Enantioselective hydrogenation of 1-phenyl-1,2-propanodione on cinchonidine-modified Rh/MCM-41 catalysts

Article abstract of DOI:10.1016/j.molcata.2014.06.002

Enantioselective hydrogenation of 1-phenyl-1,2-propanodione (PPD) on Rh/MCM-41 catalysts was studied. The catalysts were prepared using Rh(acac) ₃ as a metal precursor and metal loadings that ranged from 0.5 to 1.5 wt%. The samples were characterised by nitrogen adsorption-desorption isotherms at 77 K using ICP-MS, XRD, TEM, XRD and XPS. The reaction was performed at 298 K, and 0.01 mol L^-1 of PPD and cinchonidine (CD) was used as chiral modifier. The effect of the modifier concentration on the enantioselectivity (ee) and the conversion level in the hydrogenation reaction shows a relationship between the activity and ee as well as the CD concentration. The volcano-type curve observed for each catalyst suggests competitive adsorption of the modifier and substrate on the catalyst surface. The metal loading influences both the Rh crystallite size and catalytic behaviour. An increase in the Rh levels was accompanied by a parallel increase in both the crystallite size and number of Rh ensembles with a subsequent enhancement in both the conversion level and ee. The catalyst with the highest metal loading was 1.0 wt% Rh/MCM-41, and it displayed the highest ee among the catalysts studied. Additionally, the effects of hydrogen pressure and solvent nature on the catalytic activity were also studied. Moreover, the H₂ pressure also influenced the conversion levels with only a minor effect on the ee values. Considering the solvent effects, we observed non-linear ee dependence with an increasing solvent dielectric constant, which showed a decrease in conversion levels in the order cyclohexane > toluene > tetrahydrofurane > dichloromethane.

Full text of DOI:10.1016/j.molcata.2014.06.002

Accepted Manuscript
Title: Enantioselective Hydrogenation of
1-Phenyl-1,2-Propanodione on Cinchonidine-Modified
Rh/MCM-41 Catalysts
Author: Cecilia C. Torres Cristian H. Campos J.L.G. Fierro
Patricio Reyes Doris Ruiz
PII:
S1381-1169(14)00248-9
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.molcata.2014.06.002
MOLCAA 9138
To appear in:
Journal of Molecular Catalysis A: Chemical
Received date:
Revised date:
Accepted date:
25-3-2014
19-5-2014
1-6-2014
Please cite this article as: C.C. Torres, C.H. Campos, J.L.G. Fierro, P. Reyes, D.
Ruiz, Enantioselective Hydrogenation of 1-Phenyl-1,2-Propanodione on Cinchonidine-
Modified Rh/MCM-41 Catalysts, Journal of Molecular Catalysis A: Chemical (2014),
http://dx.doi.org/10.1016/j.molcata.2014.06.002
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Enantioselective Hydrogenation of 1-Phenyl-1,2-Propanodione on Cinchonidine-
Modified Rh/MCM-41 Catalysts
Cecilia C. Torresa, Cristian H. Camposa, J.L.G. Fierrob, Patricio Reyesa and Doris Ruiza*
aDepartamento de Físicoquímica, Facultad de Ciencias Químicas, Universidad de
Concepción, Casilla 160-C, Concepción, Chile
bInstituto de Catálisis y Petroleoquímica, CSIC, Cantoblanco, 28049 Madrid, Spain
*doruiz@udec.cl
Highlights
 Heterogeneous enantioselective hydrogenation of 1-phenyl-1,2-propanodione over
Rh/MCM-41 catalysts was generated.
 The enantioselectivity for R-1-hydroxy-1-phenyl-2-propanone was approximately 70%.
 The confinement effect of the MCM-41 substrate enhances both activity and selectivity.
 Catalytic performance is significantly affected by the H2 pressure and solvent type.
1
Page 1 of 32

Abstract
Enantioselective hydrogenation of 1-phenyl-1,2-propanodione (PPD) on Rh/MCM-41
catalysts was studied. The catalysts were prepared using Rh(acac)3 as a metal precursor and
metal loadings that ranged from 0.5 to 1.5 wt%. The samples were characterised by
nitrogen adsorption–desorption isotherms at 77 K using ICP-MS, XRD, TEM, XRD and
XPS. The reaction was performed at 298 K, and 0.01 mol L-1 of PPD and cinchonidine
(CD) were used as chiral modifiers. The effect of the modifier concentration on the
enantioselectivity (ee) and the conversion level in the hydrogenation reaction shows a
relationship between the activity and ee as well as the CD concentration. The volcano-type
curve observed for each catalyst suggests competitive adsorption of the modifier and
substrate on the catalyst surface. The metal loading influences both the Rh crystallite size
and catalytic behaviour. An increase in the Rh levels was accompanied by a parallel
increase in both the crystallite size and number of Rh ensembles with a subsequent
enhancement in both the conversion level and ee. The catalyst with the highest metal
loading was 1.0 wt% Rh/MCM-41, and it displayed the highest ee among the catalysts
studied. Additionally, the effects of hydrogen pressure and solvent nature on the catalytic
activity were also studied. Moreover, the H2 pressure also influenced the conversion levels
with only a minor effect on the ee values. Considering the solvent effects, we observed
non-linear ee dependence with an increasing solvent dielectric constant, which showed a
decrease in conversion levels in the order cyclohexane > toluene > tetrahydrofurane >
dichloromethane.
Keywords: Rh catalysts; Hydrogenation; 1-phenyl-1,2-propanedione; Enantioselectivity.
2
Page 2 of 32

1. Introduction
Recently, asymmetric hydrogenation reactions over modified heterogeneous catalysts have
been extensively studied. Complex mechanisms have been proposed for these reactions,
which consider that adding a chiral template molecule is required for chiral induction
during heterogeneous catalytic hydrogenation. The template molecule, which is
cinchonidine (CD) herein, forms chirally modified sites through an interaction with a metal
surface or, in the bulk liquid phase, through a substrate–modifier interaction [1-3]. In
addition, the architecture of the catalyst surface influences the reaction to a certain extent.
The solvent, hydrogen pressure and catalyst pre-treatment notably influence both the
reaction rate and enantioselectivity [4-11]. Heterogeneous enantioselective hydrogenation
of α-ketoesters has been intensively studied, whereas enantioselective hydrogenation of
diones is less well-studied [12-15].
Enantiomerically pure hydroxyketones are used as intermediates for synthesising
antifungals and chiral amino alcohols. Azole antifungals are interesting as drugs for AIDS-
related diseases, and chiral amino alcohols (e.g., (1RS, 2SR)-(±)-2-amino-1-phenyl-1-
propanol) are used as good vasoconstricting agents and nasal decongestants [16]. The most
representative substrate is 1-phenyl-1,2-propanedione (PPD) in Scheme 1, which is
hydrogenated to (R or S)1-hydroxy-1-phenyl-2-propanone, and (R or S) 2-hydroxy-1-
phenyl-1-propanone and has been used as a model compound for heterogeneous
enantioselective hydrogenation by a few research groups [16].
3
Page 3 of 32

Scheme 1. 1-Phenyl-1,2-propanodione (PPD) hydrogenation products. 1R-PP: (R)-1-
phenyl-1-hydroxi-2-propanone, 1S-PP: (S)-1-phenyl-1-hydroxi-2-propanone, 2R-PP: (R)-
2-hydroxi-1-phenylpropanone, 2S-PP: (S)-2-hydroxi-1-phenylpropanone, Diols: (1R,2S)-1-
phenyl-1,2-propanediol or (1R,2R)- 1-phenyl-1,2-propanediol; (1S,2R)-1-phenyl-1,2-
propanediol and (1S,2S)-1-phenyl-1,2-propanediol.
In various enantioselective hydrogenation studies, Pt is used as the metal for carbonyl
group reduction [8, 10, 17], whereas Pd is preferred for C=C hydrogenation using
cinchonidine as a chiral modifier [18-20]. Other transition metals, such as Ir, Rh or Ru,
have been most studied in homogeneous phases [21, 22]. These systems are generally chiral
phosphine–metal complexes and represent an effective and highly studied area, but
4
Page 4 of 32

recovery and separation is a highly desirable goal. Recently, it was confirmed that Rh is an
effective metal for heterogeneous enantioselective hydrogenation of certain prochiral
compounds using nanoparticles stabilised with chiral ligands. Our group reported that
chirally modified rhodium nanoparticles are effective for enantioselective hydrogenation
[23, 24]; however, the enantioselectivities appeared low or, at best, moderate for PPD
hydrogenation. The main features of the chirally modified rhodium were compared with the
corresponding reported platinum-CD systems.
To improve the performance of heterogeneous Rh catalysts, we used the mesostructured
MCM-41 system as a substrate for the catalytic component. This system may have potential
applications for catalysis as an enantioselective catalyst for asymmetric hydrogenation.
These enantioselective catalysts are essentially composed of a hexagonal arrangement of
non-connected channels with a high porous structure, narrow pore distribution, high
specific surface area and the appropriate pore diameter for processing large substrates.
Certain authors have reported using Pt and Pd nanoparticles deposited on MCM-41 as
enantioselective catalysts for asymmetric hydrogenation [25, 26]. The better catalytic
performance obtained using these catalysts was explained through the effect of the substrate
mesostructure because the straight channels yield a confinement effect that leads to
enhanced catalytic activity compared with other amorphous supported metal catalysts. The
aim of the present work is to obtain and evaluate Rh catalysts supported on a
mesostructured MCM-41 substrate to which a CD modifier and different metal loading
were incorporated. In addition, the effects of the H2 pressure and solvent on catalytic
performance were also studied using the most enantioselective catalyst.
5
Page 5 of 32

2. Experimental
2.1 Chemicals
PPD, CD, cyclohexane, toluene, tetrahydrofurane (THF), dichloromethane, rhodium
acetylacetonate (Rh(acac)3), aerosil silica (99.8%), tetramethylammonium hydroxide
(TMAOH) and cetyltrimethylammonium bromide (CTMABr) were supplied by Aldrich, as
received, and H2 (99.995%) was supplied by AGA.
2.2 Catalyst synthesis and characterisation
The support MCM-41 was synthesised using the reaction of an aqueous solution of
TMAOH and CTMABr with aerosil silica under continuous stirring. The resulting
composition (molar ratio) of the homogeneous gel (pH= 11) was SiO2:0.15 CTMABr:0.26
TMAOH:24.3 H2O and was dried as well as calcined for 1 h in N2 followed by 10 h under
dynamic air flow at 810 K. The catalysts were prepared through impregnation using a slight
excess of Rh(acac)3 solution to fill the pore volume of the supports at the required levels to
generate 0.5, 1.0 and 1.5 wt.% of Rh. All catalysts were calcined under air at 573 K and
reduced in H2 at 773 K for 1 h prior to their characterisation or catalytic tests.
The elemental content was quantitatively determined for the selected fresh catalysts
through inductively coupled plasma spectrometry using a Perkin Elmer ICP-MS instrument
after microwave-assisted digestion of the samples in a nitric/hydrochloric acid solution. The
specific area and porosity were obtained from nitrogen adsorption–desorption isotherms at
77 K performed in an automatic Micromeritics apparatus Model ASAP 2010 at a relative
pressure in the range 0.05–0.995. The BET surface areas were calculated using a linear
section of the BET plot based on IUPAC parameters. Pore-size distributions were
6
Page 6 of 32

calculated from the N2 adsorption branch using the Barrett-Joyner-Halenda model (BJH).
The powder X-ray diffraction patterns were recorded using Cu-Kα radiation (λ = 1.542 Å)
and a Rigaku diffractometer equipped with a rotating anode operated at 45 kV and 100 mA.
The counts were accumulated every 0.02° (2θ) at the scan speed 1° 2θ min−1. The
interaction of hydrogen with the rhodium catalysts was studied in temperature-programmed
reduction experiments (TPR), the catalysts were reduced under continuous flow using a 5%
H2/Ar mixture (40 cm3 min−1), at 10 K min−1 from room temperature up to 873 K in
TPR/TPD2900 Micromeritics system. Hydrogen consumption was monitored with the aid
of a TCD detector and recorded for TPR analysis.
TEM micrographs were obtained in a Jeol Model JEM-1200 EXII System. The samples
were loaded onto a holey carbon film that was supported on a copper grid by dipping the
grid into a sample suspension in ethanol. The catalyst samples were prepared using the
extractive replica procedure. Up to 300 individual metal particles were counted for each
catalyst, and the surface area-weighted mean Rh diameter (dp) was calculated using the
following equation:
3
∑
∑
ꢀꢀꢀꢀꢀꢀꢀꢀ
(1)
ꢀꢀ
ꢀꢀ =
ꢀꢀ
2
ꢀꢀꢀꢀꢀꢀꢀꢀ
ꢀꢀ
Where ni is the number of particles, and di is the diameter. The size limit for detecting the
Rh particles on the samples was ca. 1 nm.
The photoelectron spectra (XPS) were recorded using an Escalab 200R spectrometer with a
hemispherical analyser operated using a constant pass energy mode and Mg Kα X-ray
radiation (hν = 1253.6 eV) at 10 mA and 12 kV. The charge effects on the sample were
corrected by fixing the binding energies (BE) for the Si2p core-level at 103.4 eV. The
7
Page 7 of 32

surface Rh/Si ratios were estimated using the integrated intensities of Rh 3d and Si 2p lines
after the background was subtracted and atomic sensitivity factors were used for
normalisation [27].
2.3. Activity Tests
The catalytic assays for PPD hydrogenation were performed in a stainless steel, Parr-type
batch reactor. The substrate concentration was 0.01 mol L-1 in cyclohexane as the solvent.
The samples were stirred at 700 rpm under 20 bar of H2 pressure. All further experiments
were performed without external mass transfer limitations. The catalysts used were in the
form of fine powders (> 30 μm), and a negligible effect from the pore diffusion limitations
was assumed. In catalyst mass studies, the (PPD/Rh) molar ratio was varied, the PPD
concentration was constant, and the catalyst mass was varied. An insignificant difference
was obtained for the initial rates in the catalyst mass range studied (0.01-0.07 g) because
the reaction rate was proportional to the catalyst mass. These observations indicate that the
gas–liquid and liquid–solid external mass transfer limitations were absent [6, 28].
Liquid hydrogenation samples were periodically collected from the reactor and analysed
using a gas chromatograph–mass spectrometer (GCMS-QP5050 Shimadzu) with a chiral β-
dex 225 column (30 m; Supelco) and electrolytic helium as carrier gas. The activity of the
catalyst was followed by conversion levels:
(2)
where [PPD]0 correspond to initial concentration of the PPD and [PPD]t as the
concentration
at
different
times.
8
Page 8 of 32

The turnover frequency (TOF) of the catalysts was calculated as follows.
(3)
(4)
The selectivity was measured by enantiomeric excess (ee) using the following equation:

R

xx



S

xx
eeCx (%) 
100%

R
S
where [R] and [S] are the concentrations of the major and minor enantiomer, respectively.
The regioselectivity (rs) was calculated as the ratio between (R)+(S)-1-hydroxy-1-
phenylpropanone over (R)+(S)-2-hydroxy-1-phenylpropanone.
(5)

RC1





SC

SC12 
rs 

RC 2

3. Results and Discussion
3.1 Catalyst Characterisation
The structure of MCM-41 was confirmed through an XRD analysis. Figure 1 shows the
XDR profiles of the samples studied. Diffraction lines corresponding with the planes (100),
(110), (200), and (210) of a hexagonal structure with the symmetry P6 for the reflection
planes (hkl) were observed. These data indicate the hollow and hexagonal form. The
parameters of the cell (ao) for the MCM-41 hexagonal pore arrangement was 48 Å, which is
9
Page 9 of 32

consistent with our previous reported results for the Pt/MCM-41 catalyst [25]. A structural
characterisation of the solid MCM-41 is shown in Table I.
Figure 1. Diffraction pattern of MCM-41.
Table I. Structural characterization of MCM-41 and Rh catalysts.
Pore
diameter
(nm)a
Thickness
of wall
d100
SBET
VT
ao
Sample
(m2/g)
(cc/g)c
(Å)
(nm)b
(nm)
MCM-41
4.1
4.1
4.1
4.1
1151
1110
1059
1001
3.7
3.7
3.6
3.6
0.40
0.40
0.50
0.50
1.2
1.1
1.1
1.1
48
48
48
48
0.5%Rh/MCM-41
1.0%Rh/MCM-41
1.5%Rh/MCM-41
a
b
Determined by the BJH. Determined by the difference between d100 and the pore
diameter. c Volume of pore at P/Po = 0.99.
10
Page 10 of 32

Nitrogen adsorption-desorption isotherms for the support are shown in Figure 2. These
isotherms correspond to type IV (IUPAC classification) and display a H1-type hysteresis
loop [29-31] associated with regular-form compact mesopores that present a narrow pore
size distribution. The abrupt fall of the hysteresis loop indicates that the pores are have a
uniform and cylindrical shape.
Figure 2. Adsorption desorption isotherms of nitrogen on MCM-41 at 77 K.
The surface areas determined by the BET equation show that, upon Rh loading, the textural
properties only change slightly, indicating that Rh incorporation does not significantly
affect the original structure, which is consistent with the XRD results. In addition,
mesostructured catalysts have framework pore sizes in the range 3–5 nm, which are
11
Page 11 of 32

sufficiently large to accommodate the starting reagent and reaction products. Crocker et al.
[32] reported that the critical dimension of phenyl acetylene is essentially determined by
the benzene ring thickness (similar to PPD) and adopts the value 0.3 nm. Therefore, we
inferred that the apparent substrate size dependence arises from the reactant chemisorption
spatial requirements, both on the external surface sites and in the restricted internal region.
Three different catalysts were used for this study on the metal loading effect. Table 2 shows
the ICP analyses and reveals that the catalysts’ metal content is similar to the nominal
content, which indicates that all of the metal was deposited during synthesis of the
catalysts.
Table II. Elemental analysis by ICP-MS and particle mean size for all the catalyst
determinated by TEM.
Rh (wt%)
0.48
Catalyst
dTEM (nm)
2.8
DTEM
0.36
0.31
0.5% Rh/MCM-41
1.0% Rh/MCM-41
0.97
3.2
1.49
1.5% Rh/MCM-41
3.7
0.27
Figure 3 shows the TPR profiles for the calcined Rh/MCM-41 samples. Two reduction
peaks are shown. One peak is at lower temperatures and centred at 353 K, which shows a
similar intensity in all samples and can be associated with reduction of rhodium oxide
particles located on the external mesopore channels. A second peak is located at higher
temperatures (ca. 573 K), which can be due to reduction of rhodium oxide crystallites,
12
Page 12 of 32

which are located inside the mesopores and interact with the substrate. These observations
are consistent with a previous report, which argued that reduction/oxidation of particles
located inside the mesopores of the substrate always requires the most severe reduction
conditions [25]. The intensities of these second reduction peak increase with greater
rhodium loading. The quantity of H2 consumed indicates that level of hydrogen reduced is
close to the stoichiometric values for the samples with higher Rh loading. For the catalyst
containing 0.5 wt% Rh, the level of reduction was approximately 85%.
Figure 3. TPR patterns of bulk phases and supported Rh/MCM-41 catalysts post oxidation
RhxOy/MCM-41.
Figure 4 shows certain representative TEM micrographs and the metal particle size
distribution for the Rh/MCM-41 catalysts studied. A narrow particle size distribution is
13
Page 13 of 32

shown, and the average value of the metal particle sizes obtained using TEM are
summarised in Table 2. As expected, a slight increase in the particle size upon increasing
Rh loading is shown. Most of the crystallite sizes were in the range 2.0 to 4.0 nm.
Figure 4. TEM Micrographs (A) MCM-41, (B) 1.5% Rh/MCM-41 and C) 1.5% Rh/MCM-
41 particle size distribution.
Figure 5 shows the Rh 3d5/2 core level spectra for the catalysts studied. The binding
energies for Rh0, Rh+ and Rh3+ are 307.1 eV, 307.6–309.6 eV and 308.8–311.3 eV,
respectively [33], as shown Table 3.
14
Page 14 of 32

Table III. Binding energy of Si 2p, and O 1s, and Rh 3d5/2 core level for Rh/MCM-41
catalysts.
Rh/Si
(sup)
Rh/Si
(bulk)
Rh 3d5/2
Catalyst
Si2p
O1s
307.2 (70)
309.1 (30)
0.5% Rh/MCM-41
103.4
532.8
0.0018
0.0016
307.1 (80)
309.0 (20)
1.0% Rh/MCM-41
1.5% Rh/MCM-41
103.4
103.4
532.8
532.8
0.0037
0.0030
307.2 (86)
309.0 (14)
0.0049
0.0047
All Rh3d core-level profiles showed two doublets, the major Rh3d5/2 peaks for which
appeared at approximately 307.2 and 309.1 eV. These observations support the notion that
the Rh component in these catalysts exists in two different chemical environments: BE at
ca. 307.1 eV, which is associated with metallic Rh0, and the BE at 39.1 eV, which
originated from the oxidised Rhδ+ species. These values may be explained by considering
that the binding energy is affected not only by an electronic interaction between the metal
and other catalyst components but also by the small particle size of the metal nanoparticles,
as we have reported for other MCM-41 catalytic systems [25, 26]. It has also been shown
that the concentration of Rhδ+ species decreases with greater metal loading on the support
surface. Considering the metal particle sizes obtained using TEM, as shown in Table 2,
only a few changes in the BE should be expected for the 1.0% Rh and 1.5 % Rh catalysts.
Additionally, the Rh/Si atomic surface ratios obtained through XPS showed comparable
values for both samples, which is consistent with the TEM results.
15
Page 15 of 32

Figure 5. XPS spectra of Rh 3d core level for Rh/MCM-41 catalysts.
3.1 . PPD hydrogenation
Preliminary experiments using small catalyst particle sizes (< 200 μm) and a high stirring
speed (800 rpm) precluded the absence of external and internal mass transfer limitations. In
previous studies, we demonstrated that this reaction is a structure-sensitive reaction. In fact,
the ability of the support surface to disperse the metal restricts the metal particle size and
affects the activity and enantioselectivity; however, the average metal particle size can be
modified by changing the metal loading.
3.1.1 1.0% Rh/MCM-41: Modifier concentration
To improve enantioselectivity for a given system, a study on the effect of the modifier
concentration is relevant. Margitfalvi et al. [34-38] reported that the modifier concentration
16
Page 16 of 32

dependence forms a bell-type curve during ethyl pyruvate hydrogenation. Similar results
have been published by Toukoniitty et al. [9, 39].
Figure 6 displays the change in conversion level. In all cases, pseudo-first-order kinetics
with respect to PPD were observed. The reaction rate maximises as the modifier
concentration increased (Table 4). The main product was 1-hydroxy-1-phenyl-2-propanone
(1R- and 1S-PP), and the regioisomer 2-hydroxy-1-phenyl-1-propanone (2R- and 2S-PP)
was formed in minor quantities at the CD concentration > 5.0 x 10-4 mol L-1. The
regioselectivity is approximately 43%, as shown in Table 4. This behaviour is likely due to
higher steric hindrance from the spherical methyl group compared with the flat phenyl ring.
Figure 6. Activity curves based on PPD conversion for 1.0%Rh/MCM-41 catalyst at
different CD concentrations until 120 minutes. Reaction conditions: PPD concentration:
0.01 mol L-1, catalyst mass: 0.050 g, PH2: 20 bar, mixing speed: 700 rpm and solvent:
cyclohexane.
17
Page 17 of 32

Table IV. Catalytic data for PPD hydrogenation over Rh/MCM-41 catalyst at different CD
concentrations. Reaction conditions: PPD concentration: 0.01 mol L-1, PH2: 20 bar, catalyst
mass: 0.050 g, stirring speed: 700 rpm and solvent cyclohexane.
Rh loading (%)
CD
0.5
kg
1.0
kg
1.5
kg
concentration
( 10-3 mol L-1)
X(%) (min-1
rs
X(%) (min-1
rs
X(%) (min-1
rs
-1
-1
-1
gmetal
)
gmetal
)
gmetal
)
0.10
0.25
0.50
0.75
1.00
76.9
79.1
81.1
59.1
49.2
34.4
48
54.2
66.5
67.8
73.1
69.2
5.6
-
78.9
76.6
83.3
81.5
74.2
11.3
5.4
48.4
59.2
15.6
11.2
26
25
-
10.4
17.2
11.0
8.8
-
-
14.4
23.3
15.6
13.3
3.7
6.8
6.1
4.8
-
-
43
The hydrogenation rates of hydroxyketones to diols were not reported because these
reactions did not occur under the conditions herein. The maximum enantioselectivity was
observed at moderate modifier concentrations (0.5 mmolL-1) than the maximum
hydrogenation rate, as shown in Figure 7. However, it should be emphasised that the
maximum appears relatively close to the 2:1 molar ratio of modifier to surface Rh and away
from the 0.05:1 ratio of modifier to reactant in the bulk liquid. The enantiomeric excess of
18
Page 18 of 32

1R-PP was approximately 54%, and it was almost constant with increasing PPD
conversion. These results indicate that the modifier not only induces chirality but also
enhances the activity, which is consistent with Margitfalvi et al. [36, 38] for ethyl pyruvate
hydrogenation and our previous work on PPD hydrogenation using Pt/MCM-41 catalysts
[25]. At a lower CD concentration, only a small fraction of the metallic phase is covered by
the modifier, and the remaining metal sites are mainly occupied by the substrate and
dissociated hydrogen.
Figure 7. Enantiomeric excess vs PPD conversion for 1.0%Rh/MCM-41 catalyst at
different CD concentration. Reaction conditions: PPD concentration: 0.01 mol L-1, catalyst
mass: 0.050 g, PH2: 20 bar, mixing speed: 700 rpm and solvent: cyclohexane.
19
Page 19 of 32

The adsorbed modifier interacts with a PPD molecule and forms an adsorbed
supramolecule, which is hydrogenated by hydrogen spillover to produce the R-enantiomer.
Additionally, the adsorbed substrate may also be hydrogenated, which yields the racemic
mixture, and, therefore, may produce low enantioselectivity. Conversely, at a higher CD
concentration, dimer formation (CD)2 limits the quantity of the desired enantiomer. The Rh
nanoparticles with (CD)2 adsorbed on their surface cannot accept parallel adsorption of the
PPD phenyl group that controls the ee and rs. In addition, the conversion levels can be
decreased through active blockage.
3.1.2 Metal loading
The influence of Rh loading on enantioselective hydrogenation was determined using a
series of Rh/MCM-41 catalysts for which the nominal Rh levels were 0.5, 1.0, and 1.5%
Rh. As demonstrated through certain exploratory experiments, the CD concentration is an
important parameter in the target reaction; hence, it was investigated. Thus, the experiments
were conducted under standard conditions using PPD at 0.01 mol L-1. The characterisation
results are shown in Table 2 with the Rh concentrations determined using ICP-MS, the Rh
dispersion determined using TEM and the calculated Rh average particle size. The PPD
hydrogenation rates and ee values are lower than reported for the α-ketoesters [40]. This
observation can be attributed to the greater ability of diketones to form oligomers and
polymers, which are considered strong catalyst poisons. The 1.5% Rh/MCM-41 catalyst
was the most active and yielded conversion levels up to 80%. This value is promising
20
Page 20 of 32

compared with the best value obtained for asymmetric hydrogenation of PPD using 5%
Rh/Al2O3 [10].
Figure 8 shows the ee vs CD concentration for the Rh/MCM-41 catalysts. The ee is a
clearly a function of both the Rh loading and CD concentration. The best ee results (65–
70%) were obtained with the 1.0% Rh/MCM-41 catalyst. The smooth variation of the ee
maximum as a function of Rh dispersion suggests that Rh particle size is an important and
central parameter that controls enantioselectivity. If the Rh loading is too low, the particle
sizes are also smaller and ineffective at interacting with the alkaloid modifier results.
Conversely, a slightly increase in Rh loading produces large nanoparticles and lower
overall reactivity (see Table 2), but the selectivity is enhanced because, at a lower CD
concentration, the Rh sites can be effectively modified. In our case, the differences in
conversion levels and ee demonstrated by the 1.0% Rh/MCM-41 catalyst compared to its
1.5% Rh/MCM-41 homologue can be explained based on the mesostructure silica effect.
The straight channel observed in this type of support produces a confinement effect, which
enhances the catalytic activity, but the Rh nanoparticles sizes differ. For 1.5% Rh/MCM-
41, the mean size is 3.7 nm, which is similar to the MCM-41 pore diameter (see Table 1);
under these conditions, the catalyst shows diffusional resistance attributed to pore blockage
by the active phase, which is consistent with previous studies reported by Murzin et al.
[26]. This behaviour can also be due to high metal loading, which blocks the catalyst pores
and, hence, limits reactant diffusion into the pore structure.
21
Page 21 of 32

Figure 8. Enantiomeric excess vs CD concentration for Rh/MCM-41 catalysts. Reaction
conditions: PPD concentration: 0.01 mol L-1, catalyst mass: 0.050 g, PH2: 20 bar, mixing
speed: 700 rpm and solvent: cyclohexane.
In this work, the metal Rh particles ca. 3.2 nm in size developed on the 1.0% Rh/MCM-41
catalyst appeared optimal. However, further work is required to fully ascertain the optimal
Rh particle size. This observation is similar to the relationship observed for ethyl pyruvate
hydrogenation over Rh/Al2O3 catalysts with particle size distributions where in the maxima
were at 1.5 nm and 2.0 nm. The smaller particle size yielded poor enantioselection, and it
was concluded that such small particles could not interact effectively with the modifier and
the reactant [40].
22
Page 22 of 32

3.1.3 The effect of hydrogen pressure
The effect of hydrogen pressure was negligible where the hydrogen pressure was varied
from 10 to 40 bar. Figure 9 shows the TOF and selectivity as a function of H2 pressure. The
surface H2 concentration may directly influence the reactant adsorption and, thus,
enantioselectivity. An increase in H2 pressure (from 10 to 40 bar) markedly effects the TOF
but only slightly effects the ee. The TOF increases with increasing pressure. The reaction
proceeded slowly at 10 bar; the PPD conversion was only 47% after 240 min. However, at
40 bar, the PPD conversion was 89% over the same time period, which indicates that
hydrogen adsorption is mainly non-competitive. The role of hydrogen adsorption in the
hydrogenation process is interesting. On one hand, hydrogen is dissociatively adsorbed on
Rh and competes with organic molecules for metal sites; on the other hand, hydrogen
molecules are much smaller than the organic molecules.
23
Page 23 of 32

Figure 9. Selectivity and TOF for PPD hydrogenation on 1.0%Rh/MCM-41 catalyst at
different H2 pressures. Reaction conditions: PPD concentration 0.01 mol L-1, CD
concentration: 0.5x10-3 mol L-1, catalyst mass: 0.050 g, stirring speed: 700 rpm and solvent:
cyclohexane. Enantiomeric excess (ee) at 50% of conversion and diols selectivity and rs at
240 minutes of reaction.
Previous work has shown that PPD hydrogenation on a commercial Pt/Al2O3 catalyst
proceeds with a quasi-zero reaction order with respect to hydrogen over a minor pressure
range (1-5 bar) [4, 6, 10]. The approximately linear plot of TOF values versus hydrogen
pressure indicates first-order kinetics with respect to hydrogen pressure. At pressures > 20
bar, other hydrogenation products were detected. At the CD concentration used, the PPD
phenyl ring was not hydrogenated to a cyclohexyl ring; thus, undesirable products were
24
Page 24 of 32

formed both from the reactant and each product, which yielded seven hydroxyl products. It
can be concluded that C=O non-selective hydrogenation is generally more favourable at
higher pressures due to the short residence time of PPD in the straight channels.
3.1.4 Solvent Effects
In heterogeneous asymmetric hydrogenation, enantioselectivity is sensitive to the solvent
used, such as for asymmetric hydrogenation of PPD catalysed by Pt/CD [11, 41]. Solvents
with dielectric constants ranging from 2 to 8 were included in this study for PPD
hydrogenation over a 1.0% Rh/MCM-41 catalyst; these values correspond to pure solvents,
excluding the effects from low levels of reactants and modifiers (CD/PPD 0.05). We used
weakly polar solvents because CD is highly soluble therein. The results are shown in Figure
10 and correspond to a qualitative model, including the enantioselectivity and dielectric
constant of the pure solvents, similar to the proposal in Toukoniitty et al. [11, 41]. These
results show that TOF decreases with an increase in solvent polarity. The highest
conversion was observed for cyclohexane, and low reaction rates were observed for
tetrahydrofuran (THF). The side products 2R- or 2S-PP and diols were not detected. The
solvent is an essential factor because it modulates the modifier conformation, modifier–
reactant interaction and solution hydrogen concentration, among other features, and it
affects the enantioselectivity [11, 42-44]. Solvent effects over Rh are similar to Pt catalysts
(e.g., the highest and dichloromethane the lowest eeC1 are observed in cyclohexane).
25
Page 25 of 32

Figure 10. Enantiomeric excess and TOF for PPD hydrogenation on 1.0%Rh/MCM-41
catalyst using different solvents. Reaction conditions: PPD concentration 0.01 mol L-1, CD
concentration: 0.5x10-3 mol L-1, catalyst mass: 0.050 g, PH2: 20 bar and stirring speed: 700
rpm. ee at 50% of conversion.
Toukoniitty et al. [11] reported that the initial hydrogenation rate did not correlate with
hydrogen solubility. Our results for PPD hydrogenation with the 1.0% Rh/MCM-41
catalyst show a roughly linear correlation between the solvent dielectric constant and
optical yield as well as reaction rate. The highest ee was recorded using hydrocarbon
solvents with the dielectric constant < 2.3, which is qualitatively consistent with findings by
Baiker et al. [45] for ethyl pyruvate hydrogenation, showing that optical yield decreased
when solvent polarity increases. However, exceptions have been demonstrated (e.g., the
study by Toukoniitty et al. [10] on PPD hydrogenation using commercial Rh/Al2O3
26
Page 26 of 32

catalysts, which showed similar conversion levels for toluene (ε: 2.34) and acetic acid (ε:
6.2)). We show that apolar-aprotic solvents positively influence both selectivity and activity
because these solvents not only improve the reactants’ mobility but also enhance an
effective interaction between a CD-PPD complex or Rh-CD-PPD supramolecular complex,
as reported Toukoniitty et al. [11].
Conclusions
Cinchonidine-modified Rh/MCM-41 catalysts with different Rh loading ratios are effective
for enantioselective PPD hydrogenation. Here, we applied a mesostructured support to
prepare a series of Rh catalysts with 0.5 to 1.5 wt% Rh. This support provided small Rh
particles with relatively narrow particle size distributions, and the best catalyst (1.0 wt%
Rh) showed outstanding enantioselectivity for PPD hydrogenation. The Rh dispersion and,
hence, Rh particle size appear to be the dominant factor in combination with the
appropriate CD concentration for the reaction. These three Rh loading levels display
volcano-type curves for the cinchonidine concentration studies, which showed an optimum
quantity of modifier required for the maximum enantioselectivity using each catalyst. This
behaviour can be attributed to nanoparticle and dispersion differences. More importantly,
the activity and selectivity were improved using a mesostructured support. It is suggested
that the significant difference in the support structure improves the interaction efficiency
between the Rh metal surface and modifier–substrate adduct adsorption through the
confinement effect. An increase in the H2 pressures accelerates the hydrogenation rates, but
decreases the selectivity to C=O in the adjacent phenyl group. Finally, the solvent effect
yields a linear correlation between the solvent dielectric constant-hydrogenation rate and
27
Page 27 of 32

non-linear correlation; ee displayed the highest enantioselectivity when cyclohexane was
used as the solvent.
Acknowledgements
Authors thank CONICYT for financial support (FONDECYT Initiation GRANT
11121631, FONDECYT GRANT 1061001,) and Red Doctoral REDOC.CTA. C. C. Torres
and C. H. Campos thank CONICYT for Doctoral fellowship.
References
[1] H.-U. Blaser, H.-P. Jalett, M. Muller, M. Studer, Catal. Today 37 (1997) 441-463.
[2] F. Hoxha, B. Schimmoeller, Z. Cakl, A. Urakawa, T. Mallat, S.E. Pratsinis, A.
Baiker, J. Catal. 271 (2010) 115-124.
[3] A. Vargas, F. Hoxha, N. Bonalumi, T. Mallat, A. Baiker, J. Catal. 240 (2006) 203-
212.
[4] E. Toukoniitty, P. Mäki-Arvela, A.K. Neyestanaki, T. Salmi, R. Sjöholm, R. Leino,
E. Laine, P.J. Kooymane, T. Ollonqvist, J. Väyrynen, App. Catal. A: General 216
(2001) 73-83.
[5] T. Marzialetti, M. Oportus, D. Ruiz, J.L.G. Fierro, P. Reyes, Catal. Today 133-135
(2008) 711-719.
[6] E. Toukoniitty, P. Mäki-Arvela, N. Kumar, T. Salmi, D.Y. Murzin, Catal. Today 79-
80 (2003) 189-193.
[7] T. Marzialetti, J.L.G. Fierro, P. Reyes, Catal. Today 107-108 (2005) 235-243.
28
Page 28 of 32

[8] E. Toukoniitty, B. Sevciková, P. Mäki-Arvela, J. Wärná, T. Salmi, D.Y. Murzin, J.
Catal. 213 (2003) 7-16.
[9] E. Toukoniitty, P. Mäki-Arvela, J. Wärná, T. Salmi, Catal. Today 66 (2001) 411-417.
[10] E. Toukoniitty, S. Franceschini, A. Vaccari, D.Y. Murzin, App. Catal. A: General
300 (2006) 147-154.
[11] E. Toukoniitty, P. Mäki-Arvela, J. Kuusisto, V. Nieminen, J. Mol. Catal. A: Chem
192 (2003) 135-151.
[12] R.A. Sheldon, J. Chern. Tech. Biotechnol. 67 (1996) 1-14.
[13] H.-U. Blaser, H.-P. Jalett, M. Müller, M. Studer, Catal Today 37 (1997) 441--463.
[14] H.-U. Blaser, H.-P. Jalett, F. Spindler, J. Mol. Catal. A: Chem 107 (1996) 85-94.
[15] A. Taskinen, V. Nieminen, E. Toukoniitty, D.Y. Murzin, M. Hotokka, Tetrahedron
61 (2005) 8109–8119.
[16] V.B. Shukla, P.R. Kulkarni, World J. Microbiol. Biotechnol. 16 (2000) 499.
[17] D.Y. Murzin, J. Mol. Catal. A: Chem 289 (2008) 91–94.
[18] G. Farkas, A. Sípos, A. Tungler, A. Sárkány, J.L. Figueiredo, J. Mol. Catal. A: Chem
170 (2001) 101-107.
[19] N.J. Coulston, R.P.K. Wells, P.B. Wells, G.J. Hutchings, Catal. Today 114 (2006)
353-356.
[20] E. Zhan, S. Li, Y. Xu, W. Shen, Catal. Comm. 8 (2007) 1239-1243.
[21] C. Chapuis, D. Jacoby, App. Catal. A: General 22 (2001) 93-117.
[22] J. Hagen, Industrial Catalysis. A Practical Approach, Alemania, 1999.
[23] D. Ruiz, C. Mella, P. Reyes, J.L.G. Fierro, J. Chil. Chem. Soc 57 (2012) 1394-1399.
[24] D. Ruiz, M. Oportus, C. Godard, C. Claver, J.L.G. Fierro, P. Reyes, Curr. Org. Chem.
16 (2012) 2754 - 2762.
29
Page 29 of 32