11392 J. Am. Chem. Soc., Vol. 121, No. 49, 1999
Bernasconi and Ali
PNS,42 lower log ko. The π-donor ability of oxygen being
stronger than that of sulfur, there will be a stronger reduction
in log ko for 1-M than for 5-M and hence a decrease in the
difference in log ko for the two systems. The other interaction
mechanism is the preorganization of the structure of the (CO)5M
moiety in 5-M or 1-M toward its electronic configuration in
the adducts that results from the π-donor effect. As a conse-
quence, the lag in the charge delocalization into the CO ligands
at the transition state is reduced, and the intrinsic rate constant
is not as strongly depressed by the PNS effect associated with
this lag. Since the reaction of 1-M will benefit more from this
effect, this will enhance the difference between log ko(5-M) and
log ko(1-M).
On the basis of our results it cannot be decided which of
these interaction mechanisms is the dominant one; there are
precedents for either situation in similar systems as discussed
elsewhere.36 The fact that log ko for 1-M exceeds log ko for
5-M by such a large amount (more than two log units) suggests
that the steric and inductive/field effects may not be sufficient
to account for the entire difference in the log ko values, i.e.,
there may have to be a contribution by the π-donor effect. This
would imply that the preorganization factor is dominant.44 Note,
though, that this is not a firm conclusion; in other reactions the
ko-reducing resonance loss has been found to be dominant.45
ko(5-M)
ko(1-M)
kR1 S(5-M)
kR1 S(1-M)
KR1 S(5-M)
KR1 S(1-M)
log
) log
- 0.5 log
(11)
It yields log{ko(5-Cr)/ko(1-Cr)} ) -2.41 and log{ko(5-W)/ko-
(1-W)} ) -2.60, respectively.
These differences in the intrinsic rate constants are indicative
of transition state imbalances31,41 where one or several of the
factorssinductive, steric, and π-donorseither lag behind or are
ahead of bond formation. As a result of the imbalances the
relative importance of these factors in how they affect the rate
constants is different from how they affect the equilibrium
constants, and this is the reason the intrinsic rate constants differ.
The principle of nonperfect synchronization (PNS)42 provides
guidance as to what factors play a central role.
1. Steric Effect. One factor that appears to contribute strongly
toward making log ko for 5-M lower than for 1-M is steric
hindrance. According to the PNS, if development of steric
hindrance at the transition state is ahead of bond formation,
log ko is lowered. Since steric hindrance is stronger for the
reaction with 5-M, log ko for 5-M will be reduced more than
log ko for 1-M. Recent evidence from SNV reactions suggests
that early development of steric hindrance indeed appears to
be the rule.43
C. kH-1. H+-catalyzed MeO- expulsion from 4-M- (R )
n-Pr), kH2 , is much faster than H+-catalyzed RS- expulsion,
kH-1,1 and hence no k-H1 value is available for this latter process
that could provide a comparison with kH-1 for acid-catalyzed
MeS- loss from the MeS- adducts, 6-M-. But even without a
precise value for kH-1 for 4-M-, it is clear that thiolate ion
departure from Fischer carbene adducts is much less sensitive
to acid catalysis than alkoxide departure. It is this low sensitivity
to acid catalysis which makes the kH-1 values relatively low and
has allowed the determination of pKMa H values based on eq 6.
Effect of Changing the Metal. As seen in Table 1, the rate
and equilibrium constants for thiolate addition to the tungsten
carbene complexes (5-W) are larger than for addition to the
chromium analogues (5-Cr). The kR1 S(5-W)/k1RS(5-Cr) ratios
vary between 3 and 7, with an average of about 4.5, while the
KR1 S(5-W)/KR1 S(5-Cr) ratios vary from about 4.3 to 12, with an
average of about 8, except for a low ratio of 2.55 for the reaction
with MeS-. The low ratio for this latter reaction may be an
artifact of the “abnormal” k-1(5-W)/ k-1(5-Cr) ratio (>1)46
which may just reflect an unusually large experimental error,
as mentioned at the beginning of the Discussion.
It is noteworthy that the effect of changing the metal on the
electrophilic reactivity of 5-W vs 5-Cr is much larger than for
the corresponding methoxy carbenes (1-M). For example, the
average kR1 S(1-W)/kR1 S(1-Cr) ratio for thiolate ion addition to
1-M is about 1.3, the average KR1 S(1-W)/K1RS(1-Cr) ratio is less
than 3.1 The larger influence of the metal on the electrophilic
reactivity of the MeS compared to the MeO carbene complexes
parallels the larger effect of the metal on the pKa of 13-M
compared to that of 14-M: pKa ) 8.3736 for 13-W vs pKa )
2. InductiVe Effect. Even though inductive effects are thought
to develop synchronously with bond formation and, therefore,
should not affect intrinsic rate constants per se, they can
contribute to differences in intrinsic rate constants because of
imbalances arising from other factors, especially resonance/
delocalization effects.31 There is strong evidence that in adducts
such as 2-M-, 3-M-, 4-M-, etc. the negative charge is dispersed
into the CO ligands10 which provides substantial resonance
stabilization to these adducts, a point highlighted by the bracket
symbol on the adduct structures. The same must be true for T-
in Scheme 1. There is also abundant evidence that in reactions
that lead to resonance delocalized products charge delocalization
lags behind transfer or bond formation.31 This is symbolized in
16 (and 9) by placing the partial negative charge in the transition
state on the metal rather than the CO ligands. Inasmuch as the
negative charge is closer to X at the transition state than in the
adduct, the transition state derives a disproportionately strong
stabilization (relative to the degree of bond formation) from
the inductive effect of X compared to the adduct. This enhances
ko and contributes to the difference in log ko between 5-M and
1-M because of the stronger inductive effect of oxygen.
3. π-Donor Effect. The influence of the π-donor effect of X
on the intrinsic rate constants is difficult to predict because of
two opposing interaction mechanisms. One is the loss of
resonance stabilization of the reactant (10 or 11) that that is
expected to run ahead of bond formation and, according to the
(44) For examples where the preorganization effect is dominant, see:
(a) Bernasconi, C. F.; Renfrow, R. A.; Tia, P. R. J. Am. Chem. Soc. 1986,
108, 4541. (b) Bernasconi, C. F.; Zitomer, J. L.; Schuck, D. F. J. Org.
Chem. 1992, 57, 1132.
(40) Marcus, R. A. J. Chem. Phys. 1965, 43, 679.
(41) (a) Jencks, D. A.; Jencks, W. P. J. Am. Chem. Soc. 1977, 99, 7948.
(b) Jencks, W. P. Chem. ReV. 1985, 85, 511.
(42) The PNS states that if the development of a product-stabilizing factor
lags behind bond changes or charge transfer at the transition state, ko is
reduced. The same is true if the loss of a reactant-stabilizing factor runs
ahead of bond changes or charge transfer. For product-stabilizing factors
that develop early or reactant-stabilizing factors that are lost late, ko is
enhanced.31
(45) For examples where the ko-reducing resonance effect is dominant,
see: (a) Bernasconi, C. F.; Panda, M. J. Org. Chem. 1987, 52, 3042. (b)
Bernasconi, C. F.; Killion, R. B., Jr. J. Org. Chem. 1989, 54, 2878. (c)
Bernasconi, C. F.; Flores, F. X.; Claus, J. J.; Dvorˇa´k, D. J. Org. Chem.
1994, 59, 4917.
(43) Bernasconi, C. F.; Ketner, R. J.; Chen, X.; Rappoport, Z. Can. J.
Chem. 1999, 77, 584.
(46) KR1 S was calculated as k1RS/k-1. Note that the kH (5-W)/k-1(5-Cr) )
-1
0.76 is “normal,” i.e., <1.