Physical organic chemistry of transition metal carbene complexes. 16.1 Reactions of (CO)5M=C(OR)Ph (M = Cr or W; R = Me or Et) with thiolate ions in aqueous acetonitrile. Complete kinetic dissection of the two-step mechanism

Article abstract of DOI:10.1021/ja9908265

A kinetic study of the reactions of (CO)₅M=C(OR′)Ph (1a, M = Cr, R′ = Me; 1b, M = W, R′ = Me; 1c, M = Cr, R′ = Et; 1d, M = W, R′ = Et) with n-PrS^-, HOCH₂CH₂S^-, MeO₂CCH₂CH₂S^-, and MeO₂CCH₂S^- in 50% MeCN-50% water (v/v) at 25°C is reported. At low [RS^-] and relatively low pH the reaction leads to the nucleophilic substitution products (CO)₅M=C(SR)Ph without accumulation of any intermediate. At high [RS^-] and high pH formation of a tetrahedral intermediate, (CO)₅MC(OR′)(SR)Ph, is observed. Upon addition of acid the intermediate is converted into the substitution product. For the reactions of most thiolate ions a detailed kinetic analysis allowed the determination of the rate constants for nucleophilic attack on the carbene complexes (k₁) and its reverse (k_-1), the equilibrium constant for nucleophilic addition (K₁ = k₁/k_-1), and the rate constants for alkoxide ion departure from the respective intermediates catalyzed by H⁺ (k^h₂) and N-methylmorpholinium ion (k^BH₂). The dependence of these rate and equilibrium constants on the metal, the leaving group, and the thiolate ion and comparisons with reactions of thiolate ions with other electrophiles provide insights into the transition state structure, the resonance effect of the (CO)₅M moieties, the requirements for partial desolvation of the nucleophile prior to entering the transition state, and transition state imbalances.

Full text of DOI:10.1021/ja9908265

6630
J. Am. Chem. Soc. 1999, 121, 6630-6639
Physical Organic Chemistry of Transition Metal Carbene Complexes.
16.1 Reactions of (CO)5MdC(OR)Ph (M ) Cr or W; R ) Me or Et)
with Thiolate Ions in Aqueous Acetonitrile. Complete Kinetic
Dissection of the Two-Step Mechanism
Claude F. Bernasconi,* Kevin W. Kittredge, and Francis X. Flores
Contribution from the Department of Chemistry and Biochemistry, UniVersity of California,
Santa Cruz, California 95064
ReceiVed March 15, 1999
Abstract: A kinetic study of the reactions of (CO)5MdC(OR′)Ph (1a, M ) Cr, R′ ) Me; 1b, M ) W, R′ )
Me; 1c, M ) Cr, R′ ) Et; 1d, M ) W, R′ ) Et) with n-PrS-, HOCH2CH2S-, MeO2CCH2CH2S-, and
MeO2CCH2S- in 50% MeCN-50% water (v/v) at 25 °C is reported. At low [RS-] and relatively low pH the
reaction leads to the nucleophilic substitution products (CO)5MdC(SR)Ph without accumulation of any
intermediate. At high [RS-] and high pH formation of a tetrahedral intermediate, (CO)5Mh C(OR′)(SR)Ph, is
observed. Upon addition of acid the intermediate is converted into the substitution product. For the reactions
of most thiolate ions a detailed kinetic analysis allowed the determination of the rate constants for nucleophilic
attack on the carbene complexes (k1) and its reverse (k-1), the equilibrium constant for nucleophilic addition
(K1 ) k1/k-1), and the rate constants for alkoxide ion departure from the respective intermediates catalyzed by
H+ (kH2 ) and N-methylmorpholinium ion (kB2 H). The dependence of these rate and equilibrium constants on the
metal, the leaving group, and the thiolate ion and comparisons with reactions of thiolate ions with other
electrophiles provide insights into the transition state structure, the resonance effect of the (CO)5M moieties,
the requirements for partial desolvation of the nucleophile prior to entering the transition state, and transition
state imbalances.
Introduction
A major difference between the Fischer carbene complexes
and the carboxylic esters is the much higher reactivity of the
former. The enhanced reactivity can be mainly attributed to
much stronger stabilization of the negative charge in 2 by the
(CO)5M moiety compared to the stabilization of the charge by
oxygen in the corresponding ester tetrahedral intermediate. An
approximate measure of this increased stabilization can be
derived from a comparison of the pKa value of 4a (12.5)9 with
In the presence of bases or nucleophiles, transition metal
carbene complexes of the Fischer type display reactivity patterns
that are similar to those of the isolobal carboxylic esters.2,3
A
prominent example is the substitution of an alkoxy group by
nucleophiles. This process is commonly assumed to involve the
two-step mechanism shown in eq 1. Such reactions have been
that of ethyl acetate (26.5).10 Recently it was shown that in the
reaction of 1a and 1b with MeO- in methanol this enhanced
documented with OH-,4 water,4b amines,2,5 thiolate ions,2,6
carbanions,2,7 and other nucleophiles.2,8
(6) (a) Fischer, E. O.; Leupold, M.; Kreiter, C. G.; Mu¨ller, J. Chem.
Ber. 1972, 105, 150. (b) Lam, C. T.; Senoff, C. V.; Ward, J. E. H. J.
Organomet. Chem. 1974, 70, 273. (c) Aumann, R.; Schro¨der, J. Chem. Ber.
1990, 123, 2053.
(7) (a) Fischer, E. O.; Riedmu¨ller, S. Chem. Ber. 1976, 109, 3358. (b)
Fischer, E. O.; Held, W.; Kreissl, F. R. Chem. Ber. 1977, 110, 3842. (c)
Burkhardt, T. J.; Casey, C. P. J. Am. Chem. Soc. 1973, 95, 5833. (d) Fischer,
E. O.; Held, F. R.; Kreissl, F. R.; Frank, A.; Hattner, G. Chem. Ber. 1977,
110, 656. (e) Casey, C. P.; Burkhardt, T. J.; Bunnell, C. A.; Calabrese, J.
C. J. Am. Chem. Soc. 1977, 99, 2127.
(1) Part 15: Bernasconi, C. F.; Ali, M. J. Am. Chem. Soc. 1999, 121,
3039.
(2) Do¨tz, K. H.; Fischer, H.; Hofmann, P.; Kreissl, F. R.; Schubert, U.;
Weiss, K. Transition Metal Complexes; Verlag Chemie: Deerfield Beach,
FL, 1983.
(3) Bernasconi, C. F. Chem. Soc. ReV. 1997, 26, 299.
(4) (a) Aumann, R.; Hinterding, P.; Kru¨ger, C.; Goddard, R. J.
Organomet. Chem. 1993, 459, 145. (b) Bernasconi, C. F.; Flores, F. X.;
Kittredge, K. W. J. Am. Chem. Soc. 1997, 119, 2103.
(5) (a) Klabunde, U.; Fischer, E. O. J. Am. Chem. Soc. 1967, 89, 7141.
(b) Connor, J. A.; Fischer, E. O. J. Chem. Soc. A 1969, 578. (c) Fischer, E.
O.; Kollmeier, H.-J. Chem. Ber. 1971, 104, 1339. (d) Fischer, E. O.;
Leupold, M. Chem. Ber. 1972, 102, 599. (e) Fischer, E. O.; Heckl, B.;
Werner, H. J. Organomet. Chem. 1971, 28, 359. (f) Werner, H.; Fischer,
E. O.; Heckl, B.; Kreiter, C. G. J. Organomet. Chem. 1971, 28, 367. (g)
Bernasconi, C. F.; Stronach, M. W. J. Am. Chem. Soc. 1993, 115, 1341.
(8) (a) Fischer, E. O.; Kreis, G.; Kreissl, F. R.; Kreiter, C. G.; Mu¨ller, J.
Chem. Ber. 1973, 106, 3910. (b) Casey, C. P.; Brunsvold, W. P. Inorg.
Chem. 1977, 16, 391. (c) Bell, R. A.; Chisholm, M. H.; Couch, D. A.;
Rankel, L. A. Inorg. Chem. 1977, 16, 677.
(9) (a) Gandler, J. R.; Bernasconi, C. F. Organometallics 1989, 8, 2282.
(b) Bernasconi, C. F.; Sun, W. J. Am. Chem. Soc. 1993, 115, 12526.
(10) Amyes, T. L.; Richard, J. P. J. Am. Chem. Soc. 1996, 118, 3129.
10.1021/ja9908265 CCC: $18.00 © 1999 American Chemical Society
Published on Web 07/03/1999

Physical Organic Chemistry of TM Carbene Complexes. 16
J. Am. Chem. Soc., Vol. 121, No. 28, 1999 6631
stabilization is sufficient to render the equilibrium for intermedi-
ate formation (eq 3) favorable at high methoxide ion concentra-
tions.11 This allowed a spectroscopic detection of 5a and 5b
and a determination of the rate constants (k1 and k-1) for eq 3
as well as the equilibrium constants K1 ) k1/k-1
.
Based on these results it seemed likely that in a “real”
substitution reaction, i.e., one where the products are different
from the reactants, the corresponding intermediate may also be
directly observable, provided sufficiently strong nucleophiles
are used. Thiolate ions appeared to be among the most promising
nucleophiles for two reasons. (1) They are very reactive and
have recently led to the successful observation of intermediates
in nucleophilic vinylic substitution reactions.12 (2) Based on a
study of the reaction of 4a with NaSPh in methanol/benzene
by Lam et al.,6b there is a strong indication that the correspond-
ing intermediate indeed accumulates to detectable levels.
Specifically, these authors observed that the reaction leads to
an unidentified species that is quite stable but, upon treatment
with HCl, yields the substitution product, 6.
Figure 1. Conversion of 5 × 10-5 M (CO)5WdC(OMe)Ph (spectrum
A, λmax 396 nm) to (CO)5WdC(SCH2CH2CO2Me)Ph (spectrum B, λmax
452 nm) in the presence of 7.94 × 10-7 M MeO2CCH2CH2S- and
10-3 M MeO2CCH2CH2SH in an N-methylmorpholine buffer at pH
7.56. Spectra taken every 10 s for 150 s. Spectrum C corresponds to
the intermediate (2b with R ) MeO2CCH2CH2) generated in the
presence of 10-3 M MeO2CCH2CH2S- at pH 10.69.
10-3 M, [RS-] ) 7.94 × 10-7 M). The figure shows a sharp
isosbestic point at 422 nm, indicating that there is no accumula-
tion of an intermediate to detectable levels. The behavior seen
in Figure 1 is representative for the reactions of all four Fischer
carbene complexes 1a-d with n-PrS-, HOCH2CH2S-, MeO2-
CCH2CH2S-, and MeO2CCH2S-.14 For the reactions of 1b and
1c with n-PrS- the identity of the products was confirmed by
comparison with independently synthesized samples.
A different situation arises in the presence of higher thiolate
ion concentrations at a pH close to the pKRa SH of the thiol. Here
the carbene complex is rapidly converted to a new species whose
spectrum (dashed line in Figure 1) resembles neither that of
the starting material nor that of 3a or 3b, respectively. This
species is stable for a few minutes but, upon acidification yields
the substitution product. We conclude the new species is the
intermediate 2, specifically 2b (with R ) MeO2CCH2CH2) in
Figure 1.14 Its UV/vis spectrum resembles that of 5b11 and of
7.11 IR spectra of 2a and 2b (R ) CH3CH3CH2) show the
characteristic shifts of the CO stretching frequencies compared
to those in the starting materials: for 1b f 2b the bands at
2061 (s), 1996 (w), and 1947 cm-1 (s) are replaced by bands at
2035 (s), 1911 (w), and 1845 cm-1 (m), respectively, while for
1c f 2c the bands at 2069 (s), 1993 (m), and 1940 cm-1 (w)
are replaced by bands at 2045 (s), 1905 (s), and 1848 cm-1
(m), respectively.
In this paper we report that the reactions of 1a, 1b, as well
as 1c and 1d with a series of alkanethiolate ions (n-PrS-,
HOCH2CH2S-, MeO2CCH2CH2S-, MeO2CCH2S-) in 50%
MeCN-50% water (v/v) lead to substitution with spectroscopi-
cally observable intermediates, 2a-d. A detailed kinetic analysis
is presented that allows the determination of the rate constants
of the various steps in eq 1. A preliminary report of some of
our results has appeared earlier.13
Results
General Features. In the presence of low thiolate ion
concentrations at a pH below the pKRa SH of the thiol a clean
conversion of 1a, 1b, 1c, and 1d to the corresponding substitu-
tion products 3a and 3b is observed. This is illustrated in Figure
1 for the reaction of 1b with MeO2CCH2CH2S- in an N-
methylmorpholine buffer at pH 7.56 (pKRa SH ) 10.69, [RSH] )
NMR spectra of the intermediates were more difficult to
obtain, due to decomposition. The most stable intermediate was
1
2b with R ) n-Pr, which yielded relatively clean H and 13C
NMR spectra. The most characteristic changes for the conver-
sion of 1b to 2b are in the chemical shifts of the MeO groups
(1H NMR: 4.47 ppm in 1b, 4.04 ppm in 2b; 13C NMR: 71.4
ppm in 1b, 49.0 ppm in 2b) and the carbene carbon (13C
NMR: 322.1 ppm in 1b, 52.4 ppm in 2b). For more details see
the Experimental Section.
(11) Bernasconi, C. F.; Flores, F. X.; Gandler, J. R.; Leyes, A. E.
Organometallics 1994, 13, 2186.
(12) (a) Bernasconi, C. F.; Fassberg, J.; Killion, R. B., Jr.; Rappoport,
Z. J. Am. Chem. Soc. 1989, 111, 6862. (b) Bernasconi, C. F.; Fassberg, J.;
Killion, R. B., Jr.; Rappoport, Z. J. Am. Chem. Soc. 1990, 112, 3169. (c)
Bernasconi, C. F.; Ketner, R. J.; Chen, X.; Rappoport, Z. J. Am. Chem.
Soc. 1998, 120, 7461. (d) Bernasconi, C. F.; Ketner, R. J.; Chen, X.;
Rappoport, Z. Can. J. Chem. 1999, 77, 584.
Judging from the UV/vis spectra, the reactions of all four
carbene complexes with n-PrS-, HOCH2CH2S-, and MeO2-
(13) Bernasconi, C. F.; Flores, F. X.; Kittredge, K. W. J. Am. Chem.
Soc. 1998, 120, 7983.
(14) For spectra of the reaction of 1a with MeO2CCH2CH2S- see
ref 13.

6632 J. Am. Chem. Soc., Vol. 121, No. 28, 1999
Bernasconi et al.
CCH2CH2S- show similar behavior. However, in the reaction
of 1b with MeO2CCH2S- the absorption spectra of intermediate
and product suggest contamination by unidentified byproducts
while for the reactions of 1a, 1c, and 1d with MeO2CCH2S-
acidification of the “intermediate” yielded very little if any
substitution products, as judged from the absorption spectra of
the product solution.
The observations presented above indicate that the reactions
of thiolate ions with 1a-d can be described by eq 4. At high
pH and high [RS-] the first step is fast and the equilibrium
favors the intermediate, while product formation is slow. In fact,
Figure 2. Reactions of 1a-d with MeO2CCH2CH2S- at pH 10.69
leading to the corresponding intermediates. Plots according to eq 5:
(0) 1c; (9) 1d; (O) 1a; and (b) 1b.
spontaneous conversion of the intermediate to 3a or 3b (note
that 3a, the Cr derivative, is the product of both 1a and 1c,
respectively, while 3b, the W derivative, is the product of both
1b and 1d, respectively) by the k2 step is so slow that product
formation only occurs after addition of acid, indicating hydro-
nium ion and/or buffer acid catalyzed (kHaH and k2BH[BH])
+
2
alkoxide ion departure. In the absence of acid the intermediate
does not lead to 3a or 3b but slowly decomposes into
unidentified byproducts. For the reactions with MeO2CCH2S-
decomposition of the intermediate, if it is formed at all (more
on this below), occurs quite rapidly and precludes product
formation.
At low [RS-] and low pH values, 2a-d turn into steady-
state intermediates because the equilibrium of the first step is
+
no longer favorable and aH and/or [BH] are high enough to
ensure rapid product formation via the kHaH and/or k2BH[BH]
+
2
pathways. From the fact that product formation is observed even
with MeO2CCH2S- as nucleophile we conclude that the
intermediate is generated under these conditions and that at low
pH its acid-catalyzed conversion to products is faster than its
decomposition.
Figure 3. Reaction of 1c with MeO2CCH2CH2S- in various N-
methylmorpholine buffers leading to the substitution product 3c
under steady state conditions: (O) pH 6.86; (b) pH 7.26; (0) pH 7.56;
(9) pH 7.85; (4) pH 8.04; and (2) pH 8.26.
Kinetics. All kinetic experiments were run under pseudo-
first-order conditions with [RS-] held constant either by being
present in large excess over the substrate or, at pH , pKaRSH
,
suggests that nucleophilic addition does occur but that the
intermediate decomposes rapidly.
by having a large excess of RSH over the substrate. The
temperature was 25 °C throughout and the ionic strength was
maintained at 0.1 M with KCl. Three types of kinetic experi-
ments were performed.
A. Formation of Intermediate. Rates of formation of the
intermediate were measured by monitoring the loss of carbene
complex in a stopped-flow spectrophotometer. Pseudo-first-order
B. Steady-State Kinetics. Rates were determined by moni-
toring product formation by conventional spectrophotometry.
For a reaction with a given thiolate ion kobsd was determined in
buffered solutions at 7-9 different pH values. In the reactions
with n-PrS- the pH range was from 8.56 to 10.56 maintained
by triethylamine buffers, in the reactions with HOCH2CH2S-
and MeO2CCH2CH2S- the pH range was from 6.76 to 8.26
maintained by N-methylmorpholine buffers, and in the reactions
with MeO2CCH2S- the pH range was from 4.93 to 6.93
maintained by acetate buffers. Figure 3 which shows plots of
rate constants determined between 5 × 10-4 and 5 × 10-3
M
RS- at pH ) pKaRSH showed a linear dependence on [RS-]
according to eq 5 (Figure 2) from which k1 (Table 1) could be
obtained. The intercepts of the plots of kobsd vs [RS-] were too
small to yield accurate k-1 values.
k
obsd vs [RS-] for the reaction of 1c with MeO2CCH2CH2S- is
representative. All plots are linear with intercepts that are either
indistinguishable from zero or quite small; small intercepts on
the order of 5 × 10-4 s-1 for 1c and 1d and 3 × 10-3 s-1 for
1a and 1b are expected due to hydrolysis.15 The slopes of the
plots increase with decreasing pH, demonstrating the effect of
kobsd ) k1[RS-] + k-1
(5)
Interestingly, the reactions of 1a-d with MeO2CCH2S-
showed the same kinetic behavior as those with the other thiolate
ions, despite the fact that no clean spectra of the intermediates
derived from MeO2CCH2S- could be obtained (Vide supra). This
(15) Bernasconi, C. F.; Flores, F. X.; Kittredge, K. W. J. Am. Chem.
Soc. 1997, 119, 2103.

Physical Organic Chemistry of TM Carbene Complexes. 16
J. Am. Chem. Soc., Vol. 121, No. 28, 1999 6633
Table 1. Summary of Rate and Equilibrium Constants
CH3CH2CH2S-
HOCH2CH2S-
MeO2CCH2CH2S-
MeO2CCH2S-
pKRa SH ) 11.94 ( 0.04
pKRa SH ) 10.79 ( 0.03
pKRa SH ) 10.69 ( 0.02
pKRa SH ) 9.45 ( 0.03
(CO)5CrdC(OMe)Ph (1a)
(2.25 ( 0.10) × 104
(2.23 ( 0.29) × 104
(6.98 ( 0.87) × 108
k1, M-1 s-1 (eq 5)
k1, M-1 s-1 (eq 8)
(1.34 ( 0.07) × 104
a
(2.30 ( 0.04) × 104
(1.72 ( 0.68) × 104
(3.04 ( 0.24) × 107
(6.76 ( 0.20) × 104
(7.35 ( 0.67) × 104
(2.51 ( 0.32) × 106
(4.29 ( 0.62) × 108
kH/k-1, M-1
2
(5.39 ( 0.11) × 108
(3.50 ( 0.92) × 108
(1.99 ( 0.02) × 108
kH2 , M-1 s-1
k-1, s-1
1.26 ( 0.21
5.01 ( 1.93
6.55 ( 0.58
K1, M-1
(1.06 ( 0.23) × 104
(4.49 ( 1.93) × 103
(3.51 ( 0.35) × 103
(1.39 ( 0.61) × 102
(6.88 ( 0.80) × 101
kB2 H, M-1 s-1 b
(CO)5WdC(OMe)Ph (1b)
(6.52 ( 0.21) × 104
(6.29( 0.59) × 104
k1, M-1 s-1 (eq 5)
k1, M-1 s-1 (eq 8)
(4.21 ( 0.25) × 104
a
(7.26 ( 0.05) × 104
(6.06 ( 0.59) × 104
(1.26 ( 0.38) × 107
(1.73 ( 0.14) × 105
(2.39 ( 0.14) × 105
(1.38 ( 0.12) × 106
(1.83 ( 0.28) × 108
(2.69 ( 0.45) × 107
kH/k-1, M-1
2
(5.69 ( 0.25) × 108
(3.76 ( 0.98) × 108
(2.05 ( 0.03) × 108
kH2 , M-1 s-1
k-1, s-1
3.11 ( 0.61
14.0 ( 5.9
16.3 ( 5.8
K1, M-1
(1.35 ( 0.35) × 104
(4.66 ( 2.10) × 103
(4.45 ( 1.44) × 103
(2.17 ( 1.02) × 102
(1.46 ( 0.48) × 102
kB2 H, M-1 s-1 b
(CO)5CrdC(OEt)Ph (1c)
(1.20 ( 0.06) × 104
(1.30 ( 0.08) × 104
(2.94 ( 0.42) × 107
k1, M-1 s-1 (eq 5)
k1, M-1 s-1 (eq 8)
(7.05 ( 0.37) × 103
a
(1.20 ( 0.02) × 104
(1.38 ( 0.04) × 104
(1.70 ( 0.35) × 107
(3.43 ( 0.18) × 104
(3.74 ( 0.32) × 104
(1.38 ( 0.12) × 106
(3.06 ( 0.29) × 108
kH/k-1, M-1
2
(4.23 ( 0.09) × 108
(2.67 ( 0.12) × 108
(1.68 ( 0.02) × 108
kH2 , M-1 s-1
k-1, s-1
1.38 ( 0.16
9.08 ( 1.72
9.88 ( 2.15
K1, M-1
(5.11 ( 0.86) × 103
(1.32 ( 0.32) × 103
(1.21 ( 0.29) × 103
(8.40 ( 2.18) × 101
(4.33 ( 1.10) × 101
kB2 H, M-1 s-1 b
(CO)5WdC(OEt)Ph (1d)
(4.04 ( 0.15) × 104
(3.62 ( 0.45) × 104
(1.28 ( 0.25) × 107
k1, M-1 s-1 (eq 5)
k1, M-1 s-1 (eq 8)
(2.33 ( 0.11) × 104
a
(4.25 ( 0.10) × 104
(2.61 ( 0.53) × 104
(6.96 ( 0.27) × 106
(1.05 ( 0.07) × 105
(0.87 ( 0.01) × 105
(6.71 ( 0.25) × 105
(1.72 ( 0.15) × 108
kH/k-1, M-1
2
(5.12 ( 0.13) × 108
(2.65 ( 0.09) × 108
(1.84 ( 0.09) × 108
kH2 , M-1 s-1
k-1, s-1
2.98 ( 0.33
20.7 ( 4.8
26.4 ( 2.3
K1, M-1
(7.82 ( 1.25) × 103
(1.95 ( 0.52) × 103
(1.61 ( 0.18) × 103
(9.23 ( 2.74) × 101
(6.47 ( 0.91) × 101
kB2 H, M-1 s-1 b
a Intercept in plot according to eq 8 is too small for determination of k1. b BH ) N-methylmorpholinium ion; kB2 H obtained from the kB2 H/k-1 ratios
(Table 2) and k-1
.
acid catalysis on the conversion of the intermediate to products.
Applying the steady-state approximation to eq 4 yields eq 6
k1(k2 + kH2 aH+ + kBH[BH])
2
slope )
(6)
k-1 + k2 + kH2 aH+ + kBH[BH]
2
for the slopes. Equation 6 may be simplified to eq 7
k1kH2 aH
k-1 + kH2 aH
+
slope ≈
(7)
+
because independent runs showed that buffer catalysis
(kB2 H[BH]) is negligible compared to H+ catalysis (kHaH
+
)
2
under most conditions of our experiments while the lack of
product formation at high pH indicates k2 for spontaneous
leaving group departure to be very small. Inversion plots
according to eq 8 are shown in Figure 4 for the reactions of
Figure 4. Reactions of 1a-d with HOCH2CH2S-. Plot of 1/kobsd vs
+
1/aH according to eq 8: (0) 1c; (9) 1d; (O) 1a; and (b) 1b.
k-1
k1kH2 aH
1
1
with those obtained from the experiments at high pH (eq 5).
The latter will be adopted in our subsequent discussion; they
)
+
(8)
kobsd k1
+
are also the values used to calculate kH/k-1 from the slopes
2
1a-d with HOCH2CH2S-. Similar inversion plots were obtained
according to eq 8.
with the other thiolate ions. They yield the k1 and kH/k-1
A few experiments were conducted where buffer catalysis is
not negligible, to provide some representative kB2 H/k-1 ratios.
The best conditions were for the reactions of 1a-d with
2
values summarized in Table 1. Even though the k1 values have
a relatively large standard error, they are in good agreement

6634 J. Am. Chem. Soc., Vol. 121, No. 28, 1999
Bernasconi et al.
Table 2. Slopes of Plots of kobsd/[RS-] vs N-Methylmorpholinium Concentration at pH 7.56 and kBH/k-1 Ratios
2
HOCH2CH2S-
MeO2CCH2CH2S-
slope,a M-2 s-1
kB2 H/k-1,b M-1
slope,a M-2 s-1
kB2 H/k-1,b M-1
(CO)5CrdC(OMe)Ph (1a)
(CO)5WdC(OMe)Ph (1b)
(CO)5CrdC(OEt)Ph (1c)
(CO)5WdC(OEt)Ph (1d)
(6.25 ( 0.03) × 105
(1.01 ( 0.02) × 106
(1.11 ( 0.02) × 105
(1.80 ( 0.05) × 105
27.8 ( 1.5
15.5 ( 0.8
9.25 ( 0.66
4.46 ( 0.30
(2.41 ( 0.03) × 105
(6.48 ( 0.03) × 105
(5.26 ( 0.02) × 104
(1.04 ( 0.03) × 105
10.5 ( 0.3
8.93 ( 0.10
4.38 ( 0.13
2.45 ( 0.13
a Slope ≈ k1kBH/k-1 (eq 9). b Calculated using k1 determined via eq 5.
2
Discussion
The present work reports the first detailed kinetic study of a
nucleophilic substitution of Fischer carbene complexes that
allowed the direct observation of the tetrahedral intermediate
and a complete dissection of the rate constants of various steps
of the mechanism. Specifically, rate constants for nucleophilic
addition of thiolate ions (k1) and its reverse (k-1), for hydronium
ion catalyzed alkoxide ion expulsion from the intermediate
(kH2 ), and for general acid-catalyzed alkoxide ion departure by
N-methylmorpholinium ion (kB2 H) for some selected reactions
were obtained. Rate constants for spontaneous conversion of
the intermediate to products (k2) could not be determined,
though, because this process is slower than the decomposition
of the intermediate into unidentified byproducts. Rate constants
for general acid-catalyzed leaving group expulsion by Et3NH+
and acetic acid could not be obtained either because the catalytic
effect was too small against the background of the kHaH
Figure 5. Reactions of 2a-d (R ) n-Pr) with H+ in N-methyl-
morpholine buffers at constant N-methylmorpholinium ion concen-
tration (5 × 10-3 M). Plots according eq 10: (b) 2b; (O) 2a, (9) 2d;
and (0) 2c.
+
2
process to yield accurate kB2 H values.
A complete set of k1, k-1, and kH values was obtained for the
2
reactions of 1a-d with n-PrS-, HOCH2CH2S-, and MeO2-
CCH2CH2S- (Table 1). For the reactions of 1a-d with
HOCH2CH2S- and MeO2CCH2CH2S- in N-methylmorpholine
MeO2CCH2S- only k1 and the kH/k-1 ratio could be deter-
buffers at pH 7.56. At this pH the kHaH term is relatively
2
+
2
mined because the intermediate decomposes too rapidly. The
nature of the decomposition reaction is unknown.
small compared to kB2 H[BH] and kHaH + kB2 H[BH] , k-1
,
+
2
ensuring that conversion of the intermediate to products is rate
limiting. Plots of kobsd/[RS-] vs [BH] (not shown) were linear
with slopes given by eq 9. These slopes are reported in Table
2 along with the kB2 H/k-1 ratios. Control experiments in the
absence of thiolate ion showed that the observed catalysis by
The error limits associated with the various parameters
reported in Table 1 are standard deviations. For k1 they are
relatively small, typically 3-6%. For kH2 they are also quite
small in most cases, typically 2-5%, except for the reactions
of 1a and 1b with HOCH2CH2S- where they are around 26%.
These latter standard deviations reflect poor reproducibility of
the kinetic results for these two reactions. The standard
k1kB2 H
slope ≈
(9)
deviations for the kH/k-1 ratios are mostly between 12 and
k-1
2
20%. This leads to even larger errors in k-1 obtained as
kH/(kH/k-1) and K1 obtained as k1/k-1. However, these experi-
mental uncertainties do not affect the major conclusions to be
derived from the various parameters and their dependence on
structure.
N-methylmorpholine is not due to nucleophilic addition to 1a-b
or base catalysis of hydrolysis of 1a-b by N-methylmorpholine.
C. Reaction of the Intermediate with Acid. The respective
intermediates of the reactions of 1a-d with n-PrS-, HOCH2-
CH2S-, and MeO2CCH2CH2S- were generated in the presence
of 10-3 M RS- at pH ) pKaRSH; as mentioned earlier, in the
reactions with MeO2CCH2S- no intermediate could be observed
due to rapid decomposition. The solutions containing the
intermediate were then mixed with various N-methylmorpholine
buffers ranging in pH from 6.97 to 8.05 in a stopped-flow
spectrophotometer. The conversion of the intermediate to
products was monitored at λmax of the products. Plots of kobsd
2
2
Rate and Equilibrium Constants for Nucleophilic Addi-
tion. A. The (CO)5M Moiety as π-Acceptor. As pointed out
by numerous authors,2 there is a close formal analogy between
the reaction of nucleophiles with Fischer carbene complexes
and with carboxylic esters. However, quantitatively there is a
large difference in that the equilibrium constants for nucleophilic
addition to the carbene complexes are several orders of
magnitude larger than for addition to esters. For example, K1
for MeO- addition to 1a or 1b in methanol was estimated to
be 2 × 108 to 109 fold larger than for MeO- addition to methyl
benzoate.11 Similarly, thiolate ion addition to 1a or 1b is strongly
favored over thiolate ion addition to acyl carbon. Based on our
K1 values for n-PrS- addition to 1a and 1b and on Guthrie’s16
estimate of K1 ≈ 7.9 × 10-14 M-1 for EtS- addition to acetic
acid in water at 25 °C one obtains K1(1a)/K1(CH3CO2H) ≈
+
vs aH were linear according to eq 10, as shown in Figure 5 for
a representative series. The intercepts (k-1) were in general too
small to yield a precise k-1 value. However, k-1 was obtained
kobsd ) kH2 aH+ + k
(10)
-1
based on kH2 and the kH/k-1 ratios determined under steady-
2
state conditions (Table 1).
(16) Guthrie, J. P. J. Am. Chem. Soc. 1978, 100, 5892.

Physical Organic Chemistry of TM Carbene Complexes. 16
J. Am. Chem. Soc., Vol. 121, No. 28, 1999 6635
1.3 × 1017 and K1(1b)/K1(CH3CO2H) ≈ 1.7 × 1017. These
comparisons show even more dramatically that the (CO)5M
moiety is much more electron withdrawing than the carbonyl
oxygen of an ester or a carboxylic acid.17 This is, in large
measure, due to the π-acceptor ability of the (CO)5M moiety
which, in the adducts, leads to delocalization of the negative
charge into the CO ligands.
Table 3. Rate and Equilibrium Constants for the Reactions of
MeO2CCH2CH2S- with 1a and 10
(CO)5CrdC(OMe)Ph
(1a)
Ph(NO2)CdC(OMe)Ph
(10)
50% MeCN-50% H2Oa 50% DMSO-50% H2Ob,c 1a/10
k1, M-1 s-1
k-1, s-1
2.30 × 104
3.60 × 102
8.99 × 10-2
4.00 × 103
63.9
72.9
0.88
6.55
K1, M-1
3.51 × 103
Evidence for this charge delocalization comes from the IR
data which show that the CO stretching frequencies decrease
substantially upon conversion of 1b to 2b or 1c to 2c. These
frequency shifts are similar to those obtained in the formation
of 5a and 5b11 and of 7 or similar amine adducts;18 they indicate
a decrease in the CdO double bond character of the CO ligands.
Additional evidence can be based on the well-documented
inverse relationship between intrinsic rate constants19 of car-
banion forming reactions and the degree of resonance stabiliza-
tion of the carbanion. The underlying reason for this inverse
relationship is that the development of the resonance effect lags
behind bond formation or charge transfer at the transition state21
(transition state “imbalance”). According to this notion the
intrinsic rate constant for nucleophilic addition to carbene
complexes such as 1a-d should be lower than that for
nucleophilic addition to carboxylic acids or esters. This is
because in the tetrahedral ester adduct the entire negative charge
is localized on one oxygen atom, i.e., there is no resonance
stabilization. Comparison of MeO- addition to 1a and 1b in
methanol with the same reaction of methyl benzoate and
methyltrifluoroacetate indeed showed that the intrinsic rate
constant for the carbene complex reactions is about 2 orders of
magnitude lower than that for the ester reactions.11 This indicates
that the (CO)5M moiety is a moderately strong π-acceptor.
A similar conclusion is reached when comparing the reactions
of 1a or 1b with n-PrS- to the reactions of eq 11. By applying
a 25 °C. b 20 °C. c Reference 12b.
delocalized into the CO ligands, a point emphasized by the
bracket symbol in 9.
Another comparison of interest is that between the thiolate
ion addition to 1a-d with the analogous reaction of â-methoxy-
R-nitrostilbene, eq 13.12a,b This comparison is particularly
appropriate because the two reactions have similar equilibrium
constants. For example, K1 for the reaction of 10 with
MeO2CCH2CH2S- in 50% DMSO-50% water at 20 °C is
almost identical to K1 for the reaction of 1a with the same
nucleophile in 50% MeCN-50% water at 25 °C (Table 3). On
the other hand, the rate constants in both directions are almost
2 orders of magnitude higher for 1a than for 10, indicating that
the intrinsic rate19 constant is higher by the same amount. We
conclude that even though the (CO)5M moieties are moderately
strong π-acceptors, they are not as strong as the nitro group,
i.e., the resonance component that contributes to the stabilization
of 2a-d is not as strong as that for 11. The same conclusion is
reached when comparing the intrinsic rate constants for the
deprotonation of 4a or 4b with those for PhCH2NO2.3
B. Dependence on the Nucleophile. (1) In comparison with
MeO- addition to 1a and 1b, the K1 and k1 values for thiolate
addition are substantially higher despite the much lower proton
basicities of the thiolate ions. For example, for 1a K1(n-PrS-)/
K1(MeO-) ) 151 and k1(n-PrS-)/k1(MeO-) ) 174, for 1b
K1(n-PrS-)/K1(MeO-) ) 122 and k1(n-PrS-)/k1(MeO-) ) 226.
This contrasts with pKa ) 11.94 for n-PrSH vs pKa ) 18.31
for methanol in methanol.25 These comparisons demonstrate the
well-known high carbon basicity and high nucleophilicity of
sulfur bases which have been mainly attributed to their strong
polarizability26 or “softness”.27 1a and 1b may be considered
to be particularly soft, hence optimizing the favorable soft-
soft interactions with thiolate ions and accentuating the unfavor-
able soft-hard interactions with the hard methoxide ion. An
additional factor, particularly important in hydroxylic solvents,
is the weaker solvation of thiolate ions compared to MeO-.
(2) Another observation that may be understood in terms of
soft-soft/soft-hard/hard-hard interactions is the much larger
K1(1a)/K1(CH3CO2H) ratio (≈1.3 × 1017) for the addition of
n-PrS- to 1a vs addition of EtS- to CH3CO2H, compared to
the simplest version of the Marcus equation22 to Guthrie’s
estimate for K1 and k123 one obtains log ko ≈ 3.5 for R ) CH3
and log ko ≈ 5.9 for R ) CF3. Applying the same equation to
the reaction of 1a with n-PrS- yields log ko ≈ 2.1. Even though
crude, these estimates clearly show that the intrinsic rate constant
for the carbene complex reaction is substantially lower than that
for the ester reactions,24 just as is the case for MeO- addition.
As to the nature of the imbalance, the situation may be described
by eq 12, which shows the charge at the transition state mostly
localized on the metal while in the adduct the charge is mostly
(17) The reasons why these equilibrium constant ratios are so much
larger than the corresponding ratios for MeO- addition will be discussed
in section B.
(18) Kreissl, F. R.; Fischer, E. O. Chem. Ber. 1974, 107, 183.
(19) The intrinsic rate constant of a reaction with forward rate constant
(23) For R ) CH3, log K1 ≈ -13.1 and log k1 ≈ -3.1; for R ) CF3,
log K1 ≈ -5.3 and log k1 ≈ 3.3.16
(24) The actual difference between log ko for the reaction of 1a + n-PrS-
and the reactions of eq 11 is probably even greater because log ko for the
reactions of eq 11 in 50% MeCN-50% water is expected to be higher
than that in pure water.21a, c
k1 and reverse rate constant k-1 is defined as ko ) k1 ) k-1 when K1
)
k1/k-1 ) 1. The intrinsic rate constant is a purely kinetic measure of the
reaction barrier that is not influenced by any thermodynamic driving force.20
(20) Marcus, R. A. J. Chem. Phys. 1965, 43, 679.
(25) Rochester, C. H.; Rossall, B. Trans. Faraday Soc. 1969, 65, 1004.
(26) (a) Sander, E. G.; Jencks, W. P. J. Am. Chem. Soc. 1968, 90, 6154.
(b) Hine, J. Structural Effects on Equilibria in Organic Chemistry; Wiley:
New York, 1975; p 225.
(27) (a) Pearson, R. G.; Songstad, J. J. Am. Chem. Soc. 1967, 89, 1827.
(b) Pearson, R. G. SurV. Prog. Chem. 1969, 5, 1.
(21) (a) Bernasconi, C. F. Acc. Chem. Res. 1987, 20, 301. (b) Bernasconi,
C. F. Acc. Chem. Res. 1992, 25, 9. (c) Bernasconi, C. F. AdV. Phys. Org.
Chem. 1992, 27, 119.
(22) log ko ) log k1 - 0.5 log K1.20

6636 J. Am. Chem. Soc., Vol. 121, No. 28, 1999
Bernasconi et al.
Table 4. Brønsted Coefficients for the Dependence of k1, k-1, K1, and kH on pKaRSH
a
2
1a
1b
1c
1d
ânuc ) d log k1/d pKRSH
a,b
c,d
c,d
-0.28 ( 0.05
-0.55 ( 0.08
0.36 ( 0.09
0.27 ( 0.16
-0.25 ( 0.04
-0.57 ( 0.06
0.39 ( 0.06
0.27 ( 0.17
-0.28 ( 0.05
-0.70 ( 0.06
0.50 ( 0.06
0.26 ( 0.13
-0.26 ( 0.04
-0.75 ( 0.06
0.54 ( 0.06
a
âlg ) d log k-1/d pKRSH
a
âeq ) d log K1/d pKRSH
a
âpush ) d log kH/d pKRSH
b,c
0.31 ( 0.09 (0.36)
2
a
a Based on 4 points (n-PrS-, HOCH2CH2S-, MeO2CCH2CH2S-, MeO2CCH2S-). b Error limits are standard deviations plus (0.02. c Based on 3
points (n-PrS-, HOCH2CH2S-, MeO2CCH2CH2S-). d Error limits are standard deviations plus (0.04.
K1(1a)/K1(PhCO2Me) ≈ 2 × 108 to 109 for the addition of
MeO- to 1a vs addition of MeO- to methylbenzoate (see
previous section). The reaction of 1a with a thiolate is
particularly favorable because it leads to a soft-soft interaction
while the reaction of acetic acid with a thiolate ion is disfavored
because it leads to a hard-soft interaction. This tends to make
the K1(1a)/K1(CH3CO2H) ratio particularly large. On the other
hand, the reaction of 1a with MeO- is disfavored because it
represents a soft-hard combination while the reaction of methyl
benzoate with MeO- is favored because of the hard-hard
combination. Hence both factors reduce the K1(1a)/
K1(PhCO2Me) ratio.29
(3) A third point of interest in comparing thiolate ion with
methoxide ion addition to the Fischer carbene complexes is that
the k1(n-PrS-)/k1(MeO-) ratios are similar to the K1(n-PrS-)/
K1(MeO-) ratios or even slightly larger. This implies that the
intrinsic rate constants are larger for thiolate ion addition than
Figure 6. Brøsted plots for the reaction of 1b with thiolate ions.
for MeO- addition. Again applying the simplest version of the
Marcus equation,22 log ko ≈ 2.11 for 1a + n-PrS- vs log ko ≈
Scheme 1
0.96 for 1a + MeO-,11 and log ko ≈ 2.56 for 1b + n-PrS- vs
log ko ≈ 1.25 for 1b + MeO-.11 Similar findings have recently
been reported when comparing thiolate and alkoxide ion addition
to olefinic and vinylic substrates.12c,d The differences in the
intrinsic rate constants are the result of transition states that are
imbalanced with respect to the factors that are responsible for
complexes. The âlg and âeq values show no significant depen-
the differences in the equilibrium constants, i.e., the stronger
solvation of alkoxide ions and the soft-soft interactions in the
thiolate ion reactions. Specifically, following generally observed
behavior,30 the partial desolvation of the nucleophile that occurs
as it enters the transition state should be more advanced than
bond formation. This has the effect of reducing the intrinsic
rate constants, and more so for the reactions with the more
strongly solvated alkoxide ions.12c Regarding the soft-soft
interactions, it is not unreasonable to assume that they are more
advanced than bond formation at the transition state,31 which
has the effect of enhancing the intrinsic rate constants for the
thiolate ion reactions. Hence both factors reinforce each other
in increasing the difference in the intrinsic rate constants of these
reactions.
dence on the metal but some dependence on the alkoxy group;
in view of the relatively large uncertainties32 and the fact that
the âlg and âeq were obtained based on three points only, with
two points referring to very similar pKRa SH values, it is not clear
whether the apparent dependence on the alkoxy group is
significant and no interpretation will be attempted.
The most interesting features are the fact that ânuc is negative
and that âeq is much smaller than unity. Negative ânuc are
unusual but not without precedent.33 According to Jencks,34 they
may result from the requirement of partial desolvation of the
nucleophile prior to reaction. Whether this desolvation is part
of the nucleophilic attack, a point of view favored by us12a,c
and implied by the discussion in the previous section, or as a
separate step,30b is unclear. However, the following analysis is
simpler if desolvation is assumed to be a separate preequilibrium
step and hence the two-step model will be adopted here. It is
illustrated by Scheme 1.
(4) The dependence of K1, k1, and k-1 on pKRa SH for the
reactions of 1b is shown in Figure 6; the corresponding Brønsted
plots for 1a, 1c, and 1d (not shown) are very similar. The slopes,
â
nuc, âlg, and âeq are summarized in Table 4. The ânuc values
In terms of Scheme 1, the experimental k1 value corresponds
to Kdk′1, with Kd being the equilibrium constant for partial
are, within experimental error, the same for all carbene
(32) The error limits given for âlg and âeq are larger than the standard
deviations (see footnote d in Table 4) because of the relatively large errors
(28) (a) Parker, A. J. Chem. ReV. 1969, 69, 1. (b) Bordwell, F. G.;
Hughes, D. L. J. Org. Chem. 1982, 47, 3224. (c) Taft, R. W. Prog. Phys.
Org. Chem. 1983, 14, 247.
in the kH/k-1 ratios (Table 1) from which k-1 and K1 are calculated, and
2
(29) The fact that K1(1a) for the reaction with n-PrS- refers to 50%
MeCN-50% water while K1(CH3CO2H) was estimated in water cannot
account for more than a minor part of the large K1(1a)/K1(CH3CO2H)
ratio for the thiolate ion reactions.
additional uncertainties arising from experimental errors in the pKaRSH
values. This latter factor was also considered in assigning error limits to
ânuc (footnote b in Table 4).
(33) Negative ânuc values have been reported for some phosphoryl transfer
reactions to amines34 and for reactions of highly reactive carbocations
with amines;35,36 ânuc values close to zero have also been found in the
reaction of diphenylketene with amines.37
(30) (a) Hupe, D. J.; Jencks, W. P. J. Am. Chem. Soc. 1977, 99, 451. (b)
Jencks, W. P.; Brant, S. R.; Gandler, J. R.; Fendrich, G.; Nakamura, C. J.
Am. Chem. Soc. 1982, 104, 7054.
(31) Bernasconi, C. F.; Killion, R. B., Jr. J. Am. Chem. Soc. 1988, 110,
7506.
(34) Jencks, W. P.; Haber, M. T.; Herschlag, D.; Nazaretian, K. L. J.
Am. Chem. Soc. 1986, 108, 479.

Physical Organic Chemistry of TM Carbene Complexes. 16
J. Am. Chem. Soc., Vol. 121, No. 28, 1999 6637
Scheme 2
Using the same reasoning as in the derivation of eq 16, one
may derive eq 17 with â′eq defined as d(log k′1/k-1)/d pK′a ) d
âd + â′eq
âd + 1
âeq
)
(17)
log K′1/d pK′a and âd defined as before. From eq 17 it is apparent
that âeq ) â′eq if â′eq ) 1, irrespective of the value of âd, but
for âeq < 1 and âd < 0 one obtains âeq < â′eq; the difference
between âeq and â′eq becomes larger the smaller â′eq.
desolvation of the nucleophile. In the simplest approach, ânuc
may be approximated by eq 14.
d log k1 d log Kdk′1
ânuc
)
)
)
C. Dependence on the Metal and on the Alkoxy Group.
The reactions of the tungsten complexes are somewhat favored
over those of the chromium complexes. The effect on the K1
values is rather minimal but somewhat larger on the rate
constants: the K1(1b)/K1(1a) and K1(1d)/K1(1c) ratios vary
between 1.27 and 1.53, the k1(1b)/k1(1a) and k1(1d)/k1(1c) ratios
between 2.56 and 3.54 (Table 5). Similar results were found
for MeO- addition to 1a and 1b in methanol: K1(1b)/K1(1a)
) 111/70.1 ) 1.58 and k1(1b)/k1(1a) ) 186/77.1 ) 2.41. These
findings are also consistent with a report that the acidity of 4b
(pKa ) 12.36)46 is slightly higher than that of 4a (pKa )
12.50).9b
The rather minimal dependence of the equilibrium constants
on the metal may be the result of the extensive dispersion of
the negative charge into the CO ligands of the adducts.47 Since
this leaves little charge on the metal, the identity of the metal
has little effect on the stability of the adduct. An additional factor
is related to the strong contribution of 1( to the structure of 1.2
Because 1( leads to substantial negative charge on the (CO)5M
d pKRa SH
d pKRa SH
d log Kd d log k′
+
1 ) âd + â′nuc (14)
d pKRa SH d pKRa SH
Since desolvation becomes more difficult as the basicity of
RS- increases, âd < 0. If â′nuc is small, ânuc may be dominated
by âd and become negative. For |âd| > â′nuc, eq 14 underesti-
mates the effect of the desolvation step on ânuc, while for |âd|
< â′nuc, it overestimates it. This is because it neglects the effect
of the desolvation step on pKRa SH. If this is taken into account
(Scheme 2), pKaRSH corresponds to log Kd + pK′a and ânuc is
given by
d(log Kd + log k′1)
ânuc
)
(15)
d(log Kd + pK′a)
Defining â′nuc ) d log k′1/d pK′a and âd ) d log Kd/d pK′a
converts eq 15 to eq 16.38
âd + â′nuc
âd + 1
ânuc
)
(16)
Even though usually not negative, ânuc values are typically
quite low for thiolate ion addition to a variety of electro-
philes.12b,30a,31,39-41 This implies a transition state with little bond
formation.42
moiety of the carbene complex, the greater stabilization of the
negative charge by the (CO)5W moiety in the adduct is partially
offset by a similar effect on the carbene complex. Support for
this notion has been discussed elsewhere.1
The fact that âeq is smaller than unity indicates that the
carbon basicities of the thiolate ions toward the carbene
complexes are not quite proportional to their proton basicities.
This is a common pattern in reactions of thiolate ions with
electrophiles.12b,c,31,41 It may be related to the high polarizability
of the thiolate ions in the sense that, just as the absolute carbon
basicities are dominated by the soft-soft interactions and rather
insensitive to the factors that determine proton basicities, the
effect of the R group of RS- on the carbon basicity is also
attenuated relative to its effect on the proton basicity. The partial
desolvation of the nucleophile (Scheme 2) contributes to the
The fact that the k1 values show a somewhat greater
dependence on the metal than K1 and that not only k1 but also
k-1 is larger for the tungsten complexes is noteworthy; it means
that the intrinsic rate constant19 is somewhat higher for the
tungsten complexes than for the chromium complexes. This can
be understood with reference to eq 12. If the fractional charge
on the metal is greater at the transition state (8) than in the
adduct (9) and tungsten is better able to support negative charge
than chromium, the greater stabilization resulting from the
change from chromium to tungsten will affect the transition state
more than the adduct and increase k1 more than K1.
reduction in âeq in a similar way as to the reduction in ânuc
.
Regarding the dependence on the alkoxy group, the K1 values
are somewhat lower for the ethoxy carbene complexes than for
the methoxy carbene complexes: the K1(1a)/K1(1c) and K1(1b)/
K1(1d) ratios vary between 1.73 and 2.90 (Table 5). The same
is true for the k1 values with the k1(1a)/k1(1c) and k1(1b)/
k1(1d) ratios varying between 1.61 and 1.97 (Table 5). The lower
K1 and k1 values for the ethoxy carbene complexes may be
attributed to the stronger π-donation (1() by the ethoxy group
compared to the methoxy group. The same effect is seen in the
higher pKa values of 12 (12.98)46 compared to 4a (12.50).9b
Leaving Group Departure. The mechanism of acid-
catalyzed leaving group departure is probably concerted, with
(35) Richard, J. P. J. Chem. Soc., Chem. Commun. 1987, 1768.
(36) McClelland, R. A.; Kanagasabapathy, V. M.; Banait, N. S.; Steenken,
S. J. Am. Chem. Soc. 1992, 114, 1816.
(37) Andraos, J.; Kresge, A. J. J. Am. Chem. Soc. 1992, 114, 5643.
(38) For example, if âd ) -0.30 and ânuc′ ) 0.10, ânuc ) -0.29 (eq 16)
and -0.20 (eq 14).
(39) (a) Gilbert, H. F.; Jencks, W. P. J. Am. Chem. Soc. 1977, 99, 7931.
(40) Ritchie, C. D.; Gandler, J. R. J. Am. Chem. Soc. 1979, 101, 7318.
(41) (a) Bernasconi, C. F.; Schuck, D. F. J. Org. Chem. 1992, 57, 2365.
(b) Bernasconi, C. F.; Ketner, R. J. J. Org. Chem. 1998, 63, 6266.
(42) This is the traditional view,43 although this view has been
challenged44 as well as defended.45
(43) (a) Hammond, G. S. J. Am. Chem. Soc. 1955, 77, 554. (b) Leffler,
J. E.; Grunwald, E. Rates and Equilibria of Organic Reactions; Wiley: New
York, 1963; p 156. (c) Kresge, A. J. Acc. Chem. Res. 1975, 8, 354.
(44) (a) Pross, A. J. Org. Chem. 1984, 49, 1811. (b) Bordwell, F. G.;
Hughes, D. L. J. Am. Chem. Soc. 1985, 107, 4737. (c) Pross, A.; Shaik, S.
S. New J. Chem. 1989, 13, 427.
(45) Jencks, W. P. Chem. ReV. 1985, 85, 511.
(46) Bernasconi, C. F.; Sun, W. Organometallics 1997, 16, 1926.
(47) See structure 9.

6638 J. Am. Chem. Soc., Vol. 121, No. 28, 1999
Table 5. Equilibrium and Rate Constant Ratios
Bernasconi et al.
ratio
n-PrS-
HOCH2CH2S-
MeO2CCH2CH2S-
MeO2CCH2S-
alkoxy group dependence of first step
K1(1a)/K1(1c)
k1(1a)/k1(1c)
K1(1b)/K1(1d)
k1(1b)/k1(1d)
2.07 ( 0.79
1.90 ( 0.20
1.73 ( 0.71
1.81 ( 0.19
a
2.90 ( 0.98
1.92 ( 0.06
2.76 ( 1.18
1.71 ( 0.16
1.87 ( 0.17
a
1.61 ( 0.11
1.97 ( 0.13
1.65 ( 0.23
metal dependence of first step
K1(1b)/K1(1a)
k1(1b)/k1(1a)
K1(1d)/K1(1c)
k1(1d)/k1(1c)
1.27 ( 0.59
3.14 ( 0.34
1.53 ( 0.49
3.30 ( 0.33
a
1.27 ( 0.53
3.16 ( 0.07
1.33 ( 0.46
3.54 ( 0.14
2.90 ( 0.22
a
3.37 ( 0.29
2.56 ( 0.28
3.06 ( 0.36
alkoxy group dependence of second step
kH2 (1a)/kH2 (1c)
kH2 (1b)/kH2 (1d)
kB2 H(1a)/kB2 H(1c)
kB2 H(1b)/kB2 H(1d)
1.27 ( 0.05
1.11 ( 0.07
1.31 ( 0.39
1.42 ( 0.41
b
b
1.18 ( 0.03
1.11 ( 0.07
1.59 ( 0.58
2.26 ( 1.05
metal dependence of second step
kH2 (1b)/kH2 (1a)
kH2 (1d)/kH2 (1c)
kB2 H(1b)/kB2 H(1a)
kB2 H(1d)/kB2 H(1c)
1.06 ( 0.07
1.21 ( 0.06
b
1.03 ( 0.03
1.10 ( 0.07
2.12 ( 0.94
1.49 ( 0.58
0.99 ( 0.08
b
b
a Experimental errors in the K1 values are too large to give meaningful ratios. b Experimental errors in the k2H and kB2 H values are too large to give
meaningful ratios.
transition states 13 for the kH2 and 14 for the k2BH steps,
Conclusions
respectively. The k2H values are all above 108 M-1 s-1, the
(1) The stabilization of the tetrahedral intermediate by the
highest being 5.69 × 108 M-1 s-1 (1b/n-PrS-); this is only about
(CO)5M moiety in the reactions of 1a-d with nucleophiles is
1 order of magnitude below the limit for diffusion-controlled
much stronger than the stabilization of the intermediate by the
reactions.48
carbonyl oxygen in the reaction of esters with nucleophiles. This
is, in large measure, due to the π-acceptor ability of the (CO)5M
moiety. Evidence for charge delocalization in the intermediate
comes from IR data and the comparatively low intrinsic rate
constants for nucleophilic addition.
(2) The extent to which K1 for nucleophilic addition to
carbene complexes exceeds K1 for addition to acyl carbon
depends strongly on soft-soft, soft-hard, and hard-hard
interactions between nucleophile and electrophile. Soft-soft
combinations such as thiolate ions/carbene complexes are
particularly favorable, and so are hard-hard combinations such
as alkoxide ions/acyl compounds. Less favorable are soft-hard
or hard-soft combinations such as thiolate ions/acyl compounds
and alkoxide ions/carbene complexes, respectively.
(3) The intrinsic rate constants are higher for thiolate ion
addition to carbene complexes than for MeO- addition. This
appears to be the result of partial desolvation of the nucleophile
being ahead of bond formation and the fact that solvation of
thiolate ions is weaker than that of MeO-. A contributing factor
may be early development of soft-soft interactions in the
transition state.
(4) ânuc for the k1 step is negative, indicating that bond
formation has made little progress at the transition state and
partial desolvation of the nucleophile prior to reaction plays a
dominant role. The desolvation step is also partly responsible
for the small âeq values.
There is an increase in kH2 with increasing basicity of the
thiolate ion. This indicates the operation of an electronic push
by the RS group, presumably because of developing resonance
(15) in the transition state. From the slopes of plots of log k2H
vs pKRa SH (Figure 6) the âpush values summarized in Table 4 are
obtained.
The kH2 and kB2 H values are slightly larger for the methoxy
derivatives: the k2H(1a)/k2H(1c) and k2H(1b)/kH2 (1d) ratios vary
between 1.11 and 1.42, and the kB2 H(1a)/k2BH(1c) and k2BH(1b)/
kB2 H(1d) ratios vary between 1.59 and 2.26 (Table 5). In acid-
catalyzed alkoxide ion departure the relative progress of C-O
bond cleavage vs oxygen protonation determines which leaving
group leaves faster. If C-O bond cleavage is ahead of proton
transfer, the energetics of the reaction should be more reflective
of C-O bond cleavage than of the proton transfer and hence
the less basic leaving group (MeO-) should leave faster. On
the other hand, if proton transfer is ahead of C-O bond cleavage
the opposite holds and the ethoxy derivative should react faster.
Our results suggest that the former situation applies.
(5) The dependence of K1 on the metal is almost negligible,
but there appears to be a significant dependence of the intrinsic
rate constant on the metal (W > Cr). This suggests a transition
state imbalance that places a greater fractional negative charge
on the metal at the transition state than in the adduct; in the
adduct the charge is mainly dispersed into the CO ligands.
(6) K1 and k1 are somewhat lower for the ethoxy carbene
complexes compared to the methoxy derivatives. This reflects
Regarding the dependence of kH2 and k2BH on the metal, the
effect on kH2 is too small and the uncertainties in the kB2 H values
are too large to discern a significant difference in reactivity.
(48) Eigen, M. Angew. Chem., Int. Ed. Engl. 1964, 3, 1.

Physical Organic Chemistry of TM Carbene Complexes. 16
J. Am. Chem. Soc., Vol. 121, No. 28, 1999 6639
CH3), 2.70 (t, 2H, SCH2CH2CH3), 6.6 (d, 2H, Ph), 7.4 (m, 3H, Ph). 1H
NMR (3b): δ 0.91 (t, 3H, CH2CH2CH3), 1.48 (m, 2H, SCH2CH2CH3),
2.49 (t, 2H, SCH2CH2CH3), 6.7 (d, 2H, Ph), 7.5 (m, 3H, Ph). 13C NMR
(3a): δ 13.0 (SCH2CH2CH3), 21.3 (SCH2CH2CH3), 46.5 (SCH2CH2-
CH3), 126.8-128.7, 157.5 (Ph), 215.5 (cis CO), 228.6 (trans CO), 362.6
(dC). 13C NMR (3b): δ 13.3 (SCH2CH2CH3), 21.2 (SCH2CH2CH3),
46.9 (SCH2CH2CH3), 127.4-128.9, 157.9 (Ph), 197.4 (cis CO), 209.0
(trans CO), 329.4 (dC).
the stronger π-donation by the ethoxy group in the neutral
carbene complex.
(7) Acid-catalyzed leaving group departure is probably
concerted. The kH2 values increase with increasing basicity of
the thiolate ion, reflecting the increased “push” by the RS group
due to developing resonance in the product (15).
Experimental Section
NMR of Intermediate. The intermediate derived from the reaction
of 1b with n-PrS- was stable enough to be characterized by NMR. 1H
NMR: δ 1.35 (t, 3H, SCH2CH2CH3), 1.96 (m, 2H, SCH2CH2CH3),
Instrumentation. 1H and 13C NMR spectra were recorded in CDCl3
on a 500 MHz Varian Unity instrument. UV-vis spectra were obtained
on a Hewlett-Packard 8452A diode array spectrophotometer. IR spectra
were taken with a Perkin-Elmer 1600 FT IR spectrophotometer. Kinetic
experiments were performed on an Applied Photophysics DX.17MV
stopped-flow apparatus (fast reactions) or Hewlett-Packard spectro-
photometer (slow reactions).
2.44 (t, 2H, SCH2CH2CH3), 4.04 (s, 3H, OCH3), 7.66 (m, 5H, Ph). 13
C
NMR: δ 13.7 (SCH2CH2CH3), 22.5 (SCH2CH2CH3), 38.8 (SCH2CH2-
CH3), 49.0 (OCH3), 52.4 (>C<), 127.4-128.6 (Ph), 202.5 (cis CO),
207.6 (trans CO).
Kinetics. All reactions were conducted under pseudo-first-order
conditions with a large excess of thiol/thiolate or, for the experiments
2- f 3, of buffer over the substrate. Runs in which 1a-d were reacted
with RSH buffers were monitored at λ ) 396 nm, which is at or near
Materials. 1a-d were available from a previous study.15 The alkane
thiols were distilled under argon prior to use. Triethylamine and
N-methylmorpholine were refluxed over CaH2 and freshly distilled prior
to use. Acetic acid was used as received. KOH and HCl solutions were
prepared using “dilut-it” from Baker Analytical. Acetonitrile was used
as received. Water was taken from a Milli-Q ion exchange water
purification system. 50% MeCN-50% water mixtures were degassed
by three cycles of the freeze-pump-thaw method.
λ
max of 1a-d. Runs where the intermediate was reacted with an acidic
buffer were monitored at 454 (W) or 464 (Cr) nm, i.e., at or near the
max of 3a-d. These latter runs were conducted in the sequential mixing
λ
mode of the stopped-flow apparatus as follows. A 10-3 M solution of
the carbene complex was first mixed with a 10-3 M solution of the
thiolate ion. After being aged for 0.5 to 5 s to allow generation of the
intermediate, the solution was mixed with the appropriate acidic buffer.
pH and pKa Measurements. The pH in 50% MeCN-50% water
was determined according to eq 1849 with pHmeas referring to the reading
of the pH meter calibrated with standard buffers. All pH measurements
Identification of Products. All products (3a and 3b) showed the
characteristic absorption spectrum of thioalkoxy carbene complexes
(Figure 1, spectrum B for 3b and Figure 1 in ref 13 for 3a). For 3a
and 3b (R ) n-Pr) the absorption spectra obtained from kinetic runs
of the reactions of n-PrS- with 1c and 1b, respectively, were
indistinguishable from samples of 3a and 3b that were synthesized
independently. The synthesis was performed under conditions that were
as similar as possible to those in the kinetic experiments, as follows.
An 8 mmol sample of 1b or 1c was added to a mixture of 40 mL of
MeCN and 4 mL of water under Ar and stirred. n-PrS-K+ (32 mmol)
was added while stirring. After 5 min HCl was added dropwise until
the pH was ∼4. Stirring was continued for 1 h. The following
procedures were carried out in the open air. The solution was reduced
to a small volume by rotovap. Water (40 mL) was added followed by
100 mL of n-hexane. The aqueous layer was removed and the hexane
layer was extracted with 50 mL of water, collected, and dried over
Na2SO4. The solution was filtered and reduced by rotovap to leave a
dark red oily residue. This residue was filtered through a silica plug
and an HPLC filter plug (0.1 µm). After removal of most of the solvent
by rotovap the sample was dried under high vacuum at -78 °C for 24
h. 1H NMR (3a): δ 0.89 (t, 3H, SCH2CH2CH3), 1.54 (m, 2H, SCH2CH2-
pH ) pHmeas + 0.18
(18)
were done on an Orion 611 pH meter equipped with a glass electrode
and a “Sure Flow” (Corning) reference electrode. The pH of reaction
solutions for stopped-flow experiments was adjusted in mock-mixing
experiments that mimicked the stopped-flow runs. The pKRa SH values
of the thiols were determined potentiometrically.
Acknowledgment. This research has been supported by
Grant No. CHE-9734822 from the National Science Foundation.
We thank Gabriel Perez for technical assistance in taking NMR
spectra of 3a (R ) n-Pr).
JA9908265
(49) Allen, A. D.; Tidwell, T. T. J. Am. Chem. Soc. 1987, 109, 2774.