4494
COMMUNICATIONS TO THE EDITOR
VOl. 86
in eq. 3, 4. and 5 and the free-energy diagram which is a necessary concomitant of these facts is shown in Fig. 1.
+ BrCHzBr --+ I'yBr + .CH2Br
(3)
Py. f ' C H 2 B r --f I'yCH2Br
(4)
Py.
PyBr
Py+xREACTION COORDINATE
-
Fig. 1.-A free-energy us. reaction coordinate diagram for halogen abstraction by pyridinyl radical. Py. is pyridinyl free radical, P y x is the adduct of halide ion and pyridinium ion, Py+ X is pyridinium halide, P y C is the dihydropyridine formed from the radical, .C, left by halogen abstraction and the pyridinyl radical, and F, Fz *,and Fa represent the transition-state free energies for the transformations indicated in the diagram.
*,
*
acetonitrile (2 = 71.3) and isopropyl alcohol (2 = 76.3)5 to ethanol (2 = 79.6).5 The absence of a solvent effect upon the reaction rate demonstrates t h a t the transition state for the reaction has approximately the same degree of charge separation as the initial state. T h e net dipole moments of the reactants, COOCHB
+P y + B r -
(5)
The "carbonium ion stabilization" for transition states of reactions between radicals and halocarbons or hydrocarbons proposed by Huyser,'.* Walling,g and Bamford'Oa cannot be significant in the case of pyridinyl radicals in spite of the fact t h a t i t would be expected t h a t the pyridinyl radical would be unusually effective in accumulating positive charge in the transition state because of the stability of the pyridinium ion. N o spectroscopic change attributable to a chargetransfer bandlob was observed for the radical 1 in dichloromethane in the visible region, although such absorption may well occur a t shorter wave lengths. Visible light (irradiation with a tungsten lamp) or ultraviolet light (small amounts from the spectrophotometer) had no apparent effect on the course of the reaction. Strong ultraviolet irradiation led to a photochemical reaction with a halocarbon in one case." The radical 1 discriminates between different halogen-carbon bonds far more effectively than sodium12 or 1-phenylethyl r a d i ~ a 1 s . I ~ I t is also noteworthy t h a t there is so little difference between chlorine and bromine with regard to stabilization of the transition state for formation of CH2X. Disulfides react with 1. Less reactive pyridinyl radicals like methyl viologen cation radical (7) do not react appreciably with 2 a t room temperature but do react with more reactive halocarbons like tetrachloroand tetrabroinomethane. l 3
-
-
CH 3 N ( 3 - ( 3 &
H3
7
COOCHa
Further studies with pyridinyl radicals are continuing and promise much fundamental information on radical reactions.
1 and 2 , cannot be large since both are soluble in hexane, X solutions being possible for the radical 1. The transition state, 5 , cannot be highly charged and must decompose into the uncharged species, PyBr (6), the covalently bonded adduct of bromide ion, and the pyridinium ion6 and a bromomethyl radical (eq. 2 ) . cu. 0.02
[ P y . . . . Br . . . .CH*Br].-> 5
PyBr 6
+ .CH2Br
(2)
(7) E. S. Huyser, J . A m . Chem. Soc., 82, 394 (1960). (8) E . S . H u y s e r , H . Schimke, a n d R.L . B u r h a m , J . Org. C h e m . , 28, 2141 (1963). (9) C. Walling, "Free Radicals in Solution," John Wiley a n d Sons, I n c . , New Y o r k , N . Y . , 1957, p . 158. (10) ( a ) C. H. B a m f o r d , A . D. Jenkins, a n d R. J o h n s t o n , T r a n s . Faraday S o c . , 66, 418 (1959) (b) D . P. Stevenson a n d G . M . Coppinger, J . A m . Chem. Soc., 84, 149 (19621. (11) E J. Poziomek, unpublished results, A photochemical reaction between 7 a n d ethyl iodide has been observed. (12) J , N . Haresnape, J , M . Stevels, a n d E. W a r h u r s t , T r a n s . Faraday Soc., 3 6 , 465 (1940). (13) Cf.r e f . 9. p. 152. (14) Alfred P. Sloan Fellow, 1960-1964. ( 1 5 ) T h e a u t h o r s are grateful t o t h e National Science Foundation for Anancial s u p p o r t through G r a n t GP-251
EDWARD M. K O S O W E R ~ ~ DEPARTMENT O F CHEMISTRY STATE UNIVERSITY OF NEWYORKA T STOXY BROOK IRVING SCHWAGER" STONY BROOK,XE\V YORK RECEIVED MAY12, 1964
Both products disappear quickly, 6 by ionization to pyridinium bromide and . CHzBr by reaction with another molecule of 1. The mechanism is summarized
Metalation of Triphenylmethane by Organolithium Compounds
( 4 ) E . h l . Kosower. J . A n i . Chein. Soc. 8 0 , 3253 (19%). ( 5 ) T h e Z-values of t h e halocarbon-alcohol mixtures are actually somewhat lowei- t h a n those values cited for t h e put-e alcohols. a s noted in T a b l e I. ( 6 ) C,f K . Wallenfels a n d H Schtily. A n t i . , 621, 111 (1959)
Sir : The rate of metalation of triphenylmethane, i . e . , reaction 1, by different organolithium reagents varies
COMMUNICATIONS TO THE EDITOR
Oct. 20, 1964
markedly (250-fold) with the structure of RLi. k2 + RLi -+
(C6Hj)3CH
(CsH&CLi
+ RH
In
9
TABLE I RATESOF METALATION OF TRIPHESYLMETHASE BY ORGANOLITHICM REAGENTS
RLi Benzyllithium Allyllithium n-Butyllithium Phenyllithium Vinyllithium
7 6 7 6 6
1
1.7
0.10
8 9
1 5 1 9
0.020
Methyllithium
5 3
1 4
0 008
0
2 6
2 2 1 6
2 1
0.35 0.043
Relative Reactivities 250 51
13 5.5 2 5 1
Key to Fig. 1
0 0 a
0
-6 0
{ RLi ) i = the initial formal concentration of organolithiurn
compound as determined by titration. b For (CsHs)sCLiin T H F , log ej,,i, m p = 4.52. Allowance for clustering was made as follows K
(l,’n)(RLi)n
RLi and K = [RLi]/[(RLi),]””
Thus for reaction 1
d[(C6Hs)3CLi]dt
=
IC
(1)
tetrahydrofuran ( T H F ) solution, the following order of decreasing reactivity is obtained : benzyllithium > allyllithium > n-butyllithium > phenyllithium > vinyllithium > methyllithium. Kinetic data comparing metalating abilities of organolithium compounds have not been reported previously. In metalations with organolithium compounds carried out heretofore the emphasis has been on preparing particular compounds or on determining the position of metalation. The only prior information relating structure and metalating ability of organolithium reagents is based on product yields and results in a different reactivity order than t h a t communicated here. In this study the progress of reaction I was followed spectroscopically using an absorption cell described previously.* Triphenylmethane proved to be a suitable acid, because the electronic spectrum of its conjugate base triphenylmethyllithium is well-defined* and its acidity provides experimentally acceptable reaction rates. A 40-fold excess of organolithium reagent relative to triphenylmethane was used. Consequently, in the initial stage all the reactions were kinetically pseudo first order in triphenylmethane. The reaction temperature was 22’ and the solvent a t least 90y0 (vol.) T H F with diethyl ether or n-hexane making up the remainder Rate data for reaction 1 are given in Table I . Rates of formation of (C6H5)3CLiare illustrated in Fig. 1. The only reaction deviating strongly from first-order behavior prior to achieving 50% reaction is that with n-butyllithium, this being caused by the simultaneous consumption of n-butyllithium b y reaction with T H F . Rate constants determined from the initial slopes of optical density v s . time curves are identical, within
Initial concentration /RLi} [(CaHr)3X CHI, X 102,‘ Jf 108, AI ki. min.-15
4495
k2[RLil[ ( C G H ~ ) ~ C=H ]
kzK[(RLi).]””[(C~Hs)gCHI = kl [ ( C G H ~ S C H ] (1) H. Gilman and H . 4 . McNinch, J . Ovg. Chem., S I , 1889 (1962). Decreasing metalation activity based a n yields of acid obtained from t h e metalation of dibenzofuran and subsequent carhonation is reported t o b e ir-butyllithium > phenyllithium > methyllithium > benzyllithium. (2) K . Waack a n d M . A. D o r a n , J . A m . Chem. SOL.,85, 1651 (1963).
E 7
6
a s 0
8 $ 4
._ -1 0
s“ 2
3 I
8
0
\
8
0 v
2
-.
TIME. MINUTES
Fig. 1.-Rate of formation of ( CsHj)aCLiin T H F via ( C6Hj)3CH -+ (C6H5)d2Li,where RLi = C benzyllithium, Q allyllithium, 0 n-butyllithium, 0 phenyllithiurn, -0vinyllithiurn, 0 methyllithium. See Table I for details.
+ RLi
experimental error, with those derived from firstorder treatment. On the basis of structure similarities, the organolithium compounds in Table I may be divided into three classifications : (1) charge delocalized, e.g., benzyllithium and allyllithium : ( 2 ) alkyl, e.g., n-butyllithium and methyllithium ; and (3) sp?-hybridized, nondelocalized, e.g., phenyllithium and vinyllithiurn. Methyllithium, the first member of the alkyllithium series, is appreciably less reactive than n-butyllithium, which agrees with previous findings for this species. The most reactive metalating species are the resonancestabilized organolithium compounds which are also believed to be the most thermodynamically stable. Thus, no correlation between rate of metalation and expected thermodynamic equilibrium constants exists. The relative reactivities found for reaction 1 suggest that over-all nucleophilicity4 of the metalating species is rate controlling, supporting the presumption that the mechanism of metalation of compounds such as triphenylmethane is nucleophilic attack on hydrogen.j Since both basicity and polarizability contribute to over-all nucleophilicity. the high reactivity of resonance-stabilized organolithium compounds is attributed to high polarizability despite low inherent basicity. A similar high relative reactivity was observed for the resonance-stabilized organolithium com( 3 ) G . E. Coates, “Organometallic C o m p o u n d s , ” 2 n d E d . , John Wiley and S o n s . Inc., K e w Y o r k , N . I?.. 1960. p. 1 9 , H. G i l m a n and B. J G a j , J Org Ch~nz.2 , 2 , 116.5 (1937). ( 4 ) J . 0. Edwards and K.G. Pearson, J . A m Chenz. Soc.. 84, 16 (1‘3621. ( 5 ) H. Gilman, O I R Reactions, 8 , 262 ( I % i 4 J , J . 11 Roberts and D. Y . C u r t i n , J . A m . Chem. S n r . , 68, 1 6 5 8 (1946).
COMMUNICATIONS TO THE EDITOR
4496
pounds when used as initiators of vinyl polymerization.6 The high reactivity of benzyllithium in halogensubstitution reactions' is also in agreement with a high nucleophilicity for this species. T h e rates of reaction 1 do not follow the expected order of relative organolithium basicities,* although correlations between rate constants and equilibrium constants within a series of like reactions are known.g A well-studied example is t h a t rates of metalation by a given base are proportional to the acidity of the hydrogen which is replaced.I0 In contrast, rates of metalation of K H by different organolithium bases presuniably are not controlled solely by the base strength of the metalating reagent. Other nucleophilic displacement reactions on hydrogen also fail to obey the general rateequilibrium parallelism. l ' A factor also anticipated to be important in regulating the relative reactivity of organolithiurn compounds is their degree of clustering in solution.12 The existence of n-butyllithium as hexamers in benzene ~ o l u t i o n ' ~ results in a kinetic dependence for metalation'l or initiation of polymerizationlj of [n-b~tyllithiutn]'/~ in this solvent. Measurements of the order of reaction 1 with respect to organolithium reagent, within the concentration range 10V2to lo-' i U total RLi reagent, indicate that benzyllithiurn, allyllithium, and phenyllithium are reacting as monomeric species, whereas clustering appears to be dominant with n-butyllithium, Styryllithium, methyllithium, and vin ylli thium. l 6 which should behave similarly to benzyllithium, is reported to react as a monomeric species in the presence of T H F . " T h a t the reactivity of an organolithium compound may vary with concentration. which in turn may depend on its propensity to self-associate, is indicated by the observation that 0.1-0.5 Jf benzyllithium is stable in T H F solution, whereas a t 1 0 F -11 it decomposes, presumably by reaction with T H F . Since there are only small differences in organolithium concentrations used in the kinetic experiments reported here and those used by Gilman and McNinch,' i t is unlikely that varying degrees of aggregation are responsible for the differences in the orders of relative reactivities of the KLi reagents. X possible explanation for the differing reactivity scales is that the product yields of Gilman and McNinch' reflect the equilibrium situation and are not a measure of the rate of metalation. (6) I
