958 ROBINSON AND ROBINSON : L X X X I I. -Re sea IT hes on Pseudo - bases. Pa? *t 11. Note o n some Beybering Derivatives and Remayks on the Mecharhszz of the Condensation Reactions o j Pseud o-bases. By GERTRUDE MAUD ROBINSON and ROBERT ROBINSON. THE analogy between berberine and cotarnine was spectro-chemically demonstrated by Tinkler (T. 1911 99 1340) and in view of the fact that the former base contains an unreduced isoquinoline nucleus it became of interest to examine its behaviour in connexion with the formation of condensation products analogous t o the numerous anhydrocotarnine derivatives. Anhydroberberineacetone (I) (+B-CH,*COMe) is the only sub-stance of this type which has been investigated. It was prepared by Gaze (Zeit. Nuturwiss. Halle 1890 62 399) and its analogy to anhydrocotarnineacetone was pointed out by Gadamer (Arch.Pharm. 1905 243 42) and confirmed by the examination of the substance which was made by Pyman (T. 1911 99 1694). 0-CH /\A I CH*OH(OEt) OH I I -+ A/\(7/\/ Me01 \ / \ / & \ / C H 2 I RN I f-Me0 CH CH2 I I CH,-COMe I CH -+ Y<OH(O I H ~ t ) RN*OH(OEt) +- li I (111.) * Part I (T. 1914,105 1456) contains the following errata on page 1458 : for 6-nitropiperonylhydrocotarnine read 6-nitropiperonoylhydrocotarnine and for anhydrocotarnine-6-nitroveratrole and 6-nitroveratrole read anh ydro-cotarnine- 6-nitrohornoneratrole and 6-nitrohornoveratrole respectively RESEARCHES ON PSEUDO-BASES. PART 11. 959 It was found in small-scale experiments that berberine condensed readily with alcohols amides such as carbamide phthalimide, 2-methylindole acetophenone 1-hydrindone cydohexanone nitro-methane 2 4-dinitrotoluene 2 4 6-trinitrotoluene diethyl malonate ethyl acetoacetate ethyl phenylacetate phenylaceto-nitrile and indene.Berberinol (compare Tinlder Zoc. c i t . ) was dissolved in alcohol and after the addition of one of the com-pounds mentioned the mixture was gently warmed. As a rule a sparingly soluble substance separated often in a viscid condition, but crystallisable after treatment usually by washing with alcohol and the derivatives were lemon-yellow with the exception of the nitro-compounds which were orange or red. All the sub-stances were resolved into their components by acids and anhydro-berberine derivatives are therefore much less stable than the corre-sponding substances obtained from cotarnine or hydrastinine.Circumstances forced the authors to abandon the detailed examina-tion of these substances a t a time when the work had not pro-ceeded far and was quite incomplete but it nevertheless now seems desirable to record the results of the investigation. Cyano-dihydroberberine ($-B-CN) methoxydihydroberberine ($-B-OMe), anhydroberberineacetophenone ($-B-CH,*COPh) and anhydro-berberinenitromethane ($-B-CH,*NO,) are de-scribed in the experi-mental portion of the paper. The Mechanism of Pseudo-base Condensation. The fact that berberine resembles cotarnine in its reactions emphasises the probability that the reactive modification is in each case the quaternary ammonium hydroxide form since berberine has an even greater tendency than cotarnine to assume this unsaturated condition.The formation of anhydroberherine-acetone was found to occur in dilute alcoholic solution under con-ditions not favourable to the existence of berberinol and a hitherto unrecorded observation on the condensation of cotarnine with nitromethane may also be cited. A dilute aqueous acetic acid solution of cotarnine was mixed with nitromethane and on the addition of sodium acetate the condensation product was formed slowly in the cold rapidly OM heating t h O mixture and the substance was identified with anhydrocotarninenitromethane (Hope and Robinson T. 1911 99 2119). I n this acid solution the presence of the carbinol form of cotarnine is extremely improb-able.A t the same time condensation products are obtained from cotarnine and berberine under conditions which do not favour electrolytic dissociation of an ammonium hydroxide. Thus the presence of an excess of sodium hydroxide does not appear t 960 ROBINSON AND ROBINSON : inhibit the reactions. A representation of the mechanism of the Knoevenagel reaction based on the assumption of a reaction between ions was suggested by Lapworth (Hope and Robinson, Zoc. c i t . 2117) and was a great advance on the ideas existing a t that time especially as it facilitated the collation of data derived from such separate investigations as those of Knoevenagel on the use of amines particularly secondary amines as catalysts in con-densations of Dobbie Lauder and Tinkler on the spectrochemistry of pseudo-bases and of numerous workers on the chemistry of these substances.These advantages are secured by the recogni-tion that the reactive form of a carbinol-amine (11) is the un-saturated ammonium hydroxide (111) but instead of representing the further stages as due to ionisation combination of the ion with a negative residue and finally migration the present authors prefer to regard the condensations as due to a simple addition of the components as illustrated in the scheme : “”-% ....... fi c_) I fH*+ NR I I n the case of a condensation between a pseudo-base and pseudo-acid the theory of the reaction between ions demands two intra-molecular changes but on the hypothesis now advanced the carbon to carbon synthesis occurs in the first stage of the process and migrations are not required.This is illustrated in the case of anhydrocotarninenitromethane and it will be seen that the mi-modification of nitromethane and the ammonium hydroxide form of cotarnine yield a complex (IV) in which there may be a change in affinity dist’ribution which results in the separation of water and the production of anhydrocotarninenitromethane. (IV.) The essential feature of these representations is the postulation of conjugated partial dissociation as a preliminary stage od th RESEARCHES ON PSEUDO-BASES. PART 11. 961 reactions and this is a particular case of the process described by Baly as "opening up the molecular force fields." The mode of expression is however slightly differentl from that employed in it former communicat-ion (T.1916 109 1031 * e t seq. 1042). Taking methyl iodide as an example then in reactions in which the iodine becomes separated from the methyl group it is assumed that there is a partial dissociation and that the reactive molecule should be represented as . . . CH,I.. . . The present suggestion is merely that the partial valency so expressed shall be considered to be derived from the normal valencies and that the dissociation necessarily weakens the bond between the carbon and iodine atoms, so that the complete symbol is ... CH ... I... . Where a partial dissociation can be followed by complete electrolytic dissociation, there is a clue t o the polarity of the partial valencies since it may reasonably be assumed that the partial dissociation is a stage in the complete process.Further it is clear that the partial dis-sociation of latent valencies must be assumed in some cases 88, for exzmple in the combination of ammonia with hydrochloric acid : The conjugated partial dissociation of such an ammonium hydroxide as cotarninium hydroxide is a more complex example of the same kind: * In this paper the residual affinity was regardcd as additional to the I ormal valency not as a part of the latter. The theory of the reactions is not fundamentally altered and in accordance with the suggested method of expression the addition of an alkyl haloid to an unsaturated base would be represented by the scheme 962 ROBINSON AND ROBINSON : I n partial dissociation of latent valencies two partial valencies of opposite sign become available and these emanate from the same atom whereas when a normal valency is divided the two parts will be of the same sign so that the ring in IV in regard to the polarity of the partial valencies should be expressed as shown below: The logical application of schemes of partial dissociation simple and conjugated of addition and decomposition by making and breaking of partial valencies and of redistribution of affinity, demands the consideration of these questions of polarity and leads to a system of mechanism of reactions which appears to be capable of including the representation of chemical changes of the most varied type and the present authors are not acquainted with ally examples of reactions the course of mhich cannot be illustrated in the manner implied.It is true that the subdivision of units invariably supplies greater facilities for explanations but in the present instance there is the important restriction on the elasticity of the theory which is imposed by the necessity of providing the reactive complexes with two free partial valencies of opposite sign and this has introduced no difficulty in any case examined. I n order t o avoid possible misapprehension it should be stated that reactions between ions are not excluded but regarded as the limiting case and further that it is recognised that the symbols which are used to express the activated condition of molecules can represent only a first approximation to the actual distribution of affinity. We cannot deal with every instance in which it is imagined that the method of representation advocated has clear advantages in the summarisation of the experimental data and we therefore confine ourselves t o two reactions which have been the subject of comparatively recent controversy.The Bromination of Ketones. Lapworth's theory of the mechanism of the bromination of acetone and other ketones (T. 1904 85 30) has received much support from subsequent experimental work and i t may be said to be universally accepted that the essential reaction is the addi-tion of bromine to the enolic form of the ketone. Leuchs (Ber. RESEARCHES ON PSEUDO-BASES. PART 11. 963 1913 46 2435) however brominated optically active o-carboxy-2-benzyl-1-hydrindone and obtained 5-10 per cent.of an optically active bromo-derivative and since the enolic form of this ketone contains no asymmetric carbon atom it was claimed that the bromiaation was in part a direct substitution. I n accordance with the theory of partial conjugated dissociation of an enol however, the actual reactive conditions is not .... ! . . ‘ . . . . and this reactive form is seen to be intermediate as regazds its distribution of affinity between the ketone and its enolic modifi-cation. Consequently the catalytic action of hydrobromic acid on the ketone in producing the enol will involve the reactive form of the latter as an intermediate stage. I n this molecule the partial valency preserves the asymmetric environment of the carbon atom and the formation of an optically active bromo-derivative is therefore possible.The whole process may be repre-sented in the following manner: . . iWec ha n i s m of Dicrz o-coupling . K. H. Meyer (compare “Annual Reports,” 1914 11 100; 1915, 12 115) and his co-workers hold the view that diazo-coupling is due to an addition of the diazonium salt to a double bond or conjugated double bond in the second component. Other authors, as Auwers and Michaelis and Karrer (loc. c i t . ) are of the opinion that the reaction is in the first place one of addition t o the oxygen or nitrogen atom of the phenol or amine and that this is followed by migration. Both these views are experiinentaIly founded and a t present regarded as contradictory. The application of the theory of addition of partly dissociated coinplexes leads t o a representation which in the present authors’ opinion explains the whole of the facts including those relating t o the chemistry of the diazonium salts themselves.It has already been suggested (T., VOL. UXT. P 964 ROBINSON AND ROBINSON 1916 109; 1042) that the characteristic reactions of aromatic amines and phenols must be ascribed to additions t o a conjugated unsaturated system which includes the nitrogen or oxygen atoms. The neutral and reactive phases of a phenol such as m-cresol will therefore be the following : An examination of the polar properties of the partial valencies shows that the orientation rules are a direct consequence of the opposite sign of the latent valencies of elements such as oxygeii and nitrogen but it must be remembered that in dealing with amines and phenols the effects observed are considerable and well defined and that in connexion with the general problem of orientation in the benzene ring it may be necessary t o take cognisance of even more delicate influences than the conjugation of partial valencies.The reactive phases of an aromatic amine will correspond with those figured above in the case of m-cresol. Addition to the unconjugated reactive inodifications will involve the attachmentl of a group to the oxygen or nitrogen atoms, whereas addition t o the molecule in its conjugated dissociated condition will involve nuclear substitution. I F the former reac-tion is reversible and this is usually the case there may ensue an apparent transference of a group from oxygen or nitrogen to the nucleus and the transformation of diazoamino-compounds into true azo-derivatives is not improbably a reaction of this type.The phenomenon is analogous t o that involved in the productio RESEARCHES ON PSZlUDO-BASES. PART 11. 965 of mesidine from phenyltrimethylammonium chloride and as this is a simpler case the first stage of the process may be illustrated: The last reaction is a conjugated decomposition that is the reverse of addition t o a conjugated system. Such reactions are of great importance in the aromatic series and there appears to be no valid reason why the decomposition should have a more complex mechanism than the formation of the additive product. Turning to the diazonium compounds it' must be noted that these substances (for example hydroxides) are in constitution and properties strikingly analogous t o such substances as cotarninium hydroxide and owe their reactivity t o a similar partial decom-position which is expressed in V.Addition between the reactive phases of a phenol and of a diazonium hydroxide will result in VI and possibly VII may then be obtained by a redistribution of affinity, (V. 1 (VI-) (VII.) VII is clearly the oxoniuin hydrate of the keto-form of an azo-phenol but it is also the hyd_rate by conjugated addit,ion of the enolic modification and the latter may be obtained by conjugated decomposition as shown above for the precisely similar conversion of a ketone into an enol with the aid of hydrobromic acid. It should be pointed out however that i t is unnecessary to go so far as VII in rearranging the affinity of VI.If for example, the partial valency connecting the nitrogen atoms is broken the nat,ural result of the activity of the free partial valencies is indicated in VIII IX and X. P P 966 ROBINSON AND ROBINSON : The above will apply t o a phenol such as &naphthol which couples in the ortho-position. Para-substitution will involve the inclusion of an additional double bond in the conjugated system, and it is clear that the scheme is applicable to amines as well as to phenols. .... I . . . .*. .. . . I.. . . . . 0 (VIII.) E X P E R I M E N T A L . Cyanodihydrob erb epine. This substance was prepared by Henry (Annnlen 1860 115, 136) and regarded as a sparingly soluble salt which crystallised from alcohol in yellow rhombic leaflets.The analysis given is 3 per cent. lower than the theory and indeed a t that time the composition of berberine was supposed to be C,2H,,010N (C= 6 ; 0 = 8 ) . I n 1872 Fliickiger (Jahresber. 748) stated that the sub-stance did not exist, and later Pommerehne (Arch. Pharm. 1895, 233 127) reaffirmed that berberine forms a stable hydrocyanide. l'inkler (Zoc. cit.) used the substance in connexion with his spectro-chemical work. Since i t appears that this compound has not yet been accurately analysed or described we prepared a specimen by adding potassium cyanide to a solution of berberine sulphate until the yellow colour disappeared. The solid was collected and rapidly crystallised from alcohol and then from benzene in which the substance is sparingly soluble.It was found necessary t o keep the solution in the dark as the compound is decomposed by light. The pale yellow prisms melted a t 184-186O when somewhat rapidly heated : 0.1405 gave 0.3602 CO and 0.0661 H,O. C,,H,,O,N requires C = 69.6 ; H = 5.2 per cent. The substance is readily soluble in chloroform and sparingly so in ethyl acetate from which i t crystallises in yellow octahedra. It is stable towards aqueous potassium hydroxide and is not immediately decomposed by cold dilute hydrochloric acid. On gently warming hydrocyanic acid and loerberine chloride were C=69*9; H=5*0 RESEARCHES ON PSEUDO-BASES. PART 11. 967 produced. No definite evidence of the formation of a salt of the base was obtained but that substances of this type are real bases which can form salts without decomposition is evident from the behaviour of cyanohydrocotarnine.This substance dissolves in dilute sulphuric acid but on scratching the container colourless crystals of a sparingly soluble sulphate separate. Only on heat-ing does the colourless solution become yellow and hydrocyanic acid is then set free. It is a mistake therefore to term these substances pseudo-salts. They are in no sense salts but bases which can combine with acids and in this condition ar9 readily decomposable. We have also observed that cyanohydrocotarnine forms a methiodide. Me t h ox ydi hgdro h erb e rin e J/-B-O Me . Perhaps the most characteristic reaction of pseudo-bases is the formation of ethers of the carbinol form by simple treatment with alcohols and i t is interesting to note that this property is exhibited not only by the rosaniline bases but also by many t’ri-phenylcarbinol derivatives not containing nitrogen and by the xanthhydrols and other pseudo-oxonium bases.I n the presence of excess of water these ethers are as readily hydrolysed as they were formed and the conclusion to be drawn from the whole matter seems to be that the interconversion of the two forms of the pseudo-base is accomplished by addition and subtraction of water or of an alcohol. Gaze in a private communication to Beilstein’s I ‘ Handbuch,” states that berberine yields an alcoholate, C,,,H,,0,N,C2H,0 which he describes as golden crystals but there appear to be no further references in the literature to substancss of this type.Berberinol (15 grams) was added with stirring to methyl alcohol (35 grams) when the substance became more crystalline in appearance. After remaining during two hours the substance was collected and rapidly crystallised from methyl alcohol. It was obtained in pale yellow prisms melting a t 1 5 2 O : C,lH,,O,N requires @= 68.7; H=5.7 per cent. 0.1239 gave 0.3126 CO and 0.0661 H20. I n attempting to crystallise this substance from ethyl acetate, oxyberberine melting at 198 -ZOOo was obtained. The substance was also readily changed to osyberberine by boiling with sodium methoxide in methyl-alcoholic solution. It appears to be more easily oxidised than berberiiiol itself Like most dihydroberberine derivatives the dilute solution in ethyl acetate exhibits blue fluorescence.The methoxy-group was readily removed in the four Qf methyl alcohol on warming the substslice with water, C=68*8; H=5*9 968 RESEARCHES ON PSEUDO-BASES. PART 11. Ethoxycli~iydroberberine.-This derivative crystallised in golden-yellow rectangular prisms which darken a t 125O and melt a t 1 3 6 O . Its properties were similar t o those of the methoxy-compound. In absolute ethyl alcohol the substance condensed readily with acetone producing anhydroberberineacetone melting at 175O. isoAntyloxydihydro b erb wine.-This substance was obtained in circular clusters of golden needles and melted a t 157O. Its ethyl acetate solution exhibits blue fluorescence. ,4 qihydrob er b erineacetophei~one $-B-CH,*COPh.This derivative was obtained by condensing berbetinol with acetophenone in alcoholic solution or by starting with methoxydi-hydroberberine and carrying out the reaction in absolute methyl-alcoholic solution. It' is however more convenient t o operate in the following manner. Berberine sulphate (20 grams) was mixed with alcohol (100 c.c.) and acetophenone (10 grams) and after gently heating the mixture a 20 per cent. aqueous solution of potassium hydroxide was added until the orange colour became red and then yellow. The liquid wa5 vigorously stirred and after the addition of water the yellow oil gradually solidified and was collected and crystallised first from alcohol in which the sub-stance is sparingly soluble and then from ethyl acetate. The final purification was by crystallisation from benzene containing a little light petroleum and the substance was then obtained in bright yellow prisms melting at 140-141O : 0.1310 gave 0.3542 CO and 0-0658 H,O.C28H2,0,N requires C = 73.8 ; H = 5.5 per cent. This compound is a t once decomposed by acids yielding a berberine salt and acetophenone the odour of which is perceptible even when the compound is warmed with water. The substance was changed by methyl sulphate into' what appeared t o be a mix-ture of metho-salts with berberinium sulphate. 011 boiling the golden-yellow mixture with hydrochloric acid acetophenone was liberated and a salt separated which had the appearance of berberinium chloride but was coiitaminated with a considerable proportion of some analogous compound.The mixture could not be separated into its constituents hut analysis indicated that it consisted of berberinium chloride and a methylberberinium chloride. C=73'7; H=5.6. A I I Jiy1rob erb frill P ) L it row Ptlicoi e $-R-CH,*NO,. Berberine sulphsts (20 grams) was warmed with a mixture of ethyl alcohol (100 c.c.) and water (100 c.c.) and treated with MEEK THE ABSORPTION SPECTRA ETC. 969 concentrated solution of potassium hydroxide until a clear solu-tion was obtained. Nitrornethane (10 c.c.) was then added and more dilute potassium hydroxide gradually introduced care being taken not t o render the solution strorigly alkaline. A crystalline, red precipitate was formed which was collected and recrystallised from alcohol. The orange-red needles melted a t 140'. Recrystal-lisation from alcohol resulted in lowering of the melting point to 136O but from toluene orange prismatic needles melting a t 1 4 2 O were obtained : 0.1165 gave 0.2733 CO and 0.0582 H,O. C=63*9; H=5*5. This substance is changed by acids into a berberine salt. C,lH,,O,N requires C = 63.6 ; H = 5.1 per cent. It was also obtained by condensation of methoxydihydroberberine and nitromethane in methyl-alcoholic solution. UNIVERSITIES OF SYDNEY AND LIVERPOOL. [Received September 4th 1917.