Aug.

5, 1964 COMMUNICATIONS
TO THE EDITOR 3157

but Michael and Noyes5 have reaffirmed the original

suggestions. If the energy-transfer mechanism from
benzophenone outlined in reactions 1 to 3 is correct,
nN"
then the triplet state of the carbonyl compounds de-
composes to free-radical products. These results show
that it does not yield type 2 products. I t can be inferred,
therefore, that the type 2 decomposition takes place \
'C
through a singlet intermediate as was originally
suggested.
The free-radical products in the normal photode-
c 2 1 . 3 9
Y
1.13
N 7
composition may well arise from a triplet state, but it is Fig. 1.-Bond distances and angles in pyridinium dicyano-
clear from the work of Parmenter and Noyes6 that the methylide.
type 1 products (alkane and carbon monoxide) also
arise from singlet excited states. lowed by a complete full matrix isotropic weighted
Acknowledgment.-The author wishes to thank least-squares. The final R value for the 112 independ-
the D.S.I.R. and the London Chemical Society for ent (h0l) data was 8%. The R value for the 429
apparatus grants. independent three-dimensional data was 10%. The
(5) J. L. Michael and W. A. Noyes, J , A m Chem. SOL., 88, 1027 (1963) observed dimensions of the C(CN)3- ion are: C-C
(6) C S Parmenter and W. A. Noyes, ibid , 8 8 , 416 (1963). bond length, 1.40, 1.40, and 1.40 f . O 1 A. ; C-N
DEPARTMENT O F CHEMISTRY PETERBORRELL bond length, 1.15, 1.34, and 1.16 f 0.1 A.; C-C-C
UNIVERSITYOF KEELE bond angle 119'40', 119'31', and 119' 32' f 1' ; C-C-N
KEELE,STAFFORDSHIRE
GREATBRITAIN bond angle 180, 180, and 180 f 1'. If one passes a plane
RECEIVEDMAY18, 1964 through the three nitrogen atoms of this ion, the
central carbon atom is found to be 0.13 A. above this
plane and the cyanocarbon atoms are all 0.08 A.
An X-Ray Diffraction Study of Nonplanar above this plane. Each C-C-N unit makes an angle
Carbanion Structures of 3' with respect to its projection in this plane.
There are no abnormal interionic distances.
Sir :
The structure of C ~ H L N + - C - ( C N
was
) ~ refined by the
Recently, spectroscopic investigations of the com- same procedure as that employed for NHIC(CN)3.
pound KC(CN)3have been reported by Long, Carring- The final value of R,based on 325 independent (hkl)
ton, and Cravenor,' and Miller and Baer.* An assign- data was 12%. The observed dimensions of the
ment can be made, on the basis of both infrared and molecule are shown in Fig. 1. The molecule lies
kaman spectra, for either a planar trigonal or a pyram- across a crystallographic mirror plane through atoms
idal model for the anion. Both sets of investigators 1 and 4. The pyridinium ring as well as the trigonal
favor the planar structure. A preliminary X-ray carbon atom are coplanar within f 0.01 A. The two
diffraction study by Anderson and Klewe3 of KC(CN)s cyano groups are inclined with respect to this plane
also favors a planar anion within rather large experi- such that the distance from the plane to atoms 6 and 7
mental limits. are 0.08 A. and 0.13 A,, respectively. All distances
We have determined, by X-ray diffraction techniques, reported in this molecule have an associated estimated
the crystal structure of the compounds ammonium standard deviation of f 0.01 A.
tricyanomethide, NH4C(CN)3,and pyridinium dicyano- I t thus appears that in both these systems a signifi-
methylide, C6H6N +-C-(CN)?. Complete three-dimen- cant deviation from planarity of the carbanion group
sional refinements of both structures show the trigonal exists, even though the possibility of resonance sta-
carbon atom environments to be significantly non- bilization of a planar configuration is possible. De-
planar. tailed accounts of these results will be presented else-
Both compounds crystallize in the monoclinic system where.
with the unit cell constants shown in Table I . Both
structures were solved through Patterson projections Acknowledgment.-This work was supported by
down the short axis followed by a three-dimensional grants from the National Aeronautics and Space
analysis based on packing and other considerations. Administration and the Robert A. Welch Foundation.
DEPARTMENTO F CHEMI'STRY CHARLESBUGG
TABLE
I RICEUNIVERSITY ROBERTDESIDERATO
NH4C (CN) I CsHsNs HOUSTON,
TEXAS RONALDL . SASS
Space group C62h-P21/C C*.&-Pz,/m
J U N E 12, 1964
RECEIVED
a 9.055 f 0.007 A. 7.87 f 0.02k.
b 3 . 8 7 f 0.010 i(. 12.512 f 0.004 k
C 17.325 f 0.014 A. 3.86 f 0 . 0 1 k.
B 104.6 f 0 . 2 ' 114.8 f 0 . 1 " Cubane
I, 4 2 Sir:
The structure of NH4C(CN)s was refined by two- We have completed the synthesis and fundamental
dimensional Fourier and least-squares techniques fol- characterization of the hydrocarbon cubane.
( 1 ) D. A . Long, R. A. G. Carrington, and R . B. Cravenor, Nature, 196, The bromocyclopentadienone dimer I , prepared as
371 (1962). described in an earlier communication,' is converted
(2) T. A. Miller and W. K. Baer, Spectrochim. A c t a , 19, 73 (1963)
(3) P . Anderson and B. Klewe, Nalure, 200,464 (1963). (1) P E Eaton and T W Cole, J r , J A m Chem Soc , 86, 962 (1964).
3 158 COMMUNICATIONS
TO THE EDITOR Vol. 86

& (1) (cH,oH),-H+

(2) aqueous. HC1
a . 5 %
n
+ 4 - 0 hv
q#-?
Br
10;5r,

Br Br I11
0 0
I I1

(1) SOCI, cumene

(2) (CH,)&OOH-pYridine ~

95 %

HOOC (CH3)3COOC H
IV 0 VI
V

-
QgB: 26 % KOH
55 %

VI1 VI11
IX
,
150'
ca. 30%

to the bisethylene ketal by prolonged reflux in benzene Cubanecarboxylic acid (VIII) a most versatile
with ethylene glycol and excess p-toluenesulfonic acid. intermediate, forms slowly from VI1 on reflux with 25y0
Selective regeneration of the 1-carbonyl group with aqueous potassium hydroxide; 55%, m.p. (from pen-
concentrated aqueous hydrochloric acid gives the 1- tane) 124-125', : : :A 5.91 @, 7 ~ ~ -2.2
1 , (IH, singlet)
keto-&ethylene ketal 11, 85%, m.p. (from ether) 172- and +5.6 to 6.2 (7H, multiplet) p.p.m.
1Z0, 5.78 and 6.28 p. Cubane is produced by thermal decomposition a t
Ultraviolet irradiation of I1 in benzene solution leads 150' of the t-butyl perester (IX) of cubanecarboxylic
quickly to the cage compound I11 in nearly quantita- acid in diisopropylbenzene. The hydrocarbon is re-
tive yield ; m.p. (from methylene chloride-hexane) moved from the solvent as it is formed entrained in a
148-150, 5.53 (sh), 5.61, and 5.63 (sh) p . nitrogen stream and then captured in an ice trap.
Hot aqueous 1 0 ~ potassium
o hydroxide readily con- Crystallization from methanol and sublimation just
verts I11 to the acid IV, 95%, m.p. (from methylene above room temperature a t atmospheric pressure gives
chloride-hexane) 187-189', A ~ ~ ~ 5.70 " * and 5.79 p, pure material as glistening rhombs, m.p. (sealed capil-
lary) 130-131°.8
T C D C I ~ -1.1 ( l H , singlet), $6.0 (4H, symmetrical
multiplet), 6.3 (5H, multiplet), and 7.2 ( l H , multiplet) The identity of cubane follows unmistakably from
p.p.m. the parent peak in the mass spectrum a t mle 104, from
the diagnostic simplicity of the infrared spectrum in
The t-butyl perester V, prepared by reaction of the
which the only noticeable absorptions in the region from
acid chloride of IV with t-butyl hydroperoxide and pyr-
4000 to 660 cm.-' appear a t 3000, 1231, and 851 ern.-',
idine in anhydrous ether, undergoes ready radical frag-
from the single-line proton magnetic resonance spec-
mentation in boiling cumene. Some of the alkyl radi-
trum in which the one and only absorption appears a t
cals thus formed (4070)combine with the connate t-
T 6.0 p.p.m. with width a t half-height of 0.3 c.P.s.,
butoxy radicals,* but the major fraction (55%) es-
and from the threefold symmetry axis of the monomolec-
capes this and extracts hydrogen from the solvent to
ular 5.34 A. rhombohedral unit cell.
give VI, m.p. (from pentane) 64-65', T C C I , 6.0 (4H,
symmetrical multiplet), 6.5 (6H, multiplet), and 7.2 Acknowledgment.-We are indebted to Dr. Everly
(1H. multiplet) p.p.m. Fleischer of this department for the X-ray crystallo-
Hydrolysis of the ketal VI in 75g;b aqueous sulfuric graphic analysis and to Dr. Seymour Meyerson of the
acid gives the hydrate of VII. The anhydrous ketone American Oil Company for his mass determination.
is obtained after desiccation in boiling benzene and The National Science Foundation and the Alfred P.
crystallization from hexane; 9070, m.p. 90-91 ', A 2 2 Sloan Foundation each contributed generously to the
3.23, 5.32 (w), 5.45 (m), 5.60 (s), 5.65 (s) and 5.73 (w) support of this work.
p ; the complex carbonyl absorption probably derives (3) The hydrocarbon is not thermally stable; significant decomposition
from coupling with the vibrations seen a t 10.22, 10.51, is evident at 200'.
(4) Alfred P. Sloan Foundation Research Fellow.
10.77, and 10.98 p , 7cclr6.3 (6H, multiplet) and 6.9 ( l H ,
multiplet) p.p.m. DEPARTMENT OF CHEMISTRY PHILIP E. EATON'
THEUXIVERSITV OF CHICAGO THOMASW. COLE,J R .
(2) Evidently the combination with f-butoxy radical occurs within a sol- 37. ILLISOIS
CHICAGO
vent cage, the distribution of products is unchanged with variation in con-
centration. RECEIVED
J U X E 23, 1964