Sept. 20, 1963
COMMUNICATIONS TO THE EDITOR
Acknowledgments.-Lye gratefully acknowledge the support of the National Science Foundation (NSF G-15361, GP-748), the Socony Mobil Oil Co., and the Alfred P. Sloan Foundation. (19) Trubek Fellow, 1Y62-1963
DEPARTIIEYT OF CHE'MISTRY THOMAS J KATZ COLUMBIA L-\IVERSITY PETER J GARRATT" XEK YORK27, NE\$1 - 0 ~ ~ RECEIVED JCLY31, 1963
Cyclononatetraenide. An Aromatic 10-a-Electron System
Szr. One area in which both theoretical and synthetic chemists share a common interest is that of nonbenzenoid aromatic compounds The guiding principle has been the well known Huckel rule' which states that carbomonocyclic compounds having a conjugated system of (4n 2 ) a electrons should exhibit aromatic character For the case where n = 2 , there was previously only a single example where this rule had been experimentally substantiated, z e , the cyclooctatetraene dianion l y e now wish to report the synthesis of cyclononatetraenide, a 10-a carbocycle having Dgh symmetry, in two steps from cyclooctatetraene The reaction of cyclooctatetraene with methylene chloride and methyllithium under the general conditions described by Closs and Closs" for preparing chloro-
+
630,
1
Li
I
IV
I11
cyclopropanes led to a 3 : 1 mixture of syn and anti 9chlorobicyclo [6.1.r)]nona-2,4,B-triene (Ia and Ib). The structural assignment of Ia and I b in the chlorocyclopropane fraction follows from: (a) H 1 n.m.r. [CDCl3 solution, 6 in p.p.m. us. (CH3)&i]: multiplet centered a t 6.N (olefinic H ) , triplet a t 3.45 ( J = 7.6 c.p.s., syn-CHCl), triplet a t 2.52 ( J = 4.3 c.P.s., antiCHCl) and an unresolved multiplet centered a t 1.83 248 mp (E (allylic cyclopropane H ) ; (b) A:",",'"""" 30120); (c) the presence of tertiary cyclopropyl hydrogens": 1.67 (E 0.79); and (d) elemental (1) E Huckel, Z . P h y s i k , 70, 204 (1931) ( 2 ) ( a ) A R . Ghbelohde, Chem. I n d . (London), 153 ( 1 9 % ) ; (h) T. J . K a t z , J A m . Chem. Soc., 82, 3784 (1960); (c) H. P. Fritz a n d H. Keller, Z .Vatuv,fousch , 16b,231 (1961). ( 3 ) The reported' synthesis oi cyclodecapentaene has been shown t o be in error j Attempts t o prepare compounds having t h e 1,G-diazacyclodecapentaene ring system were unsuccessful,^ and recent studies' of 2,3,i;,7-dihenzo(and t h e I-oxa and 1-thia analogs) I-methyl-1,4,.5-triazacyclohepta-2,fi-diene have shown t h a t these compounds exhibit n o aromatic character attributable t o the 10-a-electron system. Recently, the synthesis of 2,3-henzo-l,-1dioxacyclimcta-2,,j,7-triene has been reported8 a n d it has been concluded t h a t the molecule is not planar. ( 4 ) W Keppe, 0 Schichting, and H. hleister, A n n . , 660, 93 (1948). (;1) ( a ) I* E . C r a i g a n d C. E. Larrabee, J . A m Chem. S O L ,73, l l Y l (1931), ( b ) .A C Culie a n d S. W Fenton, ibiil , 75, 119.5 (1951), and references therein, (c) I ) . S Withey, J . C h e m . S o c . , 1930 (1952) (li) 4 E . Blood and C R . Noller, J . Ovg. Chem., 22, 873 (1957). ( 7 ) S 1,. Allinger a n d G . A . Youngdale, J . A m . Chem. S o r . , 84, 1020 (lQ(i2) ( 8 ) W. Schroth, K K r a n k e , a n d J . Reinhardt, A n g e w . Chem., 75, 303
(19Ki). (9) Cyclononatetraenide has also been prepared independently by T . J. Katz and P J. G a r r a t t . J . A m Chem S o c . , 86, 2852 (1963). ( I O ) A theoretical study of the structure and electronic spectrum of cyclononatetraenide is being carried o u t in collaboration with I l r . H E. Simmons and will be reported shortly (11) G . 1,. Clossand I>. E. Closs, J . A m . Chem. S o c . , 82, 5723 (1960).
2853
analysis. Anal. Calcd. for CgHgC1: C, 70.82; H , 5.94; mol. wt., 152.6. Found: C, 70.76; H, 6.14; mol. wt., 152 (mass spec.)]. Upon treating the isomeric mixture Ia and I b in tetrahydrofuran with a lithium dispersion for 2 hr. at room temperature, the multiplet a t 6.0 p.p.m. in the n.m.r. disappeared and a sharp line (width a t half height no greater than 0.8 c.P.s.) was observed a t 6.72 p,p.m. The lithium cyclononatetraenide (111) thus formed in 63Tc yield (n.m.r.) was stable in an inert atmosphere, The C13 n.m.r. of the lithium salt 111 in T H F consisted of a doublet a t 19.0 p.p.m. ( J = 137 c.p.s.)to highfield from benzene.I3 hIetathesis of the lithium salt 111 with tetraethylammonium chloride produces tetraethylammonium cyclononatetraenide (IV, 97yc) as a stable (inert atmosphere) white solid which was purified by recrystallization from anhydrous acetonitrile, dec. pt. 818" (Anal. Calcd. for CIiH2gN: C, 82.5%; H , 11.81; N , 5.66. Found: C, 82.30; H , 11.55; K, 3.75); H1 n.m.r. (dfidimethyl sulfoxide) : 6.82 (aromatic H , width a t half height no greater than 0.0 c.P.s.), quartet a t 3.03 (CH2, J = 7 c.P.s.), triplet of triplets a t 1.05 (CH3, J = 7 c.p.s. and ca. 1.3 c.P.s.). The ultraviolet absorption spectrum is very simple, as expected for a molecule having a ninefold rotational axis: 250 m p (E >61,700) anddoubletat 31iand322mp (E > 6 , l 7 0 ) . l 4 Polarographic analysis of the tetraethylammonium salt IV in acetonitrile containing LiC104 (0.1 :If) as the supporting electrolyte with a rotating platinum electrode shows what appears to be an irreversible oneelectron oxidation a t E o , 5= -0.03 v. (os. s.c.e.). As expected from this result, the tetraethylammonium salt I V can be oxidized by either tetracyanoethylerie or 7,7 ,8,8-tetracyanoquinodimethan.Attempts to detect the resulting radical'j in l ,%dirnethoxyethane solution via e.p.r. a t ambient temperature or a t -80' have been unsuccessful although the spectrum of TCNE:is observed. Lithium cyclononatetraenide (111) does not undergo detectable electron exchange over a period of 45 days with cyclooctatetraene to give cyclooctatetraene dianion. However, the salt I11 reacts with cyclopentadiene to give lithium cyclopentadienide. On the basis of the H' and C13 n.ni.r. data and the strong absorption in the ultraviolet, it is concluded that cyclononatetracnide has aromatic character, Further studies are in progress. ( 1 2 ) J. Afeinwald, A Lewis, and P G. Gassman, i b i d , 84, 977 t1962), and references therein (13) We are grateful t o Dr. H Foster for the C11 N . m . r . studies I'he C L achemical shift is in excellent agreement with t h e simple linear correlation between chemical shift and r-electron density as found recently for the a r o ~ matic series C7H7-, CsHs, CsHa- and CsHs*- [ H . Spiesecke and V,' G. Schneider, Telvohedron Letleis, No 14, 408 f1YO1) 1. Calculations using t h e least-squares method for their published d a t a together with t h a t found Although the value for CsHs- gives the equation 6C.S = l ( i 7 . 8 0 - Iti9.9 reported for CaHs- appears t o be anomalous, i t has been confirmed hy I>r. Foster. As previously noted [T Schaefer and V,' G . Schneider. C a n . J Chem , 41, 906 (19A3)1 the prediction of the H ' chemical shift from t h e a-electron density is less satisfying for larger ring aromatic compounds due t o the difficulty of applying ring size corrections. T h e HI chemical shift for 1,iCsHa from internal benzene was found t o increase Upon dilution ( 2 7 :ic 11.5 a t ( a . 3 mole lo t o 30.5 C . P . S . a t ca 0 1 mole (,To) as expected f i x decreasiny ion association (11) Some difficulty was experienced in determining the exact value of the extinction coefficients since Beer'? law was n o t followed upon carrying out the necessary dilutions, presumably because of reaction with trace impurities in the solvent. (1.5) %'e are indebted t o l>r, 11.T. Jones for this experiment.
COSTRIBUTION S o . 909 E . .\. LALASCETTE CESTRALRESEARCH DEPARTYEST R.E . BENSON EXPERIMEYTAL STATION E. I . DU POSTDE SEXOL-RS ASD COMPANY WILMISGTOK, DELAWARE RECEIVED AUGUST2. 1963
