Enantioselective Hydrogenation: V. Hydrogenation of Butane-2,3-dione and of 3-Hydroxybutan-2-one Catalysed by Cinchona-Modified Platinum

Article abstract of DOI:10.1006/jcat.1998.2204

Pt/silica modified by cinchonidine and cinchonine is active for the enantioselective hydrogenation of butane-2,3-dione to butane-2,3-diol in dichloromethane at 268-298 K and 10 bar pressure. Reaction proceeds in three stages. In the first, about 85% of the butane-2,3-dione is converted to 3-hydroxybutan-2-one and 15% to three higher molecular weight products by hydrodimerisation. The initial enantiomeric excess in the hydroxybutanone is modest (20 to 40%(R) with cinchonidine as modifier, 10%(S) with cinchonine as modifier) and dependent on the amount of alkaloid used in catalyst preparation. In the second stage, 3-hydroxybutan-2-one is converted to butane-2,3-diol; a marked kinetic effect is observed whereby the minority enantiomer is converted preferentially to butanediol and the enantiomeric excess in the remaining hydroxybutanone increases dramatically to values in the range 62 to 89%(R) and to 30%(S). Under all conditions, the most abundant stereochemical form of the final product is meso-butane-2,3-dione. In the third stage the three dimers are slowly converted by hydrogenation, dissociation, and further hydrogenation to butane-2,3-diol. In the absence of alkaloid, butane-2,3-dione hydrogenation to racemic products in dichloromethane solution proceeds in two distinct stages with no dimer formation. Butane-2,3-dione hydrogenation has also been studied over Pt/silica modified anaerobically by exposure to cinchonidine in ethanol under propyne at 2 bar. This catalyst is remarkably active for the conversion of diketone to diol in ethanol at 293 K and 10 bar and kinetic selection in the second stage of reaction is again observed. The hydrogenation of racemic 3-hydroxybutan-2-one in dichloromethane over cinchonine-modified Pt/silica at 273 K and 10 to 40 bar pressure also showed kinetic selection, an enantiomeric excess of up to 70%(S) appearing in the reactant as it was consumed. Mechanisms which account for these hydrogenations and dimerisations and for the enantioselectivities observed and their variation are presented. This diketone hydrogenation provides an example of consecutive thermodynamic and kinetic control of enantioselectivity in a multistage catalytic reaction.

Full text of DOI:10.1006/jcat.1998.2204

JOURNAL OF CATALYSIS 179, 267–276 (1998)
ARTICLE NO. CA982204
Enantioselective Hydrogenation
V. Hydrogenation of Butane-2,3-dione and of 3-Hydroxybutan-2-one
Catalysed by Cinchona-Modified Platinum
�
�
�
,1
�
,2 and P. B. Wells�,3
J. A. Slipszenko, S. P. Griffiths, P. Johnston, K. E. Simons,† W. A. H. Vermeer,
�
Department of Chemistry, University of Hull, Hull HU6 7RX, United Kingdom; †Johnson Matthey Precious Metals Division,
Orchard Road, Royston, SG8 5HE, United Kingdom
Received April 15, 1998; revised June 9, 1998; accepted June 9, 1998
Key Words: Butane-2,3-dione hydrogenation; 3-hydroxybutan-2-
Pt/silica modified by cinchonidine and cinchonine is active for the one hydrogenation; enantioselective hydrogenation; enantiomeric
enantioselective hydrogenation of butane-2,3-dione to butane-2,3- excess; cinchona alkaloids; cinchonidine; cinchonine; platinum;
diol in dichloromethane at 268–298 K and 10 bar pressure. Reaction EUROPT-1; modification (of catalysts by alkaloids); kinetic selec-
proceeds in three stages. In the first, about 85% of the butane-2,3- tion; selective enantioface adsorption; diol formation; hydrodimeri-
dione is converted to 3-hydroxybutan-2-one and 15% to three higher sation of diones.
molecular weight products by hydrodimerisation. The initial enan-
tiomeric excess in the hydroxybutanone is modest (20 to 40%(R)
with cinchonidine as modifier, 10%(S) with cinchonine as modifier)
and dependent on the amount of alkaloid used in catalyst prepa-
INTRODUCTION
ration. In the second stage, 3-hydroxybutan-2-one is converted to
butane-2,3-diol; a marked kinetic effect is observed whereby the mi-
nority enantiomer is converted preferentially to butanediol and the
enantiomeric excess in the remaining hydroxybutanone increases
The enantioselective hydrogenation of butane-2,3-dione
to 3-hydroxybutan-2-one is catalysed by Pt/silica modified
by adsorption of cinchona alkaloids onto its surface (1).
dramatically to values in the range 62 to 89%(R) and to 30%(S). Cinchonidine and cinchonine (Figs. 1a and 1b) facilitate the
Under all conditions, the most abundant stereochemical form of preferential formation of R-product and S-product, respec-
the final product is meso-butane-2,3-dione. In the third stage the tively, and in this respect the reaction resembles the much-
three dimers are slowly converted by hydrogenation, dissociation,
and further hydrogenation to butane-2,3-diol. In the absence of
alkaloid, butane-2,3-dione hydrogenation to racemic products in
dichloromethane solution proceeds in two distinct stages with no
occurs in dicholoromethane, toluene, acetone, and ethanol
dimer formation. Butane-2,3-dione hydrogenation has also been
as solvent, the enantiomeric excess at 10 bar pressure and
studied over Pt/silica modified anaerobically by exposure to cin-
chonidine in ethanol under propyne at 2 bar. This catalyst is re-
markably active for the conversion of diketone to diol in ethanol at
studied enantioselective hydrogenation of methyl pyruvate
over the same cinchona-modified Pt/silica catalysts (2–6).
Preliminary communications have recorded that reaction
2
73–293 K decreasing from 38% in dichloromethane to 8%
in ethanol (1) and that the conversion of 3-hydroxybutan-
-one to butane-2,3-diol occurs readily in dichloromethane,
2
2
93 K and 10 bar and kinetic selection in the second stage of reaction
is again observed. The hydrogenation of racemic 3-hydroxybutan-2- the enantiomeric excess in the remaining hydroxyke-
one in dichloromethane over cinchonine-modified Pt/silica at 273 K tone increasing dramatically to 89% at high conversion
and 10 to 40 bar pressure also showed kinetic selection, an enan- (7). This paper describes (i) the three-stage hydrogena-
tiomeric excess of up to 70%(S) appearing in the reactant as it was
consumed. Mechanisms which account for these hydrogenations
anddimerisationsandfortheenantioselectivitiesobservedandtheir
tion of butane-2,3-dione to butane-2,3-diol over cinchona-
modified Pt/silica in dichloromethane solution, (ii) the hy-
drogenation of racemic 3-hydroxybutan-2-one under the
variation are presented. This diketone hydrogenation provides an
same conditions, (iii) the stereochemistry of the first stage
example of consecutive thermodynamic and kinetic control of enan-
of the hydrogenation of hexane-2,3-dione and of hexane-
tioselectivity in a multistage catalytic reaction.
�c 1998 Academic Press
3
,4-dione, and(iv)anovelpreparationtechniquewhichpro-
vides a catalyst capable of fast continuous hydrogenation
of butane-2,3-dione to butane-2,3-diol.
Modification by the strychnos alkaloid brucine and the
morphine alkaloid codeine is ineffective in inducing enan-
tioselectivity in butane-2,3-dione hydrogenation, although
1
Present address: Johnson Matthey, Precious Metals Division,
Orchard Road, Royston, SG8 5HE, UK.
2
Department of Chemistry, University of Reading, Reading, RG6
2
AD, UK.
3
To whom correspondence should be addressed.
267
0021-9517/98 $25.00
Copyright �c 1998 by Academic Press
All rights of reproduction in any form reserved.

2
68
SLIPSZENKO ET AL.
these alkaloids each induce enantioselectivity in methyl
pyruvate hydrogenation (8,9).
Procedure I involving ex-situ aerobic modification. This
procedure follows the original Orito methodology (12) and
the “normal modification” described in Part II (13). 50 mg
alkaloid dissolved in 5-ml dichloromethane was injected
via a septum onto 0.1 g reduced catalyst under hydrogen.
Catalyst and modifier solution were then transferred to an
open beaker which contained a further 35 ml of the modifier
solution and stirred in air for 1 h, after which the solution
was decanted and discarded and the wet catalyst washed
into the Fischer–Porter reactor with a further 20-ml solvent.
EXPERIMENTAL
Materials
Experiments were conducted using 0.1-g samples of 6.3%
Pt/silica (EUROPT-1) for which a detailed characterisation
has been reported (10,11). As-received catalyst, reduced
by the manufacturer (Johnson Matthey) but subsequently
oxidised by the action of air, was evacuated at 293 K for
1
0-ml (115-mmol) butanedione was added to the vessel,
the reactor was flushed first with nitrogen and then with
hydrogen, and subsequently pressurised to 10 bar. Stirring
was commenced immediately.
0
.5 h, reduced at 403 K for 0.5 h in two changes of hydrogen
(1 bar) and cooled to ambient temperature in hydrogen
before modification by alkaloid.
Butane-2,3-dione, 3-hydroxybutan-2-one (acetoin), and
butane-2,3-diol (all Fluka), cinchonidine and cinchonine
Procedure II involving in-situ aerobic modification. Al-
kaloid (50 mg) was dissolved in 5-ml solvent in the Fischer–
Porter reactor. To this was added 0.1-g reduced catalyst in
(Aldrich, Figs. 1a and 1b), dichloromethane and ethanol
5
2
-ml solvent, 6-ml (69 mmol) butanedione, and a further
4-ml solvent. The reactor was then flushed with hydro-
were used as received. Reference materials 3-quinuclidinol,
and RR- and SS-butane-2,3-diol were obtained from
Aldrich and 2,3,5,6-tetramethyl-1,4-dioxan-2,5-diol (ace-
toin dimer, Fig.1c) from Lancaster.
In this paper, the terms butanedione and diketone
always refer to butane-2,3-dione, hydroxybutanone and
hydroxyketone always refer to 3-hydroxybutan-2-one, and
butanediol and diol always refer to butane-2,3-diol.
gen, pressurised to 10 bar, and the stirring commenced.
The above procedure was followed for reactions in the
Baskerville reactor, except that the quantity of butane-
dione was 3 ml (34.5 mmol) and the total volume of solvent
2
7 ml. This procedure was also used for studies of hydroxy-
butanone hydrogenation.
Procedure III involving in-situ anaerobic modification
under hydrogen. This procedure has been described pre-
Reactors and Procedures
Liquid-phase hydrogenations were mostly conducted at viously (13). Solvent (ethanol), reactant, and modifier
73 K and 10 bar pressure in a glass Fischer–Porter reactor solution were each rigorously degassed by the succes-
2
of volume 200 ml fitted with a magnetic stirrer. Hydrogen sive freeze–thaw method. 0.1-g catalyst was reduced in
pressure was maintained at 10 bar and hydrogen uptake the Fischer–Porter reactor under the conditions described
recorded by computer control. Reactions at 293 K involv- above. Modifier solution containing 50-mg alkaloid in 20-ml
ing cinchonine as modifier and described in Tables 1 and ethanol was injected via a septum onto the reduced cata-
5
were carried out at pressures in the range 10 to 40 bar lyst under hydrogen and the mixture stirred for 1 h; 10-ml
in a Baskerville stirred stainless steel reactor of volume (115 mmol) butanedione was then added. Hydrogen pres-
8
0 ml.
sure raised to 10 bar and the stirrer started.
FIG. 1. Structures of the modifiers cinchonidine (a), cinchonine (b), and of the reference compound acetoin dimer (c).

ENANTIOSELECTIVE HYDROGENATION V
269
Procedure IV involving in-situ anaerobic modification
under hydrocarbon. Procedure III was followed except
that, after catalyst reduction, hydrogen was removed and
immediately replaced by 2 bar propyne or buta-1,3-diene
(
ca 17 mmol). When the slurry had been stirred for 1 h in
this hydrocarbon atmosphere, hydrogen was admitted to
0-bar pressure and the stirrer started.
1
Analysis
After filtration to remove catalyst, 0.1 �l aliquots of reac-
tion mixture were analysed by chiral gas–liquid chromato-
graphy using a 50-m permethylated beta-cyclodextrin (Cy-
dex B) column. Separation of reactant and all products was
achieved; components were eluted in the order: solvent,
diketone, hydroxyketone, diol, dimers. Complete separa-
tion of R- and S-hydroxybutanone was obtained. Two of the
FIG. 2. Variations of hydrogen uptake with time. (a) Butane-2,3-
dione hydrogenation in dichloromethane at 273 K and 10 bar over
dimerswereelutedinrapidsuccessionwithbaselinesepara- cinchonidine-modifiedPt/silica(aerobicmodification, ProcedureII)show-
ing the three stages of reaction. (b) Butane-2,3-dione hydrogenation in
ethanol at 293 K and 10 bar over cinchonidine-modified Pt/silica (upper
curve) and unmodified Pt/silica (lower curve) (anaerobic modification,
Procedure IV, 2 bar propyne).
tion as peaks of identical size; the third dimer in comparable
yield was eluted later. The enantiomers of butanediol were
eluted in the order SS-, RR-, meso-, as determined by use of
the first two compounds as reference materials. The meso-
form was separated from the SS- and RR-forms but base-
cinchonidine-modified Pt/silica at 268 to 298 K and 10-bar
line separation between RR- and SS- was not achieved. Val-
pressure occurred in three stages (Fig. 2a). A fast first stage
ues of the yield of meso-butanediol were determined from
the chromatograms on the assumption that the response
factors for the enantiomers were equal. Values of the ratio
� 1 � 1
(
typically Rm = 275 mmol h
g
at 273 K) was followed by
� 1 � 1
a less rapid second stage (Rm = 30 mmol h g ) and a very
slow third stage. Maximum rates at 273 K for the first stage
were first order in hydrogen pressure (3 to 10 bar), approx-
imately zero order in diketone, and of 0.6 order in catalyst
mass (0.025 to 0.100 g). Such rates obeyed the Arrhenius
[RR-] : [SS-] were determined by comparison of gc traces
with those of known calibration mixtures.
Analysis by gas chromatography mass spectrometry
(GCMS) was achieved by use of a conventional gc (no chiral
equation over the range 268 to 298 K, giving a value of 30 �
separation) and a Finnegan GCQ quadrupole mass spec-
trometer.
� 1
5
kJ mol for the apparent activation energy, Ea. The or-
der in catalyst mass suggests a measure of diffusion control
but the value of Ea confirms that rate was largely controlled
by chemical processes at the catalyst surface. Above 298 K
rates and values of the enantiomeric excess declined in a
manner similar to that previously described for pyruvate
hydrogenation (13).
Rate Measurements
Hydrogen uptake against time curves for reactions in
dichloromethane were similar to those reported for pyru-
vate ester hydrogenation in ethanol in Part I (14); i.e., there
was a short period during which the rate accelerated fol-
lowed by a sustained period at a constant maximum rate,
3
-Hydroxybutan-2-one (acetoin) was formed as major
product in the first stage, together with three minor prod-
ucts in comparable yields which accounted for about 15%
of reactant consumption. No butanediol was formed in the
first stage. GCMS showed the minor products to be of mass
�
1
� 1
Rm/mmol h (g catalyst) . The period of acceleration was
less pronounced than with pyruvate ester, so that values
of Rm did not differ greatly from the initial rates for most
reactions.
1
74 and of related structure. Parent ions were present at
m/z = 175 (due to proton addition in the source) and the
Molecular Modelling
major fragment ions were at m/z = 131, 88 and 89, 71, and
The procedures used for molecular modelling were ex- 43. Interpretation of the spectra of these products was as-
actly as described elsewhere (15).
sisted by comparison with that of the reference compound
acetoin dimer to which they showed a family resemblance.
Accordingly these products were assigned structures pro-
posed in the Discussion and contained in Fig. 6; they were
partially hydrogenated dimers of butanedione.
RESULTS
A. Reactions in Dichloromethane Solution
under Aerobic Conditions
The second stage of reaction commenced when most of
(a) Butanedione hydrogenation. Enantioselective hy- the butanedione had reacted, hydroxybutanone being con-
drogenation of butanedione in dichloromethane over verted to butanediol.

2
70
SLIPSZENKO ET AL.
TABLE 1
Enantiomeric Compositions of the Products during Butanedione Hydrogenation in Dichloromethane at 10 bar
Pressure over Pt/silica Modified by Cinchonidine (CD) and Cinchonine (CN)
H2 uptakeb
[Dione]i
Hydroxybutanone
Butanediol
[RR-] : [SS-]
Entry
Modifier
Temp/K
R/%
S/%
ee/%
meso/%
c
1
2
3
4
5
6
7
8
9
CD
CD
CD
CD
CD
CD
CN
CN
CN
273
273
273
273
273
273
293
293
293
0.12
0.18
0.50
0.95
1.58
1.63
0.17
1.13
1.53
60
60
61
62
81
77
44
39
35
40
40
39
38
19
23
56
61
65
20
20
22
24
62
54
12
22
30
—
—
—
—
c
c
d
63
60
2 : 1
2 : 1
c
—
62
68
1 : 1.5
1 : 1.5
a
Modification: Procedure II (50 mg alkaloid). Reactions at 273 K conducted in the Fischer–Porter reactor, reactions at
93 K conducted in the Baskerville reactor.
2
b
Units: H2 uptake = mmol; [Dione]i = mmol dione originally present (69 mmol, entries 1–6; 34.5 mmol, entries 7–9).
c
Butanediol yield = zero.
d
Butanediol yield <1%.
The third stage corresponded to the last 15% or so of the minor enantiomer was preferentially hydrogenated to
hydrogen uptake and involved mostly the conversion of butanediol. All three stereochemical forms of butanediol
the dimers to butanediol which was the only product at the were produced; the yield of meso-diol exceeded 50% and
end of the reaction.
Hydrogenation of butanedione to racemic product in whichever alkaloid was adsorbed at the catalyst surface.
dichloromethane over unmodified catalyst (Procedure II, The high values of the enantiomeric excess in the re-
the concentrations of RR- and SS-diol were comparable
no alkaloid added) at 293 K and 10 bar provided a hydro- maining hydroxyketone during stage two of the reaction
gen uptake curve similar to that shown in Fig. 2a over the recorded in Table 1 were further raised by increasing both
first stage; i.e., modified and unmodified catalysts showed the amount of modifier used in catalyst preparation from
similar values of Rm. Over an unmodified catalyst, no higher 50 to 200 mg and the amount of catalyst from 0.1 to 0.6 g
molecular weight products were formed, the second stage (Table 3).
of the hydrogenation was slow and no third stage was evi-
Towards the end of stage three, when the reaction was
dent. However, formation of the same dimers accompanied virtually complete, analysis showed the dimers to be absent
the reaction to racemic product when 3-quinuclidinol was (i.e., they had been consumed) but very small quantities of
presentinthereactionmixture(ProcedureII, 50-mgN-base both butanedione and hydroxybutanone were present. The
added in place of cinchona alkaloid).
latter was almost racemic (ee = 2%).
The enantiomeric compositions of the C4-products dur-
ing the first and second stages of reaction over alkaloid-
modified catalysts are shown in Table 1. In the first stage,
cinchonidine induced enantioselectivity in favour of R-
hydroxybutanone, whereascinchoninedirectedreactionto-
wards S-product. The enantiomeric excess was independent
of conversion in the early stages of reaction, independent
of hydrogen pressure (3 to 10 bar), fairly independent of
diketone concentration (Table 2), and increased slightly as
temperature was lowered (295 to 268 K). The enantiomeric
excess was, however, dependent on the quantity of cin-
chonidine used in catalyst modification; the value at low
conversion rose from 20% (Table 1) to 38% (Table 2, en-
try 6) as the mass of cinchonidine used in modification was
increased from 50 to 200 mg.
TABLE 2
Variation of Maximum Rate, Rm, and Enantiomeric Excess, ee,
with Reactant Concentration during Butanedione Hydrogenation
over Cinchonidine-Modified Pt/Silica in Dichloromethane at 273 K
and 10 bar Pressure
� 1 � 1
ee/%(R)b
[
Dione]/M
1.4c
Rm/mmol h
g
1
1225
1375
1350
1000
1200
1100
36
36
33
38
42
38
1
1
1.1
0.2
8
.5
5.7
.8
2
a
b
c
Modification: Procedure I (200 mg modifier).
Measured during the first stage of reaction.
Pure reactant (no solvent).
In the second stage of reaction the enantiomeric excess
in the remaining hydroxybutanone rose dramatically as

ENANTIOSELECTIVE HYDROGENATION V
271
TABLE 3
and 10 bar over unmodified and modified catalyst pro-
ceeded at comparable rates; i.e., as with the dione, there
was no rate enhancement over the modified catalyst. In-
stantaneous reaction rate was strictly first order in the re-
maining hydroxyketone. As reaction over modified catalyst
progressed an enantiomeric excess appeared in the reac-
tant; the effect was favoured by an increase in temperature
but not by an increase in pressure (Table 5). No dimers
were formed. All three enantiomers of butanediol were
again formed with the yield of the meso-form exceeding
Values of the Enantiomeric Excess in the Remaining Hydroxy-
butanone at High Conversion Observed during Butanedione
Hydrogenation to Diol over Cinchonidine-Modified Pt/silica in
Dichloromethane at 10 bar Pressure
Mass of catalyst/g
[Dione]/M
Temp/K
ee/%
0
0
0
0
0
.1
.6
.6
.6
.6
1.7
0.6
1.3
0.9
0.9
273
273
273
283
273
62
63
75
85
89
5
0% but with [RR-] > [SS-] when cinchonidine was modi-
fier and [SS-] > [RR-] when cinchonine was the modifier.
B. Reactions in Ethanolic Solution over Catalysts Prepared
under Aerobic and Anaerobic Conditions: Butanedione
Hydrogenation
Hydrogenation of the symmetrical hexane-3,4-dione to
-hydroxyhexan-3-one over cinchonidine-modified catalyst
4
at 293 K and 10 bar pressure gave an enantiomeric ex-
cess of 33%(R) in the first stage of reaction. Reaction of
the unsymmetrical hexane-2,3-dione at 273 K gave equal
yields of 3-hydroxyhexan-2-one and 2-hydroxyhexan-3-one
showing values of the ee of 35%(R) and 29%(R), re-
spectively. Comparable reactions of hexane-2,3-dione over
cinchonine-modified catalyst provided values of the ee of
Hydrogenation of butanedione in ethanol over Pt mod-
ified by cinchonidine according to Procedure II (aerobic
modification) gave hydroxyketone as the major product,
together with smaller yields of the three dimers than were
observed in dichloromethane solution. Activity for the sec-
ond stage of reaction was minimal. Hydrogenation rates
�
1
� 1
at 293 K were typically 2200 mmol h
g
over modified
2
0%(S) and 11%(S), respectively.
�
1
� 1
catalyst with ee = 15 to 20%(R) and 550 mmol h
g
for
The competitive hydrogenation of butanedione and
reaction over the unmodified catalyst; i.e., a conventional
rate enhancement (13) was observed. Anaerobic prepara-
tion by Procedure III produced modified catalysts of rela-
methyl pyruvate in dichloromethane and in ethanol over
cinchonidine-modified catalyst at 273 K and 10 bar pressure
was investigated (Table 4). The reactants competed effec-
tively with each other for the surface. The enantiomeric ex-
cess in the hydroxyketone was not affected by the presence
of methyl pyruvate, but that in methyl lactate was depressed
by 10 to 15% by the presence of the diketone.
�
1
� 1
tively low activity, 220 mmol h g , and enantioselectivity,
5
to 8%(R), and unmodified catalysts showed an activity
�
1
� 1
of 40 mmol h
g
(the rate enhancement was again ob-
served). Anaerobic preparation under 2-bar propyne (Pro-
cedure IV) gave catalysts, both unmodified and modified,
(b) Hydrogenation of 3-hydroxybutan-2-one. Hydro- of comparable high activity (i.e., no rate enhancement for
genation of racemic 3-hydroxybutan-2-one in dichlo- modified reaction) as shown in Fig. 2b. Reactions showed
romethane over modified and unmodified Pt/silica was in- a very short induction period after which fast reaction in
�
1
� 1
vestigated under a range of conditions. Reactions at 273 K the range R = 1300 to 2000 mmol h
g
occurred. Reac-
m
tion over unmodified catalyst ceased after the formation of
3
-hydroxybutan-2-one, but over modified catalyst rapid hy-
TABLE 4
drogenation extended to diol formation and complete con-
Competitive Hydrogenation of Butane-2,3-dione to 3-Hydroxy- version was achieved. The initial enantiomeric excess in the
butan-2-one and of Methyl Pyruvate to Methyl Lactate in hydroxyketone was about 10%, lower than the values ob-
Dichloromethane (DCM) and in Ethanol at 273 K and 10 bar over served for reactions in dichloromethane over catalysts pre-
a
Cinchonidine-Modified Pt/silica
pared by Procedure II. During butanediol formation over
cinchonidine-modified catalyst the preferential removal of
S-hydroxybutanone was again observed. Thus, as butane-
diol yields progressed from 10, to 30, to 50, to 75%, so the
enantiomeric excess in the remaining hydroxyketone rose
from 8, to 11, to 26, to 40%(R).
Anaerobic modification with cinchonidine under 2 bar
buta-1,3-diene (Procedure IV) gave a fast first stage re-
action without an induction period. Catalysts so modified
were not active for the conversion of hydroxyketone to
diol. The enantiomeric excess in the 3-hydroxybutan-2-one
againincreasedwiththeamountofthemodifierpresent, the
Enantiomeric excess/%(R)
Single reactant Competitive
Solvent
Reactant(s)
Conversion/%
reaction
reaction
DCM
DCM
DCM
Butanedione
95
95
95
38
86
Pyruvate ester
Diketone + ester
40, 75
26, 60
Ethanol Butanedione
Ethanol Pyruvate ester
Ethanol Diketone + ester
60
60
60
22
75
a
Modification: Procedure I (200-mg modifier).

2
72
SLIPSZENKO ET AL.
TABLE 5
Enantiomeric Composition of the Reactant and Product and Values of kS/kR Observed during the Hydro-
genation of a Racemic Mixture of R- and S-Hydroxybutanone in Dichloromethane over Pt/silica Modified by
a
Cinchonidine (CD) and Cinchonine (CN)
Enantiomeric composition /%
H2
pressure
/bar
in hydroxybutanone
b
Modifier
Temp/K
Conv /%
R/%
S/%
ee/%
meso-diol/%
50c
kS/kR
None
CD
CN
CN
CN
CN
CN
10
10
10
10
20
30
40
273
273
273
293
293
293
293
30
50
90
80
90
90
95
50
54
44
20
15
27
30
50
46
56
80
85
73
70
0
8
1.0d
1.3
1.1
2.2
2.0
1.8
1.3
d
—
65
e
12
60
70
46
40
56
60
59
55
a
Modification: Procedure II (50 mg cinchonidine). Reactions at 273 K conducted in the Fischer–Porter reactor, reactions
at 293 K conducted in the Baskerville reactor.
b
(
H2 uptake/mmol)/(mmol 3-hydroxybutan-2-one initially present).
c
d
e
[
RR-] : [SS-] = 1 : 1.
RR-] : [SS-] � 8 : 1, meso-diol not measured.
SS-] : [RR-] � 8 : 1.
[
[
values at 90% conversion, for reactions at 293 K and 10 bar alkaloid used in the modification step (compare Tables 1
pressure, being 12%, 18% and 35% when the amounts of and 2), whereas in pyruvate hydrogenation maximum ee
modifier present were 5 mg, 50 mg, and 100 mg, respec- was achieved at very low alkaloid concentrations (19). The
tively.
second and third features suggest that the diketone is more
strongly adsorbed than pyruvate and may displace some al-
kaloid from the platinum surface. Certainly, butanedione
competed effectively for the surface with pyruvate in the
competitive reaction (Table 3). Moreover, the distinct tran-
sition from stage 1 to stage 2 of the reaction (Fig. 2a and
DISCUSSION
A. Reactions in Dichloromethane Solution
The first stage of the reaction, the hydrogenation of Table 1) shows that the concentration of diketone had to
butan-2,3-dione to 3-hydroxybutan-2-one, shows many of be reduced to a very low level before its surface coverage
the characteristics of the more extensively studied hydro- fell sufficiently for hydroxyketone hydrogenation to diol to
genation of pyruvate to lactate esters (13). Thus, each re- become significant.
action is zero order in the organic reactant and first or-
Our interpretation of the observed sense of the enan-
der in hydrogen, indicating strong adsorption of the former tioselectivity in the conversion of butanedione to hydroxy-
and relatively weak adsorption of the latter. The apparent butanone during the first stage of reaction follows that pub-
�
1
activation energy of 30 kJ mol for diketone hydrogena- lished for the hydrogenation of pyruvate to lactate (15) and
�
1
tion compares with 38 kJ mol for pyruvate ester hydro- is therefore presented here in summary only. H/D exchange
genation (13). In each system, cinchonidine induces enan- in cinchona alkaloid over Pt/silica has shown that adsorp-
tioselectivity in favour of R-product, whereas cinchonine tion occurs via the quinoline moiety (20). The minimum
favours preferential S-product formation but as tempera- energy conformations of cinchonine and cinchonidine that
ture is raised above 300 K the activity and enantioselectiv- are involved in the creation of enantioselective sites on their
ity are progressively lost. The reactions show some inter- adsorption at the Pt surface have been identified (15). The
esting differences. First, for catalysts capable of providing 1 : 1-interaction between cinchona alkaloid and pyruvate
hydroxyketone hydrogenation, enantioselective diketone ester that results in (i) selective enantioface adsorption of
hydrogenation is not rate-enhanced; by contrast single- reactant and (ii) conversion to enantiomerically enriched
stage pyruvate hydrogenation is normally rate-enhanced product has been modelled and visual representations of
over modified catalysts. Second, the enantiomeric excess in the proposed molecular arrangements presented (15). An
diketone hydrogenation is not conversion dependent over analogous modelling study for the 1 : 1-interaction of bu-
the first 20% or so of reaction as is the case in pyruvate tanedione with these alkaloids in their appropriate lowest
hydrogenation (14, 16–18). Third, the initial enantiomeric energy states has been undertaken (21) and again the inter-
excess in dione hydrogenation varied with the amount of action energies reveal that selective enantioface adsorption

ENANTIOSELECTIVE HYDROGENATION V
273
diol production was much slower in reactions in alkanols,
acetone, toluene, and chlorobenzene (1,21). Second, kinetic
selection was observed whereby, with cinchonidine as mod-
ifier, S-hydroxyketone was preferentially hydrogenated to
diol, whereas with cinchonine the R-hydroxyketone was
selectively removed. Thus, the kinetic selection was also
manifested in the appearance of an enantiomeric excess
in 3-hydroxybutan-2-one when a racemic mixture was hy-
drogenated over modified catalyst. Third, dimers formed
in stage 1 of diketone hydrogenation were consumed in
stage 3. These characteristics are discussed in detail.
Choice of solvent. Hydroxybutanone was weakly ad-
sorbed by comparison with butanedione, as evidenced by
the fact that its hydrogenation occurred only after near-
complete removal of the diketone. In principle, the adsorp-
tion of reactant and all intermediates occurs in competition
with solvent, and the failure to obtain useful rates of diol
formation in other solvents may signify ineffective read-
sorption of hydroxybutanone onto the catalyst surface un-
der those conditions. At the commencement of this work
FIG. 3. Schematic representation of the adsorption of butane-2,3-
dione as its enantiofaces (A) and (B) on the enantioselective site adja-
cent to an L-shaped adsorbed cinchonidine molecule. N represents the
quinuclidine-N atom. The greater steric interaction between enantioface there was no evidence in the literature for the dissociative
(
B) and the alkaloid disfavours this adsorbed state, and hence the for-
adsorption of dichloromethane on Pt; it was considered in-
ert and a good solvent for the purpose (1). More recently,
reports have appeared of the dissociative adsorption of
haloethanes on copper (22,23) but to our knowledge there
are still no reported investigations of dichloromethane ad-
sorption on Pt. It remains our view that the effectiveness
of dichloromethane as a solvent for this reaction is related
to the ease with which hydroxybutanone can readsorb from
this solvent onto the Pt surface after the removal of butane-
dione.
mation of S-product. The diketone is adsorbed by both carbonyl groups;
adsorption sites are envisaged as Pt atoms in a (111)-surface and are rep-
resented by asterisks.
of the reactant is expected to occur in the vicinity of ad-
sorbed cinchonidine and cinchonine. A schematic repre-
sentation in Fig. 3 shows that adsorption by enantioface
A at the site adjacent to the quinuclidine-N atom of L-
shaped adsorbed cinchonidine involves less steric repulsion
than that by enantioface B. Adsorption by enantioface A is
Kinetic selection. Dynamic kinetic resolution of race-
therefore preferred over that by enantioface B on energetic mic mixtures has been discussed widely in the literature
grounds, but not to the exclusion of B. The reverse situa- (24–26).
tion applies to the cinchonine-modified surface, where the
(
a) Kinetic selection in hydroxybutanone hydrogenation.
An enantiomeric excess, modest at 273 K but substantial at
93 K, appeared in the hydroxybutanone as it was hydro-
alkaloid is L-shaped in the opposite sense (not shown). On
hydrogenation at the cinchonidine-modified surface, enan-
tioface A provides R-hydroxybutanone and enantioface B
provides the S-enantiomer; the observed sense of the enan-
tioselectivity is thereby interpreted. Any kinetic distinction
in the rates of H-atom addition to enantiofaces A and B
is insufficient to invalidate thermodynamic control of the
enantioselective outcome of this first stage of reaction.
The enantioselective hydrogenation of hexane-2,3-dione
2
genated (Table 5). On the assumption that the usual inte-
grated first-order rate equations apply, the ratio of the rate
coefficients for the removal of R- and S-hydroxybutanone
is given by Eq. [1], where [R]0 and [S]0 are the
ln{[R]0/[R]t }/ln{[S]0/[S]t } = kR/kS
[1]
and hexane-3,4-dione provided values of the enantiomeric concentrations of R- and S-hydroxybutanone at time zero
excess that concurred with those obtained for butane-2,3- and [R]t and [S]t are the values at a later time t. Substitution
dione under comparable conditions. The mechanism pre- of values from Table 5 into Eq. [1] gives kS = 1.3kR at 273 K
sented in Fig. 3 accommodates these observations since the for cinchonidine as modifier and kR = 1.1kS for cinchonine
1
: 1-interaction is only mildly sensitive to a lengthening of as modifier. The higher value of kS/kR observed for cincho-
the linear hydrocarbon chain.
The most novel features of this reaction reside in the (and with the accompanying increasing reaction rate).
second and third stages of the hydrogenation. First, signifi- Figure 4 shows that the reactions having the rate coeffi-
nine at 293 K (Table 5) diminished with increasing pressure
cant hydrogenation of hydroxybutanone to butanediol oc- cients kR and kS each contribute two pathways to products,
curred in reactions conducted in dichloromethane, whereas and the fraction of product formed along each pathway, f1,

2
74
SLIPSZENKO ET AL.
(
b) Kinetic selection in the second stage of butanedione
hydrogenation. The rate at which the enantiomeric ex-
cess changed in the second stage of butanedione hydro-
genation at 273 K was greater than that observed in the
hydrogenation of hydroxybutanone at the same tempera-
ture (compare Tables 1 and 5); i.e., the kinetic effect was
more pronounced. By use of Eq. [1] and information in
Table 1, the values of kS/kR for entries 5 and 6 (cinchoni-
dinemodification)are2.5and1.9andthevalueforkR/kS for
entry 9 (cinchonine modification) is 1.7. Table 6 shows that
the fraction factors that accompany these ratios do not dif-
ferentiate between f1 and f2 for cinchonidine modification
but favour f3 over f4, and correspondingly do not differ-
entiate between f3 and f4 for cinchonine modification but
favour f2 over f1. It is notable that there is more severe dis-
crimination in hydroxybutanone hydrogenation than in the
FIG. 4. Processes for the conversion of R- and S-3-hydroxybutan- second stage of butanedione hydrogenation. That is, the ad-
2
-one (HB) to RR-, meso, and SS-butane-2,3-diol.
layer of alkaloid is more discriminating towards hydroxy-
ketone if it has not been previously subject to the effects
of the more strongly adsorbed diketone. This suggests that
f2, f3, and f4 can be determined from a knowledge of the the diketone damaged or disorganised the original alkaloid
enantiomeric composition of the butanediol. Table 6 shows ad-layer during the first stage of reaction.
the f-values used in these calculations and the calculated
Dimer formation and removal. The three partially hy-
and observed values of the yield of meso-butanediol and
drogenated dimers were initial products of butanedione
of the [RR-] : [SS-] ratio. At the enantioselective site ad-
hydrogenation over cinchona-modified catalysts and were
jacent to cinchonidine, R-hydroxybutanone is more likely
formed in reactions over unmodified catalysts to which the
to undergo the process creating a second chiral centre of
base3-quinuclidinolhadbeenadded. Theywerenotformed
R-configuration than one of S-configuration although the
in simple solutions of cinchonidine, nor in solutions of bu-
tanedione in dichloromethane under hydrogen in the ab-
sence of catalyst. Furthermore, they were not formed in
butanedione hydrogenation to racemic products over un-
distinction is not great, f1 : f2 = 3 : 1. However, the adsorp-
tion of S-hydroxybutanone at the same site adjacent to (R-
directing) cinchonidine results in a very weak propensity to
create a further chiral centre of S-configuration, most of the
modified catalysts or in hydroxybutanone hydrogenation
molecules undergoing the R-directed step to form meso-
under any conditions. The mechanism set out in Fig. 5 is
butanediol, f3 : f4 = 14 : 1. Remembering that the chiral out-
consistent with these observations. It is proposed that pairs
of adsorbed butanedione molecules (I) couple by a base-
inducedH-transferinvolvingonemoleculeintheenolform,
to give (II). Repetition of this process would give (III). Ad-
dition of one mole of hydrogen to (II) gives two dimers
of closely related structure, (IV) and (V). The third dimer,
comes are reversed for cinchonine-modified reaction, there
is little distinction between paths 3 and 4 with this modifier
(
f4/f3 = 1.5) but still the marked preference for path 2 by
comparison with path 1 (f2/f1 = 14).
(
VI), may be formed by hydrogenation of (III) or, possibly,
TABLE 6
directly from (V). The C4- and C2-ions formed by fragmen-
Calculated and Observed Values of the [RR-] : [SS-] Ratio and tation of the dimers in the mass spectrometer are consistent
of the Meso-Butanediol Yield Based on the Experimental Values
with the structures shown in (IV), (V), and (VI), and the in-
terpretation of the ion fragmentations is consistent with the
a
of kR/kS and Optimised Values of fi
measured spectrum for acetoin dimer which is the molecule
formed by hydrogenation of (VI) (compare Fig. 1c).
f1
f2
f3
f4
[RR-] : [SS-]
Meso-diol/%
Obs Calc
Initial
reactant Modifier
b
/%
/%
Obs
Calc
Conjugation involving the carbon-carbon double bond
and the lone pairs on the adjacent O-atom serves to sta-
bilise intermediates (II) to (VI). When (IV), (V), and (VI)
undergo saturation of the carbon–carbon double bond con-
jugative stabilisation is lost and the molecule becomes
susceptible to dissociation. As noted above, species (VI),
on such hydrogenation becomes “acetoin dimer”; gc and
HB
HB
BD
BD
CD
CN
CD
CN
75 25 93
7
50 50 80 20
20 80 50 50
7
8 : 1
9.6 : 1
62
62
66
64
66
93 40 60
1 : 8
1 : 9.1 65
2 : 1
2.6 : 1
63
1 : 1.5
1 : 2.2 68
a
The values fi represent the fractions of molecules undergoing step i
in Fig. 4.
b
1
HB = hydroxybutanone; BD = butanedione.
NMR experiments with an authentic sample showed that

ENANTIOSELECTIVE HYDROGENATION V
275
tained were indeed enantioselective and of high activity,
from which it is concluded that oxygen and propyne play
common roles in these aerobic and anaerobic procedures,
respectively. The removal of adsorbed propyne as propene
and propane in the early stages of reaction has been con-
firmed in the context of pyruvate hydrogenation (26) and
will be reported in a later paper in this series.
The effects observed are several and inter-related. First,
one effect of preconditioning the surface with propyne was
to increase the rate of racemic diketone hydrogenation over
�
1
� 1
unmodified catalyst from 40 to 1500 mmol h g . It is well
known that ethyne and propyne adsorb dissociatively on
Pt and other Group 8 metals to give permanently retained
hydrocarbonaceous residues that promote H-atom transfer
between the surface and reactive hydrocarbon intermedi-
ates (27–29). This fast hydrogenation to racemic product
over propyne-conditioned catalyst is attributable to the ef-
fect of such a hydrocarbonaceous ad-layer. Second, catalyst
modified under propyne was active for the second stage of
butanedione hydrogenation, a feature not shared by un-
modified catalyst. Thus, hydroxybutanone hydrogenation
required the presence of adsorbed alkaloid and the pro-
cess was specifically facilitated at the enantioselective site.
However, the process was also dependent on conditioning
by propyne specifically, because modification under buta-
FIG. 5. Proposedmechanismforthedimerisationofbutane-2,3-dione
and for the subsequent hydrogenation and dissociation of the dimers.
1
,3-diene did not provide second stage activity. Third, the
(
VI) instantly and completely dissociates to hydroxybu- two enantioselective catalyst systems reported here that
tanone in dichloromethane solution at room temperature. provide rapid second stage activity in diketone hydrogena-
Thus, in the third stage of reaction, (V) and (VI) are hy- tion do not show the normally observed rate enhancement
drogenated and dissociate to hydroxybutanone, and (IV) in respect to activity for the first stage. One factor is that
is analogously converted to butanedione and butanediol. conditions for fast reaction to racemic product have been
The latter accounts for the traces of butanedione present achieved. However, thisfeaturemeritsfurtherinvestigation
in the very final stages of reaction when the dimers have because the observations and interpretations of enantiose-
disappeared. The hydroxybutanone that accompanied this lectivity and rate enhancement have, until now, been closely
butanedione was racemic which indicates that hydrodimeri- linked, and the reactions described here appear to provide
sation, which involved the creation of various chiral centres, exceptions to a pattern of behaviour that has come to be
was not directed by the chiral environment of the alkaloid accepted as the norm.
modifier.
Note added in proof. After submission of this paper an
investigation of butane-2,3-dione hydrogenation in toluene
solution at 273–298 K and 25 to 135 bar hydrogen pres-
B. Reactions in Ethanol Solution
Reactions were fast when catalyst samples were modified sure catalysed by Pt/alumina modified by dihydrocinchoni-
under aerobic conditions and when solvent and organic re- dine was reported in Chemical Communications by Studer,
actant contained normal concentrations of dissolved air. Okafor, and Blaser (30). The reactions described showed
Stringent exclusion of air caused a 90% reduction in rate enantioselectivity in the first stage of reaction and simi-
and a 60% reduction in enantiomeric excess. In this respect, lar kinetic resolution in the second stage. However, special
diketone hydrogenation followed pyruvate hydrogenation procedures had to be adopted to achieve convenient rates
as reported in Part II (13). As a hypothesis, we suppose and kinetic resolution in the second stage (addition of fresh
that competitive co-adsorption of oxygen and alkaloid pre- modifier and catayst) and no dimer formation was observed
vents alkaloid achieving coverages at which it acts as a poi- (31).
son. Enantioselective sites then become available as ad-
sorbed oxygen is removed by hydrogen in the early stages
ACKNOWLEDGMENTS
of reaction. In developing the process of modification un-
der propyne, the objective was to provide an alternative
The authors thank EPSRC for three studentships (to JAS, SPG and
WAHV), Johnson Matthey for financial, material, and scientific support,
strong co-adsorbent for the alkaloid. The catalysts so ob- and Dr I Theaker for valuable assistance.

2
76
SLIPSZENKO ET AL.
REFERENCES
14. Sutherland, I. M., Ibbotson, A., Moyes, R. B., and Wells, P. B., J. Catal.
25, 77 (1990).
. Vermeer, W. A. H., Fulford, A., Johnston, P., and Wells, P. B., J. Chem. 15. Simons, K. E., Meheux, P. A., Griffiths, S. P., Sutherland, I. M.,
1
1
Soc. Chem. Commun., 1053 (1993).
Johnston, P., Wells, P. B., Carley, A. F., Rajuman, M. K., Roberts,
2
. Blaser, H-U., and Muller, M., in “Heterogeneous Catalysis and Fine
M. W., and Ibbotson, A., Recl. Trav. Chim. Pays-Bas 113, 465 (1994).
Chemicals II” (M. Guisnet et al., Eds.), p. 73. Elsevier, Amsterdam, 16. Wehrli, J. T., Baiker, A., Monti, D. M., Blaser, H-U., and Jalett, H. P.,
1
991.
J. Mol. Catal. 57, 245 (1989).
3
4
5
. Webb, G., and Wells, P. B., Catal. Today 12, 319 (1992).
. Baiker, A., J. Mol. Catal. 115, 473 (1997).
. Blaser, H-U., Jalett, H. P., Muller, M., and Studer, M., Catal. Today 37,
17. Singh, U. K., Landau, R. N., Sun, Y., LeBlond, C., Blackmond, D. G.,
Tanielyan, S. K., and Augustine, R. L., J. Catal. 154, 91 (1995).
18. Wang, J., Sun, Y., LeBlond, C., Landau, R. N., and Blackmond, D. G.,
J. Catal. 161, 752 (1996).
4
41 (1997).
6
7
. Pfaltz, A., and Heinz, T., Topics Catal. 4, 229 (1997).
. Vermeer, W. A. H., Fulford, A., Johnston, P., Wells, P. B., and de Wit,
19. Bond, G., Simons, K. E., Ibbotson, A., Wells, P. B., and Whan, D. A.,
Catal. Today 12, 421 (1992).
A. M., poster presentation, in “Koninklijke Nederlandse Chemische 20. Bond, G., and Wells, P. B., J. Catal. 150, 329 (1994).
Vereniging ‘Zomer Congres’, Amsterdam, 1993.” 21. Vermeer, Ph.D. thesis, University of Hull, 1995.
. Griffiths, S. P., Johnston, P., Vermeer, W. A. H., and Wells, P. B., J. 22. Kerkar, M., Walter, W. K., Woodruff, D. P., Jones, R. G., Ashwin,
Chem. Soc. Chem. Commun., 2431 (1994). M. J., and Morgan, C., Surf. Sci. 268, 36 (1992).
. Griffiths, S. P., Wells, P. B., Griffin, K. G., and Johnston, P., Paper 23. Kadodwala, M. F., Davis, A. A., Scragg, G., Cowie, B. C. C., Kerkar,
8
9
No. 8, in “17th Conference of the Organic Reactions Catalysis Society,
New Orleans, March 1998.”
M., Woodruff, D. P., and Jones, R. G., Surf. Sci. 392, 199 (1997).
24. Kagan, H. B., and Fiaud, J. C., Topics Stereochem. 18, 249 (1988).
1
0. Geus, J. W., and Wells, P. B., Appl. Catal. 18, 231 (1985). [Associated 25. Noyori, R., Tokunaga, M., and Kitamura, M., Bull. Chem. Soc. Jpn.
papers in the same volume] 68, 36 (1995).
1
1. Jackson, S. D., Keegan, M. B. T., McLellan, G. D., Meheux, P. A., 26. Wells, P. B., and Wilkinson, A. G., Topics Catal. 5, 39 (1998).
Moyes, R. B., Webb, G., Wells, P. B., Whyman, R., and Willis, J., in 27. Webb, G., Catal. Today 7, 139 (1990).
“
Preparation of Catalysts, V” (G. Poncelet et al., Eds.), p. 135. Elsevier, 28. Berndt, G. F., Thomson, S. J., and Webb, G., J. Chem. Soc. Faraday
Amsterdam 1991. Trans I 79, 195 (1983).
2. Orito, Y., Imai, S., and Niwa, S., Nipon Kagaku Kaishi, 670 29. Somorjai, G.A., and Zaera, F., J. Phys. Chem. 86, 3070 (1982).
1
1
(
1980).
3. Meheux, P. A., Ibbotson, A., and Wells, P. B., J. Catal. 128, 387
1991).
30. Studer, M., Okafor, V., and Blaser, H-U., J. Chem. Soc. Chem. Com-
mun., 1053 (1998).
(
31. Blaser, H-U., personal communication, 1998.