COMMUNICATIONS TO THE EDITOR
3516
VOl. 85
The evidence cited includes the "small" low-field shifts of the proton resonances in hexamethyldisiloxane' and tris(trimethylsilyl)amine2 as against tetramethylsilane; the smaller low-field shifts with increasing n in the series (CHa)4-,,lZC1, and (CHa)a-I,lIHCl,, when hl is Si than when M is C s , r ; the decrease in shielding along the series (CH3);,SiX,X = F, C1. Rr, I s ; and the increase in J(I3CH3) in (1RCH3)SiX(12CH:,), and in J(2gSiH)in (12CEI:ij32ySiX along the series CH3 < F < C1 < Br < I . 5 W'e have recently studied several substituted methylsilanes and related alkanes; from the results, taken with published work, we conclude t h a t many of the so-called anomalies in the spectra of silicon compounds are also to be found in the spectra of similar derivatives of carbon, and t h a t it is a t present unjustified to use these effects as evidence in favor of the occurrence of n-bonding in silicon conipounds. The 0-proton shieldings in the series of compounds / I CH3SiH2Xdecrease along the series X = H . N , 0, F (which is consistent with increasing inductive deshielding), but increase along the series X = I , Br, 080 OBI 0.90 0.95 1.00 I/[ I eip(-AG./NkTi] C1, F6; a similar effect is observed both in the dimethylsilyl' and t r i m e t h y l ~ i l y l halides, ~ and has been exFig. 1. -Plot o f eq. 1 for various trial values of AGO (kcal./mole). plained for the last named derivatives in terms of A variation of Iioz with temperature can also be increasing (p d ) n-bonding between silicon and the found in situations where rigidity of the molecular halogen atoms in the order I < Br < C1 < F. The same framework would seem to preclude the possibility of effect, however, has been observed in e t h ~ l , iso~,~ radical conformational changes. For example, we have propyl,y,g and cyclohexyl'" halides. presobserved such temperature variations for ROTin camphor ent there is no satisfactory explanation for this, b u t and norcainphor. In these cases, no temperaturesince it occurs in compounds in which the a-atom is independent value for AGO would yield a straight line carbon, i t is unlikely to be caused by (p + d ) r-bonding. It is now clearly established that SiHXYZ resonance TABLE I chemical shifts are less sensitive than are C H X U Z ----Yc Diequatorial conformer-shifts to changes in the rest of the molecule6,'; this is Present UltraSolvent Ro2BR X lO*O C.L, d a t a violet" O.R.IXa true not only when the substituents are potentially Methanol + 1 40 97 f 2 100 99 strongly a-bonding groups, such as --OK or -C1. but Dioxane +0 .80 96 3z 2 88 100 also for others (like -I, -Br, and -SR) which are unEI'A +0 42 95 f 2 likely to be involved in strong n-bonds. It is therefore CCI, -1 21 90 f 3 not surprising that CH3SiXYZ resonances are less Isooctane - 1 87 89 f 3 82 82 sensitive than CH3CXUZ resonances to changes in From J , Allingcr, S . L. Allinger, L . E. Geller, arid C . Djerassi, X, U,and 2 . Substitution of -I (or -Br) for (Si)H in J . O i g . Chcni., 26, 3321 (1961). These authors indicate an acthe three methylsilanes, for instance, shifts the curacy of about i10' I p-proton resonance 0.7 (or 0.5) p.p.m. to low field, while in alkanes the analogous substitution shifts are in the plot suggested for eq. 1, and these results are about 0.9 and 0.7 p.p.m., r e ~ p e c t i v e l y . ~For polar, indicative of nonnegligible A S o values, such as one might expect in the case of asymmetric s o l ~ a t i o n . ~ potentially strongly a-bonding groups, the shifts follow the same pattern : -F or -OR substitution shifts The analysis of these data in terms of this hypothesis in the methylsilanes are about 0.15 and 0.05 p.p.m. and the problems posed by a multiplicity of conformers to low field, respectively,'j as against substitution shifts will be treated in a later paper. for the same substituents in alkanes of some 0.3 and 0.1 ( Y ) A. hloscowitz. K . M. Wellman, and C. I)jerassi, Pvoc. iVoil. Acad. S c i . L ' S . . 60 ( N o v . , 1963). p,p,m,g There is no reason to conclude from these ( I O ) 1 7 e I l w of the Alfred P Sloan Foundation. data that there is any unusual bonding in the silicon ( I 1 ) National Institutes of Health Postdoctriral Fellow, 19li2-1903. compounds. ALBERTM O S C O W I T Z ~ ~ 1) EPART41 E S T 0 F c H E M ISTRY Rather more convincing evidence indicating (p d) UXIVERSITY OF MISSESOTA n-bonding conies from the high-field shifts of the SiH M I N T E APOLI s 14, M I N s E SOT A proton resonances of fluorosilane (0.05 p.p.m.),l 1 DEPARTSIEST OF CHEMISTRY KEITHWELL MAN^^ difluorosilane (0.20)," and methylfluorosilane (0.0:3)7; STANFORI) USIVERSITY CARLDJERASSI and of the P-proton resonances of methylfluorosilane STASFORD, CALIFORNIA RECEIVED AUGUST 30, 1963 (0.05)7 and dimethylfluorosilane (0.02)' on substitution of one of the (Si)H atoms by -F. Even this, however, is by no means decisive. I t is very dangerous Nuclear Magnetic Resonance Spectroscopy and to draw conclusions from changes in S i H chemical (p d) a-Bonding in Silicon Compounds (3) M ,P. Brown and I ) . E Webster, J . P h y s . Chem., 64, 658 (1900) 20
/
,
I 1
-+
-+
-
Sir: In several studies of the n.m.r. spectra of substituted methylsilanes, it has been concluded t h a t the results can best be explained in terms of (p --t d ) a-bonding between silicon and electronegative atoms or groups. l P 5 [ I ) H Schmidbaur a n d 51. Schmidt, J .4m C h e n t . Soc., 84, 10G5 (1962). ( 2 ) H Schmidbaur a n d M Schmidt, A n g e w . Chem. Inlerit E d . Etigl , 1, 327 (19(i2)
( 4 ) D . E. Webster, .I. Chem Soc., ,5132 (1900). S o c . , 86, 2331 (1903). (a) E . A . V . Ebsworth and S . G . Frankiss, T r a n s F a r a d a r Soc., 6 9 , 1.518 (1903) 17) E . A . V. Ebsworth a n d S.G Frankiss, unpublished observations ( 8 ) A . A . Bothner-By a n d S C S a a r - C o l i n , J . A m . Chem. Sor , 8 0 , 1728 (19.58). (9) J . R . Cavanaugh and B. P. Dailey, J C h e m . P h y s , 3 4 , IO99 ( I O i i l ) (10) W C . Neikam and B . P. Dailey, ibid., 38, 445 (1903) (11) E . A , V . Ebsworth and J . J . Turner, J. P h y s . Chem , 6'7, 805 ( 1 9 6 3 )
(a) H. Schmidbaur, J A m C h e m .
Nov. 5 , 1963
COMMUNICATIONS TO THE EDITOR
shifts when i t is still not known why SiH4 gives a resonance to low field of CH4.'l The high-field shifts of the p-protons are very small. In unsaturated carbon compounds the high-field shifts produced by -F or -OR substituents, which are likely to be due at least in part to a-bonding effects, are much greater, being some 0.5 p.p.m. (-F substitution) in acetylene,13 1.0 (-F) and 1.3 (-OR)in ethylene,l4.l5and 0.9 (-OR)in thiophene. I G We have previously drawn attention to the rough correlation between increasing J(13CH) in CH3MXYZ and decreasing 7(CH3) in '*CH311XYZ, which appears to hold whether 11 is C or Si.6 It therefore seems as unreasonable to invoke ( p + d ) a-bonding to explain changes in J(13CHaSiXVZ)as i t is to explain the pproton chemical shifts in the same way. Moreover, although the available data are not very precise, it seems that J(I3CH3)in substituted ethanes may change with substituent in much the same way as in methylsilanes; J(I3CH3)in CH3CHBr2, for instance, is 131.0 i 0.3 c.p.s. as against 128.1 i 0.5 in CH3CHF2.10 These values may be compared with J(13CH3) of 118.8 + 1 c.p.s. in (CH3)&iF, and of 121.0 1 in (CHa)$3Br .j Finally, it must be emphasized that in molecules of formula 12CH329SiXYZ neither the sign of J(29SiH)nor the coupling mechanism is yet known. At present it is by no means established t h a t this coupling constant [or the analogous J(H-12C-13C) in carbon compounds] must depend only on the s-character in the intervening bonds.
3517
some general way, our evidence is t h a t only one E t 2 0 is specifically associated with it, probably because when this one is on one side of the four-membered ring of the dimer, alkyl groups (C3H7 in I) on the other side of the ring shield i t from association of solvent there. Ordinate values in Fig. 1 show the mole ratio of E t 2 0 to hexane in the vapor phase at 25' in equilibrium with 50 ml. of hexane to which increments of EtzO were added, curve A for ordinary hexane and curve B for hexane containing 0.08 mole of BuLi. It. is seen t h a t the vapor pressure of E t 0 2 is depressed by the BuLi until a 1 : 2 respective molar ratio is reached, i . e . , curve B shows a discontinuity at 0.04 mole of Et20 added. This discontinuity is even more distinct it1 curve C , which is E t 2 0 added L I S . the ratio of ordinate values from A and R.
10
*
(12) W. G . Schneider, H J . Bernstein, a n d J. A . Pople, J . Chem Phys , 28, (io1 (19%). (1.0 W . 1)renth and A . Loewenstein, Rec. l i a v . chim., 81, 635 (1562). ( 1 4 ) C. N . Banwell and S . Sheppard, Mol. P h y s . , 3, 361 (156,O). (15) K . 'I' Hobgood, G . S. R e d d y , and J. H . Goldstein, J . P h y s . Chem., 67, 110 ! 19liS) (10) S Gronowitz and R . A. H o f f m a n ,A r k i v K e m i , 16, 539 (1960).
USIVERSITYCHEMICAL LABORATORY E . A . V. EBSWORTH S . G. FRANXISS LENSFIELD ROAD CAMBRIDGE, ESGLAND RECEIVED SEPTEMBER 16, 1963
The Structure of Butyllithium in Ether. Dimer
A Solvated
Sir: Reported here are data which establish beyond reasonable doubt t h a t n-butyllithium (BuLi) exists in ether (Et20) solution as a solvated dimer, Et20:(BuLi)*. The conce-t, favored b y some,? of lithium alkyls as carbanions would seem to be incomplete a t best, b u t the concept, rejected by othersj3of three-center bonding would seem t o be applicable. Hence, we propose structure I for the complex of butyllithium with ether. U-hile excess ether may cluster about the complex in
s'
1 (1) Paper 111 in the series "Solvent Effects i n Organometallic Reactions." Paper I1 J . F Eastham and G U' Gibson, J A m . Che???.Sor , 81, 2 1 7 1 (1903) (2) G . Fraenkel, U . G . Adams, and J Williams, I'elrahedron L e l l e i s , 7117 (1963); R. E . Uessy and F. Paulik, J . Chem. Edfrc., 40, 185 (19C,3) ( 3 ) hl. Weiner and R. West, J . A m . Chem. Soc., 81, 18.5 (1903); T. 1, B r o w n , 11. W Ilickerhoff, and L) A. Bafus, i b i d , 84, 137 (1902); L). E . Appleouist and 0 . F. O'Brien, ibid., 81. 743 (15G3).
Fig. 1.-Curve A shows mole ratio of ether to hexane in vapoi over 50 ml. of hexane a t 25" to which increments of ether were added. Curve B shows same except 0.08 mole of BuLi was ii. the hexane. Curve C is the ratio of A to B.
Shown in Fig. 2 are representative traces of proton magnetic spectra (GO Mc.) from EtzO, BuLi, and their mixtures in hexane. The methylene signal from EtzO at any concentration in ordinary hexane is +a04 c.p.s. (downfield from TMS), but with BuLi present (cf. center curve) the Et20 methylene signal is +21h.5 c.P.s., so long as the Et20-BuLi mole ratio is 0.5 (ie., Et10 in excess of that needed to form the complex, cj' lower curve) is a single average value from complexed and uncomplexed ether, even a t low temperatures, presumably because of rapid equilibration between the two forms. The methylene signal from BuLi (protons CY to Li) a t any concentration in ordinary hexane is -50 C.P.S. (cf. upper curve), b u t with sufficient ether present t o form the 1 : 2 complex, the BuLi methylene signal is -59 C . P . S . (cf. lower curve). The 9-c.p.s. upfield shift of the BuLi methylene protons is a rational consequence of a decrease in the electronegativity of adjacent lithium when it is complexed with Et.0. When the E t 2 0 present is insufficient to convert all BuLi to the I : 2 complex ( c j . center curve), the positior:
