fairly readily so in 50 per cent. acetic acid. The c7ihydrochloride forms a pale yellow powder soluble in water very sparingly so in methyl alcohol or ethyl alcohol and quite insoluble in acetone or ether. The aqueous solution reacts acid to litmus, but neutral to methyl-orange.After drying in a vacuum over-night it retains solvent equivalent to two molecules of water. Found N = l l . 3 ; As=30*2; C1=13*8. C,IH,oN,As.,2RC1,2TT,0 (493.1) requires N = 11 - 4 ; As= 30.4 ; C1= 14.4 per cent. 3 4-nincc.tylnniil.~op7~en?/larsinic Acid and 2-Methyl-l 3-henzo-diazole-5-a~sinic A cid (VI). Experiments in which diamiiiophenylarsinic acid was heated with glacial acetic acid with acetic anhydride with acetyl chloride and pyridine or with acetic anhydride in the presence of a trace of sulphuric acid were unsuccessful owing to the considerable amount of decomposition which occurred. It was found however that i f the acid dissolved in methyl alcohol was treated with a mixture of acetic acid and acetic anhydride and the methyl alcohol then removed as completely as possible by distillation on the water-bath, the resulting solution after heating to active boiling for four hours, gave an excellent yield of a product which from its composition, proved to be the diacetyl derivative of diaminophenylarsinic acid.3 ; 4-Diacetylami~iol~hen.yJn~sin i c acid crystallises from water i ACIDS AND THEIB REDUUTION PRODUOTS. 1317 which it is fairly readily soluble as a felted mass of fine needles which retain from 2 to 2.5 per cent. of solvent. Found loss at l l O o (two specimens) 2.6 2.2. I n dried material N=8*8; As=23*8. C,oH1,O,N,As (316.1) requires N,= 8.9 ; As = 23.7 per cent. As the product did not lose acetic acid when heated to 250°/ 20 mm. it was heated with ten times its weight of water in a sealed tube for four hours a t 130° when the desired reaction took place almost completely.2-Met hyl-1 3-benzodiazole-5-arsinic acid crystallises from water, in which it is somewhat sparingly soluble in minute needles contain-ing two and a-half molecules of water of crystallisation the) last half molecule being removed with difficulty a t l l O o but fairly readily a t 120O. After drying a t l l O o it darkens on heating above 250° and decomposes a t about 270° (corr.). Found loss a t 110°=15*5. I n dried material C = 37.3 ; H = 3.8 ; N = 10.7 ; As = 29.3. C,Hg0,N2As,2~H20 requires H,O = 15.0 per cent. C,H,O,N,As (256.1) requires C = 37.5 ; H = 3-5 ; N = 10.9 ; As = 29-3 per cent. 2 21-Dimethyl-5 51-ameno-1 3 11 3~-benzodiazole.This was prepared in the same manner as 5 5'-arseno-l 3 1' 31-benzodiazole which it resembles very closely in its properties, The dikydrochloride forms a pale yellow powder soluble in water, but almost completely insoluble in methyl alcohol ethyl alcohol, ether or acetone. After drying in a vacuum for twenty-four hours it retains solvent approximating t o three molecules of water. In aqueous solution it reacts acid to litmus but neutral to methyl-orange. Found N=10*3; C1=12*7; As=28*0. C~6H14N4As,,2HCl,3H20 (539.1) requires N = 10.4 ; C1= 13.1,; As = 27.8 per cent. 5 6-Diamino-i~-toZylarsin.ic A cid (I) A solution of 8-6 grams of 5-nitro-6-amino-~n-tolylarsinic acid in 75 C.C. of water and 6.2 C.C. of ION-sodium hydroxide was cooled t o -lo and treated in one operation while stirring vigorously, with 20.4 grams of sodium hyposulphite.When the reaction was complete as shown by the change in colour the solution was filtered treated with 8.1 C.C. of hydrochloric acid (D 1-12) and the re~ulting acid which separated in the course of a few minutes 1378 BAXTER AND FARGHflR I ~-BENZODIAZOSfEArRSMIC collected washed with water and recrystallised from watef in which it is sparingly soluble in the cold but fairly readily so on warming. It separates in colourless needles which contain between one and a-half and two molecules of water of crystallisation. It dissolve3 fairly readily in methyl alcohol or acetic acid but is very sparingly soluble in ether benzene or petroleum. I t s solution in dilute hydrochloric acid gives a characteristic deep violet color-atiofi with a drop of a dilute solution of potassium dichromate.It appears to decompose slowly on keeping. Loss a t l10°=11*2. C7H,,0,N,As,l~H20 requires 9.9 per cent. C,H,,03N,As,2H,0 requires 12.8 per cent. C,R,,O3N,As (246.1) requires N = 11.4 ; As=30.5 per cent. In dried material hT=11.3; As=30.4. 7-Methyl-1 2 3-benzotriazoEe-5-arsinic acid (11) was obtained by treating a solution of 3 4-diamino-m-tolylarsinic acid in dilute hydrochloric acid with a molecular proportion of sodium nitrite. It forms a crystalline powder very sparingly soluble in water but m p e readily so in alcohol. It crystallises from 50 per cent. alcohol in minute colourless glistening needles which are anhydrous and gradually decompose on heating above 280O.Found N = 16.6. C7H,0,N,As (257.1) requires N = 16.4 per cent. 4-Met h ybphennn t Ii rn ph encl(2ine-2 -urshic acid (111) was obtained by the condensation of molecular proportions of phenanthraquinone and 3 4-diamino-v~-tolylarsinic acid in acetic acid solution. It forms an amorphous yellow powder sparingly soluble in water and the usual organic solvents. With concentrated sulphuric aeid it develops an eosin-red colour with nitric acid a cherry-red and with hydrochloric acid an insoluble red compound. The colour is dis-charged in all three cases by dilution with water. It dissolves in dilute sodium carbonate or sodium hydroxide but a flocculent pre-cipitate of the sodium salt is thrown down on the addition of excess of the reagent.Found N = 6 * 4 . C2,H,,CJ3N,As (418.2) requires N = 6.7 per cent. 7-Methyl-1 3-he.nzodiazole-5-a~sinic Acid (IV). Two grams of 3 4-diamino-m-tolylarsinic acid were boiled wit,h 20 C.C. of glacial formic acid for six hours. The isolation of the product was carried out as in the case of 1 3-benzodiazole-5-arsinic acid save thatq the treatment with charcoal was found to be mor AOIDS AND THlIR REDUCTION PRODUCTB. 1379 effeative in acid solution. The air-aried product contains only traces of moisture and resembles the lower homalogue very closely in its properties. It dissolves sparingly in water and crystallism in minute prismatic needles which darken when heated above 280° and melt and decompose a t about 300° (mrr.), Found loss a t 110°=1.3.I n dried material N = 10.8 ; As = 29.4. C,H,O,N,As (256.1) requires N = 10.9 ; As= 29.3 per cent, 7 71-Dimethyl-5 51-arseno-1 3 If 31-benzodiazole. This was obtained by the reduction of the above acid with sodium hyposulphite in the manner previously described (p. 1376) as a pale yellow powder insoluble in water and the usual organic solvents but moderately soluble in acetic acid. The corresponding dihydrochloride forms a pale yellow granular powder soluble in water but insoluble in methyl alcohol ethyl alcohol ether or acetone. After drying for twenty-four hours under greatly reduced pressure it retained solvent corresponding with two molecules of water. Found N=10.6; As=28*6; C1=13*1. C,,H14N4As~,2HC1,2H,0 (521.1) requires N = 10.75 ; As= 28.8; Cl= 13.6 per cent.2 7-Dirnethyl-1 3-benzodiuzole-5-arsinic ,4 cid (VII). This acid was obtained by the prolonged action (six to eight hours) of boiling acetic acid on 5 6-diamino-m-tolylarsinic acid, the isolation being carried out in the usual manner. It is some-what soluble in boiling water but. sparingly so in cold and separates from the former in minute prismatic needles which contain two molecules of water of crystallisation. Found loss a t 110°=12'2. I n dried material N = 10.6 ; As = 27.9 27.5. C9Hl1O3N,As,2H,O requires 11.8 per cent. C,H,,QsN,As (270.1) requires N= 10.4 ; As= 27.75 per cent. 2 7 21 71-Tetranaethyl-5 5'-a.rseno-l 3 1/ 31-benzodiazoZe. The above acid was readily reduced to the corresponding arseno-benzene by sodium hyposulphite under the conditions previously employed. The urseno b enzene closely resembles those already described in its physical and chemical properties. The dihydro-chloride forms a pale yellow powder soluble in water but insoluble 3 $ 1380 COWABD AND WILSON THE EQUILIBRITTM in the usual organic solventa. After drying for twenty-four hours in a vacuum it retains solvent corresponding with approximately two molecules of water. Found N = 10.0 ; As = 27.3 ; Cl = 12.3. C,sH,sN4As,,2HC1,2H,0 (549.1) requires N = 10.2 ; As = 27.3 ; C1= 12.9 per cent. WELLUOME CREMICAL RESEARCH LABORATORIES, LONDON E.C. 1. [Received October 27th 1919.