[1950] Cddysed .Hydrogett Peqoxide Oxidation etc. Part I. 12. C’uiuly~~.I Hydrogen Peroxide Oxidation of Aromatic By J. W. COOKand (Miss) R. SCHOENTAL. A stndy has been made of the products formed by oxidation of a series of aromatic hydro- carbons (benzene naphthalene phenanthrene anthracene 1 :Zbenzanthracene 1 :2-5 :6-di-benzanthracene chrysene pyrene and 3 :4-benzpyrene) with Milas’s reagent (hydrogen peroxide in terf.-butanol catalysed by osmium tetraxide). The reagent was diluted with acetone to promote solubility of the polycyclic hydrocarbons which are sparingly soluble in tert.-butanol ; the presence of acetone appeared to have no significant influence on the course of the reaction. The products isolated included acis-dioi (in the case of phenanthreneand 1 2-beazanthracene), quinones and acids formedby ring-fission.1 :2-Benzanthracene and 1:2-5 :6dibenzanthracene gave not only the 9 10-quinones but also the 3 :4-quinones (XI1 and XTV) and in spite of the relatively low yields the method furnishes a convenient method of preparation of these ortho-quinones. Benzene gave together with other products dlomucic acid an interesting example of a simpb and direct route from an aromatic hydrocarbon to a sugar ad. The mechanism of these transformations is discussed and attention is drawn to their bearing on the mode of biochemical oxidation of the hydrocarbons for which an exact parallel by purely chemical methods has yet to be found. THEauthors have shown (Cook and Schoental J. 1948 170; Nature 1948 161 237) that oxidation of polycyclic aromatic hydrocarbons with Criegee’s osmium tetroxide reagent (osmium tetroxide in benzene-pyridine followed by hydrolysis of the resulting ohc ester) leads to smooth perhydroxylation with formation of homogeneous a-glycols.Fromthe work of Criegee (Annalen 1936 522 75 ; 1942 550 99) these were clearly cis-a-glycols. In nearly every case the positions attacked were different from those at which oxidation is effected by other reagents. The products of osmium tetroxide oxidation were similar in type to some of the compounds which have been isolated from the excreta of animals to which the hydrocarbons were administered (Boyland and Levi Bkherut. J. 1935,29 2679 ; Boyland and Shoppee J. 1947 801 ; Boyland and Wolf Biockem. J. 1948 42,XXXII; Booth and Boyland did.1947 41 XXIX; 1949 44 361). Such metabolic glycols are probably intermediates in the formation of the phenols which have been isolated in other cases (see e.g. Weigert and Mottram Cancer Res. 1946 6 109; Berenblum and Schoental ibid. p. 699; Bzochmz. J. 1949 44 604). There are however important differences between the osmium tetroxide oxidations and the biochemical oxidations. Usually the positions in the molecules at which oxidation occurs are different in the two cases. Moreover evidence is accumulating that the metabolic diols which have been isolated are trans-a-glycols. The experiments now recorded were initiated in an endeavour to find a reagent 48 Cook and Schoental Catalysed Hydrogen Peroxide Oxidation of which .would provide a closer parallel to the biochemical perhydroxylations of aromatic hydrocarbons As long ago as 1912 Dakin (" Oxidations and Reductions in the Animal Body," New York) pointed out that hydrogen peroxide alone of all the known chemical oxidising agents could bring about the same types of oxidation of a variety of compounds as do enzyme systems and the widespread occurrence of catalase and peroxidases is a pointer to the importance of hydrogen peroxide in some types of biochemical oxidation.The oxidation of aromatic hydrocarbons by hydrogen peroxide was first reported by Leeds (Bey. 1881,14,975 1382) who obtained phenol and oxalic acid from benzene ,P-naphthol from naphthalene and anthraquinone from anthracene. In recent years knowledge of the variety of conditions under which oxidations by hydrogen peroxide and by peroxides and per-acids can be effected has been widely extended (see e.g.Waters Ann. Reports 1945 42 145) and the reaction mechanisms have been extensively studied. It was decided therefore to re-examine the oxidation of these hydrocarbons by hydrogen peroxide under varying conditions and also to examine some more complex polycyclic hydrocarbons which are of interest on account of their carcinogenic activity. The present communication deals mainly with oxidations by hydrogen peroxide in tert.-butanol in presence of osmium tetroxide as catalyst. For comparison some preliminary attention has also been devoted to oxidation by hydrogen peroxide alone and by hydrogen peroxide catalysed by light by chloride ions and by haematin.The osmium tetroxide reagent was introduced by Milas and Sussman (J.Amer. Chem. SOC. 1936 58 1302) who showed that olefins are not attacked by a solution of anhydrous hydrogen peroAxide in tert.-butanol at room temperature until osmium tetroxide is added whereupon they become oxidised to a-glycols. This reagent was found unsatisfactory in many of the cases which we studied on account of the low solubility of the polycyclic hydrocarbons in tert.-butanol. Acetone was therefore used as a solvent in conjunction with the Milas reagent. Treibs (Ber. 1939 72 7) had successfully used acetone as a solvent in catalysed hydrogen peroxide oxidations but under the conditions which he used he did not apparently observe the formation of the cyclic peroxide of acetone (Wolffenstein Ber.1895 28 2265) which we regularly encountered. This appeared to have no significant influence on the course of the reaction. Parallel experiments on the oxidation of benzene by the Milas reagent led to the isolation of the same products both in presence and in absence of acetone. RESULTS. Benzene.-The hydroxylation of benzene by hydrogen peroxide in tert.-butanol catalysed by osmium tetroxide has already been reported by Milas and Sussman (J.Amer. Chem. SOC.,1937 59 2345) who obtained phenol in 23% yield. A somewhat higher yield (30%) of phenol was obtained by Milas (ibid. p. 2342) when vanadium pentoxide was used as the catalyst. We also observed the formation of phenol when benzene was treated with the osmium tetroxide catalysed reagent but if the experiment was continued for several months with excess of hydrogen peroxide a deposit separated and this was identified as allomucic acid ; from the mother-liquors there were isolated also mesotartaric and oxalic acids.The transformation of benzene into a sugar acid is noteworthy. It is uncertain whether phenol is an intermediate or whether benzene becomes fully hydroxylated to a hexahydroxycyclohexane (I or 11)which then undergoes ring- OH OH OH fission. This is a point which we are examining. Boeseken and Engelberts (Proc. Akad. Wetensch. Amsterdam 1931 34 1292; 1932 35 750; compare Boeseken and Metz Rec. Tvav. chim. 1935 54 345) have shown that oxidation of phenol by peracetic acid gives with other products a 35% yield of cis-cis-muconic acid (111).cis-Hydroxylation of the double bonds of this acid should give a mixture of allomucic (V) and mannosaccharic (VI) acids [compare Behrend and Heyer Annalen 1919 418 294 who obtained idosaccharic (VII) and mucic (VIII) acids by oxidation of trans-tram-muconic acid (IV)with sodium chlorate in presence of osmium tetroxide]. If a hexahydroxycyclohexane is formed in our conversion of benzene into allomucic acid then this probably has the completely cis-configuration (I). The alternative configuration (11)which could arise from cis-addition in pairs of hydroxyl groups to the benzenoid double bonds (see [19501 Aromatic Hydrocarbons. Part I. later) could give on ring-fission not only allomucic acid (V) but also mucic (VIII) DL-manno-saccharic (VI and enantiomorph) and DL-talomucic (IX and enantiomorph) acids and there was no indication of the formation of such a complex mixture of hexahydroxyadipic acids.yo& H0.Y.H HO. .H SH HO.7.H H0-T-H zE -+ HO.7-H -k H0.Y-H CH CO,H CO,H (111.) w.1 (VI.1 (IV.1 (VII.) (VIII.) (IX.) This degradation of benzene by oxidation with the Milas reagent is of interest in connection with the metabolic conversion of benzene into trans-trans-muconic acid (IV) which is possibly an artefact formed by isomerisation of the cis-cis-acid (111)which should be the primary product of ring fission (Jaff6 2. physiol. Chem. 1909 62 58; Drummond and Finar Biochem. J. 1938 32,79; Bernhard and Greesley Helv. Chim.Acta 1941 24 83). Such isomensation is known to be readily effected (Grundmann and Trischmann Ber. 1936 69 1755). Naphthalene Phenanthrene Chrysene Anthracene and Pyrene.-The products formed from these hydrocarbons by oxidation with the Milas reagent call for little comment. In the case of naphthalene the only product isolated was phthalic acid also obtained by Charrier and Moggi (Gazzetta 1927,57,736) by oxidation with hydrogen peroxide in an acid medium. Less extensive degradation occurs when naphthalene a-or @-naphthol or (4-naphthaquinone is oxidised with peracetic acid o-carboxyallocinnamic acid being formed in each case (aeseken and Slooff Rec. Trav. chim. 1930 49 100; Boeseken and von Konigsfeldt ibid. 1935 54 313). Phenanthrene gave cis-9 10-dihydroxy-9 10-dihydrophenanthrene phenanthraquinone and diphenic acid.Charrier and Moggi (Zoc. cit.) had obtained diphenic acid by oxidising phenanthrene with hydrogen peroxide in boiling acetic acid a result which we confirmed. Chrysene was oxidised to 1 :2chrysaquinone and anthracene to 9 10-anthraquinone which are the normal products of oxidation with chromic acid. Pyrene (X) was rapidly attacked and after about a week the products were separated into a mixture of quinones which appeared to be the 3 8-and 3 10-quinones which are the normal oxidation products. When the action of the Milas reagent ,was prolonged for a year the product was naphthalene-1 4 5 8-tetracarboxylic acid characterised as its methyl ester. There was no evidence of attack at the 1:2-positions of pyrene which are the positions exclusively attacked when pyrene is treated with osmium tetroxide in benzene-pyridine (Cook and Schoental Zoc.cit.). 1:2-Benzanthracene and 1:2-5 :6-Dibenzanthracene.-After a few days 1 :2-benzanthracene gave the 3 :4-quinone (XII)(isolated in 15% yield) and the 9 10-quinone (XI) (isolated in 20% yield). The o-quinone (XII) which is conveniently prepared in this way is evidently formed through the intermediary of cis-3 :4-dihydroxy-3 :4-dihydro-1 :2-benzanthracene (Cook and Schoental loc. cit.) a small amount of which was isolated by chromatography. When 1 :2-benzanthracene was heated on the water-bath with the reagent acidic products were formed more rapidly than in the cold. Phthalic acid and 2+henyZnaphthaZene-3 :2'-dicarboxyZic acid (XIII)were isolated and there was also evidence of the formation of keto-acids.With 1 :2-5 :6-dibenzanthracene the reaction took a similar course. On account of the low solubility of the hydrocarbon a more dilute solution was used and the reaction was very 1 50 Cook and Schoental Catalysed Hydrogen Peroxide Oxidation of sluggish at room temperature. When the reactants were heated on the water-bath the dark red 1 :2-5 6-dibenzanthra-3 4-quinone (XIV) separated after a few days and was isolated in 35% yield. From the liquors the 9 10-quinone was also isolated and in addition 2-phenyl- phenanthrene-3 :2’-dicarboxylic acid (XV) which was also obtained directly from the o-quinone (XIV) by oxidation with alkaline hydrogen peroxide.3 4-Benzpyrene (XVI).-This was readily oxidised and the solution became deep-red. A mixture of quinones was then separated from other products of oxidation by chromatography on alumina. These quinones did not account wholly for the intensity of colour (noted also in the case of some of the other hydrocarbons). Possibly the high colour intensity was due to formation of hydroxylated quinones. Warren (Biochem. J. 1943 37 338) who studied the atmospheric oxidation of aqueous acetone solutions of 3 4-benzypyrene in presence of ascorbic acid also observed the formation of deeply coloured products in addition to known quinones. Chromatography of the quinone fraction from benzpyrene resulted in only partial separation and the individual constituents were not isolated.Evidence was obtained however of the presence of the known 5 8-and 5 10-quinones (Vollmann et aZ.,Annalen 1937 531,1)and of a new o-quinone. The latter is probably 3 :4-benzpyrene-6:7-quinone which was prepared for comparative purposes by oxidation of the dihydroxydihydro-3 :4-benzpyrene of Cook and Schoental (Zoc. czt.) and characterised by formation of a quznoxaline by reaction with o-phenyl- enediamine. The diol is believed to be the 6 7-dio1 although the alternative 1 2-structure is not excluded (see numbering on formula XVI). Oxidation of 1 2-Benzanthracene with Hydrogen Peroxide in Absence of Osmium Tetroxide.-These experiments were carried out for comparison with the action of the Milas reagent. A solution of the hydrocarbon in acetone and tert.-butanol containing hydrogen peroxide remained unchanged for several weeks when kept in the dark.If methylamine hydrochloride was also present (used as a source of chloride ions) then oxidation took place and after two months a considerable proportion of the hydrocarbon had been oxidised to 1 :2-benzanthra-9 10-quinone (XI) and phthalic acid. There was no evidence of the formation of the 3 :4-quinone (XII). A similar result was obtained when haematin was used in place of methylamine hydrochloride. DISCUSSION. The results of this investigation are mainly of interest on account of the additional evidence which they furnish as to the mechanism of oxidation by the Milas reagent and they also perhaps provide one further span towards bridging the gap between biochemical and purely chemical oxidation of the aromatic hydrocarbons.Our earlier results with Criegee’s osmium tetroxide reagent showed that it is possible to bring about selective chemical attack at positions in the molecules of polycyclic aromatic hydrocarbons which are not normally the most reactive. They also showed how it is possible to simulate the “ perhydroxylation ” reactions which are a char-acteristic feature of several of the biochemical oxidations. The positions attacked by osmium tetroxide are not those at which biochemical hydroxylation takes place but in this connection it has been postulated that combination between the hydrocarbon and the oxidising enzyme may immobilise some of the more reactive centres in the hydrocarbon molecule.* The diols and the related phenols which have been isolated in metabolic experiments represent only a small proportion of the hydrocarbon administered and it is clear that most of the material undergoes more extensive degradation.Except in the case of benzene (see above) the products of this degradation have not been identified. Using 1 2-5 6-dibenzanthracene containing isotopic carbon Heidelberger and his collaborators (Cancer 1948 I,252 261) have shown that acids are formed by ring-fission of the hydrocarbon and also that some of the material is com- pletely oxidised to carbon dioxide. Metabolic studies by one of us (R. S.) with chrysene 1:2-benzanthracene 1 2-5 6-dibenzanthracene and 3 :4-benzpyrene have likewise indicated that in all these cases acidic degradation products of the hydrocarbons are present in the excreta of the animals.To this extent oxidation by the Milas reagent imitates the metabolic oxidation for prolonged treatment with the reagent always led to extensive conversion of the hydrocarbons into acidic degradation products only some of which have been identified. There remains the question as to how far the chemical perhydroxylations really follow the * Berenblum and Schoental (Cancer Res. 1943 3,686) pointed out that in the cases of 1 2-benz-anthracene and 1 :2-5 6-dibenzanthracene the positions of the molecules where metabolic oxidation occurs are those at which sulphonation takes place in the corresponding quinones. This has been mis- interpreted by Neish (Biochem.J. 1948 43,534) who states incorrectly that Berenblum and Schoental observed that benzanthracene and dibenzanthracene are metabolically attacked at points where sulphonation of these hydrocavbons occurs. Aromatic Hydrocarbons. Part I. pattern of the corresponding biochemical processes. Osmium tetroxide oxidation of olefins and of phenanthrene has been shown by Criegee (Zoc. cit.) to give cis-diols through the inter- mediary of cyclic osmic esters. It is quite conceivable that metallic catalysts could function in a similar way in biochemical " perhydroxylation." But if it can be shown that one method of oxidation consistently gives cis-diols and the other trans-diols then it is a reasonable inference that the reaction mechanisms are different. It is for this reason that great importance attaches to the determination of the configurations of the metabolic diols and as already stated evidence is accumulating that these are trans-isomers.An alternative possibility is that biochemical perhydroxylation is effected by free hydroxyl radicals and there is much current opinion in favour of the view that free radicals play an important role in biological oxidations (see Waters '' Chemistry of Free Radicals," Oxford 1946 p. 259; LuValle and Goddard Quart. Review BioZ. 1948 23 197). Milas (J. Amer. Chem. SOL.,1937 59 2342) was inclined to regard oxidation by osmium tetroxide-catalysed hydrogen peroxide as involving attack by free hydroxyl radicals formed by the action of perosmic acid on hydrogen peroxide as in its dissociation by light (compare Spring Ann.Reports 1943 40 108). That free hydroxyl radicals can oxidise benzene to phenol was shown by Stein and Weiss (Nature 1948 161 650; cf. Loebl Stein and Weiss J. 1949 2074) who found phenol among the products of the action of X-rays on water and benzene. To this extent the Milas reagent produces the same result as free hydroxyl radicals. Moreover Milas Kurz and Anslow (J.Amer. Chem. SOC. 1937 59 543) observed the oxidation of maleic acid to mesotartaric acid by hydrogen peroxide in presence of light and attributed this to addition of free hydroxyl radicals. This cis-addition rather than the trans-addition which might have been expected by such a mechanism excited no comment. Recently Merz and Waters (J. 1949 S 15) have shown that free hydroxyl radicals produced by the action of ferrous salts on aqueous hydrogen peroxide do not add to the double bond of maleic and fumaric acids.Hence the photochemical oxidation of maleic acid is not to be ascribed to the action of free hydroxyl radicals. The same type of argument may be adduced to show that the perhydroxylation of double bonds by hydrogen peroxide in presence of osmium tetroxide is not due to the action of free hydroxyl radicals. It is true that in the Milas reagent part of the hydrogen peroxide is present as tert.-butyl hydroperoxide (compare Milas and Surgenor J. Amer. Chem. SOC. 1946 68 205; Criegee and Dietrich Annalen 1948,560 135) and that an alkyl hydroperoxide has been shown to break down to give free radicals (Robertson and Waters Trans.Faraday SOC. 1946 42 201; J.,1948 1578). But the Milas reagent does not normally bring about the extensive degradations which Merz and Waters (Zoc. cit.) observed with free hydroxyl radicals; it does effect perhydroxylation of the double bonds in compounds of the maleic acid type and moreover the method leads to cis-addition to the double bond (Milas and Sussman J. Amer. Chem. SOC. 1937 59 2345; Milas Sussman and Mason ibid. 1939 61 1844). It is clear in fact that the oxidation involves the intermediate formation of cyclic osmic esters as in the Criegee method in which such esters are isolated. Our own results with polycyclic hydxocarbons show that in part oxidation at a " potential " double bond takes place by the same mechanism.The isolation of cis-a-diols from phenanthrene and 1 2-benzanthraceneJ with no evidence of the presence of stereoisomeric trans-diols points clearly to the intermediate formation of cyclic osmic esters. These diols undoubtedly represent the first stage also in the reactions which lead to 1:2-benzanthra-3 :4-quinone (XII) and the acid (XIII) from 1:2-benzanthracene and the allied compounds (XIV) and (XV) from 1 :'2-5 :6-dibenzanthracene. Moreover the quite appreciable amounts of these products isolated indicate that the reactions by this route represent a considerable proportion of the whole. However quite a different mechanism must operate in the oxidation of anthracene 1 2-benzanthracene and 1 2-5 6-dibenzanthracene to the 9 :10-quinones (e.g.XI) and in the oxidation of pyrene and 3 :4-benzpyrene to their heteronuclear quinones. These oxidations may be due to the action of free hydroxyl radicals formed by dissociation of hydrogen peroxide but there is no evidence bearing on this point ,and judgment must be reserved. We are extending this work to study the effect on polycyclic hydrocarbons of free hydroxyl radicals and also of hydrogen peroxide in the presence of other catalysts. In this connection it may be noted that Sequin (Compt. vend. 1943 216 667) obtained only the #vans-a-glycol by oxidising cyclohexene with hydrogen peroxide in acetone in presence of selenium dioxide. A similar result was reported by Treibs (Angew. Chem. 1939,52,698) with pervanadic acid as catalyst. 52 Cook and Schoental Catalysed Hydrogen Peroxide Oxidation of EXPEWMENTAL.General.-A solution of the hydrocarbon (0.001 g.-mol.) in acetone was treated with a stock solution of hydrogen peroxide (2-10 c.c.) prepared by diluting 90% hydrogen peroxide to 20-40% with tert.-butanol and with a 1-2% solution of osmium tetroxide in tert.-butanol so that the final concen- tration of the tetroxide varied between O.Olyoand 0.5%. The solution was kept in a stoppered flask at room temperature in diffused light. Addition of the ted-butanol solutions usually resulted in partial separation of the crystalline hydrocarbons but the crystals redissolved in the course of the reaction. The colourless solutions slowly developed yellow or red colours of increasing intensity. There was usually an induction period when highly purified material was used.After the hydrogen peroxide initially added had been consumed the osmium tetroxide became reduced so that the solution became almost black. This colour was discharged by addition of more hydrogen peroxide solution. The rapidity with which this darkening of colour occurred furnished an index of the rate of oxidation of the hydrocarbons except in the cases of pyrene and 3 :4-benzpyrene when very dark oxidation products were formed and the dark colour then persisted for some months. In view of the complex mixtures of oxidation products formed and the difficulty of isolating individual constituents a quantitative study was not attempted. Yields are recorded only when reaction products actually crystallised from soh tion.For isolation of the products any solid which had separated was collected and the filtrate concentrated under reduced pressure at 40-40". By this means solvents osmium tetroxide and volatile products were removed. Colourless crystals which collected in the still-head and condenser were identified as the trimeric cyclic peroxide of acetone m. p. 9695" (Wolffenstein Zoc. cit.). Attempts to isolate the products without this preliminary concentration resulted in formation of dark colloidal suspensions of lower oxides of osmium and as far as could be judged the distillation process did not bring about any alteration in the oxidation products. The residue after distillation was a yellow or brownish viscous oil which was treated with dilute sodium hydrogen carbonate solution and extracted with benzene.Dark resinous material remained insoluble in both phases but dissolved when sodium hydroxide solution was added after removal of the sodium hydrogen carbonate layer. Acidification of the dark sodium hydroxide extracts gave brownish material which did not crystallise and gave no crystalline derivatives by reductive methylation or acetylation. The benzene extracts were submitted to chromatography on alumina and were thus separated into unchanged hydrocarbon diol (in twocases). and quinones. In all cases highly coloured zones of stronglv adsorbed material were visible on the upper part of the column but pure products could not be obtained after these were eluted with acidified alcohol. The sodium hydrogen carbonate extract was acidified and the resulting acids separated by fractional crystallisation from various solvents or by vacuum distillation of their methyl esters formed by treatment with ethereal diazomethane.Benzene.-Two parallel experiments were made. In one the reaction was carried out in acetone solution as in the other cases recorded in this paper. In the other tert.-butanol alone was used as solvent. Oxidation took place more rapidly in acetone solution but in both cases the same products were isolated. Thereaction mixtures prepared in each case by using 10 C.C. of benzene became yellow in a few days and phenol was detected by its smell by the preparation of tribromophenol m. p. 93" from a sample of distillate and by the blue colour given with 2 :6-dichloroquinonechloroimidein neutral or alkaline solution (Porteous and Williams Biochem.J. 1949 44 56). The colour of the solution of reactants deepened on storage but after some months during which several additions of hydrogen peroxide solution were made the colour became pale yellow and a white precipitate was formed. This (35 mg.) was recrystallised from hot water and formed a microcrystalline powder which decomposed at 197". Analysis showed the compound to be a tetrahydroxyadipic acid (Found C 34.0; H 5.0. Calc. fofC&&8 C 34-3; H 4.8%). Its ethyl ester prepared by heating it under reflux with ethanol containmg a few drops of concentrated hydrochloric acid had m. p. ca. 150". These data are in approximate agreement with those given for allomucic acid by Posternak (Helv.Chim. Acta 1935 18,1283) (acid m. p. 197-198" ; ethyl ester m. p. 153-154") and by Lapsley Robertson and Patterson (J. 1940 862) (acid m. p. 199-200" ethyl ester m. p. 155"). Through the kindness of Mr. J. Robertson who generously gave us a sample of O-tetra-acetylallomucic acid prepared by the method described by himself and his collaborators we have been able to establish that our oxidation product is in fact allomucic acid. The methyl ester of our acid had m. p. 170" and gave an O-tetra-acetyl derivative m. p. 20A206". The m. p.s of the acid and its three derivatives were not depressed by mixing with the acid (prepared by acid hydrolysis of the authentic O-tetra-acetylallomucic acid) and its corresponding derivatives.On the other hand a mixture of the methyl ester of the tetra-acetate of our acid and methyl O-tetra-acetylmucate m.p. 196-197" (Simon and Guillaumin Compt. vend. 1924 179, 1324) melted at 180-196". Pure methy2 allomucate formed transparent prisms (from methanol) m. p. 175" after softening (Found C 40.6; H 5-9. C8H1408 requires C 40-3; H 5.9%). Methyl O-tetra-ucetylallomucate prepared from methyl allomucate by acetylation with acetic anhydride formed transparent rhombic prisms m. p. 205-206" (Found C 47.3 ; H 5.6. Cl6HSzOlzrequires C 47.3 ; H. 5.5%). In another experiment a mixture of benzene (10 c.c.) and acetone (10 c.c.) was treated with 20/ osmium tetroxide (2 c.c.) and 40% hydrogen peroxide (60 c.c.) both in tert.-butanol. After nine months the filtered solution was concentrated under reduced pressure at 40-50".A solution in water of the residual oil was partly neutralised with sodium carbonate and the solution heated under reflux for 2# hours with p-bromophenacyl bromide (0.5 g.) in ethanol (10 c.c.). After concentration and cooling the resulting p-bromophenacyl ester was recrystallised from ethanol and formed glistening leaflets m. p. 198-200". This was shown to be p-bromophenucyl mesotartrate a sample of which was prepared from authentic mesotartaric acid and had m. p- 198-200" alone or mixed with the above product (Found C 44-25; H 3-2. C20H,60,Br requires C 44.1 ; H 3*0p/,). [19501 A romatic Hydrocarbons. Part I. In yet another experiment the filtrate from the crude allomucic acid was concentrated under reduced pressure the residue dissolved in water and almost neutralised with sodium hydroxide and the boiling solution then treated with saturated calcium acetate solution.Without cooling the white precipitate was collected washed and dried. It was shown by titration with warm acid permanganate solution to consist of calcium oxalate. The dltrate was concentrated and gave large crystals which were redissolved in hot water ; the solution was acidified and then extracted with ether. After removal of the ether the residue was treated with ethereal diazomethane and gave colourless prisms m. p. 111-112" shown by mixed m. p. to consist of methyl mesotartrate. In yet another experiment the crude mixture of acids freed from volatile substances was estefied with diazomethane and the esters were distilled.The fraction b. p. 110'/0-3 mm. gave again methyl mesotartrate. The higher-boiling esters could not be separated ; much of the material underwent decomposition during distillation. Although the hydroxyl groups of mesotartaric acid were not affected by diazomethane it may be noted that Schmidt and Zeiser (Bev.,1934,67 2120) found that tartaric and other hydroxy-acids undergo etherifcation as well as esterification with diazomethane. Naphthalene.-A solution of the hydrocarbon (5 g.) in acetone (25 c.c.) with 1 yoosmium tetroxide in tert.-butanol(2-5c.c.) and hydrogen peroxide solution (15 c.c.) immediately became yellow and oxidation proceeded rapidly ; further quantities of hydrogen peroxide solution were added as required.After 18 days the solution was worked up in the usual way. The acidic fraction was mainly soluble in water. It was extracted with ether recovered from the extract and recrystallised several times from ethyl acetate and aqueous methanol. The resulting acid had m. p. ca. 200" and on sublimation gave long needles m. p. 137" shown by mixed m. p. to be phthalic anhydride. When the solution of the reactants was worked up after shorter periods there were shown to be present phenolic material and also coloured neutral products probably naphthaquinones (compare Milas U.S.P. ,2,395,638). Phenanthrene.-The hydrocarbon (3 g.) in acetone (15 c.c.) was treated with the Milas reagent (2 C.C. of osmium tetroxide solution and 10 C.C. of hydrogen peroxide solution) at room temperature.Further additions of hydrogen peroxide solution were made from time to time about 60 C.C. of a 20% solution being used in all. After 26 days the red solution was concentrated in vucuo and then deposited crystals (55 mg.) m. p. 204-206" which were shown to consist of phenanthraquinone. The filtrate was worked up3n the usual way and on passing the benzene solution through a column of alumina there were obtained from the first benzene filtrate some unchanged phenanthrene and then a fraction which gave fine colourless needles (from light petroleum) of cis-9 10-dihydroxy-9 10-dihydrophenanthrene (25 mg.) m. p. 178" alone or mixed with a specimen prepared as described by Criegee Marchand and Wannowius (Annalen 1942,550 99). A yellow zone was cut from the alumina column and gave on elution with benzene more phenanthraquinone m.p. 206-207". The acidic fraction was repeatedly crystallised from aqueous methanol and ethyl acetate. This gave colourless prisms m. p. 228-230". shown by mixed m. p. to be diphenic acid. Identification was completed by conversion into the anhydride m. p. and mixed m. p. 217". The sodium hydroxide extract on acidification gave a product which was purified by passing its benzene solution through a column of alumina. This gave a small amount of 9-phenanthrol m. p. and mixed m. p. 148". Chvysene.-In this case reaction was slow but the hydrocarbon in suspension (using 330 mg. in 50 C.C. of acetone) dissolved during the course of 3 weeks. The red solution was concentrated in vucuo and gave red needles of 1:2-chrysaquinone m.p. 237-238" not depressed by mixing with an authentic specimen. Small amounts of acidic and other products were also formed. Anthracene.-The hydrocarbon (100 mg.) in acetone (10 c.c.) was treated with 1yo osmium tetroxide solution (1 c.c.) and 20% hydrogen peroxide solution (2 c.c.). After 4 days long needles of anthraquinone began to separate (m. p. and mixed m. p. 275"). In another experiment 250 mg. of anthracene gave after 4 months 170 mg. of anthraquinone and traces of unidentified acidic products. Pyrene.-A solution of this hydrocarbon (350 mg.) in acetone (10 c.c.) became pink immediately on the addition of the Milas reagent. After a week the deep-red solution was worked up. Chromatography of the benzene extract led to recovery of some unchanged hydrocarbon and gave several coloured bands.Chloroform eluted a mixture of quinones which were not completely separated by repeated chromato- graphy in benzene. The colours given by the various fractions with concentrated sulphuric acid indicated the presence of the 3 8-and 3 10-quinones (Vollmann et al. loc. cit.). There was no definite evidence of the presence of the 1 :2-quinone. In another experiment the solution of reactants was kept for about a year. The deep colour slowly faded and a pale yellow solid was deposited. This was recrystallised from much ethanol and then sublimed under reduced pressure giving yellowish needles of naphthalene-1 :4 :5 8-tetracarboxylic dianhydride m. p. above 360".Treatment of the silver salt with methyl iodide gave the methyl ester as colourless leaflets m. p. 196-198" not depressed by mixture with a sample similarly prepared from Cook and Hewett's tetracarboxylic acid (J. 1933 405) (Found C 60.2 ; H 4-7. C18Hl,08 requires C 60.0 ; H 4.5%). 1 :2-~enzun~hvucene.-(~) A solution of the hydrocarbon (0.5 9.) in acetone (20 c.c.) was treated with 40% hydrogen peroxide (2 c.c.) and 1% osmium tetroxide (1 c.c.) both in tert.-butanol. The solution soon became red and in the course of 3-4 days red crystals separated (80 mg.). After crystallisation from chloroform these had m. p. 260-262". not depressed by admixture with a sample of the quinone (XII) prepared by chromic acid oxidation of cis-3 :4-dihydroxy-3 :4-dihydro-1 2-benzanthracene.The product gave a purple solution in concentrated sulphuric acid in agreement with Fieser and Dietz (J. qmer. Chem. Soc. 1929 51 3141). The quinone reacted with o-phenylenediamine in acetic acid solution to give the quinoxalzne as pale yellow needles m. p. 256-257' (Found C 86.9; H 4-3. C,,H,,N requires C 87.2; H 4.3%). The mother-liquors from which this quinone had been separated were worked up in the normal manner. Chromatography of the benzene extract showed that very little unchanged hydrocarbon was present. The benzene eluate gave 1 :2-benzanthra-9 10-quinone (XI)(120 mg.) as orange needles m. p. 167-168'. Chromatographic purification of the material from the mother-liquors of this gave a Catalysed Hydrogen PHoxide Oxidation etc.Part I. small amount of cis3 4-dihydroxy-3 4-dihydro-1 :2-benzanthracene m. p. 201-205O not depressed by admixture with a sample prepared as described by Cook and Schoental (Zoc. cit.). The acidic products insoluble in water (i-e. precipitated by acidification of the sodium hydrogen carbonate extract) were recrystallised several times from ethyl acetate and gave a small amount of colourless crystals m. p. 245". This acid was shown to be identical with 2-phenyZnuphthrrlene-3 :2'4-carboxylic acid (XIII) colourless prisms. m. p. 249-250' (from ethyl acetate) formed by oxidation of 1:2-benzanthra-3 :4-quinone (XII) with alkaline hydrogen peroxide (compare Weitz Schobbert and Seibert Bar. 1935 88,1163) and the same acid was also formed from the quinone (XII)by treatment with the Milas reagent for several weeks (Found C 74-3 ; H 4-1.C1,Hi,O requires C 74.0 ; H 4.1%). Its methyZ ester prepared by heating it with hydrochloric acid in methanol formed colourless plates m. p. 86-87.5" (Found C 75.05 ; H 5.0. C,,H,,O requires C. 75.1 ; H 5.0%) In another experiment on the oxidation of benzanthracene by the Milas reagent the acidic products were esterified with diazomethane and distilled in vucuo. The first fraction b. p. 90-100"/0~3 mm was mostly methyl phthalate identified by conversion into the anhydride. The fraction b. p. 120-180'/0~3 mm. gave on hydrolysis the acid (XIII). Treatment of the crude acid mixture with 2 :4-dinitrophenylhydrazine gave a small amount of solid indicating the presence of keto-acids but a pure substance could not be isolated.(b) 1 2-Benzanthracene (0.5 g.) in acetone (20 c.c.) was treated with 50% hydrogen peroxide in tert.-butanol (3 c.c.) and methylamine hydrochloride (140 mg.). The mixture was kept in the dark for 2 months and gave 1 2-benzanthra-9 10-quinone (170 mg.) and acidic products from which phthalic acid was isolated by distillation of the esterified material. (c) A similar experiment with haematin (prepared from haemin as described by Fischer et al. 2. ?hysioZ. Chem.,1930,193,156) in place of methylamine hydrochloride gave 1 :2-benzanthra-9 :10-quinone in 50% yield and some phthalic acid. 1 :2-5 :6-Dzbenzanthracene.-The slight solubility of this hydrocarbon necessitated the use of a very dilute solution and the reaction was exceedingly slow at room temperature.It was thsrefore accelerated by heating. The hydrocarbon (500 mg.) in acetone (50 c.c.) was treated with 50% hydrogen peroxide (10 c.c.) and 1%osmium tetroxide (10 c.c.) both in tert.-butanol and the solution was heated at 70" for 3-4 days. The purple crystals which separated (160 mg.) were recrystallised from acetic acid and formed dark red needles m. p. 325". This substance was shown to be identical with 1 :2-5 :6-dibenz-anthra-3 4-quinone (XIV) prepared as described by Stephenson (1.. 1949 2620) who gives m. p. 327- 329" (compare Cook J. 1933 1594). Concentration of the original filtrate gave a yellow solid which crystallised from benzene in orange-yellow needles m. p. 246" consisting of 1 2-5 :6-dibenzanthra-9 10-quinone (Clar Ber.1929 62,350). The sodium hydrogen carbonate fraction OD acidification gave a solid precipitate which after several recrystallisations formed a cream microcrystalline powder (from acetone) m. p. 307-309". This was 2-phenylphenanthrene-3 :2'-dicarboxylic acid (XV) a sample of which was prepared for comparison by oxidation of the quinone (XIV) with hydrogen peroxide in boiling methanolic sodium hydroxide (compare Weitz et aZ. Zoc. cit.). The resulting acid had m. p. 316' and did not depress the m. p. of the acid obtained by oxidising the hydrocarbon with the Milas resgent or that obtained by another route by Stephenson (Zoc. cit.) who gives m. p. 311-313". For characterisation. the methyl ester of (XV) was prepared by the Fischer-Speier method.It formed almost colourless leaflets (from ethanol) m. p. 1dA 145" (Found C 77.3 ; H 5.1. C,,H,,O requires C 77.8 ; H 4.9%). 3 :4-Benzpyrene (XVI) .-The oxidation of this hydrocarbon with the Milas reagent proceeded as with pyrene. Chromatography on alumina of the benzene extract after concentration (a large volume of benzene was required) gave a red zone. The column was cut and this was eluted with chloroform. Repeated chromatography from benzene of the concentrated eluate resulted in partial separation into a lower (yellow) band and an upper (red) band. These were cut eluted with chloroform and concentrated. The residue from the yellow band gave a cherry-red solution in concentrated sulphuric acid whereas that from the red band gave an olive-brown solution.These colours are consistent with the View that the yellow quinone is the 5 10-quinone and the red quinone the 5 :8-quinone obtained by Vollmann et uZ. (Zoc. cit.) by chromic acid oxidation of 3 4-benzpyrene (XVI). Some of the fractions from the chromatogram gave a deep violet colour with sulphuric acid and reacted with o-phenylenediamine in acetic acid to give a yellow precipitate indicating an o-quinone. For comparison the dihydroxydihydro-3 4-benzpyrene of Cook and Schoental (Zoc. cit.) was oxidised with chromic acid and the resulting pinone probably the 6 7-quinone crystallised from glacial acetic acid. It formed orange-red needles m. p. 240" (decomp.) and gave the expected deep-violet solution in concentrated sulphuric acid (Found C 85.0 ; H 3-6.C&,,O requires C 85.1 ; H 365%). The quimxdine prepared by treatment with o-phenylenediamine in acetic acid formed small yellow needles m. p. 318" (decomp.) not depressed by admixture with the precipitate similarly formed from fractions of the chromatogram (Found C 87-8; H 4-2; N 7.8. C,,H,,N requires C 88.1 ; H 4.0 ; N 7.9%). We thank the British Empire Cancer Campaign for a grant which has enabled one of us (R. S.) to take part in this work. We are also much indebted to hlr. J. Robertson for the gift of a specimen of O-tetra-acetylallomucic acid. Microanalyses were by Miss R. H. Kennaway. UNIVERSITY OF GLASGOW. [Received September 20th 1949.1