Analysis of the enantioselectivities and initial rates of the hydrosilylation of acetophenone catalyzed by [Rh(cod)Cl]2/(chiral diphosphine). The quantitative analysis of ligand effects

Article abstract of DOI:10.1016/S0022-328X(02)02221-0

Through the application of the quantitative analysis of ligand effects (QALE) method to the study of the hydrosilylation of acetophenone, we have shown, for the first time, that the initial rate and enantioselectivity of a complicated catalytic system responds in a rational manner to the variations in the stereoelectronic properties of the silane. The reactions (in benzene-d₆ at 63 °C) were catalyzed by [Rh(cod)Cl]₂/(chiral diphosphine) (chiral diphosphine=(R)-BINAP [(R)-(+)-2,2′-bis(diphenylphosphino)- 1,1′binaphthyl], (R,R)-tolyl-BINAP [(R)-(+)-2,2′-bis (di-p-tolylphosphino)-1,1′-binaphthyl], (R,R)-Me-DUPHOS [(R,R )-(-)-1,2-Bis-2,5-dimethylphospholano)benzene], (R,R)-DIOP [(R,R)-(-)-2,3-O-isopropylidine-2,3-dihydroxy-1,4-bis(diphenylphosphino) butane], and (R)-QUINAP [(R)-(+)-1-(2-diphenylphosphino-1-naphthyl) isoquinoline]. The ee's (R) of the hydrosilylation products (CH₃CH(OSiR₃)Ph) range between - 9 and 53% with the (R)-QUINAP giving the poorest enantioselectivity. The QALE analyses of log( R/S) for (R)-BINAP, (R)-tolyl-BINAP, (R,R)-Me-DUPHOS, and (R,R)-DIOP reveal that the steric effects associated with the silanes are not monotonic.

Full text of DOI:10.1016/S0022-328X(02)02221-0

Journal of Organometallic Chemistry 671 (2003) 13ꢃ26
/
www.elsevier.com/locate/jorganchem
Analysis of the enantioselectivities and initial rates of the
hydrosilylation of acetophenone catalyzed by [Rh(cod)Cl]2/(chiral
diphosphine). The quantitative analysis of ligand effects
Clementina Reyes, Alfred Prock *, Warren P. Giering *
Department of Chemistry, Metcalf Science and Engineering Center, Boston University, Boston, MA 02215, USA
Received 27 June 2002; received in revised form 16 December 2002; accepted 20 December 2002
Abstract
Through the application of the quantitative analysis of ligand effects (QALE) method to the study of the hydrosilylation of
acetophenone, we have shown, for the first time, that the initial rate and enantioselectivity of a complicated catalytic system
responds in a rational manner to the variations in the stereoelectronic properties of the silane. The reactions (in benzene-d6 at 63 8C)
were catalyzed by [Rh(cod)Cl]2/(chiral diphosphine) (chiral diphosphineꢀ
1,1?binaphthyl], (R,R)-tolyl-BINAP [(R)-(ꢁ)-2,2?-bis(di-p-tolylphosphino)-1,1?-binaphthyl], (R,R)-Me-DUPHOS [(R,R)-(ꢂ
1,2-Bis-2,5-dimethylphospholano)benzene], (R,R)-DIOP [(R,R)-(ꢂ)-2,3-O-isopropylidine-2,3-dihydroxy-1,4-bis(diphenylphosphi-
no)butane], and (R)-QUINAP [(R)-(ꢁ)-1-(2-diphenylphosphino-1-naphthyl)isoquinoline]. The ee’s (R) of the hydrosilylation
products (CH3CH(OSiR3)Ph) range between ꢂ9 and 53% with the (R)-QUINAP giving the poorest enantioselectivity. The QALE
/
(R)-BINAP [(R)-(ꢁ
/
)-2,2?-bis(diphenylphosphino)-
/
/)-
/
/
/
analyses of log(R/S) for (R)-BINAP, (R)-tolyl-BINAP, (R,R)-Me-DUPHOS, and (R,R)-DIOP reveal that the steric effects
associated with the silanes are not monotonic.
# 2003 Elsevier Science B.V. All rights reserved.
Keywords: Asymmetric catalysis; Ketone; Hydrosilylation; Kinetics; Ligand effects; Quantitative analysis of ligand effects
1. Introduction
that the improvement of asymmetric catalysis appears
‘to be modern alchemy’. Bosnich has also been noted
The demand, utility, and commercial value of single
isomer materials has fueled the explosive growth of
catalytic asymmetric syntheses. A survey of the 2073
chiral ligands and over 13,000 enantioselective reactions
listed in Brunner and Zettlmeier’s Handbook of En-
antioselective Catalysis [1] reveals that most of these
catalytic reactions give only modest stereoselectivity and
only very few give stereoselectivities that approach
100%. In a recent review, Vogl et al. [2] point out that
that ‘It is somewhat disconcerting that after ꢀ20 years
/
of intensive study, success in asymmetric catalysis is
largely dependent on a mixture of luck, intuition, and
perseverance’ [3]. In fact, in order to achieve the most
efficacious asymmetric catalyst, the chemical commu-
nity has focused on synthesizing chiral ligands with the
expectation that luck and intuition will lead to at least
one member of the set that will give high stereoselectiv-
ity for a given reaction. The problem is that variables
such as the charge and oxidation state of the metal,
chiral ligand, solvent, substrate, and temperature as well
as added enhancers all can profoundly influence the
outcome of a catalytic reaction. Obviously, the number
of possible combinations is enormous.
although 90ꢃ95% enantioselectivities have been
/
achieved in the laboratory, this is usually not sufficient
for the pharmaceutical industry where certification
requires enantiomeric purity of greater than 99%.
Hence, the transference of asymmetric catalytic reac-
tions to industry has been slow. They also go on to note
Consider the catalytic hydrosilylation of ketones
(Scheme 1) [4ꢃ10]. The reaction, which is an attractive
/
alternative to high pressure catalytic hydrogenation, is
technically simple, is readily scaled up, and yields
* Corresponding authors.
E-mail address: prock@bu.edu (A. Prock).
0022-328X/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0022-328X(02)02221-0

14
C. Reyes et al. / Journal of Organometallic Chemistry 671 (2003) 13ꢃ
/26
Scheme 1.
directly protected alcohols in the form of silylethers,
which are easily converted to the alcohols. The ee’s for
2. Systematic studies of the hydrosilylation of ketones
We found in the literature only a few quantitative and
systematic studies of the stereoselectivity of hydrosilyla-
tion reactions as a function of either the stereoelectronic
properties of the chiral ligand or other reactants. Most
this reaction range from a few percent to ꢀ
depending on the catalyst and chiral ligand [1].
/
90%
This reaction has been extensively studied in terms of
metals, chiral ligands, silanes, stereoselectivities, che-
moselectivity, solvent, and mechanism. Despite all this
effort, this reaction has not yet lived up to its full
potential. This asymmetric catalytic reaction generally
gives modest enantioselectivities although enationselec-
tivities as high 99% have been reported [1,9]. It is often
accompanied by the undesired extensive dehydrogena-
tive hydrosilylation to form the silylenol ethers. The
tertiary silanes react sluggishly, raising the problem of
catalyst degradation over long reaction times.
Catalytic hydrosilylation does offer the luxury of
having separately tunable components most notably
the chiral ligand, the silane, and additives. These ideas
are illustrated in Scheme 1 (vide supra). Thus, it is
possible that we could tune the stereoselectivity of this
reaction by rationally manipulating these components.
But even if we perform a study where we keep the metal,
chiral ligand, solvent, temperature, ketone and additive
constant and vary only the silane, we are dealing with a
library of silanes that numbers in the thousands. (The
number of possibilities is mind boggling when the other
components are considered as well.) In principle, the
number of combinations can be reduced via the
quantitative analysis of ligand effects method (QALE,
vide infra) where analysis of the data for a handful of
silanes can be used to predict the properties of a large
library of reactants.
of these studies are of the Hammett type [11ꢃ17]. The
/
most extensive study was reported by Kagan et al.;
unfortunately, a quantitative analysis of the stereose-
lectivities is not possible since the data were collected at
different temperatures [16]. Only a handful of reports
describe systematic (but still qualitative) examinations
of the ligands or silanes [8,18ꢃ20]. The only relevant
/
QALE analyses were performed by Marciniec and
coworkers [21,22] who found that the rate of oxida-
tive-addition of silanes to RhCl(cod)PPh3 as well as the
rate of hydrosilylation of 1-hexene could be described in
terms of the QALE model (vide infra). In their study of
the rhodium catalysis of the hydrosilylation of ketones,
Waldman et al. found that the stereoselectivity and rate
were dependent on the nature of added monophosphine
as well as silane [8]. Ito et al. have used derivatized
TRAP ligands in the hydrosilylation of a variety of
ketones [23]. None of these aforementioned studies
report data amenable to quantitative analysis. In related
studies, we have shown that the QALE model can be
used to interpret the dihydroxylation of chiral allyl
silanes [24] and the hydride addition to cationic iron
carbene complexes [25].
Herein, we report the results of our study of the
hydrosilylation of acetophenone (Scheme 2) in which we
sought to answer the question ‘Can the QALE model be
used to interpret the results of a chiral catalytic
reaction.’ The answer, as we will see, is a resounding yes.
As far as we know there have been no reports of
studies showing that enantioselectivities of a catalytic
reaction respond in a quantifiable manner to variations
in the stereoelectronic properties of any of the reactants.
It is the purpose of this study to test if such a
quantification is possible for a system as complicated
as the catalytic hydrosilyation of ketones. The generally
low enantioselectivities of this reaction is actually a
desirable property because, in principle, statistically
significant changes in enantioselectivities can be ob-
served.
3. Experimental
3.1. General considerations
The following compounds were purchased from
Aldrich, United Chemical Technologies, and STREM:
(R)-BINAP
[(R)-(ꢁ)-2,2?-bis(diphenylphosphino)-
/

C. Reyes et al. / Journal of Organometallic Chemistry 671 (2003) 13ꢃ
/26
15
Scheme 2.
1,1?binaphthyl] 99%/HPLC, (R,R)-DIOP [(R,R)-(ꢂ
2,3-O-Isopropylidine-2,3-dihydroxy-1,4-bis(diphenyl-
phosphino)butane] 98%, (R)-tolyl-BINAP [(R)-(ꢁ)-
2,2?-bis(di-p-tolylphosphino)1,1?-binaphthyl] 98%,
(R,R)-Me-DUPHOS [(R,R)-(ꢂ)-1,2-bis((2R,5R)-2,5-
/
)-
in a dry-ice acetone bath (ꢂ
to a vacuum pump via Swagelok connection where it
was subjected to four cycles of freezeꢃpumpꢃthaw
degassing before being flame-sealed. After taking the
initial measurement of the reaction, the tube was placed
in a water bath at 63 8C. The absolute yields of the
products and the amount of residual reactants in the
/
78 8C) and then connected
/
/
/
/
dimethylphospholano)benzene] 99%, (R)-acetyl mande-
lic acid (98%ee/GLC), Bu3SiH, Pr3SiH, EtMe2SiH,
Et2MeSiH, CyMe2SiH, Et3SiH, MePhSiH2, Ph2Me-
SiH, Et2SiH2, Ph2SiH2, Ph3SiH, acetophenone-d3, and
acetophenone. All the silanes and acetophenone were
distilled or recrystallized prior to use and were stored
under an argon atmosphere. The identity of the silanes
was verified by NMR spectroscopy. (p-ClC6H4)3SiH,
(p-CF3C6H4)3SiH, (p-MeC6H4)3SiH, (p-MeOC6H4)3-
SiH, n-Bu2SiH2 and Cy2SiH2 were synthesized using
the procedure described below. Rhodium trichloride was
obtained from Alfa Aesar and from the generous
donation of Professor Pericles Stavropoulos (Boston
University). Benzene-d6 was purchased from Cambridge
Isotopes and was dried using molecular sieves and
stored in a desiccator. [Rh(cod)Cl]2 was prepared
following the procedure of Crabtree and Giordano [26]
1
final reaction mixtures were determined by H-NMR
spectroscopy by comparison of the integrations of the
appropriate resonances to the that of the internal
standard.
3.3. Kinetic studies
Kinetic studies were performed as described elsewhere
[29].
3.4. Hydrolysis of CH3CH(OSiR3)Ph
The reaction was terminated when it was complete or
showed no further progress. No attempt was made to
separate the final products (CH3CH(OSiR3)Ph,
(p-XC6H4)3SiH (Xꢀ
Bu, cyclohexyl) were prepared as described by Benkeser
and Riel [27] and West and Rochow [28], respectively.
/
OMe, CF3) and H2SiR2 (Rꢀ
/
CH3CH(OH)Ph and CH2ꢀ/CH(OSiR3)Ph). The final
reaction mixture was hydrolyzed using a modification
of the procedure reported by Guillard and Morton [30].
The NMR tube was opened and the solution transferred
to a 25-ml round bottom flask. Benzene-d6 was evapo-
rated and 1.23 mmol (three equivalents) of tetrabuty-
lammonium fluoride in 1.5 ml tetrahydrofuran was
added to the residue. The solution was stirred for 24 h
at r.t. before 2.5 ml of saturated aqueous NH4Cl
solution was added. After the addition of the NH4Cl
solution, the solution was stirred for another 3 h. The
organic layer was extracted three times with 10 ml of
ethyl acetate and the combined extracts were dried
overnight over MgSO4. Filtration followed by evapora-
tion of the filtrate gave a yellowish oil, which was then
passed through a plug of silica gel (3 cm high) using a
3.2. Hydrosilylation of acetophenone
To a 10-ml Schlenk flask connected to a vacuum line
were added (R)-BINAP (0.0041 M, 0.0025 g), benzene-
d6 (to make the total volume of the reaction 1 ml),
Bu3SiH (0.82 M, 211 ml), and acetophenone (0.41 M, 48
ml). The mixture was freezeꢃ
/
pumpꢃthaw degassed after
/
each addition using liquid nitrogen. After these reac-
tants were added, the mixture was allowed to stand for 1
h at room temperature (r.t.) under an argon atmosphere.
During that time, [Rh(cod)Cl]2, (0.0041 M in Rh, 1 mg)
was weighed into a separate tube and purged with
argon. Then the mixture was added to [Rh(cod)Cl]2 and
immediately transferred to a previously argon purged
NMR tube containing an internal standard. The stan-
dard was 7 ml of trichloroethylene in a sealed 10 micro-
pipette tube (Drummond Scientific Company). The
NMR tube containing the mixture was promptly placed
90:10 petroleum etherꢃethyl acetate solution. The
/
solvent was evaporated. The NMR spectrum of the
residue taken in benzene-d6 showed only the presence of
acetophenone and 1-phenylethanol.

16
C. Reyes et al. / Journal of Organometallic Chemistry 671 (2003) 13ꢃ26
/
3.5. Determination of the enantiomeric purity of
CH3CH(OH)Ph
4. Results and discussion
We have studied the catalytic hydrosilylation of
acetophenone using the [Rh(cod)Cl]2/(chiral dipho-
sphine) system in benzene-d6 at 63 8C. Both secondary
and tertiary silanes were employed. The rates of the
reactions with secondary silanes were too fast to
measure at 63 8C. The product distributions, initial rates
and ee’s were found to be dependent on both the nature
of the silane and the chiral diphosphine ligand. The data
are presented in Table 1. The (R)-BINAP, (R)-tolyl-
BINAP, (R,R)-Me-DUPHOS and (R,R)-DIOP systems
give generally high yields of CH3CH(OSiR3)Ph accom-
panied by the formation of smaller amount of the
The procedure followed was described by Panek and
Sparks [31]. The mixture of (R)-CH3CH(OH)Ph and
(S)-CH3CH(OH)Ph obtained from the hydrolysis of
CH3CH(OSiR3)Ph was added to 3 ml of freshly distilled
methylene chloride. The mixture was transferred to a 25-
ml round bottom flask containing (R)-acetyl mandelic
acid (1.3 equivalents, 0.1034 g). This flask was placed in
an ice-bath before adding dicyclohexylcarbodiimide
(DCC), (1.206 equivalents, 0.1020 g), and 4-dimethyla-
minopyridine (DMAP) (0.090 mmol, 0.011 g). The
reaction was stirred at r.t. for 48 h. The resulting
mixture of diastereomers was purified by passing the
dehydrogenative hydrosilylation product (CH2ꢀ
/
CH(O-
SiR3)Ph and still smaller amounts of 1-phenylethanol.
The combined 1-phenylethanol formed during the
hydrosilylation reaction and the 1-phenylethanol
formed by hydrolysis of CH3CH(OSiR3)Ph exhibited
mixture through a plug of silica gel (ꢁ
/
3 cm high) using
a 90:10 petroleum etherꢃethylacetate solution. Evapora-
/
tion of the solvent gave the diastereomeric esters with a
100% conversion. The ratio of the diastereomers was
1
determined by H-NMR spectroscopy. The error was
ee’s that spanned a range of ꢂ5 to 50%. (R)-QUINAP
/
on the other hand gave highly variable yields of (CH2ꢀ
/
estimated to be 91%.
/
CH(OSiR3)Ph and CH3CH(OSiR3)Ph, which had ee’s
that were small and showed no particular pattern.
3.6. The QALE model1
We and others [32ꢃ52] have correlated a large body of
4.1. Insights into the mechanism of hydrosilylation via the
QALE model
/
data (over 400 sets of thermodynamic, electrochemical,
kinetic, stereochemical (vide infra), bond length, UV,
IR, NMR, PES, Mossbauer data) with the four stereo-
electronic parameters (xd, u, Ear and pp) via the
following QALE equation and its variants (Eq. (1)).
The most commonly accepted mechanism for the
hydrosilyation of ketones was proposed by Ojima nearly
thirty years ago (Scheme 3) [55]. Kolb and Hetflejs’ [56]
kinetic study of the hydrosilylation of t-butylphenyl
ketone by diphenylsilane using [{Rh(ꢂ
)(diop)}]ꢁClO4ꢂ
/
largely supports this mechanism. Significantly, they
observed a saturation effect as the concentration of the
ketone was increased. In their view, this observation
suggested that prior to the oxidative-addition of the
silane (Scheme 3) there is reversible complexation of the
ketone that deactivates the rhodium. Waldman et al.
also observed that the rate of hydrosilylation of 2,2-
dimethylcyclopentanone with Rh(diphos)(PPh3)2ꢁ (PF6ꢂ
or BF4ꢂ) is first-order in both silane and catalyst but is
propertyꢀaxd ꢁbuꢁb?(uꢂust)lꢁcEar ꢁdpp ꢁe (1)
xd describes the s donor capacity of the ligand; u is
Tolman’s cone angle; ust is the steric threshold, which is
the cone angle where the steric effect changes; l is the
switching function which equals zero when u is less than
ust and equals one when u is equal to or greater than ust;
Ear is a secondary electronic parameter originally
thought to be associated with aryl groups but has now
been found to be more general; and pp is a parameter
that describes the p acidity of the ligand. This general
form of the QALE equation allows for a steric threshold
with a steric effect on the low u side and a different
steric effect on the high u side of the steric threshold. We
and others have demonstrated the transferability of the
phosphorus stereoelectronic parameters to silanes and
silyl groups [21,22,24,53,54].
independent of ketone concentration for [ketone]ꢀ0.5
/
M [8]. They also suggested that the ketone adds to the
rhodium before the oxidative-addition of the silane.
Zheng and Chan studied the rhodium catalyzed hydro-
silylation of a,b-unsaturated carbonyl compounds [57].
They concluded that the Ojima mechanism could not
account for their observation that tertiary silanes give
1,2-addition, whereas, primary and secondary silanes
give 1,4-addition. They also observed no significant
kinetic isotope effect, suggesting that neither the oxida-
tive addition of the silane or the elimination of the silyl
ether from the rhodium complex could be the turnover
limiting step. They went on to suggest that the turnover
limiting step is the coordination of the ketone with a
silylhydridorhodium species.
1
See our QALE web site (www.bu.edu/qale) for a full description
of the QALE model.

C. Reyes et al. / Journal of Organometallic Chemistry 671 (2003) 13ꢃ
/26
17
Table 1
Chiral ligands and silanes used in the hydrosilylation of acetophenone catalyzed by [Rh(cod)Cl]2 in benzene-d6 at 63 8C

18
C. Reyes et al. / Journal of Organometallic Chemistry 671 (2003) 13ꢃ26
/
Table 1 (Continued)

C. Reyes et al. / Journal of Organometallic Chemistry 671 (2003) 13ꢃ
/26
19
Table 1 (Continued)
The yields of CH3CH(OSiR3)Ph, CH2ꢀ
/
C(OSiR3)Ph and CH3CH(OH)Ph (formed before hydrolysis) are given in columns 3ꢃ
/6. The ee’s of the
combined CH3CH(OH)Ph formed during the reaction and by the hydrolysis of CH3CH(OSiR3)Ph are given in the last column.
Recently, we reported the results of an exhaustive
kinetic study of the hydrosilylation of acetophenone
with HSiBu3 catalyzed by (R)-BINAP/[Rh(cod)Cl]2 [29].
The rate of formation of the silylether [PhCH(O-
SiR3)CH3] increased linearly with the concentration of
HSiBu3, whereas the rate of formation shows a satura-
tion effect as the concentration of acetophenone in-
creases in accordance with the observations of Waldman
et al. [8] and Kolb and Hetflejs [56].
this particular model can give a saturation effect without
the need to include a deactivating complexation of the
ketone. All is that is needed is the turnover limiting
oxidative addition of HSiBu3. These simulations and
experimental data are shown in Fig. 1.
We now supplement our kinetic study with the QALE
analysis of the initial rates of the formation of
CH3CH(OSiR3)Ph, Eq. (2). (Data are presented in the
first column of Table 1.) We could not obtain initial
rates for the secondary silanes because of the high
reactivity of these silanes.
Our simulations [29] of the production of PhCH(O-
SiR3)CH3 based on the Ojima mechanism, show that

20
C. Reyes et al. / Journal of Organometallic Chemistry 671 (2003) 13ꢃ
/26
PhC(ꢂO)H3 ꢁHSiZ3
[Rh(cod)Cl]2=R-BINAP
Benzene-d6 (63 ꢀC)
PhCH(OSiZ3)CH3
(2)
It turns out that the results of this analysis also
support the oxidative-addition of the silane to a
rhodium complex as the turnover limiting step. We start
with graphical analysis of the kinetic data (Fig. 2). In the
QALE model, u, Ear and pp (see Eq. (1)) are constant
for HSi(p-XC6H4)3, with the values 145, 2.7 and 0,
respectively. The stereoelectronic parameters of the
silanes are listed in Table 2. Thus, a plot of the data
versus xd for HSi(p-XC6H4)3, affords an estimate
(0.149
/
0.06) of the coefficient of xd obtained by regres-
sion analysis of the full set of data [32]. This estimate
Scheme 3.
Fig. 1. (A) Simulation and experimental results of the time dependence of the concentration of PhCH(OSiBu3)CH3 for a nominally 2:1 acetophenone
to HSiBu3 mixture. (B) Simulation and experimental results of the time dependence of the concentration of PhCH(OSiBu3)CH3 for a nominally 1:2
acetophenone to HSiBu3 mixture. In (A) and (B) the experimental points are shown as filled squares. (C) Experimental initial rates of formation of
PhCH(OSiBu3)CH3 as a function of acetophenone and HSiBu3 concentrations. (D) Simulation of the data shown in (C).

C. Reyes et al. / Journal of Organometallic Chemistry 671 (2003) 13ꢃ
/26
21
u vary for HSiR3; furthermore xd and u are linearly
related. The larger silanes (larger u) are better electron
donors (smaller xd). Since the positive slope of the line
defined by the data for HSiR3 is steeper than the line
defined by P(p-XC6H4)3 we see that u is playing a
significant inhibitory role. An important consequence of
the linear correlation between xd and u for PR3 is that
when there is no Ear effect, the line for the isosteric P(p-
XC6H4)3 ligands with uꢀ
at xdꢀ4.8 corresponding also to uꢀ
hypothetical HSiR3. The lines in Fig. 2 cross at a value
slightly less than xdꢀ4.8 suggesting a small Ear effect.
/
1458 crosses the line for PR3
/
/
1458 for a
/
The full statistical analysis, however, shows that this
effect is not significant. The full regression analysis is
displayed in Table 3, entry 5.
The xd and u profiles are shown in Fig. 3. The profile
for any parameter is obtained by taking the experi-
mental result and subtracting the calculated contribu-
tions of all the parameters except of the one interest. The
resulting difference is then plotted versus the parameter
of interest. For example, the u profile would be
generated in the following manner
Fig. 2. Plots of log(initial rate) vs. xd for HsiR3 and HSi(p-XC6H4)3 in
the [Rh(cod)]2/(R-BINAP) catalyzed hydrosilylation of acetophenone.
must be and is statistically indistinguishable from the
value (0.1390.02) obtained for the coefficient of xd via
/
regression analysis of the total set. The equivalence of
the xd coefficients then serves as a check on the full
analysis. The positive slope indicates that the poorer
electron donor silanes (larger xd) react more rapidly.
Insight into the importance of aryl (Ear) and steric
effects (u) can be obtained by combining plots of the
data for PR3 with the data for P(p-XC6H4)3 versus xd as
is shown in Fig. 2 [39]. In the QALE model, only xd and
log(initial rate)u ꢀ log(initial rate)ꢂaxd ꢂcEar ꢂd (3)
These profiles give a snapshot of how a property
responses to variations in a single parameter. By
construction each profile contains the total error.
This QALE analysis is strikingly similar to the QALE
analysis of Marciniec’s kinetic data for the oxidative-
addition of HSiZ3 to [RhCl(cod)(PPh3)] [22,54].
Table 2
The stereoelectronic parameters of the silanes used in this study
HSiZ3 ꢁ[RhCl(cod)PPh3]
ꢀ[RhHCl(cod)(SiZ3)PPh3]
(4)
(5)
No.
HSiZ3
xd
u
Ear
pp
log kꢀ(0:09890:004)xd ꢂ(0:03890:005)u
ꢁ(0:390:6)
1
2
HSiHEt2
HSiHMePh
HSiHBu2
9.90
117
117
121
122
123
126
127
132
132
134
135
136
136
140
143
144
145
145
145
145
150
160
0.0
1.0
0.0
1.0
0.0
2.0
0.0
0.0
0.0
0.0
0.0
0.0
2.2
1.4
0
1.2
1.2
1.2
0.0
0.0
1.2
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.2
0.0
0.0
0.0
0.0
1.2
0.0
12.90
9.20
10.60
7.80
14.50
7.05
6.30
6.80
5.40
6.20
5.25
12.60
17.00
5.7
3
where nꢀ
/
11, r2ꢀ
0.987 outliers HSi(OEt)3.
/
4
HSiPhMe2
HsiEtMe2
HSiHPh2
5
For direct comparison, we show below the results of
the regression analysis of our data (presented in entry 5
in Table 3) for reaction (2).
6
7
HSiEt2Me
HSiEt3
8
9
HSiMe2(i-Pr)
HSiPr3
10
11
12
13
14
15
16
17
18
19
20
21
22
log(initial rate)
HSiCyMe2
HSiBu3
ꢀ(0:1390:02)xd ꢂ(0:0590:01)uꢁ(592)
(6)
HSiMePh2
HSiMe2(C6F5)
HSi(i-Bu)3
HSiHCy2
where nꢀ 0.923
/
9, r2ꢀ
/
The results of this QALE analysis, like the analysis of
the Marciniec data, are consistent with our early
assertion [29] (based on kinetic data) that the oxida-
tive-addition of the silane to a rhodium complex is the
turnover limiting step in the R-BINAP system. We
cannot reconcile our results with the results of the
deuterium isotope effects. Clearly, this issue is deserving
of further study.
6.60
10.50
13.25
16.8
20.50
5.70
3.45
0.0
2.7
2.7
2.7
2.7
0.0
0
HSi(p-MeOC6H4)3
HSiPh3
HSi(p-ClC6H4)3
HSi(p-CF3C6H4)3
HSiH(t-Bu)2
HSi(i-Pr)3
The values of the parameters are taken from Ref. [54].

22
C. Reyes et al. / Journal of Organometallic Chemistry 671 (2003) 13ꢃ26
/
Fig. 3. Electronic and steric profiles for log(initial rate) of the
[Rh(cod)]2/R-BINAP catalyzed hydrosilylation of acetophenone.
4.2. QALE analysis of the enantioselectivities (log(R/
S)) of the 1-phenylethanol obtained after the hydrolysis
of the reaction mixtures
We analyzed, via the QALE model, log(R/S) for the
1-phenylethanol formed after the hydrolysis of the
reaction mixtures. We made no attempt to separate
the initially formed 1-phenylethanol from the
CH3CH(OSiR3)Ph. We do not believe that this influ-
ences the overall ee’s in a significant way since the yields
of the initially formed 1-phenylethanol were generally
small. The possible origins of 1-phenylethanol are
discussed in reference [29].
To begin our analysis, we plotted log(R/S) versus xd
for HSiR3 and HSi(p-XC6H4)3 for the (R)-BINAP
system (Fig. 4A). The shallow slope of the line defined
by the three HSi(p-XC6H4)3 points relative to the line
defined by HSiR3 indicates that the electron donor

C. Reyes et al. / Journal of Organometallic Chemistry 671 (2003) 13ꢃ
/26
23
tertiary silanes with u ꢀ/136 were unreactive and we
could not further probe the putative steric threshold.
It is noteworthy that the line defined by the tertiary
alkylsilanes is significantly separated from the line
defined by the secondary alkylsilanes. We know from
the analysis of the plots shown in Fig. 4A that xd plays
only a small role in determining log(R/S). If only u and
Ear were involved then a plot of log(R/S) for the
secondary and tertiary alkylsilanes versus u would yield
a single line since Ear is zero for both groups of silanes.
Therefore, we conclude that the presence of the second
SiꢃH bond in the secondary silanes incrementally
/
enhances the stereoselectively of the reaction. The
origins of this incremental effect are obscure. We can
express this incremental effect by the inclusion of a
parameter that equals the number of hydrogens minus
one.
We performed regression analyses, using Eq. (1), for
all the systems except the (R)-QUINAP system. The
stereoelectronic parameters of the silanes are listed in
Table 2. The coefficients and appropriate statistical
information for all the analyses are given in Table 3. For
the (R)-BINAP system (Table 3, entry 1), we determined
that there is, indeed, a steric threshold (ustꢀ1438) with
/
steric effects operative on either side of the threshold.
The small xd effect observed graphically is found to be
statistically insignificant for the full set of data. When
the ranges of the parameters are considered it follows
that the variations in log(R/S) are dominated by
changes in u and in the number of SiꢃH bonds.
/
The steric profile for the (R)-BINAP system is given
in Fig. 5A. This profile is quite striking. It shows a
region below uꢀ
/
1438 where R/S increases as the size of
silane increases. A steric threshold is observed at uꢀ
/
1438 followed by a region where R/S decreases as the
size of the silane increases. There are sufficient data in
this set that we can see the curvature as the system
transforms from one steric domain to the other. This
Fig. 4. (A) Plot of log(R/S) vs. xd for the [Rh(cod)]2/(R)-BINAP
catalyzed hydrosilylation of acetophenone for HSiR3 and HSi(p-
XC6H4)3 (B) Plot of log(R/S) vs. u for HSiR3 for H2SiR2. The dashed
line in (B) is drawn to suggest the existence of a steric threshold near
1438.
type of curvature had been anticipated by Poe and
¨
coworkers [58] for kinetic data. The (R)-tolyl-BINAP
and (R,R)-Me-DUPHOS systems show similar profiles
(Fig. 5B,C) and analyses (Table 3, entries 2 and 3).
The question arises as to the origin of steric profiles
capacity of the silane does not play an important role in
determining R/S. The steep slope of the HSiR3 line,
however indicates a very significant role for steric effects
as is expected for a stereodifferentiating step. The
such as those shown in Fig. 5AꢃC. The QALE model
/
actually allows for regions of different steric effects;
these regions depend on where steric effects turn on in
the different transition states or intermediates [59]. In
the systems considered herein, we are dealing with two
different stereo-differentiating transition states, one that
leads to the R isomer of CH3CH(OSiR3)Ph and one that
leads to the S isomer. In Fig. 6, we present curves that
show how the steric component of log k for the two
stereo-differentiating steps might respond to variations
in the size (u) of the silanes. For very small silanes there
should be no steric discrimination between the two
transition states (region ‘a’ in Fig. 6A and B). If the S
intersection of the two lines is well removed from xdꢀ
/
4.8 (vide supra) thereby indicating a significant Ear
effect that diminishes the log(R/S).
In Fig. 4B, we have plotted log(R/S) versus u for the
secondary and tertiary alkylsilanes. With the exception
of the datum for H2Si(t-Bu)2, the data for the secondary
silanes defines a good line. The deviation of H2Si(t-Bu)2
from this line suggests a steric threshold (i.e. there is a
change in the influence of size on the property). The
data for HSiR3 also define a good line. Unfortunately,

24
C. Reyes et al. / Journal of Organometallic Chemistry 671 (2003) 13ꢃ26
/
Fig. 5. Plot of log(R/S)ua vs. u (steric profiles) for the [Rh(cod)]2 catalyzed hydrosilylation of acetophenone using (A) (R)-BINAP, (B) (R)-tolyl-
BINAP, (C) (R,R)-Me-DUPHOS or (D) (R,R)-DIOP.
transition state is more congested than the R transition
state, then steric effects will turn on sooner in the S
transition state, and the rate of formation of the S
isomer will fall (region ‘b’ in Fig. 6A). Since the rate of
formation of the R isomer has not yet been affected by
the size of the silane, the R/S ratio will increase (region
chiral phosphines. The regression analysis via Eq. (1)
reveals that as the size of the silanes increases, R/S
decreases until a steric threshold is reached at uꢀ1238.
/
The steric profile (Fig. 5D) shows that after the thresh-
old, R/S appears to become independent of u. The
QALE equation (Table 3, entry 4) describes this
behavior by having equal but opposite steric effects
after the steric threshold. This behavior is accounted for
by the QALE model in the following manner. In this
case, after a region of steric effects for the small silanes
(region ‘a’ in Fig. 6C), steric effects turn on first for the
more congested R transition state, which results in the
preferential formation of the S isomer (region ‘b’ in Fig.
‘b’ in Fig. 6B). At uꢀ1438, steric effects become
/
operative in the R transition state and the rate of
formation of the R isomer falls (region ‘c’ in Fig. 6A and
B). Since R/S falls after the steric threshold at uꢀ1438,
/
it appears that the R transition state is less flexible than
the S transition state. (The issue of congestion versus
flexibility has been discussed elsewhere [59,60].) Thus,
6C and D). At uꢀ1238, steric effects turn on for the S
/
we suggest what we are seeing in Fig. 5Aꢃ
/C are the
isomer after which its rate of formation declines (region
‘c’ in Fig. 6C). Since R/S subsequently appears to
become independent of u (region ‘c’ in Fig. 6D), we
regions ‘b’ and ‘c’ in Fig. 6B.
The result of the analysis of the (R,R)-DIOP system is
significantly different from the analyses of the other

C. Reyes et al. / Journal of Organometallic Chemistry 671 (2003) 13ꢃ
/26
25
Fig. 6. Cartoons showing how the positions of the respective steric thresholds for the formation of the R and S isomers of CH3CH(OSiR3)Ph
determine the shape of steric profiles of the enentioselectivity of the reaction shown in Scheme 2. (A) and (B) refer to the (R)-BINAP, (R)-tolyl-
BINAP and (R,R)-Me-DUPHOS systems. (C) and (D) refer to the (R,R)-DIOP system.
envision that the R and S transition states have similar
flexibilities once their respective steric thresholds have
been exceeded.
It is curious and potentially important that in all the
systems we studied, H2SiPhMe was an outlier; its
stereoselectivity was higher than expected. This deviant
behavior might arise because this is the only silane that
we studied where the silicon becomes a stereogenic
center upon oxidative-addition to the rhodium. This
certainly deserves further study.
initial rates of this reaction are similar to the results
obtained from the QALE analysis of the oxidative
addition of silanes to [RhCl(cod)(PPh3)] as reported by
Marciniec. These analyses suggest that the turnover
limiting step is the oxidative addition of the silane to a
rhodium complex. The QALE analysis of the stereo-
chemistry of the reaction reveals that steric effects are
very non-linear and that monotonically increasing the
size of the silane in this asymmetric hydrosilylation of
acetophenone does not monotonically increase the
enantioselectivity of the reaction.
5. Conclusions
We have demonstrated that the QALE model can be
used to analyze the initial rates and enantioselectivities
of a catalytic reaction namely, the hydrosilylation of
acetophenone. The results of the QALE analysis of the
References
[1] H. Brunner, W. Zettlmeier, Handbook of Enantioselective
Catalysis, VCH Publishers, New York, 1993.

26
C. Reyes et al. / Journal of Organometallic Chemistry 671 (2003) 13ꢃ26
/
[2] E.M. Vogl, H. Groger, M. Shibasaki, Angew. Chem. Int. Ed.
Engl. 38 (1999) 1570.
[31] J.S. Panek, M.A. Sparks, Tetrahedron: Asymmetry 1 (1990) 801.
[32] M.R. Wilson, D.C. Woska, A. Prock, W.P. Giering, Organome-
tallics 12 (1993) 1742.
[3] B. Bosnich, Acc. Chem. Res. 31 (1998) 667.
[4] J.D. Morrison, I. Ojima, K. Hirai, Asymmetric Hydrosilylation
and Hydrocarbonylation. Aysmmetric Synthesis, Ch. 4, vol. 5,
Academic Press, Orlando, FL, 1985.
[33] M.R. Wilson, H. Liu, A. Prock, W.P. Giering, Organometallics 12
(1993) 2044.
[34] A.A. Tracey, K. Eriks, A. Prock, W.P. Giering, Organometallics 9
(1990) 1399.
[5] J.L. Speier, Adv. Organomet. Chem. 17 (1979) 407.
[6] B. Marciniec, J. Gulinski, J. Organomet. Chem. 446 (1993) 15.
[7] B.M. Bode, P.N. Day, M.S. Gordon, J. Am. Chem. Soc. 120
(1998) 1552.
[35] J. Panek, A. Prock, K. Eriks, W.P. Giering, Organometallics 9
(1990) 2175.
[36] A.L. Fernandez, C. Reyes, M.R. Wilson, D.C. Woska, A. Prock,
W.P. Giering, Organometallics 16 (1997) 342.
[37] A.L. Fernandez, A. Prock, W.P. Giering, Organometallics 15
(1996) 2784.
[8] T.E. Waldman, G. Schaefer, D.P. Riley, ACS Symp. Ser. 517
(1993) 58.
[9] I. Ojima, Z. Li, J. Zhu, in: Z. Rappoport, Y. Apeloig (Eds.),
Recent Advances in the Hydrosilylation and Related Reactions:
Chemistry of Organic Silicon Compounds, vol. 2;, Wiley, New
York, 1998, p. 1687.
[38] A.L. Fernandez, A. Prock, W.P. Giering, Organometallics 13
(1994) 2767.
[39] J. Bartholomew, A.L. Fernandez, B.A. Lorsbach, M.R. Wilson,
A. Prock, W.P. Giering, Organometallics 15 (1996) 295.
[40] A.J. Poe, R.H.E. Hudson, Organometallics 14 (1995) 3238.
[41] A.J. Poe, D.H. Farrar, Y. Zheng, J. Am. Chem. Soc. 114 (1992)
5146.
[10] I. Ojima, in: S. Patai, Z. Rappoport (Eds.), The Hydrosilylation
Reaction: the Chemistry of Organosilicon Compounds, Wiley,
New York, 1989, p. 1479.
[11] H. Brunner, A. Kurzinger, J. Organometal. Chem. 346 (1988) 413.
[12] T. Ohkuma, H. Ooka, S. Hashiguchi, T. Ikariya, R. Noyori, J.
Am. Chem. Soc. 117 (1995) 2675.
[42] A. Neubrand, A.J. Poe, R. van Eldik, Organometallics 14 (1995)
3249.
[13] T. Hayashi, K. Yamamoto, K. Kasuga, H. Omizu, M. Kumada,
J. Organomet. Chem. 113 (1976) 127.
[43] Farrar, D.H.; Poe, A.J.; Zheng, Y.J. Am. Chem. Soc. 1994, 116,
6252.
[14] Y. Nishibayashi, J.D. Singh, Y. Arikawa, S. Uemura, M. Hidai, J.
Organomet. Chem. 531 (1997) 13.
[44] L.Z. Chen, A.J. Poe, Coord. Chem. Rev. 143 (1995) 265.
[45] N.M.J. Brodie, A.J. Poe, Can. J. Chem. 73 (1995) 1187.
[46] R. Romeo, G. Arena, L.M. Scolaro, Inorg. Chem. 31 (1992) 4879.
[47] X. Luo, G.J. Kubas, C.J. Burns, J.C. Bryan, C.J. Unkefer, J. Am.
Chem. Soc. 117 (1995) 1159.
[15] R. Noyori, S. Hashiguchi, Acc. Chem. Res. 30 (1997) 97.
[16] J.F. Peyronel, J.C. Fiaud, H.B. Kagan, J. Chem. Res. M (1980)
4057.
[17] M.B. Carter, B. Sechiott, A. Gutierrez, S.L. Buchwald, J. Am.
Chem. Soc. 116 (1994) 11667.
[48] C.A. Bessel, J.A. Margarucci, J.H. Acquaye, R.S. Rubino, J.
Crandall, A.J. Jircitano, K.J. Takeuchi, Inorg. Chem. 32 (1993)
5779.
[18] I. Matsuda, Y. Fukuta, T. Tsuchihashi, H. Nagashima, K. Itoh,
Organometallics 16 (1997) 4327.
[49] C. Moreno, M.J. Macazaga, S. Delgado, Organometallics 10
(1991) 1124.
[19] B.T. Gregg, A.R. Cutler, J. Am. Chem. Soc. 118 (1996) 10069.
[20] J. Halpern, P.F. Phelan, J. Am. Chem. Soc. 94 (1972) 1881.
[21] W. Duczmal, B. Maciejewska, E. Sliwinska, B. Marciniec,
Transition Met. Chem. 20 (1995) 162.
[50] R.S. Herrick, R.R. Duff, Jr, A.B. Frederick, J. Coord. Chem. 32
(1994) 103.
[51] D.H. Farrar, J. Hao, A.J. Poe, T.A. Stromnova, Orgnaometallics
16 (1997) 2827.
[22] W.S. Duczmal, B. Maciejewska, B. Marciniec, H. Maciejewski,
Transition Met. Chem. 20 (1995) 435.
[52] R. Romeo, G. Alibrandi, Inorg. Chem. 36 (1997) 4822.
[53] D.M. Hester, J. Sun, A.W. Harper, G.K. Yang, J. Am. Chem.
Soc. 114 (1992) 5234.
[23] R. Kuwano, M. Sawamura, J. Shirai, M. Takahashi, Y. Ito, Bull.
Chem. Soc. Jpn. 73 (2000) 485.
[24] B.A. Lorsbach, A. Prock, W.P. Giering, Organometallics 14
(1995) 1694.
[54] www.bu.edu/qale
[55] I. Ojima, T. Kogure, M. Kumagai, S. Horiuchi, Y. Sato, J.
Organometal. Chem. 122 (1976) 83.
[25] B.A. Lorsbach, D.M. Bennett, A. Prock, W.P. Giering, Organo-
metallics 14 (1995) 869.
[56] I. Kolb, J. Hetflejs, Coll. Czech. Chem. Commun. 45 (1980) 2808.
[57] G.Z. Zheng, T.H. Chan, Organometallics 14 (1995) 70.
[58] K. Dahlinger, F. Falcone, A.J. Poe, Inorg. Chem. 25 (1986) 2654.
[59] E. Eriks, H.-Y. Liu, A. Prock, W.P. Giering, Inorg. Chem. 28
(1989) 1759.
[26] R.H. Crabtree, G. Giordano, Inorg. Syn. 19 (1976) 218.
[27] R.A. Benkeser, J.F. Riel, J. Am. Chem. Soc. 73 (1951) 3472.
[28] R. West, E.G. Rochow, J. Org. Chem. 18 (1953) 303.
[29] C. Reyes, A. Prock, W.P. Giering, Organometallics 21 (2002) 546.
[30] J.W. Guillard, R. Fortin, H.E. Morton, C. Yoajim, C.A.
Quesnelle, S. Daignault, Y. Guindon, J. Org. Chem. 53 (1988)
2606.
[60] M.N. Golovin, M.M. Rahman, J.E. Belmonte, W.P. Giering,
Organometallics 4 (1985) 1981.