A kinetic treatment of heterogeneous enantioselective catalysis

Article abstract of DOI:10.1039/a607661k

Kinetic data are reported for the liquid phase racemic hydrogenation and enantioselective hydrogenation of methyl acetoacetate (MAA) over Ni/SiO₂ and tartaric acid (TA)-treated Ni/SiO₂, respectively, conducted in a slurry-type reactor. The contribution of mass transport to the overall reaction rate was considered by studying the effects of varying a number of process variables, such as catalyst particle size, catalyst weight, stirring speed and reaction temperature. Reaction conditions are identified wherein contributions due to mass transfer are slight and the hydrogenation proceeds under surface kinetic control. Reaction orders with respect to MAA, hydrogen and the product, methyl 3-hydroxybutyrate (MHB), were determined and activation energies over the temperature range 318 ≤ T ≤ 383 for the bare and TA-modified supported nickel are presented. The reaction rates were observed to pass through maxima as the MAA concentration was increased. The rate data are fitted to a Langmuir-Hinshelwood model where both reactants adsorb competitively on the surface; the experimentally determined rates agree to ±15% with the values predicted by the kinetic model. Modification of the catalyst with TA not only induced enantioselectivity but also promoted the rate of hydrogenation and this effect is discussed in terms of surface interactions and the proposed kinetic model. The ratio of MAA concentration in bulk solution to that at the catalyst surface under reaction conditions is examined and the dependence of enantioselectivity, in the case of the TA-treated sample, on MAA concentration is presented.

Full text of DOI:10.1039/a607661k

View Article Online / Journal Homepage / Table of Contents for this issue
A kinetic treatment of heterogeneous enantioselective catalysis
Mark A. Keane¤
Department of Chemical Engineering, University of L eeds, L eeds, UK L S2 9JT
Kinetic data are reported for the liquid phase racemic hydrogenation and enantioselective hydrogenation of methyl acetoacetate
(MAA) over Ni/SiO and tartaric acid (TA)-treated Ni/SiO , respectively, conducted in a slurry-type reactor. The contribution of
2
2
mass transport to the overall reaction rate was considered by studying the e†ects of varying a number of process variables, such
as catalyst particle size, catalyst weight, stirring speed and reaction temperature. Reaction conditions are identiÐed wherein
contributions due to mass transfer are slight and the hydrogenation proceeds under surface kinetic control. Reaction orders with
respect to MAA, hydrogen and the product, methyl 3-hydroxybutyrate (MHB), were determined and activation energies over the
temperature range 318 O T O 383 for the bare and TA-modiÐed supported nickel are presented. The reaction rates were observed
to pass through maxima as the MAA concentration was increased. The rate data are Ðtted to a LangmuirÈHinshelwood model
where both reactants adsorb competitively on the surface; the experimentally determined rates agree to ^15% with the values
predicted by the kinetic model. ModiÐcation of the catalyst with TA not only induced enantioselectivity but also promoted the
rate of hydrogenation and this e†ect is discussed in terms of surface interactions and the proposed kinetic model. The ratio of
MAA concentration in bulk solution to that at the catalyst surface under reaction conditions is examined and the dependence of
enantioselectivity, in the case of the TA-treated sample, on MAA concentration is presented.
The surge in the worldwide market for single-enantiomer
forms of chiral drugs has stimulated a Ñurry of activity as
drug companies, Ðne-chemical/chirotechnology Ðrms and aca-
demic research groups explore new ways to make single-
isomer compounds.1 Biocatalysis employing enzymes has
emerged as a possible route for chiral molecule synthesis2h4
but the use of enzymes in environments di†ering signiÐcantly
from physiological conditions, as is the case in industrial pro-
cesses, has limited the commercial applications.5,6 The via-
bility of heterogeneous enantioselective catalysis, employing
inorganic materials, has not yet been fully explored and
research7h9 has largely focused on two systems, namely the
asymmetric hydrogenation of ethyl pyruvate and methyl ace-
toacetate (MAA) using cinchonidine-modiÐed platinum and
tartaric acid (TA)-modiÐed nickel catalysts, respectively. The
latter reaction involves an asymmetric addition of hydrogen
to generate methyl 3-hydroxybutyrate (MHB), the (R)-([)-
form of which is an important intermediate in the synthesis of
the carbonic anhydrase inhibitor MK-0507 used in the treat-
ment of glaucoma.10 Conventional nickel catalysts possess no
inherent chirality and the bare nickel metal generates a
racemic product.11h15 A chiral environment is achieved by
treating the activated catalysts with an optically pure reagent
or catalyst modiÐer.14h19 Such a modiÐcation has been shown
not only to induce enantioselectivity but also to increase the
overall reaction rate generated by the nickel17,18 and
platinum20h22 systems. Variations in reaction conditions have
been related to the ultimate enantioselectivity and the e†ects
of such changes on the reaction rate have, to a great extent,
remained unreported. The concentration of MHB produced in
a batch liquid phase system has, however, been shown to
increase with an increase in reaction time7,23 and
temperature7,10 and to inhibit the progress of the reaction at
high concentrations.24h26 The e†ect of increasing the sub-
strate concentration, the presence of metal poisons and the
role of the reaction solvent have also been considered.26h31
Reactions over a range of supported and unsupported nickel
catalysts at temperatures in the range 313 O T /K O 423 have
been reported to yield a zero or near zero dependence on
hydrogen concentration,32,33 a Ðrst order32 or partial33 (0.3)
dependence on MAA concentration and apparent activation
energies12,32,33 in the range 44È66 kJ mol~1. Although the
hydrogenation of MAA and ethyl acetoacetate (EAA) have
now been reported for
systems,7h9,14,34,35 and
a
range of unsupported nickel
nickel supported on
amorphous14h19,33,34,36h38 and crystalline11h13 carriers, the
reaction mechanism is still not fully understood. In this paper
a kinetic model is proposed that accounts for the observed
rate dependence on MAA concentration in the absence of
strong mass transport limitations for both an unmodiÐed and
TA-modiÐed Ni/SiO catalyst.
2
Experimental
An 11.6 wt.% Ni/SiO catalyst was prepared by the impregna-
2
tion of a non-porous microspheroidal Cab-O-Sil 5 M silica
with nickel nitrate to incipient wetness as described in detail
previously.39 The nickel loading was determined by atomic
absorption (Perkin-Elmer 360 AA spectrophotometer) using
the experimental procedure outlined elsewhere15,17 and the
water content of the catalyst precursor was measured by
thermogravimetry.40 The hydrated supported precursor was
reduced, without a precalcination step, by heating in a 150
cm3 min~1 stream of dry hydrogen at a Ðxed rate of 5 K
min~1 to a Ðnal temperature of 723 ^ 2 K which was main-
tained for 18 h. The temperature-programmed reduction
proÐle was obtained as discussed previously.17 The reduced
catalyst was then Ñushed in a puriÐed stream of nitrogen at
200 cm3 min~1, cooled to room temperature and used directly
in the “unmodiÐedÏ state to promote the hydrogenation of
MAA, adopting the procedure outlined below. Alternatively,
the cooled, activated sample was “modiÐedÏ by contacting
with a 0.01 mol dm~3 aqueous solution of R-(])-TA (TA/Ni
molar ratio \ 0.8). The pH of the aqueous modiÐer was pre-
adjusted to the isoelectric point for TA (pH \ 5.1) with 1 and
0.01 mol dm~3 NaOH and the resultant solution was thor-
oughly purged with helium to remove any entrapped air
bubbles. The solid catalyst/aqueous modiÐer suspension was
kept under constant agitation at 600 rev min~1 (Stuart Scien-
tiÐc SM26 Ceramic Magnetic Digital Stirrer/Hotplate) for 2 h
¤ E-mail address: chemaak=leeds.ac.uk
J. Chem. Soc., Faraday T rans., 1997, 93(10), 2001È2007
2001

View Article Online
at 298 ^ 1 K. The tartaric acid was of AnalaR grade and the
prepared solutions were found to be better than 99% pure by
HPLC. After modiÐcation, the catalyst was decanted and
washed with deionized water (1 ] 25 cm3), methanol (2 ] 25
cm3) and n-butanol (2 ] 25 cm3); the reagents used in the
washing step were thoroughly degassed in helium. Both the
modiÐed and unmodiÐed samples were stored in n-butanol
prior to use. The nickel content of the post-modiÐer solution,
resulting from acid leaching of the catalyst, was determined by
AA spectrophotometry.15,17 Nickel metal dispersions, before
and after modiÐcation, were determined by carbon monoxide
chemisorption15,17 at 273 K, assuming a nickel surface atom
to adsorbed carbon monoxide stoichiometry of 2 : 1.41 The
amount of TA adsorbed was determined by analysing the pre-
and post-modiÐcation solutions by HPLC (Spectra-Physics)
as outlined elsewhere.15,17,19
The liquid phase hydrogenation of MAA in n-butanol was
conducted over the temperature range 318 O T /K O 383
(^2). The apparatus and procedure have been described in
some detail previously.15,17,18 The MAA/n-butanol molar
ratio was varied (in the range 4 ] 10~3È2.0) but the total
volume was maintained at 50 cm3 and a 60 cm3 min~1 stream
of dry hydrogen was bubbled through the suspension. The
possible contribution of mass transport to the overall reaction
rate was monitored by (a) varying the weight of catalyst to
give MAA/Ni molar ratios in the range 20È400, (b) using six
separate batches of catalysts with particle diameters in the
overall range 50È1680 lm and (c) varying the stirring speed in
the range 100È1000 rev min~1. The extent of hydrogenation
was determined by HPLC using a Pirkle type 1A 5 l
reversible column (250 ] 4.6 mm) with a 10% IPAÈ90%
hexane mixture as the mobile phase. Optical yields (O.Y.) were
determined from measurements of optical rotation (AA-10
Automatic Digital Polarimeter) using the equation,
Results and Discussion
Catalyst characterization
The results reported in this paper refer to the treatment of an
11.6 wt.% Ni/SiO catalyst with R-(])-TA which yielded
2
R-([)-MHB as the predominant product. Replacement of the
modiÐer by its antipode has been shown15,17,19 to generate
identical values of enantiomeric excess but with a reversal of
the sign while modiÐcation with optically inactive meso-TA
yielded virtually racemic (e.e. \ 2%) product mixtures. The
pertinent physical characteristics of the catalyst precursor, the
activated unmodiÐed and activated modiÐed material are
given in Table 1. The temperature-programmed reduction
proÐle17 is characterized by four peaks: a small ill-deÐned
peak at T \ 528 K which has been ascribed42 to an endo-
max
thermic phase transition from rhombohedral to cubic NiO
which occurs simultaneously with reduction; a minor (573 K)
and a major (610 K) peak due to the reduction of weakly and
strongly support-interacting NiO to Ni0 ] H O; a broad
2
peak with T \ 693 K due to the more difficult reduction of
max
a surface nickel silicate species. The total hydrogen consump-
tion in the activation step exceeded the theoretical value for a
complete reduction of the nickel content and the excess may
be attributed to a spillover of hydrogen from the nickel metal
which is held on the silica carrier. The nickel metal dispersion
of the unmodiÐed reduced catalyst, expressed as
(Ni
/Ni ) ] 100%, equals 17%. The supported nickel
surface total
particle size is estimated to equal 5.9 nm from the
relationship43 d \ 101/D, where D is dispersion and d rep-
resents the surface-weighted-average crystalline diameter
assuming (spherical) particles O1 nm to be 100% dispersed.
Transition electron microscopic analysis has revealed a broad
particle size distribution with supported particles of up to 20
nm in diameter detected.17 The dispersion or dimension of the
supported nickel crystallites in the unmodiÐed catalyst
remained unchanged after repeated catalysis cycles. It has
been demonstrated elsewhere14,15,17h19,34 that the generation
of enantioselective sites is a corrosive process and results in
the extraction of nickel atoms from the metal lattice. Under
the stated modiÐcation conditions 29% of the initial nickel
content was leached into solution. In TA treatment the
smaller nickel particles were preferentially removed.15,17 with
a consequent shift in the average supported crystallite diam-
eter to the higher value of 8.4 nm. It is evident that the com-
position of the modiÐed catalyst di†ers appreciably from that
of the freshly reduced catalyst. In this kinetic study, the reac-
tion rates are quoted per gram of supported nickel which
takes into account the loss of nickel during the modiÐcation
[a]T
100a
D
O.Y. \
\
(1)
[a]T [a]T lc
0
0
where [a]T is the speciÐc rotation of the product solution
D
measured at the sodium D-line and temperature T \ 293 ^ 3
K, [a]T is the speciÐc rotation of the pure enantiomer under
0
the same conditions [[22.95¡ for R-([)-MHB], a is the mea-
sured optical rotation, l is the path length (20 cm) of the cell
and c is the solute concentration. In this paper, optical selec-
tivity is expressed in terms of enantiomeric excess (e.e.) or
percent optical purity which, in the case of an excess of R-([)-
MHB, is given by,
([R-([)-MHB] [ [S-(])-MHB])
([R-([)-MHB] ] [S-(])-MHB])
%e.e. \ 100
(2)
In many cases the reaction was repeated under identical con-
ditions and all the rate and enantioselectivity data are plotted
to illustrate reproducibility.
step. The treated Ni/SiO sample retained 1.3 ] 1019 TA mol-
2
ecules per gram of catalyst which represents a fractional
coverage (h ) of 0.32 of the available surface nickel metal.
TA
Table 1 Physical characteristics of the catalyst precursor and the activated unmodiÐed and TA-modiÐed material
precursor
support
Ni (wt.%)
Cab-O-Sil 5M SiO (194 m2 g~1)
2
11.6
H O (wt.%)
8.0
2
activated unmodiÐed catalyst
T
a/K
528, 573, 610, 693
red
H
consumption/10~2 dm3 g~1
Ni0 dispersion (%)
db/nm
5.7
17
5.9
2
activated TA-modiÐed catalyst
Ni leached (%)
Ni0 dispersion (%)
db/nm
29
12
8.4
N
c
1.3 ] 1019
TA
h
d
0.32
TA
a Temperature corresponding to a maximum in hydrogen consumption. b Average nickel particle diameter. c Number of TA molecules retained
per gram of catalyst. d Fractional surface coverage by TA.
2002
J. Chem. Soc., Faraday T rans., 1997, V ol. 93

View Article Online
Mass transport limitations
The observed inÑuence of catalyst particle diameter on reac-
tion rate is shown in the inset to Fig. 1. In the size range
150È1680 lm the rate is proportional to particle diameter and
mass transport limitations inhibit the progress of the reaction.
As the particle size was decreased from 150 to 50 lm the rate
was largely una†ected which is indicative of a reaction oper-
ating within the kinetic control regime. Catalytic rate is
directly proportional to the weight of catalyst for reactions
controlled by the surface kinetics.44,48,50 Such a relationship,
using the criteria set down by Roberts,48 is illustrated in Fig. 2
where the variation in catalyst weight spans the MAA/Ni
molar ratio range 20È400. It may be noted that the rates of
hydrogenation are appreciably higher for the modiÐed
samples, an observation that is in direct conÑict with earlier
reports24,52,53 where the adsorption of TA on Raney nickel
was viewed as inhibiting hydrogenation activity by occluding
part of the metal surface. In recent studies, rate increases due
to modiÐcation have, however, been reported for a number of
asymmetric systems17,18,20h22 and a possible source of such
an enhancement is proposed in the kinetic model discussed
below.
The rate of a number of liquid phase reactions has been
shown to be controlled by the transport of reactants and pro-
ducts in the bulk solution to and from the solid catalyst
surface.44,45 Indeed, a number of workers29,33,37 have noted
that MAA hydrogenation is limited, to varying degrees, by
mass transport constraints and the observed reaction rate is a
combination of reactant transport and catalytic activity. As
part of a thorough kinetic study it is essential to ascertain the
extent to which such physical processes control the rate of the
overall reaction and to establish the reaction conditions
wherein such rate limitations are negligible and the reactor is
operating in the “reaction-limitedÏ regime. In this particular
reactor conÐguration, hydrogen is bubbled through the MAA/
n-butanol solution which contains suspended solid catalyst
particles and the gaseous reactant must overcome the follow-
ing resistances which are viewed as acting in series: (i) H
2
must cross the gas Ðlm to reach the gas/liquid interface; (ii) H
2
must cross the liquid Ðlm to reach the main body of liquid;
(iii) H must cross the liquid Ðlm surrounding the catalyst par-
2
ticle to reach the solid surface; (iv) H then reacts with MAA
2
on the surface of the catalyst particle. Since pure hydrogen is
Kinetic model
used the gas phase resistance can be ignored but the relative
resistance at the gas/liquid interface or at the surface of the
catalyst particles must be determined.
The hydrogenation of MAA to MHB may be represented by
the power equation
In this study three established diagnostic criteria,44h48 stir-
ring speed, catalyst particle size and catalyst weight, are used
to gauge the contribution of mass transport resistance. A
more e†ective agitation of the solution around the solid cata-
lyst does not a†ect chemically controlled or surface-controlled
processes but markedly increases di†usion rates in the bulk
liquid.44,46 The inÑuence of stirring speed on the observed
reaction rate is illustrated in Fig. 1. The hydrogenation rate
clearly responds to increased agitation of the MAA/n-butanol/
catalyst suspension which serves to extend the gas/liquid/solid
interface and minimize the distance that a hydrogen molecule
has to di†use through the liquid phase in order to reach the
catalyst surface. The rate is, however, essentially independent
of stirring speed at values in excess of 600 rev min~1. Singh et
al.49 have examined the e†ect of bulk di†usion limitations in
the enantioselective hydrogenation of ethyl pyruvate over
dihydrocinchonidine-modiÐed platinum and likewise observed
a marked increase in rate on raising the agitation speed from
200 to 600 rev min~1 but little dependence at higher stirring
speeds. In a catalytic reaction where external mass transfer
resistance is high, the reaction rate has been shown to be
dependent on the catalyst particle shape and size.44,47,48,50,51
R \ k[H ]a[MAA]b[MHB]c
(3)
2
where R represents the reaction rate, k the rate constant and
a, b and c the orders of the reaction with respect to the con-
centration of H , MAA and MHB, respectively. The inclusion
of increasing MHB concentrations (9.3 ] 10~3È4.5 ] 10~2
mol dm~3) in the reactant mixture was observed to have a
negligible e†ect on the overall rate and the reaction can be
considered to be zero order with respect to [MHB]. A small
dependence (\0.1) of rate on the partial pressure of hydrogen
(in the range 0.3È1 atm) was observed at 343 K which com-
pares with the zero32 and 0.2 order54 dependences that have
been quoted elsewhere. The slopes of the logarithmic plots
that relate reaction rate to MAA concentration are shown in
Fig. 3 and yield a reaction order with respect to [MAA] of 0.4
for both the modiÐed and the unmodiÐed system. Taking the
2
reaction to be zero order in terms of [H ] and [MHB] and
2
assuming that the observed partial order with respect to
[MAA] holds over the temperature range 318 O T /K O 383,
the rate constant was calculated from the experimentally
determined rates and the resultant Arrhenius relationships for
the unmodiÐed and modiÐed samples are illustrated in Fig. 4.
The apparent rate was increased in both cases by a factor of
Fig. 1 The e†ect of stirring speed on the rate of MAA hydrogenation
over the TA-modiÐed (>) and unmodiÐed (=) Ni/SiO ; T \ 343 K;
Fig. 2 The inÑuence of TA-modiÐed (>) and unmodiÐed (=)
2
particle diameter \ 75È125 lm; MAA/Ni molar ratio \ 47È66. Inset:
Ni/SiO catalyst weight on the hydrogenation rate; T \ 343 K;
2
the variation of hydrogenation rate with the unmodiÐed catalyst par-
particle diameter \ 75È125 lm; stirring speed \ 800 rev min~1;
ticle size; stirring speed \ 800 rev min~1.
MAA/Ni molar ratio \ 20È400
J. Chem. Soc., Faraday T rans., 1997, V ol. 93
2003

View Article Online
larger than the 10È16 kJ mol~1 activation barrier that is char-
acteristic of purely di†usion-controlled processes44,46,48,52
and lie within the range of values provided elsewhere.12,32,33
Hydrogenation of MAA over the TA-modiÐed catalyst there-
fore proceeds at a higher rate and with a lower associated
energy barrier.
The relationship between reaction rate and MAA concen-
tration is shown in Fig. 5. It can be seen that the rate passes
through a maximum (in the range 0.1È0.2 mol dm~3 MAA)
for both the modiÐed and unmodiÐed catalysts as [MAA] is
increased which is typical of a Langmuir dependence of rate
on concentration. According to the LangmuirÈHinshelwood
model, MAA and hydrogen react in the adsorbed state on the
catalyst surface and the associated rate expression assumes
di†erent forms depending on whether the adsorption is com-
petitive or not.46,47 The surface conditions which prevail
during hydrogenation were probed by considering the follow-
ing three pretreatments: (a) pre-contacting the activated
catalyst/n-butanol suspension with a 60 cm3 min~1 purge of
Fig. 3 Rate of MAA hydrogenation as a function of MAA concen-
tration over TA-modiÐed (>) and unmodiÐed (=) Ni/SiO ; T \ 343
2
K; particle diameter \ 75È125 lm; stirring speed \ 800 rev min~1
hydrogen for
1 h prior to the introduction of MAA
(9.3 ] 10~2 mol dm~3) where the activated catalyst had been
swept with helium (400 cm3 min~1) for 24 h at 723 K; (b)
pre-contacting the helium-Ñushed activated catalyst with a
MAA/n-butanol mixture (mole fraction of MAA \ 0.02) for 1
h prior to the introduction of hydrogen; (c) simultaneous
introduction of MAA (9.3 ] 10~2 mol dm~3) and hydrogen
to the helium-Ñushed activated catalyst/n-butanol suspension
without any pretreatment. Pre-contacting the catalyst with
hydrogen or MAA resulted in a lower reaction rate (by a
factor of up to 1.2) than that observed for the simultaneous
addition of both reactants. Such an observation suggests that
there is competition for adsorption sites between MAA and
hydrogen which suppresses the reaction rate when one of the
reactants is pre-adsorbed on the surface. Taking the adsorp-
tion of hydrogen, MAA and n-butanol to be reversible and
up to 75 in going from 318 to 383 K, which is certainly diag-
nostic of a surface reaction proceeding with minimal mass
transport constraints. The author could Ðnd no available
experimental data on di†usion of MAA in n-butanol but the
di†usion coefficient (cm3 s~1) may be calculated using the
WilkeÈChang estimation which is represented by55
7.4 ] 10~8(/M )0.5T
B
D
\
(4)
AB
k V 0.6
B
A
where A corresponds to the solute (MAA), B the solvent (n-
butanol), / an association factor, M the solvent molecular
B
mass (g mol~1), T the absolute temperature (K), k the solvent
B
viscosity (cp) and V the critical molar volume of solute A.
A
The temperature dependence of D reduces to a constant
AB
competitive, ratios of the fractional coverages of MAA (h
)
multiplied by the ratio of reaction temperature to n-butanol
MAA
and n-butanol (h
) to that of hydrogen (hH2) are given by46
viscosity. Solvent viscosities at 318 K and 383 K were esti-
mated using the temperature related constants provided by
Yaws et al.56 and the di†usion coefficient is calculated to be
four times greater at the higher temperature. Gerrard57 has
reported that the solubility of hydrogen in alcohols is also
raised with increasing temperature. Nevertheless, the observed
hydrogenation rate increase is considerably greater than the
relative improvement in the transport of reactants to the cata-
lyst surface and can be attributed predominantly to kinetic
phenomena. The Arrhenius relationships yield apparent acti-
vation energy values (with 95% conÐdence limits) of 59 ^ 4 kJ
mol~1 and 49 ^ 5 kJ mol~1 for the unmodiÐed and modiÐed
nvbut
h
h
A
[MAA]B A[d*G
B
MAA
MAA
\
exp
(5)
(6)
[H ]
RT
H2
2
h
h
A
[n-but]B A[d*G
B
nvbut
nvbut
\
exp
[H ]
RT
H2
2
where d*G
(or d*G
) is the di†erence of the Gibbs ener-
MAA
nvbut
gies of adsorption between MAA (or n-butanol) and hydrogen.
The reaction was conducted at a constant pressure of hydro-
gen and assuming that the solubility of hydrogen is indepen-
systems, respectively. These values of E
are considerably
app
Fig. 5 Plot of MAA hydrogenation rate vs. [MAA] for the TA-
modiÐed (>) and unmodiÐed (=) Ni/SiO ; T \ 343 K; particle
2
Fig. 4 Arrhenius plots for the hydrogenation of MAA over TA-
diameter \ 75È125 lm; stirring speed \ 800 rev min~1. Note: the
modiÐed (>) and unmodiÐed (=) Ni/SiO ; particle diameter \ 75È
solid line represents the rate expression R \ 2.85[MAA]/
(1 ] 8.39[MAA])2 and the dashed line the expression
R \ 1.47[MAA]/(1 ] 7.67[MAA])2.
2
125 lm; stirring speed \ 800 rev min~1; Ni/MAA molar
ratio \ 47È66
2004
J. Chem. Soc., Faraday T rans., 1997, V ol. 93

View Article Online
dent of the mole fractions of MAA and n-butanol the liquid
phase concentration of hydrogen can be considered constant
and
mined rates agree to ^15% with the values predicted by the
kinetic model and both experimental and calculated rates
exhibit maxima in the [MAA] range 0.1È0.2 mol dm~3.
There is strong spectroscopic evidence2 to suggest that TA
is adsorbed on the nickel metal surface as a carboxylate and
the active centres in the modiÐed system may then be con-
sidered to be composed of both zero-valent nickel metal which
serves to dissociate the adsorbed hydrogen and the nickel ions
in the surface nickel/modiÐer “complexÏ where the latter
environment is essential to promote enantioselectivity. The
MAA reactant is adsorbed on a bare nickel site as an O-
bonded chelate58 but on the modiÐed surface MAA can also
interact with adsorbed TA via hydrogen bonding between the
OH group of the hydroxy acid and the methoxy oxygen atom
of MAA. In the batch liquid phase reaction, the free reactants
as well as the site attached reactants are assumed to be in
h
h
MAA
H2
\ c
[MAA]
(7)
MAA
\ c
h
h
nvbut
H2
[n-but]
(8)
nvbut
where
c
\ (1/[H ]) exp([d*G
/RT ) and
c
\
MAA
2
MAA
nvbut
(1/[H ])exp([d*G
/RT ). In the absence of strong di†u-
2
nvbut
sional limitations, the rate-determining step is the surface
reaction between adsorbed MAA and hydrogen where the rate
is proportional to the fractional surface coverage,
R \ kh
h
(9)
MAA H2
and assuming that the catalyst has at least a monolayer cover-
age
equilibrium. The term c
given in the rate expression in eqn.
MAA
(14) can be considered to be a measure of the driving force or
displacement from this equilibrium. The values of c evalu-
h
] hH2 ] h
\ 1
(10)
MAA
MAA
nvbut
ated from the data plotted in Fig. 6 are 8.4 dm3 mol~1 and 7.7
dm3 mol~1 for the TA-modiÐed and unmodiÐed samples,
respectively. The product MHB once formed is not converted,
under the stated reaction conditions, back to MAA and the
The fractional coverage by each component will depend on
di†erences in adsorption strength. Substituting eqn. (7) and (8)
into eqn. (10) yields
parameter c
refers to the equilibrium formation of the
active site/MAA complex shown below for the unmodiÐed
system
1
MAA
h
\
(11)
H2
(1 ] c
[MAA] ] c
[n-but])
MAA
nvbut
and substitution of eqn. (11) into eqn. (9) gives a Langmuir-
type expression of the form
Ni ] MAA H NiwMAA
(15)
where the greater the value of c
the greater is the equi-
librium shift to the right. The additional interaction of MAA
with the nickel-supported TA is proposed as the source of the
kc
[MAA]
MAA
MAA
[MAA] ] c
nvbut
R \
(12)
(1 ] c
[n-but])2
MAA
Contacting the activated catalyst with n-butanol [n-butanol:
Ni molar ratio \ (2.6È6.5) ] 102] for extended periods (up to
24 h) prior to the introduction of MAA and/or hydrogen had
no observable e†ect on rate which suggests that the solvent
higher c
values. Moreover, this interaction must induce a
MAA
further polarization of the carbonyl function, thereby render-
ing it more susceptible to attack by hydrogen with a conse-
quent enhancement of the hydrogenation rate where the value
of k, the kinetic term, evaluated from Fig. 6 is greater for the
interacts weakly with the catalyst and the term c
[n-but]
nvbut
can be ignored in the denominator of eqn. (12). Eqn. (12) pre-
dicts a linear relationship between ([MAA]/R)0.5 and [MAA]
where
modiÐed system, i.e. 0.34 [ 0.19 mol h~1 (g )~1. The TA
modiÐer may also serve as a template through which activated
Ni
hydrogen is transferred to the substrate as distinct from
hydrogen transfer from the bare metal surface or spillover
hydrogen from the silica carrier. This supply of hydrogen may
be more e†ective than that provided by the unmodiÐed cata-
lyst or it may merely serve as an additional source of reducing
agent with the result that the overall conversion rate is
increased. The e†ect of varying [MAA] on the enantiomeric
excess generated by the TA-treated sample is illustrated in
Fig. 7. Enantioselectivity can be seen to increase with reactant
concentration, tending towards a limiting value of 22 ^ 2%
which is independent of concentration over the [MAA]/Ni
A
[MAA]
B
0.5
A
1
B
0.5
A
c
B
0.5
MAA
\
]
[MAA]
(13)
R
kc
k
MAA
and the straight line plots are shown in Fig. 6. The graphically
evaluated values of k and c
were used to calculate reaction
MAA
rate over the [MAA] range using the expression
kc
[MAA]
MAA
(1 ] c
R \
(14)
[MAA])2
MAA
and the calculated R vs. [MAA] relationship is compared with
the experimental values in Fig. 5. The experimentally deter-
Fig. 6 Plot of ([MAA]/R)0.5 vs. [MAA] used to evaluate k and c
Fig. 7 The variation of enantiomeric excess with increasing [MAA];
T \ 343 K; particle diameter \ 75È125 lm; stirring speed \ 800 rev
min~1
MAA
for the TA-modiÐed (>) and unmodiÐed (=) Ni/SiO ; T \ 343 K;
2
particle diameter \ 75È125 lm; stirring speed \ 800 rev min~1
J. Chem. Soc., Faraday T rans., 1997, V ol. 93
2005

View Article Online
Algebraic manipulation of eqn. (21) gives a quadratic for x, so
that an analytical solution is available
b
1
x \
^
)(b2 [ 4g2i2)
(22)
2gi2 2gi2
where
b \ 2ig [ i2 [ 2i [ 1
(23)
Taking the e†ectiveness factor at the highest recorded reaction
rate for both systems to be equivalent to unity, the apparent
ratio of surface to bulk [MAA] under the reaction conditions
was calculated and is plotted as a function of stirring speed in
Fig. 8. The value of x is clearly a function of the stirring speed
and it is immediately evident that a vigorous agitation of the
catalyst/reactant suspension is required to ensure that the rate
of supply of the [MAA] reactant approaches the rate of con-
sumption. The unmodiÐed and modiÐed systems exhibit the
same dependence of [MAA] di†usion on agitation where
mass transfer contributions are minor at stirring speeds in
excess of 600 rev min~1.
Fig. 8 Relationship between stirring speed and the dimensionless
parameter
x
(\[MAA]*/[MAA]) for the TA-modiÐed (>) and
unmodiÐed (=) Ni/SiO ; T \ 343 K; particle diameter \ 75È125 lm;
2
MAA/Ni molar ratio \ 47È66
Conclusions
range 1È65 mol
(g )~1. In contrast, Lipgart et al.31
MAA Ni
observed distinct concentration related maxima in the asym-
metric hydrogenation of ethyl acetoacetate while Nitta et al.59
reported an increase in selectivity as the mole fraction of
MAA in ethyl acetate was raised to one. On the other hand,
optical yields have been reported27,60,61 to be independent of
The liquid phase hydrogenation of MAA catalysed by
Ni/SiO in a slurry-type reactor is characterized by an appre-
2
ciable mass transport retardation of reaction rates (i) at stir-
ring speeds \600 rev min~1, (ii) when using catalyst particles
[150 lm in diameter and (iii) at MAA/Ni molar ratios [400.
The reaction is essential zero order with respect to [MHB]
reactant to catalyst ratios in the range 70È170 mmol
MAA
(g )~1. The extent of nickel leaching was not addressed in the
Ni
and [H ] but exhibits a partial (0.4) dependence on [MAA].
2
other studies with the result that the quoted reactant/metal
ModiÐcation of the surface with TA not only increases the
stoichiometry may not in fact prevail during catalysis.
The inÑuence of mass transfer on the overall reaction rate is
often represented by use of the e†ectiveness factor, g, which is
deÐned physically as
reaction rate but lowers the apparent activation energy from
59 to 49 kJ mol~1. Such an e†ect is attributed primarily to the
additional polarization of the carbonyl function due to
TA/MAA hydrogen bonding which renders the CxO bond
more susceptible to attack. The catalytic data are consistent
with a LangmuirÈHinshelwood model in which both hydro-
gen and MAA adsorb competitively on the surface to generate
[MAA] related activity maxima and the rate-determining step
is the surface reaction between the two adsorbed species. The
dependence of rate on MAA concentration over the TA-
modiÐed and unmodiÐed surface can be represented by the
rate expressions R \ 2.85[MAA]/(1 ] 8.39[MAA])2 and
R \ 1.47[MAA]/(1 ] 7.67[MAA])2, respectively. Enantio-
selectivity, on the other hand, increases with an increase in
[MAA]/Ni to attain a plateau value over the range 1È6.5
observed reaction rate
g \
(16)
rate obtained with no mass transfer resistance
At steady state the rate of MAA supply by mass transfer is
counterbalanced by the rate of MAA consumption in the reac-
tion at the catalyst surface. The e†ectiveness factor equals
unity where the surface reactant concentration ([MAA]*)
equals the bulk solution concentration ([MAA]) and the
observed reaction kinetics is the same as the true, intrinsic
kinetics at the liquid/gas/solid interface. In terms of the pro-
posed rate expression
mol
(g )~1.
MAA Ni
kc
(1 ] c
kc
[MAA]*
MAA
This work was supported by a grant from the Royal Society.
[MAA]*)2
[MAA]
MAA
MAA
g \
(17)
References
(1 ] c
[MAA])2
MAA
1
2
Chem. Eng. News, October 9, 1995, 44.
which reduces to
g \
H. Yamada and S. Shimizu, Angew. Chem., Int. Ed. Engl., 1988,
27, 622.
[MAA]*(1 ] c
[MAA](1 ] c
[MAA])2
MAA
MAA
(18)
(19)
(20)
2
4
C.-H. Wong and G. M. Whitesides, Enzymes in Synthetic Organic
Chemistry, Elsevier Science, London, 1994.
[MAA]*)2
L. Poppe and L. Novak, Selective Biocatalysis: A Synthetic
Approach, VCH, New York, 1992.
Introducing the dimensionless variables
i \ c [MAA]
5
6
7
8
9
F. H. Arnold, Curr. Opin. Biotechnol., 1993, 4, 450.
J. Hodgson, Bio/T echnology, 1994, 12, 789.
Y. Izumi, Adv. Catal., 1983, 12, 215.
H. U. Blaser, T etrahedron: Asymmetry, 1991, 2, 843.
H. U. Blaser and M. Muller, Stud. Surf. Sci. Catal., 1991, 59, 73.
*MAA+
and
[MAA]*
x \
10 R. A. Holt and S. R. Rigby, in Proceedings of the Chiral Europe
Ï95 Symposium, Spring Innovations Ltd., London, 1995, p. 13.
11 M. A. Keane, Zeolites, 1993, 13, 14.
12 M. A. Keane, Zeolites, 1993, 13, 22.
13 M. A. Keane, Zeolites, 1993, 13, 330.
14 A. Bennett, S. Christie, M. A. Keane, R. D. Peacock and G.
Webb, Catal. T oday, 1991, 10, 1619.
[MAA]
eqn. (17) takes the form
x(1 ] i)2
(1 ] ix)2
g \
(21)
2006
J. Chem. Soc., Faraday T rans., 1997, V ol. 93

View Article Online
15 M. A. Keane and G. Webb, J. Catal., 1992, 136, 1.
16 M. A. Keane and G. Webb, J. Mol. Catal., 1992, 73, 91.
17 M. A. Keane, Can. J. Chem., 1994, 72, 372.
18 M. A. Keane, L angmuir, 1994, 10, 4560.
42 A. Jones and B. P. MacNicol, T emperature Programmed
Reduction for Solid Materials, Marcel Dekker, New York, 1986.
43 J. S. Smith, P. A. Thrower and M. A. Vannice, J. Catal., 1981, 68,
270.
44 M. Spiro, in Comprehensive Chemical Kinetics, Reactions at the
L iquid-Solid Interface, ed. R. G. Compton, Elsevier, Amsterdam,
1989, p. 69.
19 M. A. Keane, Catal. L ett., 1993, 19, 197.
20 I. M. Sutherland, A. Ibbotson and P. B. Wells, J. Catal., 1990,
125, 77.
21 P. A. Meheux, A. Ibbotson and P. B. Wells, J. Catal., 1991, 128,
387.
45 P. L. Freund and M. Spiro, J. Chem. Soc., Faraday T rans. 1,
1983, 79, 491.
22 J. T. Wehrli, A. Baiker, D. M. Monti and H. U. Blaser, J. Mol.
46 G. C. Bond, in Heterogeneous Catalysis, Principles and Applica-
tions, 2nd edn. ed. P. W. Atkins, J. S. E. Hoker and A. K. Holli-
day, Oxford Science, Clarendon Press, Oxford, 1987.
47 J. M. Thomas and W. J. Thomas, Introduction to the Principles of
Catalysis, Academic Press, London, 1967.
Catal., 1990, 61, 207.
23 T. Harada, M. Yamamoto, S. Onaka, M. Imaida, H. Ozaki, A.
Tai and Y. Izumi, Bull. Chem. Soc. Jpn., 1981, 54, 2323.
24 D. O. Hubbell and P. Rys, Chimia, 1970, 24, 442.
25 N. D. Zubareva, V. V. Chernysheva, Yu. A. Grigorev and E. I.
Klabunovskii, Bull. Soc. Acad. Sci. USSR, Div. Chem. Sci., 1987,
476.
48 G. W. Roberts, in Catalysis in Organic Synthesis, ed. P. N. Rylan-
der and H. GreenÐeld, Academic Press, New York, 1976, p. 1.
49 U. K. Singh, R. N. Landau, Y. Sun, C. leBlond, D. G. Black-
mond, S. K. Tanielyan and R. L. Augustine, J. Catal., 1995, 154,
91.
26 V. V. Chernysheva, I. P. Murina and E. I. Klabunovskii, Bull.
Acad. Sci. USSR, Div. Chem. Sci., 1983, 689.
27 G. V. Smith and M. Musoiu, J. Catal., 1979, 60, 184.
28 T. Harada, Y. Haraki, Y. Izumi, J. Muraoka, H. Ozaki and A.
Tai, in Proceedings of the 6th International Congress on Catalysis,
ed. G. C. Bond, P. B. Wells and F. C. Tompkins, The Chemical
Society, London, 1977, p. 1024.
29 Y. Nitta, T. Imanaka and S. Teranishi, J. Catal., 1983, 80, 31.
30 E. I. Klabunovskii, Russ. J. Phys. Chem., 1973, 47, 765.
31 E. N. Lipgart, Yu, I. Petrov and E. I. Klabunovskii, Kinet. Katal.,
1971, 12, 1491.
32 I. Yasumori, M. Yokozeki and Y. Inoue, Faraday Discuss., 1981,
71, 385.
33 A. Hoek, H. M. Woerde and W. M. H. Sachtler, in Proceedings of
the 7th International Congress on Catalysis, ed. K. Tanabe and T.
Sieyama, Kodansha/Elsevier, Tokyo/Amsterdam, 1981, p. 376.
34 M. A. Keane and G. Webb, J. Chem. Soc., Chem. Commun., 1991,
1619.
50 J. Struijk, M. dÏAngremond, W. J. M. Lucas-deRegt and J. J. F.
Scholten, Appl. Catal. A, 1992, 83, 263.
51 R. M. Koros and E. J. Nowak, Chem. Eng. Sci., 1967, 22, 470.
52 Y. Nitta, F. Sekine, J. Sasaki, T. Imanaka and S. Teranishi, J.
Catal., 1983, 79, 211.
53 H. Ozaki, A. Tai, S. Kobatake, H. Watanabe and Y. Izumi, Bull.
Chem. Soc. Jpn., 1978, 51, 3559.
54 T. Osawa, T. Harada and A. Tai, J. Catal., 1990, 121, 7.
55 R. C. Reid, J. M. Prausnitz and T. K. Sherwood, T he Properties
of Gases and L iquids, 3rd edn., McGraw-Hill, New York, 1977.
56 C. L. Yaws, J. W. Miller, P. N. Shah, G. R. Schorr and P. M.
Patel, Chem. Eng., 1976, 83, 153.
57 W. Gerrard, Gas Solubilities, W idespread Applications, Pergamon
Press, Oxford, 1980.
58 J. A. Groenewegen and W. M. H. Sachtler, in Proceedings of the
6th International Congress on Catalysis, ed. G. C. Bond, P. B.
Wells and F. C. Tompkins, The Chemical Society, London, 1977,
p. 1014.
35 H. Brunner, K. Amberger and J. Wiehl, Bull. Soc. Chim. Belg.,
1991, 100, 571.
36 A. Hoek and W. M. H. Sachtler, J. Catal., 1979, 58, 276.
37 F. Fu, H. H. Kung and W. M. H. Sachtler, J. Mol. Catal., 1987,
42, 29.
59 Y. Nitta, F. Sekine, J. Sasaki, T. Imanaka and S. Teranishi, Chem.
L ett., 1981, 541.
60 G. Wittmann, G. B. Bartok and G. V. Smith, J. Mol. Catal., 1990,
60, 1.
38 L. J. Bostelaar and W. M. H. Sachtler, J. Mol. Catal., 1984, 27,
387.
39 B. Coughlan and M. A. Keane, Zeolites, 1991, 11, 2.
40 B. Coughlan and M. A. Keane, J. Catal., 1990, 123, 364.
41 W. Romanowski, Highly Dispersed Metals as Adsorbents, Halsted
Press, Wiley, New York, 1987, p. 171.
61 H. Brunner, M. Muschiol, T. Wishert and J. Wiehl,
T etrahedron : Asymmetry, 1990, 1, 159.
Paper 6/07661K; Received 11th November, 1996
J. Chem. Soc., Faraday T rans., 1997, V ol. 93
2007