NOTES
April, 1951
18'77
Aromatic Cyclodehydrogenation. X. Studies on Ring Oxygen Compounds. Dinaphtho[ 1,2,1',2']-
CHrOH
I
furan'
BY MILTONORCHINAND LESLIEREGGEL I I n connection with another problem,2 i t became HCOH of interest to synthesize 1-(0-tolyl)- 1-(2'-naphthyl) I HCOH HCOH 1,2-dihydronaphthalene (IV) , A possible route to IV appeared to be the reaction of o-tolylmagHLOH nesium bromide with 1-keto-2-(1'-tetraly1idene)ICHiOH I CH21,2,3,4-tetrahydronaphthalene(I) ; 1,2-addition to I1 I the carbonyl group would give the carbinol, 11, whereas 1,4-addition to the conjugated system would give the desired intermediate ketone, 111. It was thought that if 1,4-addition took place, CH20H I11 might be separated via the semicarbazone and I then be reduced and dehydrogenated to yield IV. The anticipated results were not obtained, however. AH CHzOH About 61% of the starting ketone, I, was recovered I when the Grignard reaction mixture was decomCH2OH posed and the solvents removed. The balance CHzOH CHzOH of the product was treated with semicarbazide, and I the ketonic material was recovered by hydrolysis , HAHCO-' CHzOH and reduced by the Huang-Minlon modification I I CH20H C H 2 e of the Wow-Kishner method; the crude reduced 1. material was dehydrogenated, but no pure product IY could be isolated. The material that did not form upon treatment with fi-nitrobenzoyl chloride fur- a semicarbazone, presumed to be the carbinol, 11, nished a 30% yield (based on the weight of the re- formed by 1,2-addition, or its dehydration product, duced dialdehyde) of glycerol tri-(p-nitrobenzoate) was catalytically dehydrogenated. A small of m.p. 193-195'. The fate of the remainder of the amount of 1-(o-tolyl)-2-(1'-naphthyl) -naphthalene molecule has not been determined, but i t does not (V) was isolated, but the main product was a colorseem possible that i t could have been converted to less substance, C20H120, melting point 155.0-156.5', glycerol under the experimental conditions. which formed a red 2,4,7-trinitrofluorenonecomIsolation of the p-nitrobenzoate of glycerol rather plex, an orange s-trinitrobenzene complex, and an than that of ethylene glycol (m.p. 141'), which orange dipicrate, m.p. 164.5-165.2'. Interestingly would be expected if sedoheptulosan had formula I, enough, the ultraviolet absorption spectrum of this can be explained only on the basis of formula I1 or compound proved to be identical with that of a 111. Of these we prefer formula 111because of the compound we had previously shown to be present failure of tetratosylsedoheptulosan to react with in a mixture obtained by direct cyclodehydrogenasodium iodide a t 100' even though it must contain a tion of the ketone, 1. primary tosyloxy group. This sluggish behavior In the earlier work i t was shown that treatment has been reported as a charadteristic of the primary of the ketone, I, with a palladium-on-charcoal tosyloxy group adjacent to the carbonyl group in 1- catalyst resulted3 in the formation of a mixture tosyl-2,3,4,5-diisopropy~dene-~-fructose,~ 1,4-di- from which i t was possible to isolate dibenzo [c,kl]tosyl-2,3-isopropylidene-~-xylulose,~ and l-tosyl- xanthene (VIII) ; ultraviolet absorption spectra On the other showed the presence of another substance, which 2,3 :4,6-diisopropylidene-~-sorbose.~ hand, a normal replacement of the primary tosyloxy could not be obtained pure, to which the structure group by a n iodine atom should be expected from of dinaphtho [1,2,1',2']furan (VII) was assigned. the tetratosyl derivative of 11. The spectrum of VII, as previously obtained from From these experiments, and certain other data mixtures of VI1 and VIII, was identical with the that are in themselves inconclusive, i t now appears spectrum of the compound of melting point 155.0that sedoheptulosan should be represented by 156.5' obtained in the present work; no trace of formula 111, and that i t has the same ring structure VI11 could be detected. The isolation of the dithat occurs in the known 1,6-anhydrides of glucose, naphthofuran, VII, in the present investigation, mannose, altrose, galactose, and idose. A repeti- strengthens the case for the assigned structure tion of the methylation studies is now under way but it cannot as yet be considered as completely in an attempt to resolve the contradiction with the substantiated. earlier work. In attempting to account for the formation of the furan, VII, i t was thought that the Grignard XATIONAL INSTITUTE OF ARTHRITIS reagent might have reduced I to VI, and that VI, AND METABOLIC DISEASES NATIONAL INSTITUTES OF HEALTH present in the fraction that did not form a semiPUBLICHEALTHSERVICE carbazone, might have produced VI1 on dehydroFEDERALSECURITY AGENCY
4
