4042
COMMUNICATIONS TO THE EDITOR
VOl. 85
TABLE I Compoundo
JIaa(gem)’ (c P s
CHz=N-OH (or D ) CHz=N-OCHa CHz=N--N( CHa)z
7 63 t o 9 9.Y 6 96 to 9 22e 103102 117to12OiO2 116102
,
P
11.0t o 11.4O
i I
MezCO
12. i 0.2
J
I!
16.11 16.52 16.08 16.20 16.34 16,86 16.97
MeCR‘ (neat j Me2SO MeCS 16.5 ClCHzCHzCl CCla (neat) a All compounds were preDared in standard fashion by the reaction of formaldehyde with the appropriate Hth7-R compound. All .I values were measured on carefully calibrated Varian Associates Model A-60 spectrometers, radiofrequency 60 Mc. /see., sample temp. 36 1 2”. Unless otherwise specified, the P . E . of the values given is 0.05 C.P.S. or less. Unless otherwise noted, dilute (-5% or less) solutions were used. A rough “average” value, convenient for indicating the general sizes of the couplings. e Strongly solvent dependent. For a more detailed account, see B. L. Shapiro, S. J. Ebersole, and R . M. Kopchik, J . Mol. Spectry., 11, 326 (1963). Small concentration dependence. Small solvent dependence. See G. J. Karabatsos, B. L. Shapiro, F. M . \’am, J , S. Fleming, and J . S. Ratka, J . A m . Chem. Soc ,85,2784(1963). We thank Dr. P. L. de Benneville of the Rohm and Haas Co., Bristol, P a , for generous samples of this compound and for helpful discussions
side the range 0 f 3.5 C.P.S. only for the cases of substitution by elements of very low electronegativity, such as AI, Li, and 31g,3where values in the range (+) 6 to 8 C.P.S. are observed. In Table I, we report our observations of ~JHH,(gem) in a number of CH2=Ncompounds, in which very different values are obtained. Among other interesting features4 of the spectra, the magnitudes and (in some cases) solvent dependence~ of ~the J values are noteworthy. In these CH2= N--Y compounds, lJ1 increases markedly as t h e electronegativity of Y decreases. The trend observed here, taken toget.her with that reported for the olefins,2h suggests that the sign of these CH2=N-Y J values is negative, and various sign determination experiments are in progress. I t may be noted at this stage, however, t h a t the range of magnitudes of these sp2-typeJ(gem) values overlaps extensively with the range of values hitherto associated with sp3-type J(gem) values.5 Thus, regardless of the signs of the J values reported here, a theoretical picture which gives major importance to the H-C--H bond angles6 will clearly be inapplicable, since it is evident that in our sp2-typecases, substituert efects (to put the matter in the most general terms) are dominant, as has already been pointed out for sp3-type CH2 systems and suggested strongly for vinylic sp2cases as well.’ Regardless of the detailed nature of the dominant factors controlling the spin-spin coupling constants, the observed monotonic trend in iJl(gem) with p( 3 ) (a) 1) W hlotire and J A . Happe, J . Phya Chem , 66, 224 (1961); ( b ) C . S Johnson, Jr , M A Weiner, J S. Waugh, and 17 Seyferth, J . ,4m. Chem. Sor , 8 3 , 1306 (1961); ( c ) K . T. Hobgood, J r , and J H . Goldstein, Ypertvorhim A r l o , 18, 1280 (1962); (d) G Fraenkel. 1) C;. Adanis, and J Williams, T?lvahedvo?t L e t t e r s . No. 1 2 , 767 (1063) (4) To he discussed a t length elsewhere (a) See for example, b! Barfield and I ) . M . G r a n t , J A m . Chem. Soc., 86, 1x09 (1963), and reference.. cited therein (6) Cf M Karplus and I). H. Anderson, J . C h e m . P h y s . , 30, 6 (1959); H S. Gutowsky, M. Karplus, and I) M . Grant. i b i d . , 31, 1278 (1959); H. S . Gutowsky, V. I). Mochel, and B G. Somers, i b r d . , 36, 1153 (1962); M . Karplus. J 4 1 n (’hem S o c - . , 84, 2458 (1962). ( 7 ) H J Rernstein and N Sheppard. J . Chem. P h y s . , 37, 3012 (1962).
atom electronegativity (cf. vinyl-XZh) suggests that resonance form B, which is certainly unimportant in the azomethines, is also probably not very significant in the oxime and hydrazone derivatives. Resonance contributors of type A, however, could well be involved to some significant extent in all three types of compound. One possible and plausible implication of
+
-
CHz-X-Y A
CH 2-
+
N=Y B
this is that the more ‘(positive”the carbon atom of the CH2, the larger the ~ J l . However, other factors (such as the involvement of the n-system, and especially the availability of electron pairs on X of the CH2=X system) may well be important. These and other matters are currently being investigated. Acknowledgment.-It is a pleasure to acknowledge stimulating and helpful discussions with Drs. A . A . Bothner-By and P. C. Lauterbur. MELLON INSTITCTE B L SHAPIRO PITTSBURGH, PENNSYLVANIA S J . EBERSOLE KEDZIECHEMICAL LABORATORY G J KARABATSOS MICHIGAN STATECNIVERSITY F M VANE EASTLANSING, MICHIGAN SPACESCIENCE DIVISIOX S L MANATT JET PI?OPULSIO~X LABORATORY OF TECHNOLOGY CALIFORNIA INSTITUTE PASADENA, CALIFORNIA RECEIVED OCTOBER 18, 1963
Complexes of Organolithium Compounds with Vacant Orbital Acceptors. TI. Determination of ElectronDensity Changes by Proton Magnetic Resonance
Sir: We wish to report that changes in electron density of an aromatic organolithium compound, brought about by dative-bond formation with a Lewis acid, can be determined from changes occurring in its proton magnetic resonance ( p , m , r , ) spectra on complexing. Organolithium compounds form reversible dative con~plexes
COMMUNICATIONS TO THE EDITOR
Dec. 20, 1963
I
I
I
8.0
7.0
6.0
Fig. 1.-Proton
I 5.0
0
PPM
magnetic resonance spectra of 1.1-diphenyl-n-hexyllithium and its diethylzinc complex.
+
with weak Lewis acids, e.g., R-Li+ EtlZn e [R + ZnEt2-]Lif. The chemistry of complexes of this type has been investigated principally b y Wittig and coworkers,' but there is little information on their physical properties. Recently, we have shown t h a t such properties as association constant and bond energy can be obtained from the ultraviolet and visible spectrum of the complex.2 Since nuclear screening in molecules is determined by electron distribution in the vicinity of the nucleus, the protons of the organolithium compound move to lower field on complexing and those of the Lewis acid to higher field. We believe this to be the first quantitative observation of electron-density changes occurring in dative-bond formation. We report data for the p.m.r. of the complex formed between 1,l-diphenyl-n-hexyllithum(I) and diethylzinc in tetrahydrofuran solution. In this odd alternant organolithium compound the negative charge is extensively delocalized onto the phenyl ringsa Because the electron densities at the ortho, meta, and para positions are different,4 the aromatic protons at each position experience a different upfield chemical shift. The spectrum of I is shown in Fig. l a with internal T M S zero. The downfield shift of the aromatic protons owing to the electron density lost to diethylzinc on complex formation is shown in Fig. l b for the ratio of ZnEt2/RLi = 0.82. The spectral pattern of I is identical with t h a t of triphenylmethyllithium.6 Several worker^',^ have established t h a t there is a (1) G. Wittig, Anpew. Chrm., 10, 65 (18.58). and references therein. See ref. 2 for a more complete summary of prevlous preparative laark. (2) R . Waaek and M . A. Dorao, J. A m , Cham. Sor., 86, 2861 (1963). (3) Evidence indicates I exists in T H F rolution BI free io00 Or as very looorion~pairrhavingooly weak intersdion. (4) Huckel LCAO calculntions predict cqval charge densities at the 0 ~ o-podtionr and zero at the m-position. but cs1culation.i including electron repulsion predict different chnrye densities at each pusitloo. which inrrearr io the order 0.. m-. o..* ( 5 ) A . Brickstock and 1. A. Pople, Trans. F o r a d w Soc.. 60.801 (1854). (6) V. R . Sandel and H . H . Frccdmeo, J. Am. Chcm. Sor.. 86, 2328
(1963). 17) G.Fraenkel, R. E. Carter. A. McLachlin. end 1. H. Richards, ikid.. SI. 5846 (1960). ( 8 ) T. Sehaefer and W . G.Sehneidcr. Con. J. Ckcm., 11, 96 (1863). end rcf~reocestherein.
4043
direct relation between proton chemical shift ( 6 ) and electron density at the carbon atom to which the proton is attached. From the differevce between 6 of free RLi and completely complexed, RLi, the actual amount of electron density change (Ap) occurring a t each of the aromatic protons on dative-bond formation can be determined. Or similarly, absolute charge densities ( p ) may be determined from A6 benzene6-8 (benzene 6 = 7.25 p.p.m. in THF). These values are summarized in Table I. Because the central carbon atom in I has no attached proton, p at this position cannot be experimentally obtained. Complexes of other organolithinm compounds having a-protons are under study TABLE 1 NEGATIVE CHAKGE. DENS~TIES DETERMINED FROM CHEMICAL
SHIFTS" 1.1-
Position
orlko
Diphenyl- I d i e f h y l - Adproton n~herylzinc on conlithium complex pledng
0.040
me&
,077
para
,158
0,019 ,043 ,089
0.021
Total A 0 on =om~lerin~
0.084
,034
,136
,069
,138
Aobirar~ timd
0.53 .44 .44
Total charge ,358 on rines ,784 ,426 Using the relationship. A& = 1 0 . 7 . A ~from reference 8. Ap (on complexing) Aprrutionai = p(freeRLi) '
.
The total observable electron-density change caused by complexing with diethylzinc is 0.358 electron. Since i t seems" unlikely t h a t the central C-atom will have a higher incipient charge in the dative complex than in the free ion, a t least this amount of electron density is acquired by the diethylzinc during dative bonding. The largest change in electron density occurs at the para position, which in the free organolithium is the position having the largest fraction of negative charge The largest fractional change in charge density on complexing occurs a t the ortho position. The charge densities are in reasonable agreement with SCF calculated values,s although no corrections were
COMMUNICATIONS TO THE EDITOR
4044
made for contributions other than p to 6.8 Presumably the protons of I and its complex are subject to the same effects so their difference is independent of correction factors. The order of chemical shift with increasing field strength, i . e . , increasing charge density, is 0-,m - , 0 - , and as predicted by MO theory is a mirror image of triphenylmethyl carbonium ion.B By difference, the charge on the central carbon in I is 0.216 electron. When there is insufficient Et2Zn to complex all the RLi, a discrete spectrum is not observed for the complexed and uncomplexed RLi. There must, therefore, be rapid exchange between complexed and uncomplexed species. Because the average spectrum is sharp, the lifetimelo in each state is