Grignard reaction with chlorosilanes in THF: A kinetic study

Article abstract of DOI:10.1021/jo0498361

Kinetics of the reactions of phenylmagnesium chloride and bromide and diphenylmagnesium with chlorosilanes were investigated in tetrahydrofurane (THF) and in THF-hydrocarbon mixtures. The reaction in THF is much faster than that in diethyl ether. Assuming coordination of magnesium halides with three molecules of THF, concentrations of all the species involved in Schlenk equilibrium were calculated. In the Grignard reaction, species R₂Mg and RMgX react competitively accompanied by additional reaction paths involving electrophilic catalysis by magnesium halide. This conclusion also proved to be valid for the Grignard reaction with a ketone and probably can be expanded to any Grignard reaction. When Schlenk equilibrium is shifted far to the RMgX species, the catalytic pathways are insignificant. Substituents at the silicon center control the rate of the reaction through their inductive and steric effects.

Full text of DOI:10.1021/jo0498361

Gr ign a r d Rea ction w ith Ch lor osila n es in THF : A Kin etic Stu d y
Ants Tuulmets,*,† Binh T. Nguyen,‡ and Dmitri Panov†
Institute of Organic and Bioorganic Chemistry, University of Tartu, Tartu 51014, Estonia, and
Dow Corning Corporation, Midland, Michigan 48686-0995
ants.tuulmets@ut.ee
Received J anuary 28, 2004
Kinetics of the reactions of phenylmagnesium chloride and bromide and diphenylmagnesium with
chlorosilanes were investigated in tetrahydrofurane (THF) and in THF-hydrocarbon mixtures.
The reaction in THF is much faster than that in diethyl ether. Assuming coordination of magnesium
halides with three molecules of THF, concentrations of all the species involved in Schlenk
equilibrium were calculated. In the Grignard reaction, species R2Mg and RMgX react competitively
accompanied by additional reaction paths involving electrophilic catalysis by magnesium halide.
This conclusion also proved to be valid for the Grignard reaction with a ketone and probably can
be expanded to any Grignard reaction. When Schlenk equilibrium is shifted far to the RMgX species,
the catalytic pathways are insignificant. Substituents at the silicon center control the rate of the
reaction through their inductive and steric effects.
In tr od u ction
centration in THF. Although the importance of the
Schlenk equilibrium in Grignard reactions with various
substances has been stressed,3b,c no quantitative analysis
of the contribution of Schlenk equilibrium in reaction
kinetics is available.
In this paper we report the results of a kinetic in-
vestigation of the Grignard reaction with chlorosilanes
in THF and in some THF-hydrocarbon mixtures. Also,
an analysis of kinetic data involving the data for the
Schlenk equilibrium will be presented. In the kinetic
investigation, phenylmagnesium chloride and bromide
served as model Grignard reagents.
Although the production of organosilanes by direct
methods is greater than that with Grignard processes,
the latter remain essential for many of the specialty
silanes. The versatility of the Grignard process allows
the introduction of a broader range of functionalities than
alternative technologies.1 Despite the viability of the
Grignard process in organosilane production the quan-
titative aspects of the reaction have been little investi-
gated. Only a small number of kinetic studies of the
Grignard reaction with silanes, carried out mainly in
diethyl ether solutions, have been published.2
Of the two most common solvents for Grignard re-
agents, tetrahydrofurane (THF) reveals advantages over
diethyl ether. Apart from lower fire-hazard risks owing
to the higher boiling point, THF allows the reaction to
run at more elevated temperatures. One of the most
significant differences between THF and diethyl ether
arises from the occurrence of Schlenk equilibrium (eq 1)
in solutions of Grignard reagents.3
Resu lts a n d Discu ssion
P u r e THF . As found in our previous work,2e kinetics
of the Grignard reaction with chlorosilanes obeys the
second order law, therefore, the pseudo-first-order rate
constants determined with a great excess of the Grignard
reagent should depend linearly upon the concentration
of the reagent. However, this is not the case in THF, as
seen in Figure 1. The upward curvature of the plot in
Figure 1 may indicate a higher kinetic order in the
Grignard reagent for the reaction. A similar kinetic
feature has been found for the reaction of n-propylmag-
nesium bromide with pinacolone in THF.4 According to
the reaction mechanism deduced from our previous
results for the reaction with silanes,2d,e inclusion of two
Grignard reagent molecules in the transition state is not
2RMgX h R2Mg + MgX2
(1)
While the equilibrium is shifted far left in diethyl
ether, all three species are present in appreciable con-
* Corresponding author. Phone: +3727-375-238. Fax: +3727-375-
245.
† University of Tartu.
‡ Dow Corning Corporation.
(1) Arkles B. In Handbook of Grignard Reagents; Silverman, G. S.,
Rakita P. E., Eds.; Marcel Dekker: New York, 1996; Chapter 32.
(2) (a) Reid, A. F.; Wilkins, C. J . J . Chem. Soc. 1955, 4029-4034.
Corriu, R. J . P.; Henner, B. J . Organomet. Chem 1975, 102, 407-416.
(b) Tuulmets, A.; Ho˜rak, M.; Ko˜opere, T.; Ruotsi, J . Org. React. (USSR)
1982, 19, 102-119. (c) Sassian, M.; Panov, D.; Tuulmets, A. Appl.
Organomet. Chem. 2002, 16, 525-529. (d) Tuulmets, A.; Panov, D.;
Sassian, M. Tetrahedron Lett. 2003, 44, 3943-3945. (e) Tuulmets, A.;
Nguyen, B. T.; Panov, D.; Sassian, M.; J a¨rv, J . J . Org. Chem. 2003,
68, 9933-9937.
(3) (a) Lindsell, W. E. In Comprehensive Organometallic Chemistry;
Wilkinson, G., Ed.; Pergamon Press: Elmford, NY, 1982; Vol. 1,
Chapter 4. (b) Cannon, K. C.; Krow, G. R. In Handbook of Grignard
Reagents; Silverman, G. S., Rakita. P. E., Eds.; Marcel Dekker: New
York, 1996; Chapter 13. (c) Holm, T.; Crossland I. In Grignard
Reagents: New Developments; Richey, H. G., J r., Ed.; J . Wiley: New
York, 2000; Chapter 1.
(4) Koppel, J .; Margna, L.; Tuulmets, A. Reakts. Sposobn. Org.
Soedin. 1968, 5, 1041-1052.
10.1021/jo0498361 CCC: $27.50 © 2004 American Chemical Society
Published on Web 06/30/2004
J . Org. Chem. 2004, 69, 5071-5076
5071

Tuulmets et al.
F IGURE 1. Dependence of pseudo-first-order rate constants
on Grignard reagent concentration for the reaction of phenyl-
magnesium chloride with methyltrichlorosilane in THF at
20 °C.
F IGURE 3. Dependence of pseudo-first-order rate constants
for the reaction of 0.5 M PhMgCl with MeSiCl3 in THF at 20
°C upon the molar fraction of cosolvents (X co-solv).
gradual decrease of the reaction rate with increasing
content of toluene in the Grignard reagent solution.
Although we were able to show the absence of nonspecific
solvent effects for the Grignard reaction with silanes in
diethyl ether,2e this could not be expanded directly to the
reaction in THF. In a check experiment dichloromethane
was used instead of toluene. Additions of dichloromethane
to THF raise the medium polarity and do not change the
polarizability. In Figure 3, points for dichloromethane
additions fall on the curve for toluene-THF mixtures.
Consequently, no nonspecific solvent effects are present
in this case and both cosolvents act as indifferent diluting
agents. Thus, the observed decrease of the reaction rate
should be related to the changes in the Grignard reagent
structure.
Sch len k Equ ilibr iu m . Consideration of the structure
of Grignard reagents in THF is largely facilitated by the
absence of association of the species involved in Schlenk
equilibrium 1.7,8 However, in contrast to diethyl ether,
the Schlenk equilibrium in THF is concentration depend-
ent, the dominant effect being associated with the sol-
vation of magnesium halides.3b The alternative equilib-
riums 2 and 3 emphasize the dependency of the Schlenk
equilibrium on solvent concentration.
F IGURE 2. Dependence of pseudo-first-order rate constants
on the concentration of diphenylmagnesium for the reaction
with methyltrichlorosilane and methylvinyldichlorosilane in
THF at 20 °C.
very plausible. The same is believed to be valid for the
reaction with carbonyl compounds.3c,5 Since such a kinetic
feature has not been found for the reactions in diethyl
ether, it should be assigned to a peculiarity of Grignard
reagents in THF solutions.
To start unraveling the enigma we determined pseudo-
first-order rate constants kI for reactions of chlorosilanes
with a great excess of diphenylmagnesium in THF. The
plots of kI vs diphenylmagnesium concentration in Figure
2 can be approximated to straight lines yielding second-
order rate constants kII ) 0.26 ( 0.01 L mol-1 s-1 for
methyltrichlorosilane and kII ) 0.101 ( 0.006 L mol-1
s-1 for methylvinyldichlorosilane.
R2Mg‚2THF + MgX2‚4THF a
RMgX‚2THF + 2THF (2)
R2Mg‚2THF + MgX2‚3THF a
RMgX‚2THF + THF (3)
THF -Hyd r oca r bon Mixtu r es. The next step of the
study was an investigation of the reaction in THF-
toluene mixtures. Replacement of THF by additions of
nondonating toluene does not change the solvation of the
Grignard species in equilibrium 1; however, this can lead
to shifts in the equilibrium as will be subsequently
shown.
As THF is sterically less demanding than diethyl
ether, magnesium halide can probably coordinate up to
four THF molecules, e.g. MgBr2‚4THF, based on crys-
tallographic data.9 However, the crystal structure of
MgBr2‚2THF also has been described.10 It must be
realized that the crystals do not necessarily reflect the
Addition of nondonating species to Grignard reagents
has been used in our previous works to alter the polarity
and polarizability of the reaction medium.6 Additions of
toluene to the Grignard reagent prepared in THF should
lower considerably the polarity of the solution and
increase slightly the polarizability. Figure 3 indicates a
(6) (a) Koppel, J .; Tuulmets, A. Reakts. Sposobn. Org. Soedin. 1970,
7, 911-918. (b) Koppel, J .; Loit, J .; Luuk, M.; Tuulmets, A. Reakts.
Sposobn. Org. Soedin. 1971, 8, 1155-1164. (c) Koppel, J .; Tuulmets,
A. Reakts. Sposobn. Org. Soedin. 1972, 9, 399-411. (d) Viirlaid, S.;
Kurikoff, S.; Tuulmets, A. Reakts. Sposobn. Org. Soedin. (USSR) 1974,
11, 73-80. (e) Tuulmets, A.; Pa¨llin, V.; Tammiku-Taul, J .; Burk, P.;
Raie, K. J . Phys. Org. Chem. 2002, 15, 701-705.
(7) Walker, F. W.; Ashby, E. C. J . Am. Chem. Soc. 1969, 91, 3845-
3850.
(5) Blomberg C. In Handbook of Grignard Reagents; Silverman, G.
S., Rakita, P. E., Eds.; Marcel Dekker: New York, 1996; Chapter 11.
(8) Ashby, E. C.; Laemmle, J .; Neumann, H. M. Acc. Chem. Res.
1974, 7, 272-281.
5072 J . Org. Chem., Vol. 69, No. 15, 2004

Grignard Reaction with Chlorosilanes in THF
TABLE 1. Com p osition of P h en ylm a gn esiu m Ch lor id e
(m ol/L) in THF Solu tion sa
components in solution. Our density functional theory
calculations11a indicate that trans-dihalotetrakis(tetrahy-
drofurano)magnesiums are only by 3 kcal per mol or less
stabilized over the tris(tetrahydrofurano) complexes while
cis-tetrakis complexes are much more unstable.12 The
tris(tetrahydrofurano) complexes in their turn are by 4
to 5 kcal per mol more favored than the MgX2‚2THF
complexes.11 However, the calculations were performed
for the gas phase. In the liquid phase, small differences
can be detracted due to the presence of the second
solvation sphere. Inferences from our DFT calculations
encouraged us to adopt equilibrium 3, in particular, after
the calculations described below failed when tetrasol-
vated magnesium chloride was assumed.
PhMgb
PhMgCl
Ph2Mg
MgCl2
0.2
0.4
0.5
0.7
0.8
0.084
0.171
0.215
0.304
0.350
0.058
0.115
0.142
0.200
0.225
0.073
0.144
0.185
0.250
0.285
a
b
For W ) 1.15. Concentration of the Grignard reagent.
TABLE 2. Com p osition of 0.5 M P h en ylm a gn esiu m
Ch lor id e in THF -Tolu en e Mixtu r esa
b
c
toluene vol %
Xtol
PhMgClc
Ph2Mgc
MgCl2
0
20
40
60
75
0
0.215
0.235
0.258
0.291
0.337
0.142
0.132
0.120
0.105
0.080
0.185
0.182
0.170
0.155
0.130
According to equilibrium 3, the Schlenk equilibrium
constant can be expressed as
0.18
0.38
0.60
0.79
[RMgX]2[THF]
KS )
(4)
a
b
For W ) 1.15. Molar fraction of toluene. c Mol/L.
[R2Mg][MgX2]
In experiment, e.g. from the NMR spectra, the Schlenk
equilibrium constant is usually determined as3b
the reagent consumed for solvation of the species accord-
ing to equilibrium 3. In the first approximation, [THF]S
was calculated as
2
[RMgX]
KS,exp
)
(
)
[R2Mg]
[THF]S ) [RMg] + [MgX] + (W - 1)[RMg] )
2W[RMg]
assuming that [R2Mg] ) [MgX2].
Thus, in calculations for various Grignard reagent
concentrations, the true equilibrium constant, KS
S,exp[THF], should be used provided KS,exp is determined
in a dilute solution, and [THF] is the concentration of
free THF in the solution.
slightly underestimating the value. Then provisional
concentrations of RMgX, R2Mg, and MgX2 were calcu-
lated, the corrected value for [THF]S was found, and
ultimately the final values of the species were calculated.
Actually, the corrections done by iteration were relatively
small, ranging a few percent.
Somewhat problematic is the estimation of [THF]F. For
correct calculations, densities of Grignard reagent solu-
tions are necessary. However, we have observed that an
increase in the specific gravity of a Grignard reagent up
to 1 M concentration usually does not exceed 10% of that
of the pure solvent. Tentative calculations indicated that
the probable error is small.14
Principally the same approach was used in the calcula-
tion of concentrations of the species in THF-toluene
binary solvents. Because no appreciable volume effects
were observed by mixing the solvents, the final volume
of the mixture was considered as additive in the compo-
nents.
Calculated concentrations of the species in equilibrium
3 for phenylmagnesium chloride in THF and THF-
toluene mixtures are presented in Tables 1 and 2,
respectively.
)
K
The Schlenk equilibrium constant for phenylmagne-
sium chloride has been determined13 as 1.66 in dilute
THF solutions (0.1-0.15 M). Taking the free THF
concentration approximately equal to the pure THF
concentration (12.3 M) we further used in calculations
KS ) 20 mol/L.
For calculations, experimentally determined concen-
trations of the basic magnesium, [RMg], and of ha-
lide ion, [X-] (or [Mg-X]), are available. The ratio W )
[MgX]/[MgR], incidentally expressing the contribution of
the Wurtz reaction during the preparation of the Grig-
nard reagent, was convenient to use in the derivation of
eqs 5 and 6.
[RMgX] + 2[R2Mg] ) [RMg]
(5)
(6)
[RMgX] + 2[MgX2] ) [MgX] ) W[RMg]
Eventually, from eqs 4, 5, and 6, values of [RMgX],
[R2Mg], and [MgX2] can be calculated for any [RMg] and
W. Equation 4 involves the concentration of free THF in
the solution, [THF]F, which was found as
Tr ea tm en t of Kin etic Da ta . In general, competitive
reactions of PhMgCl and Ph2Mg species with the silane
can be expected:
(11) (a) Tammiku-Taul, J .; Burk, P.; Tuulmets, A. J . Phys. Chem.
A 2004, 108, 133-139. (b) For example, the calculated solvation
enthalpies in kcal mol-1 are for MgCl2‚2THF -43.7, PhMgCl‚2THF
-35.1, and Ph2Mg‚2THF -27.8.
[THF]F ) 12.3 - [THF]S
where [THF]S is the molar amount of THF per liter of
(12) Calculations were performed for chlorides and bromides.11a The
cis-form is
a deformed octahedron where two THF ligands are
(9) (a) Schro¨der, F.; Spandau, H. Naturwissenschaften 1966, 53,
360-362. (b) Perucaud, M.; Ducom, J .; Vallino, M. C. R. Acad. Sci.
1967, 264, 571-573.
(10) Sarma, R.; Ramirez, F.; McKeever, B.; Chaw, Y. F.; Marecek,
J . F.; Nierman, D.; McCaffrey, T. M. J . Am. Chem. Soc. 1977, 99, 5289-
5295.
equatorial and the other two are axial. The trans-form is also an
octahedron where equatorial and axial positions are indistinguishable.
(13) Smith, M. B.; Becker, W. E. Tetrahedron 1967, 23, 4215-4219.
(14) For example, if the density of 0.5 M PhMgCl is by 6% greater
than that of pure THF, the error in [THF]F is +2%, and if it is greater
by 10%, the error is -3%.
J . Org. Chem, Vol. 69, No. 15, 2004 5073

Tuulmets et al.
kobs ) k1[Ph2Mg] + k2[PhMgCl]
From independent experiment, k1 was obtained as 0.26
L mol-1 s-1 (see the section Pure THF). Thus, k2 should
be calculated as
k2 ) (kobs - k1[Ph2Mg])/[PhMgCl]
However, k2 calculated in this way was not a constant,
obtaining values ranging from 0.23 L mol-1 s-1 for 0.2 M
Grignard reagent up to 0.49 L mol-1 s-1 for 0.8 M
Grignard reagent. Moreover, the difference k′ ) kobs
-
F IGURE 4. Plot of the residual rate constant k′′ versus the
product of concentrations [Ph2Mg][MgCl2] for the THF-toluene
system.
k1[Ph2Mg] revealed a trend similar to that of the con-
centrations of Ph2Mg and MgCl2, increasing with an
increase in Grignard reagent concentration in pure THF,
and decreasing with an increase in the cosolvent molar
fraction in THF-toluene mixtures, respectively. Conse-
quently, an additional term was to be included in the
CHART 1
expression for kobs
,
kobs ) k1[Ph2Mg] + k2[PhMgCl] + k3[Ph2Mg][MgCl2]
CHART 2
(7)
In Figure 2, extrapolation of the curve to molar fraction
of toluene 1.0, corresponding to a solution of the complex
PhMgCl‚2THF in toluene, yields kobs ) 0.07 s-1, and
respectively k2 ) 0.14 L mol-1 s-1. The residual rate
constant, k′′ ) kobs - k1[Ph2Mg] - k2[PhMgCl], correlates
with the product [Ph2Mg][MgCl2] as seen in Figure 4,
yielding k3 ) 1.8 L2 mol-2 s-2
.
or Ph2Mg molecules.15 It appears that electrophilic as-
sistance provided by the magnesium center in Chart 1
can be replaced by true electrophilic catalysis as depicted
in Chart 2.
Application of estimated rate constants to the kinetic
data for pure THF led to a good fit between kobs and kcalc
within 4% or better.
Primarily we avoided using a formal statistical treat-
ment of the data, because imperfect orthogonality be-
tween the columns for concentrations in Tables 1 and 2
could bring about irrelevant results. However, correlation
of the rate data with concentrations of the species in pure
THF according to eq 8, where k′ was calculated as k′ )
kobs - k1[Ph2Mg],
Whether the magnesium halide is bound to the orga-
nomagnesium species through a halide bridge forming a
six-center transition state is not yet clear.
In this context, participation of PhMgCl species in the
catalytic pathway cannot be excluded. Higher reactivity
of Ph2Mg and considerable concentrations of symmetrical
species in the solution make the term with k3 in eq 8
significant while the last term probably remains shaded
by experimental uncertainties. Consequently, the general
eq 9 for the kinetics of the Grignard reaction can be
inferred:
k′ ) k2[PhMgCl] + k3[Ph2Mg][MgCl2] +
k4[PhMgCl][MgCl2] (8)
indicated the insignificance of the last term in the
kobs ) k1[R2Mg] + k2[RMgX] + k3[R2Mg][MgX2] +
k4[RMgX][MgX2] (9)
equation and gave for rate constants the values k2
)
0.124 ( 0.003 and k3 ) 2.02 ( 0.02 (R ) 0.999, s )
0.0003). From similar treatment of the data for the THF-
toluene mixtures we obtained k2 ) 0.137 ( 0.013 and k3
) 1.83 ( 0.18 (R ) 0.838, s ) 0.003), the last term in eq
8 being insignificant. Thus, the rate constants from two
sets of experiments were not statistically different.
Mech a n ism of th e Rea ction . It appeared in the
previous section that our kinetic data were well fitted
by eq 7. While participation of species Ph2Mg and
PhMgCl in the reaction could be anticipated, the uni-
gnorable contribution of the last term in eq 7 was
somewhat unexpected on the background of the mecha-
nism suggested for chlorosilanes in our previous paper2e
(Chart 1).
In common ethers, where the Schlenk equilibrium
(equation 1) is shifted far left, only the noncatalytic
pathway through RMgX species is significant provided
the reactivity of diorganomagnesiums does not much
exceed that of RMgX compounds. At least for the reac-
tions with silanes, this seems to be the case.2d
Rea ction w ith a Keton e. Consequences from the
discussion above challenged us to expand the approach
to an entirely different Grignard reaction in THF, viz. to
the reaction between n-propylmagnesium bromide and
pinacolone. The same trend in the pseudo-first-order rate
constants as in Figure 1 was found for that reaction.4 An
explanation, similar to the mechanism suggested for the
However, the Lewis acidities of the species vary largely
ranging in the order MgCl2 > PhMgCl > Ph2Mg.11b
Therefore, complexation of MgCl2 with the chlorine center
of the silane is even more favored than that for PhMgCl
(15) For that reason we preferred the term k3[Ph2Mg][MgCl2] in eq
8, although [Ph2Mg][MgCl2] is roughly proportional to [PhMgCl]2.
5074 J . Org. Chem., Vol. 69, No. 15, 2004

Grignard Reaction with Chlorosilanes in THF
TABLE 3. Com p osition of n -p r op ylm a gn esiu m Br om id e
CHART 3
(m ol L-1) in THF Solu tion sa
n-PrMgb
n-PrMgBr
n-Pr2Mg
MgBr2
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.118
0.178
0.238
0.299
0.361
0.423
0.489
0.041
0.061
0.081
0.100
0.120
0.139
0.155
0.056
0.084
0.111
0.123
0.165
0.191
0.216
TABLE 4. P seu d o-F ir st-Or d er Ra te Con sta n ts, k × 103
s-1, for th e Rea ction s of Ch lor osila n es w ith 0.5 M
P h en ylm a gn esiu m Ha lid es in THF a t 20 °C
a
b
For W ) 1.16. Concentration of the Grignard reagent.
silane
PhMgCl
PhMgBr
MeSiCl3
MePhSiCl2
MeViSiCl2
115
1.28
32
44.7
0.86
23.4
reaction with chlorosilanes in the present work, was
proposed, however, without quantitative verification.4
The Schlenk equilibrium constant for n-propylmagne-
sium bromide in THF is not available. Smith and Becker
determined KS,exp ) 5.09 in 0.1 M ethylmagnesium
bromide in THF13 and Holm estimated KS,exp ≈ 9 in about
2 M THF solution of n-butylmagnesium bromide.16
Recalculation of the latter for a dilute solution gave KS,exp
≈ 6, so KS,exp ) 5.6 was taken for n-propylmagnesium
bromide in THF. Concentrations of the species in equi-
librium 3 were calculated as described above. The results
are presented in Table 3.
tigation. While with methyltrichlorosilane formation of
methylphenyldichlorosilane was investigated, with meth-
ylphenyldichlorosilane and methylvinyldichlorosilane rate
constants for formation reactions of methyldiphenylchlo-
rosilane and methylphenylvinylchlorosilane respectively
were determined. The pseudo-first-order rate constants
in Table 4 provide a comparison of the reactivities of the
chlorosilanes.
From an independent experiment17 the second-order
rate constant for dipropylmagnesium of 0.156 ( 0.019 L
mol-1 s-1 was found. Similarly to the calculations for the
silane reaction a statistical analysis of the rate data from
ref 4 was performed for the concentration range of the
Grignard reagent 0.2 to 0.8 M. Rate constants in eq 9
were found to be k2 ) 0.704 ( 0.017 L mol-1 s-1 and k4
) 2.04 ( 0.09 L2 mol-2 s-2 (R ) 0.999, s ) 0.0003), the
term with k3 being statistically insignificant.
Some differences in comparison with the reaction of
silanes occasionally do not appear. Preponderant reactiv-
ity of n-PrMgBr over dipropylmagnesium, reverse to
phenylmagnesium compounds, can be rationalized re-
garding the steric requirements of the species in relation
to the sterically encumbered ketone, pinacolone. As a
consequence of the reactivities, together with higher
proportion of RMgX species in the reagent (cf. Tables 1
and 3), the last term in eq 9 becomes significant at the
expense of that with k3.
The reaction in THF is much faster than that in diethyl
ether, e.g. the corresponding rate constant for the reac-
tion of PhMgBr with MeSiCl3 in diethyl ether is as small
as 1.62 × 10-3 s-1 2e
.
Also, higher reactivity of phenyl-
magnesium chloride is evident. In chlorosilanes, both
polar and steric effects of substituents seem to be
efficient. Indeed, in Charton terms,18a the phenyl group
and a chlorine substituent exhibit close steric demands;
however, substitution of a chlorine atom for the phenyl
group reduces the reactivity of methylchlorosilanes by
two powers of 10 in accordance with their inductive
effects.18b Similarly, close inductive effects of vinyl and
phenyl groups allow their different steric effects to
become evident, although a contribution from the reso-
nance effect cannot be excluded. A considerable contribu-
tion from the steric effect has been observed also for the
Grignard reaction with alkoxysilanes.2c
Con clu sion s
The most important conclusion ensuing from the
reasonings above is the general validity of the mechanism
expressed by eq 9. It is remarkable that the electrophilic
catalysis by magnesium halides was not considered
before, particularly in the case of carbonyl compounds.
While transition states for the Grignard addition reaction
incorporating two organomagnesium species were dis-
carded for numerous reasons long ago,3,5 it appears that
similar six-member transition states including a mag-
nesium halide molecule (Chart 3) can occur, being
important in the case of Grignard reagents with the
Schlenk equilibrium shifted toward symmetrical species.
Rea ctivity of Ch lor osila n es. As this work was the
first comprehensive kinetic study of the Grignard reac-
tion with silanes in THF, it was of interest to involve
some other chlorosilanes in addition to methyltrichlo-
rosilane exploited as the model compound in the inves-
We have carried out the first comprehensive kinetic
investigation of the Grignard reaction with chlorosilanes
in THF. The reaction in THF is much faster than that in
diethyl ether.
The Schlenk equilibrium is concentration dependent
in THF. Assuming coordination of magnesium halides
with three molecules of THF, concentrations of all the
species involved in the Schlenk equilibrium can be
calculated for any Grignard reagent concentration. In the
Grignard reaction, species R2Mg and RMgX react com-
petitively accompanied by additional reaction paths
involving electrophilic catalysis by magnesium halide.
This conclusion is also valid for the Grignard reaction
with ketones and probably can be expanded to any
Grignard reaction. However, when the Schlenk equilib-
rium is shifted far to the RMgX species, e.g. in diethyl
(16) Holm, T. Acta Chem. Scand. 1966, 20, 2821-2828.
(17) Koppel, J .; Vaiga, S.; Tuulmets, A. Reakts. Sposobn. Org. Soedin
1970, 7, 898-910.
(18) (a) Charton, M. J . Am. Chem. Soc. 1975, 97, 1552-1556. (b)
Taft, R. W. In Steric Effects in Organic Chemistry; Newman, M. S.,
Ed.; J . Wiley: New York, 1956; Chapter 13.
J . Org. Chem, Vol. 69, No. 15, 2004 5075

Tuulmets et al.
equilibrium was set, 0.05 mL of silane was added providing a
(20-40)-fold excess of Grignard reagent, and the temperature
change of the reaction solution (usually up to 1 °C) was
recorded as a plot of temperature versus time. Because the
system was nearly adiabatic, the heat exchange with the
internal part of the calorimeter caused only a little heat loss
to an extent of a few percent.
ether, the other pathways, particularly the catalytic ones,
are insignificant and hardly observable in experiment.
We were able to show that substituents at the silicon
center control the rate of the reaction through their
inductive and steric effects.
First-order rate constants were usually calculated from the
first two or three half-periods of the reaction (30 to 300 s).
Use of a differential method20 for calculation of rate constants
practically eliminated the contribution of heat exchange with
the reaction vessel.
The essence of the method consists of replacement of
differentials by intervals in the equation for a first-order rate
constant
Exp er im en ta l Section
The reagents and solutions were handled under dry argon
and transferred by use of cannulas or syringes. Grignard
reagents were obtained by conventional methods.19
To obtain diphenylmagnesium in THF, phenylmagnesium
bromide was prepared in diethyl ether, magnesium bromide
was precipitated with dioxane, diethyl ether was removed from
the supernatant solution under reduced pressure, and the solid
residue was dissolved in tetrahydrofurane.
Grignard reagents in THF-hydrocarbon mixtures were
prepared by dilution of concentrated stock solutions in THF
with appropriate amounts of THF and/or hydrocarbon (toluene
or dichloromethane).
Kin etic Mea su r em en ts. The reaction was carried out in
a glass vessel mantled with foam plastic and placed in a
thermostated housing. The equipment was sealed with a
thermostated lid. The reaction cell was provided with a
magnetic stirrer and a thermistor, which was connected
through a bridge circuit to a recording potentiometer.
All parts of the equipment and the reagents were thermo-
stated. The reaction vessel was purged thoroughly with pure
argon, 15 mL of the Grignard reagent was cannulated into
the cell, and the stirring was started. After the thermal
d
dT
k ) - ln
dt dt
where T denotes temperature. Thus the rate constant can be
found from the linear plot of ln(∆T/∆t) vs tm, where ∆T ) Ti+1
- Ti, ∆t ) ti+1 - ti, and tm ) ti + (ti+1 - ti)/2.
Ack n ow led gm en t. This work was financially sup-
ported by the Dow Corning Corporation and the Esto-
nian Science Foundation (Grant no. 4630).
J O0498361
(19) Wakefield, B. J . Organomagnesium Methods in Organic Syn-
thesis; Academic Press: New York, 1995.
(20) Rudakov, E. C. Kinet. Katal. 1960, 1, 177-183.
5076 J . Org. Chem., Vol. 69, No. 15, 2004