Enantioselective hydrogenation of 1-phenyl-1,2-propanedione on immobilised cinchonidine-TiO2 catalysts

Article abstract of DOI:10.1016/j.cattod.2014.04.030

The enantioselective hydrogenation of 1-phenyl-1,2-propanodione (PPD) was investigated using cinchonidine-immobilised Pt/TiO₂ catalysts. Prior to metal deposition, TiO₂ was chirally modified by the direct anchoring of cinchonidine (CD) using trimethoxysilane as coupling agent (TMS-CD). The catalysts were prepared using a high H₂ pressure reduction-deposition method and were characterised by elemental analysis (C, H and N), TG, DRIFT, ¹³C and ²⁹Si solid-state NMR, N ₂ adsorption-desorption isotherms, XRD, XPS and HR-TEM. The catalytic activity was evaluated in a batch reactor at 298 K and 40 bar using cyclohexane as solvent with various cinchonidine concentrations. The results indicate that the enantioselectivity was sensitive to the CD surface concentration and the enantiomeric excess of the target product, 1-R-phenyl-1-hydroxy-2-propanone, was in the range of 25-51%. The best catalyst was the one supported on TiO ₂ with a nominal content of 10 wt% TMS-CD. The effect of the H ₂ pressure, the concentration of substrate, solvent and recyclability of the catalyst were studied. The results obtained confirmed that the variation of reaction conditions affects both the activity and enantioselectivity due to the substrate adsorption on the metal active sites. Concerning the solvent effect, the enantiomeric excess decreased non-linearly upon increasing the solvent dielectric constant; this was attributed to the interactions between solvents and TMS-CD on the surface. In the catalyst recycling studies, the enantiomeric excess decreased up to 40% after the 3rd reuse. The drop of activity and enantiomeric excess was attributed to the hydrogenation of the immobilised CD.

Full text of DOI:10.1016/j.cattod.2014.04.030

Catalysis Today 235 (2014) 226–236
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Enantioselective hydrogenation of 1-phenyl-1,2-propanedione
on immobilised cinchonidine-TiO catalysts
2
a
a
a
a
Cristian H. Campos , Cecilia C. Torres , Ana B. Dongil , Doris Ruiz ,
b
a,∗
José L.G. Fierro , Patricio Reyes
Edmundo Larenas 129/Departamento de Fisicoquímica, Facultad de Ciencias Químicas, Universidad de Concepción, Concepción, Chile
Instituto de Catálisis y Petroleoquímica, CSIC, Marie Curie 2, Cantoblanco 28049, Madrid, Spain
a
b
a r t i c l e i n f o
a b s t r a c t
Article history:
The enantioselective hydrogenation of 1-phenyl-1,2-propanodione (PPD) was investigated using
cinchonidine-immobilised Pt/TiO2 catalysts. Prior to metal deposition, TiO2 was chirally modified by
the direct anchoring of cinchonidine (CD) using trimethoxysilane as coupling agent (TMS-CD). The cat-
alysts were prepared using a high H2 pressure reduction-deposition method and were characterised
by elemental analysis (C, H and N), TG, DRIFT, C and Si solid-state NMR, N2 adsorption–desorption
isotherms, XRD, XPS and HR-TEM. The catalytic activity was evaluated in a batch reactor at 298 K and
Received 14 November 2013
Received in revised form 19 April 2014
Accepted 29 April 2014
Available online 12 June 2014
1
3
29
Keywords:
TiO2 catalysts
4
0 bar using cyclohexane as solvent with various cinchonidine concentrations. The results indicate that
the enantioselectivity was sensitive to the CD surface concentration and the enantiomeric excess of the
target product, 1-R-phenyl-1-hydroxy-2-propanone, was in the range of 25–51%. The best catalyst was
the one supported on TiO2 with a nominal content of 10 wt% TMS-CD. The effect of the H2 pressure, the
concentration of substrate, solvent and recyclability of the catalyst were studied. The results obtained
confirmed that the variation of reaction conditions affects both the activity and enantioselectivity due
to the substrate adsorption on the metal active sites. Concerning the solvent effect, the enantiomeric
excess decreased non-linearly upon increasing the solvent dielectric constant; this was attributed to the
interactions between solvents and TMS-CD on the surface. In the catalyst recycling studies, the enan-
tiomeric excess decreased up to 40% after the 3rd reuse. The drop of activity and enantiomeric excess
was attributed to the hydrogenation of the immobilised CD.
Immobilised cinchonidine
1
-Phenyl-propane-1,2-dione
Enantioselective hydrogenation
©
2014 Elsevier B.V. All rights reserved.
1
. Introduction
main product is (R)-1-hydroxy-1-phenylpropanone, an important
intermediate in the synthesis of drugs [13].
Asymmetric catalysis has become a challenging subject in the
The commercial application of these systems requires an effec-
tive method of catalyst recovery and re-use because of economic
and environmental concerns. Several methods for the recycling
of chiral catalysts have been described [14–18] and a number of
reactions have been conducted with this new type of catalysts,
especially since the successful immobilisation of CD [19–21]. In
previous papers, we investigated the enantioselective hydrogena-
tion of PPD over two types of CD immobilisation: (1) using it
as a stabiliser on Pt nanoclusters supported on SiO2 [22,23], and
field of organic synthesis and catalysis over the past three decades,
both from the theoretical and practical points of view [1–3].
Among the numerous enantioselective heterogeneous reactions,
the Pt/cinchona-catalysed hydrogenation of ␣-ketoesters is one of
the few successful examples with more than 95% enantiomeric
excess (e.e.) under favourable reaction conditions [4]. Another typ-
ical reaction is the hydrogenation of ␣-diketones, which was first
reported by Vermeer et al. in 1993 [5]. Due to the conjugated keto
groups of the reactant, the hydrogenation consists of two consecu-
tive steps. 1-Phenyl-1,2-propanedione (PPD) is a suitable molecule
to investigate the enantioselective hydrogenation of conjugated
carbonyl bonds [6–12]. It contains a phenyl ring and two carbonyl
groups. Under the reaction conditions, in the presence of CD, the
(2) anchoring it on SiO2 and ␥-Al O3 oxides surfaces [20,21]. The
2
catalysts not only exhibited similar enantioselectivity (i.e., 50%
enantiomeric excess for the hydrogenation of PPD) but also exhib-
ited novel catalytic properties, i.e., high regioselectivity to C
O
adjacent to phenyl groups, which differs from their conventional
heterogeneous counterparts.
As an extension of these previous studies, we describe in this
paper the catalytic performance of Pt supported on CD modified
TiO2-anatase in the enantioselective hydrogenation of PPD. We
∗ Corresponding author.
E-mail addresses: ccampos@udec.cl (C.H. Campos), preyes@udec.cl (P. Reyes).
http://dx.doi.org/10.1016/j.cattod.2014.04.030
0
920-5861/© 2014 Elsevier B.V. All rights reserved.

C.H. Campos et al. / Catalysis Today 235 (2014) 226–236
227
attempt to show some insight on the role of the support on activity
and selectivity. The study is focused on the direct tethering of mod-
ified CD molecules on pure anatase TiO2 using CD functionalised
with trimethoxysilane (TMS-CD). The active phase deposition was
the immobilisation of TMS-CD, as reported in our own previous
works [20,21]. Activation of the TiO2 surface was performed in a
round-bottom flask in which 3.0 g of TiO was mixed with 24 mL of
2
−
1
1,4-dioxane and 3 mL of aqueous solution of 3.1 mol L HCl. This
system was stirred for 30 min at 353 K, filtered, and dried under
vacuum (lower than 10 mbar) for 4 h at 393 K.
performed using H PtCl6 as a precursor at high H2 pressure and
2
−
1
room temperature. Some important reaction parameters, such as
the surface modifier concentration, PPD concentration, H pressure
2
and solvent, which have been less studied in the literature, were
investigated in detail to optimise the enantioselectivity as well as
the activity.
2
.3. Preparation of modified CD
The modification of CD and the subsequent later hydroxylation
were performed as reported in our previous works [20,21]. Briefly,
an ice-cooled solution of CD in THF containing TEA was added drop-
wise to the TMCS. The reaction mixture was stirred for 20 h at room
temperature and then for 2 h at 333 K. The product was extracted
with chloroform and washed with water. The obtained product
was hydroxylated with the Pt(COD)Cl2 catalyst precursor and TMS
(TMS/catalyst = 120, mole ratio) at 313 K using toluene as the sol-
vent. The reaction mixture was stirred for 5 h at 363 K under a N2
atmosphere. Purification was performed by flash chromatography
(hexane–acetone–TEA = 40:18:1).
2
. Experimental
2.1. Materials
All air-sensitive reactions were performed under an inert atmo-
®
sphere (Ar) in a Schlenk flask. Tetrahydrofuran (THF, Merck ) and
toluene (Merck ) were dried in metallic sodium/benzophenone;
triethylamine (TEA, Aldrich ) was distilled at reduced pressure
prior to use. Cinchonidine (96% Aldrich ), trimethylchlorosi-
lane (TMCS 98%, Merck ), trimethoxysilane (TMS 98%, Merck ),
platinum cyclooctadienyl (II) chloride (Pt(COD)Cl , Aldrich ),
anhydrous sodium sulphate (Na SO , Merck ), titanium(IV) oxide,
®
®
®
®
®
®
2
2.4. Synthesis of modified TiO2
®
2
4
®
®
TiO -anatase (99% AlfaAesar ), H PtCl ·6H O (40%Pt, Merck ),
2
2
6
2
Five supports were prepared with different amounts of CD (1,
, 10, 15, and 20 wt%), simulating the CD fractions that are added
®
NaOH (Merck ), and the solvents (cyclohexane, ethanol, ethyl
5
®
acetate, Merck ) were used as received.
in traditional systems. A 2.0 g sample of activated TiO2 and the
appropriate amount of TMS-CD in toluene were added into a round-
bottom flask, and the volume of the system was then brought to
2.2. Activation of TiO2
5
0 mL with solvent (see Scheme 1). The reaction was refluxed for
TiO activation was performed primarily to increase the surface
concentration of Ti–OH groups because they are responsible for
12 h and later filtered and washed twice with 40 mL toluene and
20 mL chloroform. The solid was dried in a furnace for 1 h at 353 K.
2
Scheme 1. Synthetic route for CD modification and catalyst preparation.

2
28
C.H. Campos et al. / Catalysis Today 235 (2014) 226–236
Subsequently, the solid was treated by refluxing in a mixture of
methanol/THF for 20 h and was then washed with 100 mL of n-
pentane. Finally, the solids were dried under vacuum for 4 h at
where n is the number of particles of diameter d . The size limit for
i i
the detection of Pt particles on samples was ca. 0.1 nm.
Photoelectron spectra (XPS) were recorded using a VG Escalab
200 R spectrometer equipped with a hemispherical analyser and
using non-monochromatic Mg K␣ X-ray radiation (hꢀ = 1253.6 eV).
The binding energies (BE) were calculated with respect to the C-
component of the C1s peak fixed at 284.8 eV. Data analysis was
performed with the “XPS Peak” software. The peaks were decom-
posed into a sum of Gaussian/Lorentzian (G/L = 90/10) after the
subtraction of a Shirley type baseline. The surface Pt/Ti and N/Ti
atomic ratios were estimated from the integrated intensities of the
Pt 4f, Ti 2p, C 1s, and N 1s lines after background subtraction and
were corrected by the atomic sensitivity factors [26].
3
93 K. All the samples of the modified supports were denoted as
TiCD[x], where [x] is the nominal TMS-CD wt% and, in the case of 0,
corresponds to TiO -activated support.
2
2.5. Catalyst preparation
The catalysts (1.0 g) were prepared to obtain 1 wt% Pt loading
using the five modified supports and activated TiO . These were
2
labelled as 1%Pt/TiCD[x], with [x] = 1–20. The methodology is sim-
ilar to that reported in our previous work [21]. The support was
dispersed in 50 mL of deionised water in a Teflon container, and
the appropriate amount of H PtCl ·6H O was added. After 1 h,
2.7. Catalytic activity
2
6
2
−
1
:
0
.5 mol L NaOH was added (OH Pt = 6) and sealed in a stainless
steel batch reactor. The pressure was adjusted to 40 bar H2 for 2 h
under constant magnetic stirring, obtaining the reduced supported
metal. Then, the solid was filtered off and washed to constant val-
ues of pH and conductivity. Finally, it was dried in a vacuum oven at
The catalytic assays of PPD hydrogenation were performed in a
stainless steel (100 mL) Parr-type batch reactor at a concentration
−
1
of 0.01 mol L of substrate using cyclohexane (25 mL) as solvent
and stirring at 700 rpm under 40 bar of H2 pressure. All catalytic
runs were conducted in the absence of external mass transfer lim-
itations. The catalysts were in the form of fine powders (>30 ␮m),
assuming a negligible effect of pore diffusion limitations. In the
catalyst mass studies, the (PPD/Pt) molar ratio was varied, the con-
centration of PPD was kept constant and the mass of the catalyst
was modified. No important differences in the initial reaction rates
were noted upon varying the catalyst mass in the 0.01–0.07 g range
because the reaction rate was proportional to the catalyst mass.
This indicated that gas–liquid and liquid–solid external mass trans-
fer limitations were absent [9,27]. The pseudo-kinetic constants (k)
were calculated using a pseudo-first-order kinetic model for a batch
reactor in similar conditions, as was reported by Toukoniitty et al.
[10]. In our case, only 4 hydrogenation products were detected, as
shown in Scheme 2.
3
73 K for 1 h and stored in a desiccator under N2 atmosphere prior
to the catalytic test.
2.6. Catalyst characterisation
Elemental analyses of C, H, and N were performed on a LECO
CHNS-932 analyser. TG studies were conducted in a Mettler Toledo
−
1
Thermogravimetric TGA/SDTA 851 using an O flow of 25 mL min
2
−
1
and a heating rate of 1 K min from 298 to 1000 K. NMR spectra
1
13
1
for H and C{ H} were obtained on a Bruker AMX-300 spectrom-
1
13
eter (300 MHz for H, 75 MHz for C) using trimethylsilane as an
1
13
internal standard. The results obtained from the NMR of H and
C
were compared with those reported by Leino et al. [24,25]. Solid-
state 13C and Si CP NMR spectra were recorded at 100.6 MHz
and 79.49 MHz, respectively, using a Bruker AV 400 WB spectrome-
ter. Diffuse reflectance infrared Fourier transform (DRIFTs) spectra
were performed on a JASCO FT/IR-6300 fitted with an MCT-A detec-
29
Reactants and products were analysed by gas chromatography
and mass spectrometry using a GC–MS device (Shimadzu GCMS-
®
QP5050) with a 30 m chiral ␤-Dex 225 column (Supelco ) and
−
1
tor and a KBr beamsplitter within the range of 4000–650 cm
.
helium as the carrier gas. The recycling assays were performed by
filtering the catalyst from the reaction medium. The filtered catalyst
was washed three times consecutively with chloroform to clean the
surface and was then dried at 373 K for 24 h.
In relation to the selectivity of the catalysts, enantiomeric excess
and regioselectivity have been defined as:
Approximately 0.020 g of the sample was loaded into a diffuse
reflectance (DRIFTS) reaction cell provided by Harrick (Praying
Mantis model). Before any measurements were performed, the
−
1
sample was heated for 1 h at 423 K in He flowing at 30 mL min
and then cooled to room temperature. For all the measurements,
−
1
spectra were a composite of 64 scans with a resolution of 4 cm
.
eeCx(%) = [
R] − [S]
[RC1] + [SC1]
[RC2] + [SC2]
x
x
× 100 and rs =
Prior to the run, backgrounds were collected by flowing He (at
[
R] + [S]
−
1
x
x
3
0 mL min ) at the respective reaction temperature. XRD pat-
terns were recorded in a RigakuD/max-2500 diffractometer with
Cu K␣ radiation at 40 kV and 100 mA. N2 adsorption–desorption
isotherms at 77 K were performed in a Micromeritics ASAP 2010
apparatus. Specific surface areas were determined by the BET
where [R] and [S] correspond to the concentration of the respective
enantiomers, x is the number of the carbonyl group, and in the case
of rs, of the alcohols of the different carbonyl groups.
(
Brunauer–Emmett–Teller) equation, using adsorption data in the
3. Results and discussion
relative pressure range of 0.05–0.3, and pore-size distributions
were estimated using the BJH method. TEM images of the catalysts
were obtained using a Philips electron microscope CM200 with an
energy dispersive analyser and digital camera coupled to a high
speed TVIPS FastScan F-114 model of 1024 × 1024 pixels and 12 bits
and a scanning electronic microscope JEOL JSM-6380 LV with a high
resolution of 3.0 nm. The samples for analysis were prepared by
3.1. Support synthesis and characterisation
Table 1 summarises the elemental analysis of the supports.
The total content of TMS-CD anchored on the surface was lower
than the nominal content, and the effect was more pronounced
upon increasing the nominal content of the inducer. The calcu-
lations were based on the N(%) because it is the only vector of
the real TMS-CD content anchored on TiO2, as reported in our
previous works [20,21]. These observations can be explained in
terms of the nature of the Ti–OH groups developed on the acti-
vated TiO2 surface and diffusion phenomena, respectively. The
1,4-dioxane/HCl-treatment increases the amount of hydroxyls on
the titanium surface. These Ti–OH groups exhibit different sur-
face structural arrangements, with different surface reactivity: (1)
dispersion in ethanol/H O (1:1) and were deposited on a holey car-
2
bon/Cu grid (300 Mesh). Up to 300 individual metal particles were
counted for each catalyst, and the surface area-weighted mean Pt
diameter (dp) was calculated from:
ꢀ
3
nid
i
i
dp = ꢀ
inidi2

C.H. Campos et al. / Catalysis Today 235 (2014) 226–236
229
Scheme 2. Simplified hydrogenation route of PPD.
Table 1
range, which corresponds to the dehydroxylation and accounts for
Elemental analysis (at%) and TMS-CD content determined by EA and TGA of the
synthesised supports.
−1
1
.2 mmolOH
g
. After the activation treatment, the TiCD(0) dis-
played a new slope in the range 400–1000 K corresponding to 4 wt%
Sample
C(%)
H(%)
N(%)
TMS-CD(wt%)
Yielda (%)
−1
mass, approximately 2.4 mmolOH
g
. The supports modified with
TMS-CD showed three steps of mass losses in the ranges 298–350 K,
50 K and 650–750 K. The first mass loss corresponds to desorp-
EA
TGA
5
TiCD(1)
TiCD(5)
TiCD(10)
TiCD(15)
TiCD(20)
1.69
2.71
3.43
4.06
4.72
0.63
0.84
0.95
1.03
1.01
0.07
0.25
0.42
0.54
0.62
1.0
3.5
6.0
7.8
8.8
6.0
7.4
8.4
9.0
9.2
99.3
70.0
60.0
51.4
43.9
tion of the water absorbed when the samples were exposed to the
atmosphere prior to TG measurements. The second mass change is
associated to the degradation of alkyl chains and the third corre-
sponds to the quinoline aromatic ring oxidation [20]. The TMS-CD
content obtained by TGA shows an excess with respect to the EA, as
shown in Table 1. For the TiCD[x] supports, the TMS-CD immobili-
a
Yield of elemental analysis: (TMS-CDReal/TMS-CDNominal) × 100%.
sation process (the reflux treatment with toluene and THF/CH OH
and washing with n-pentane) can be attributed to strong solvent
adsorption.
hydroxyls groups, with different acid–base character, that can
form hydrogen bonds from multilayer arrangements; and (2) free
hydroxyl groups characterised by lower coordination interactions,
such as that located at the corners and defects on the catalyst sur-
face, usually considered to be the most basic ones [28]. In our
system, the first arrangement would decrease the ability to form
Ti–O–Si bond during the TMS-CD immobilisation.
3
Fig. 2 shows the DRIFT spectra of the TiO -anatase and the modi-
2
fied supports. The parent TiO showed intense bands at 3674, 3630,
2
−
1
and 1624 cm , corresponding to vibrations of hydroxyl groups
on the surface of TiO , and at 1624 cm , due to the bending
−
1
2
mode of molecularly adsorbed water molecules. After the activa-
tion treatment, TiCD (0.0) displayed a broad band between 4000
Fig. 1 displays the thermogravimetric mass changes for the dif-
ferent materials. For TiO , a mass loss of 2 wt% in the 400–1000 K
2
−
1
and 3000 cm due to the increase of the population of hydroxyl
groups, as shown in Fig. 2(a). The intensity and width of this band
can be explained by the different type of Ti–OH groups on the sur-
face, in agreement with the TG results. With respect to the spectra
profiles in Fig. 2(b), TiCD[x] with [x] =/ 0 displayed bands in the
−
1
−1
range 2956–2863 cm and 1540–1200 cm . The first is associated
with the C–H stretching vibrations of the quinuclidine heterocyclic
ring, whereas in the second, the fingerprint bands of the TMS-CD
are expected to occur. Upon increasing the nominal content of TMS-
CD, the intensity of these bands increased, whereas the intensity of
−
1
the band in the 4000–3000 cm range follows the opposite trend.
Another band was detected at 1115 cm , which can be assigned
to the stretching vibration of the Ti–O–Siorganic bonds, and it would
−
1
confirm the covalent immobilisation of TMS-CD on TiO .
2
To support the DRIFT results, 13C and Si CP/MAS-NMR spectra
29
were recorded and are displayed in Fig. 3(a) and (b), respectively.
13
C NMR spectra displayed, for all the supports TiCD[x] with [x] ꢀ 1,
resonances in the range of 24–67 ppm, which are characteristic of
the quinuclidine moiety of TMS-CD [19–21]. Additional resonances
at 118, 128.8 and 137.7 ppm were also observed, which correspond
to the aromatic quinoline ring. In our previous works, an unusual
C signal was observed at ca. −2.0 ppm, due to the Si-CH3 moi-
ety, but it was not detected in the TiCD[x] support. This could be
Fig. 1. TG curves for the synthesis of supports. Conditions: O2 atmosphere to
13
−
1
2
5 mL min flow and temperatures in the range of 298–1000 K. (1) TiO2; (2) TiCD[0];
(
3) TiCD[1]; (4) TiCD[5]; (5) TiCD[10]; (6) TiCD[15]; (7) TiCD[20].

2
30
C.H. Campos et al. / Catalysis Today 235 (2014) 226–236
Fig. 2. DRIFT of the supports modified with TMS-CD. (A) TiO2-anatase and TiO2-
activated support; (B) TiCD[x] supports where the TiO2-activated support spectra
was subtracted.
Fig. 3. Solid state CP-MAS NMR of the supports modified with TMS-CD. (a) 13C for
TiCD[x] and (b) Si for TiCD[20].
2
9
explained by the acid–base nature of the Ti–OH groups of the pre-
pared support.
Fig. 3(b) shows the 29Si NMR for the TiCD(20). This technique
absence of impurities. The above characterisation results confirm
the proposed reaction suggested in Scheme 1.
is very sensitive to Si-atoms near –OH and/or –OR groups, and
it is the standard characterisation method for silicate materials.
Two signal types can be distinguished with this technique. The
first type of signal arises from the anchored TMS-CD species that
form bonds with the support/silicon alkoxides (−45 to 60 ppm
denominated Tx signal, where x = 1, 2, or 3). The second signal
type corresponds to siloxane/silane (−90 to −115 ppm denomi-
3.2. Catalyst synthesis and characterisation
Table 2 summarises the elemental analysis of the catalysts and
the most relevant physicochemical properties. It can be observed
that the TMS-CD content decreased with respect to the metal free
counterparts in Table 1. This is due to a slight leaching of the inducer
under the preparation conditions, as discussed in previous work
y
nated Q signals, where y = 2, 3 or 4) that remain after the anchoring
process. Because this support possesses the major TMS-CD content
2
on the surface, the characteristic anchoring signals of the T and
[20,21]. The values of the BET surface areas of pure TiO -anatase
2
3
T
type at −50 and −80 ppm were detected, in agreement with
and TiCD[x] are also summarised in Table 2. All the samples show
type IV isotherms with type H4 hysteresis loops, according to IUPAC
classification, corresponding to slit-shaped pores [29]. The surface
the results we reported previously [20,21]. For the Q signals, two
silanols were detected, with a similar intensity between the T and Q
3
2
−1
signals. The two Q signals, which were attributed to free silanols,
area of pure TiO -anatase was 142 m g , which is larger than the
2
occurred at −101.5 and −108.3 ppm in different chemical environ-
TiCD[x] samples. All of them displayed bi-modal pore-size distri-
butions. This information allows us to infer that as the TMS-CD
content increased, the porosity and surface area were influenced
by the anchoring of TMS-CD species on the surface.
ments. TiO is an amphoteric metal oxide with acid–base pairs of
2
4+
2−
Ti (Lewis acid) and O (Lewis base) ions on its surface [28]. These
acid–base pairs can interact with the remaining free silanol groups
and cause the signals that exhibit different chemical shifts. This
interaction is very strong. Consequently, we assumed that the res-
onance at −108.3 ppm corresponds to free silanol groups, whereas
the resonance at −101.5 ppm can be assigned to the strong inter-
action with the inorganic phase. The characteristic resonance of
an Si atom coordinated with one O atom from the TiO2 network
Fig. 4 shows the XRD patterns of the 1%Pt/TiCD[x] powdered
◦
◦
◦
samples. The dominant peaks at 2ꢁ 25.4 , 37.9 and 48.2 , which
represent the (1 0 1), (0 0 4) and (2 0 0) planes, respectively, cor-
respond to the anatase phase. The absence of diffraction lines at
◦
◦
◦
27.4 , 36.2 and 41.2 , which correspond to the rutile phase, con-
firm that the prepared Pt/TiCD[x] powders consist of pure anatase.
HR-TEM was used to estimate the average Pt particle size, and some
images are shown in Fig. 5, where the dark spots on the large TiO2
(
13 ppm) was not observed (Si impurities). This result is in agree-
1
3
ment with the C NMR characterisation and it would confirm the

C.H. Campos et al. / Catalysis Today 235 (2014) 226–236
231
Table 2
Elemental analysis and physicochemical properties: SBET, pore diameter and metallic nanoparticles diameter of 1%Pt/TiCD(x) catalysts.
CD maintaineda (%)
SBET (m2
−1
g )
Porediameter(nm)
dp HR-TEM (nm)
Sample
N (at%)
TMS-CD (wt%)
d1
d2
TiO2-anatase
–
–
–
–
–
–
142
139
139
134
131
123
123
4.1
4.0
3.9
3.6
3.6
3.3
3.2
9.8
9.5
9.4
9.2
9.1
9.0
8.8
–
1
1
1
1
1
1
%Pt/TiCD(0)
%Pt/TiCD(1)
%Pt/TiCD(5)
%Pt/TiCD(10)
%Pt/TiCD(15)
%Pt/TiCD(20)
2.6
2.1
2.5
2.6
2.6
2.8
0.07
0.14
0.22
0.30
0.34
0.9
2.0
3.2
4.3
4.9
93.5
57.1
52.4
55.6
55.3
a
CD Maintained: (TMS-CDCatalyst/TMS-CDSupport) × 100%.
surface were identified as Pt particles. The average Pt particle size
in Table 2 is smaller than 3 nm for all the catalysts. The size of Pt
particles on TiCD[x] displayed a similar distribution size, indepen-
dent of the TMS-CD content, which differs from the results recorded
using Al O3 and SiO2 materials [20,21]. The stoichiometric addi-
2
tion of NaOH during the Pt nanoparticles synthesis allowed for a
basic medium during the reduction process. In this case, TiO2 was
activated using an aprotic solvent medium to improve hydroxyl for-
mation and to avoid Ti–OH alkoxylation with protic solvent [30,31].
These hydroxyl groups enhance the metal retention on the TiO sur-
2
face for complexing with the PtCl62 metal precursor. In this case,
the supports were contacted with a precursor solution at pH ∼ 2.7.
This phenomenon may be rationalised in terms of the surface prop-
−
2−
erties of TiO2 and the hydrolytic behaviour of PtCl6 in aqueous
solution. In the case of TiO -anatase, the zero point charge (pH
2
zpc) is 6.25 [32]. Hence, at more acidic pH values, the surface is
Fig. 4. XRD patterns of the catalysts synthesised on modified supports. ꢁ corre-
sponds to the anatase-TiO2 phase (JCPDS 89-4921).
positively charged, whereas at pH values above 6.25, it is nega-
2−
tively charged [33]. In an aqueous solution, PtCl6
undergoes a
Fig. 5. HR-TEM micrograph of the 1 wt%Pt catalysts supported on the modified anatase-TiO2 (a) TiCD[0]; (b) TiCD[1]; (c) TiCD[5]; (d) TiCD[10]; (e) TiCD[15] and (f) TiCD[20].

2
32
C.H. Campos et al. / Catalysis Today 235 (2014) 226–236
Scheme 3. Schematic deposition of metal [PtCl6]2 complexes onto TiO2 surfaces.
−
for larger sized Pt clusters, the contribution of Pt␦ species was more
+
sequential hydrolysis to give various hydrolytic products, as men-
tioned by some authors [34–36]:
significant [42]. The Pt 4f7/2 BE appears at ∼73.5 eV. This value is
close to that reported by Huang et al. for similar TiO -supported
2
PtCl 2− + xOH− → [Pt(OH)xCl6−x
2−
+ xCl−
Pt particles and corresponds to a negative shift of ∼0.2 eV com-
]
6
0
pared to the most common value measured for bulk Pt at 73.7 eV
Under a basic medium, the negatively charged surface of TiO2
would explain the anchoring behaviour of the platinum complex.
[41,43]. Such negative shifts can be explained by strong support-
metal interactions. Charge transfer from the support to the particle,
especially in systems with metallic particles on reducible metal
oxides, can alter the electronic properties of the particle [44]. Small
particles also have a large number of surface atoms with reduced
coordination number. This feature is known to induce negative
binding energy shifts relative to the value measured for the bulk
metal [45]. The main ꢂBE to Pt peak (4f7/2) detected for the cata-
lyst supports on TiCD[x] with [x] > 0 showed a major decrease for
the peak positions, compared to those for bulk Pt (∼0.4–0.5 eV),
suggesting that the Pt clusters are negatively charged because of
the electron-donating nature of TMS-CD on the surface. A similar
behaviour has been reported for Pt-PVP stabilised nanoparticles
[46]. Conversely to our previously reported results for SiO2 and ␥-
Al2O3 [20,21], TiO2-anatase modified catalysts showed negatively
charged Pt clusters, and an increase of TMS-CD on the support sur-
face provide an electronic variation of Pt species.
2−
The increase of pH allows the reaction of the PtCl
precursor to
6
give mixed [PtCl6 (OH)x]2 species [36,37]. A series of ligand sub-
−
−x
2−
stitution reactions of [PtCl ] in aqueous hydroxide medium and,
6
at the same time, the TiO surface change, result in anionic species,
2
−
such as Ti–O . This charged oxygen is basic and allows the substi-
tution of the OH groups from the [PtCl6 (OH)x] complex for
Ti–O [37–39], as shown in Scheme 3. However, during the pre-
−
2−
−x
−
cursor reduction, the H2 is consumed and consequently leads to
a decrease of the pH. This is the reason for using Pt:OH = 6 or a
stoichiometric quantity of hydroxyl groups to give a pH neutral
solution.
On a negatively charged surface of TiO , the concentration of
2
2−
[
Pt(OH)xCl6
]
hydrolysed products is expected to be adsorbed on
−x
the surface by the Ti–O-complex bond, which in turn leads to a fast
deposition rate that results in a smaller size of platinum particles. In
alkaline reduction conditions (pH ≥ 9), the shape of the Pt nanopar-
ticles appeared spherical, and the agglomeration of platinum was
not observed. The catalyst prepared on the activated support
The N1s is another important signal in determining the nature of
2
3
TMS-CD on the surface. For aromatic sp or aliphatic sp -hybridised
nitrogen (Nneutral), characteristic peaks at 399.0–399.8 eV were
found [47]. In our case, two contributions for N 1s peak were
detected at 399.8–399.9 and 401.3–401.6 eV. The first corresponds
to Nneutral typical for non–adsorbed TMS-CD molecules [48,49]. The
second can be attributed to positively charged nitrogen (Nionic), in
agreement with previous results reported by Bonello et al. [50] and
Evans et al. [51]. The Nneutral and Nionic signals correspond to both
nitrogen atoms from TMS-CD, and these were not resolved so that
only an average value was determined. In addition, the chemical
shift for the second Nionic signals can be attributed to the Ti–OH
surface interaction with quinuclidinic and quinolinic N atoms due
to an acid–base effect induced by H bond formation (see Table 3).
In previous works, SiO2 and ␥-Al2O3 oxides were used as sup-
ports for the immobilisation of TMS-CD and Pt deposition by the
same synthetic routes. In both supports, a different BE for N 1s was
detected [20,21]. Fig. 6 displays a comparison of ꢂBE for the N 1s
(
2
TiCD[0]) displayed small particle sizes, reaching approximately
.6 nm, and a narrow distribution of particle size values. The organic
molecule does not influence the Pt cluster formed during the metal
reduction. Given these conditions of catalyst preparation, a random
deposition towards immobilised TMS-CD sites was expected.
The oxidation states of the species on the catalysts surface were
investigated by XPS. Table 3 compiles the Pt 4f7/2, N 1s, O 1s and Ti
2
p3/2 binding energies (BE) for all the catalysts under study with the
corresponding contribution to the overall signal in parentheses. The
Pt 4f spectra of the Pt/TiCD[x] catalysts showed only one doublet,
with the most intense Pt4f7/2 component placed at 73.3–73.5 eV.
After precursor reduction in H at room temperature and 40 bar, the
2
metallic clusters displayed small particle sizes. The electronic prop-
erties of these very small Pt clusters on TiO regularly displayed a
2
0
␦+
convolution of mainly Pt and Pt (where 0 < ı < 1) species [40,41];
Table 3
Binding energies (eV) of internal electrons and atomic surface ratio of 1 wt%Pt/TiCD(x) catalysts.
Catalyst
Pt 4f7/2 (eV)
73.6
N 1s (eV)
–
Ti 2p3/2 (eV)
458.6
O 1s
Pt/Ti at
0.0030
N/Ti at
–
1%Pt/TiCD(0)
1%Pt/TiCD(1)
1%Pt/TiCD(5)
1%Pt/TiCD(10)
1%Pt/TiCD(15)
1%Pt/TiCD(20)
529.9 (79)
531.5 (21)
73.4
73.3
73.3
73.3
73.3
399.7 (82)
458.6
458.6
458.6
458.6
458.6
530.0 (84)
531.6 (16)
530.0 (84)
531.5 (16)
530.0 (84)
531.6 (16)
529.9 (84)
531.5 (16)
529.9 (83)
531.5 (17)
0.0046
0.0045
0.0060
0.0088
0.0079
0.035
0.048
0.053
0.067
0.077
4
01.5 (18)
399.8 (75)
01.5 (25)
399.8 (73)
01.4 (27)
399.8 (73)
01.3 (27)
399.9 (71)
01.4 (29)
4
4
4
4

C.H. Campos et al. / Catalysis Today 235 (2014) 226–236
233
Fig. 7. Activity curves based on PPD conversion for 1%Pt/TiCD[x] catalysts. Reac-
−
1
tion conditions: PPD concentration: 0.01 mol L , catalyst mass: 0.050 g, PH2: 40 bar,
stirring speed: 700 rpm, solvent: cyclohexane.
Fig. 6. ꢂBE for the N 1s determinate by XPS for different TMS-CD wt% anchored
on different supports. The data for the SiO2 and ␥-Al2O3 supports was extracted
from [20] and [21], respectively. Percentage contributions to the Pt overall signal in
parentheses.
relation to the carbonyl groups without consecutive hydrogenation
to obtain di-hydroxylated products. All of the catalysts were enan-
tioselective for (R)-1-hydroxy-1-phenylpropan-2-one (1R-PP) and
(
R)-2-hydroxy-1-phenylpropan-1-one (2R-PP), no other hydro-
obtained in these three different supports. The SiO2 modified sup-
ports show a constant ꢂBE attributed to the retro-donation effect
from TMS-CD to Pt nanoparticles because, in absence of TMS-CD,
genated products were detected.
With respect to enantioselectivity, the reaction with activated
and bare TiO2 displayed the racemic mixtures, although the con-
version reached 89%. On the catalysts supported on TiCD[x] with
␦+
the catalyst shows 9% Pt on the surface, and with the increase of
the immobilised-modifier content, an increment of Pt␦+ species was
[
x] > 0, an increase in the conversion to PPD was detected up to
detected. The ␥-Al O modified supports showed a different trend.
2
3
TiCD(5), reaching 95% of PPD conversion and suggesting that the
hydrogenation was a “ligand-accelerated” reaction similar to pre-
vious studies reported by Margitfalvi et al. [52–54]. According to
these authors, this feature can be explained by the formation of a
weak modifier–substrate complex for the supported metallic cat-
alysts where the inducer was added in situ. Table 4 shows the
conversion levels and the pseudo-constant rate, kg. An increase of
the TMS-CD content resulted in higher surface coverage, and the
supported catalysts on TiCD[x] with [x] > 5 showed lower conver-
sion levels. This suggests a change of the metallic site environment.
However, these catalysts showed a regular and uniform metal
size particle distribution, independent of the quantity of anchoring
modifier on the surface. An increase of TMS-CD caused a decrease
in the catalytic activity by surface blockage of the active sites by
repulsive steric interactions with the immobilised inducer. This
behaviour is similar to that previously reported by Toukoniitty et al.
These catalysts displayed a strong Pt–Al O interaction, which pro-
2
3
␦+
vides more positively charged Pt clusters (15% Pt contribution
on ␥-Al O without TMS-CD). This interaction increases with the
2
3
increase of the TMS-CD content because the modifier molecules
cover the support surface; this coverage promotes the effective
interaction between N atoms and Pt clusters. Both supports dis-
played a decrease in the Pt nanoparticles size distribution with
the increase of TMS-CD, in agreement with the strong interac-
tion detected by XPS [45]. In TiO2 supports, a decrease of ꢂBE
was detected with the increase of TMS-CD on the support sur-
0
␦+
face. In these catalysts, convolution of the Pt /Pt contribution in
Pt 4f7/2 signals was not detected because the strong metal sup-
port interaction provides negatively charged Pt nanoparticles. The
active phase did not show a TMS-CD–Platinum retrodonation, and
the Nionic species contribution could only correspond to inducer-
TiO2 acid–basic interaction. This is in agreement with the decrease
of ꢂBE in the TiO2 catalyst because the TMS-CD surface coverage
decreases the direct interaction between Ti–OH and N atoms by
Ti–O–Si bond formation. In addition, the metal particle size dis-
tributions did not show a variation with the TMS-CD content on
surface, as shown by HR-TEM.
[
8,10–12] and Reyes et al. [6,7,55,56] in the same reaction where
the CD was added in situ.
Fig. 8 shows the evolution of the enantiomeric excess (ee) vs.
conversion. All the catalysts were selective for the hydrogenation
However, at higher TMS-CD contents, the surface Pt/Ti atomic
ratio increased, and the entire N/Ti atomic ratio showed a higher
value in relation with the bulk atomic ratios, as shown in Table 3,
mainly due to the surface coverage of the TMS-CD on the TiO2-
anatase support.
3.3. Enantioselective hydrogenation of PPD
3.3.1. TMS-CD Effect
We initially investigated the influence of the TMS-CD amount
on the catalytic performance in the asymmetric hydrogenation
of PPD with Pt/TiCD[x] catalysts and evaluated the optimal con-
centration of TMS-CD. Fig. 7 shows the profiles of conversion
vs. time in PPD hydrogenation. A pseudo-first-order kinetics
with respect to PPD was found in all cases, reaching conver-
sion levels over 85% at 240 min of reaction for the 1%Pt/TiCD[x]
with 0 < [x] < 5 and under 80% for those with [x] < 5. Scheme 2
depicts the hydrogenation reaction of the substrate only in
Fig. 8. Enantiomeric excess vs. PPD conversion for 1%Pt/TiCD[x] catalysts. Reac-
tion conditions: PPD concentration: 0.01 mol L , catalyst mass: 0.050 g, PH2: 40 bar,
stirring speed: 700 rpm, solvent: cyclohexane.
−1

2
34
C.H. Campos et al. / Catalysis Today 235 (2014) 226–236
Table 4
Catalytic data for PPD hydrogenation over 1%Pt/TiCD(x) catalysts. Reaction conditions: PPD concentration: 0.01 mol L−1, PH2: 40 bar, catalyst mass: 0.050 g, stirring rate:
00 rpm, solvent: cyclohexane at 298 K.
7
Conversiona (%)
eeC1b (%)
eeC2b (%)
rsb
kgc (min−1
g
−1
)
k1R (min
−1 −1
g
)
k1S (min
−1 −1
g
)
k2R (min
−1 −1
g
)
−1 −1
k2S (min g )
TMS-CD nominal (%)
0
1
5
0
5
0
89.1
96.3
94.6
77.7
61.7
54.1
0.4
24.9
38.7
51.4
36.5
37.5
0.6
10.3
26.6
–
–
–
62
62
49
–
–
–
0.358
0.642
0.428
0.284
0.128
0.064
–
–
–
–
0.422
0.271
0.217
0.088
0.044
0.199
0.114
0.067
0.040
0.020
0.012
0.028
–
–
–
0.010
0.016
–
–
–
1
1
2
a
b
c
2
40 min of reaction.
Average values.
kg = k1R + k1S + k2R + k2S.
Table 5
Catalytic data for PPD hydrogenation on 1%Pt/TiCD(10) catalyst at different H2 pressures. Reaction conditions: PPD concentration: 0.01 mol L−1, catalyst mass: 0.050 g and
stirring rate: 700 rpm, solvent: cyclohexane at 298 K.
Conversiona (%)
eeC1b (%)
kgc (min−1
g
−1
)
k1R (min
−1 −1
g
)
−1 −1
k1S (min g )
Pressure (bar)
20
40
60
80
64.4
77.7
90.7
91.8
47.3
51.4
51.3
52.0
0.172
0.284
0.382
0.508
0.127
0.217
0.289
0.386
0.045
0.067
0.093
0.122
a
b
c
2
40 min. of reaction.
Average values.
kg = k1R + k1S.
of the carbonyl groups (see Scheme 2). 1R-PP was the main reaction
product, and the eeC1 was similar to those reported for systems in
which the chiral inducer was added in situ and adsorbed on the sur-
face of the catalyst [55,56]. When increasing the TMS-CD amount
on the TiO2 surface, the ee was further improved to a maximum of
Increasing the liquid phase substrate concentrations led to a
significant effect on the catalytic behaviour. Table 6 shows the
kinetic results for all the used concentrations. The conversion and
ee decreased with increasing substrate/metal molar ratio while
keeping the catalyst mass constant. The conversion decreased
drastically with further increase in the PPD concentration, whereas
a different trend in the ee was detected. For all the PPD concen-
trations, the reaction only showed hydrogenation of C1, but the
substrate/metal mole ratio affect the enantioselectivity of this cat-
alyst. At lower concentrations of PPD, the Pt nanoparticles on the
surface located on unattached TMS-CD sites are the most available
for the adsorption of PPD and give racemic mixtures. In contrast, at
higher PPD concentrations the amount of substrate on the Pt surface
should increase. The mechanistic studies reported by Toukoniitty
et al. [10,12] demonstrated the influence on the parallel adsorption
of phenyl groups on Pt clusters controls the ee. When the PPD
concentration increases, different adsorption modes can be occur
on the Pt clusters and provide non-enantioenriched mixtures.
5
1% for the catalyst 1%Pt/TiCD(10). At higher TMS-CD contents, the
conversion and ee value decreased.
The most active catalysts (TiCD[x] with [x] < 10) showed hydro-
genation of the carbonyl of the acetyl group (C2), with 2R-PP being
the preferred product. All these supports show a low loading of
TMS-CD (see Table 2) and possess different type of acid–base sites:
4+
2−
Ti (Lewis acid) and O (Lewis base) ions [28]. The acid sites
can promote the hydrogenation of non-planar groups by interac-
tions with the non-bonding electrons from the carbonyl groups. The
4
+
Lewis acid Ti centres exposed at the surface can interact with the
carbonyl of the acetyl group through linear adsorption and decrease
the ability of parallel adsorption by steric hindrance of the methyl
group, decreasing the effective formation of modifier–substrate
complexes over metallic modified sites and, consequently, the
enantioselection capacity.
3.3.4. Effect of the solvent
The rs values showed a high hydrogenation rate of the carbonyl
adjacent to the phenyl group (C1), with values of 67 for the cata-
lyst 1% Pt/TiCD(0). This is due to the high adsorption capacity of
the phenyl group of PPD at the active sites (for ␩ bond forma-
tion between the active phase and the ␲ aromatic cloud of the
phenyl group), in line with studies by Toukoniitty et al. [8,10–12].
Pt nanoparticles are negatively charged, and this property could
enhance the regioselectivity.
In heterogeneous asymmetric hydrogenation, the enantioselec-
tivity is very sensitive to the solvent used, and some examples of
this effect can be found in literature for the hydrogenation of PPD
catalysed by Pt–CD [21,52,57,58]. The effect of solvents with dielec-
tric constants ranging from 2 to 25 (values of pure solvents) have
been included in this study for the hydrogenation of PPD catalysed
by 1%Pt/TiCD(10). The results are displayed in Fig. 9 and corre-
spond to a qualitative model including the enantioselectivity and
the dielectric constant of the pure solvents, similar to that pro-
posed by Toukoniitty et al. [58]. The highest activity was observed in
ethyl acetate, and very low reaction rates were obtained in tetrahy-
drofuran, whereas relatively high rates were obtained in the other
solvents in this study. Toukoniitty et al. [58] reported that the initial
hydrogenation rate does not correlate with the hydrogen solubil-
ity. In terms of enantioselectivity, among the solvents investigated,
cyclohexane led to the best e.e. values of 51% at 78% of conver-
3.3.2. Effect of H2 pressure
After finding the most enantioselective catalytic system, i.e.,
Pt/TiCD(10), the effect of H2 pressure over the PPD hydrogenation
was studied, and the results are shown in Table 5. The surface H2
concentration may directly influence the adsorption of the reactant
and thus, the enantioselectivity. An H2 pressure increase (from 20
to 80 bar) had a marked effect on the conversion levels but only a
discreet effect on the eeC1 value. The best eeC1 value was achieved
at 40 bar, similarly to that previously reported [20,21].
2
sion. Conversely to previous reported results for TMS-CD–␥-Al O3
supports, toluene showed a lower reaction rate in comparison to
3
.3.3. Effect of PPD concentration
The enantioselective 1%Pt/TiCD(10) catalysed hydrogenation
cyclohexane [21]. Both solvents have similar dielectric constants
and have shown high conversion and ee in PPD hydrogenation
using the same reaction conditions [21]. In our case, the electronic
reaction was performed with different PPD initial concentrations.

C.H. Campos et al. / Catalysis Today 235 (2014) 226–236
235
Table 6
Catalytic data for PPD hydrogenation on 1%Pt/TiCD(10) catalyst at different PPD concentrations. Reaction conditions: PH2: 40 bar, catalyst mass: 0.050 g and stirring rate:
00 rpm, solvent: cyclohexane at 298 K.
7
Concentration (mol L 1
−
)
Mole ratio PPD/Pttotal
Conversiona (%)
eeC1b (%)
kgc (min
−1 −1
g
)
k1R (min
−1 −1
g
)
−1 −1
k1S (min g )
0.005
0.010
0.025
0.100
25
50
125
500
94.5
77.7
68.7
51.3
38.7
51.4
47.6
43.4
0.412
0.284
0.172
0.08
0.299
0.217
0.129
0.058
0.113
0.067
0.043
0.022
a
b
c
2
40 min of reaction.
Average values.
kg = k1R + k1S.
Table 7
EA, HR-TEM mean Pt sizes and XPS data for recycles of the catalyst 1%Pt/TiCD(10). Reaction conditions: PPD concentration: 0.01 mol L−1, catalyst mass: 0.070 g per cycle, PH2:
0 bar, stirring rate: 700 rpm at 298 K.
4
b
Cycle
Fresh
N(%)a
0.22
TMS-CD (wt %)
3.2
Pt 4f7/2 (eV)
73.3
N1s (eV)
Pt/Ti at
0.0060
N/Ti at
0.053
dp HR-TEM (nm)
399.8 (73)
2.6
–
4
01.3 (27)
399.9 (74)
01.3 (26)
400.0 (75)
01.3 (25)
400.0 (75)
01.4 (25)
400.1 (76)
01.4 (23)
400.1 (76)
01.4 (23)
1
0.22
0.22
0.22
0.22
0.22
3.2
3.1
3.2
3.1
3.1
73.3
73.4
73.3
73.4
73.4
0.0056
0.0064
0.0058
0.0062
0.0049
0.057
0.063
0.058
0.052
0.055
4
2
–
4
3
–
4
4
–
4
Final
2.5
4
a
Average value from two measurements.
Only to fresh catalyst and after the last run was measured.
b
nature of Pt nanoparticles provided a competition between PPD and
toluene for ␩ bond interactions formation on the active phase sur-
face (␲ aromatic bonds), decreasing the conversion levels and ee.
The ee was almost 34% in ethanol and decreased to 22% in toluene.
Finally, ee decreased non-linearly with increasing solvent dielec-
tric constant, which was attributed to the interactions between
solvents and TMS-CD on the surface. On the other hand, these
results do not show a clear correlation between the catalytic per-
formance and solvent polarity, in agreement with similar studies
chloroform to remove the organic matter that could have been
adsorbed on the support and the TMS-CD from leaching [59]. Fig. 10
shows the variation of the conversion and ee levels as a function of
the number of reaction cycles. All the cycles show a pseudo-first-
order trend during the hydrogenation reaction. As the number of
reaction runs increased, the activity decreased, reaching values of
30% in the last cycle.
Table 7 shows the catalyst characterisation after each cycle. N(%)
shows constants values during the cycles, suggesting that non-
leaching of the inducer was detected. The active phase showed
constant values for the Pt BE and Pt/Ti surface atomic ratios after
all the cycles. The nanoparticle distribution sizes displayed a sim-
ilar trend and regular shapes as fresh 1%Pt/TiCD(10) catalyst. For
both reasons, non-leaching and/or sintering of nanoparticles was
assumed in all runs. The XPS shows the presence of two types
of surface nitrogen, Nneutral and Nionic, as was discussed in pre-
vious sections. The remaining anchored chiral inducer is slightly
reported by Toukoniitty et al. for commercial Pt, Pd and Rh/␥-Al O3
2
supported catalysts [8].
3.3.5. Recycling
The catalyst 1%Pt/TiCD(10) catalyst was used to perform
recycling studies. Between cycles, the catalyst was washed with
Fig. 9. Curves of the enantiomeric excess and PPD conversion vs. dielectric constant
for PPD hydrogenation on 1%Pt/TiCD[10] at different solvents. Reaction conditions:
PPD concentration 0.01 mol L , PH2: 40 bar, catalyst mass: 0.050 g and stirring
Fig. 10. Activity and ee as a function of the number of reaction cycles for the catalyst
1%Pt/TiCD[10]. Reaction conditions: PPD concentration: 0.01 mol L , catalyst mass:
−1
−1
speed: 700 rpm.
0.070 g, PH2: 40 bar, stirring speed: 700 rpm, solvent: cyclohexane.

2
36
C.H. Campos et al. / Catalysis Today 235 (2014) 226–236
hydrogenated in the quinoline cycle of the TMS-CD because a slight
Nneutral BE shift to a higher value was detected. The N 1s for aliphatic
N (piperidine contribution) at 400.0 eV has been reported by Luo
et al. [60,61]. These N 1s BE values are in agreement with Nneutral and
Nionic contributions because in all the cycles, these values were con-
stant. For quinoline hydrogenation, in the absence of leaching, the
contribution of ionic species should not increase. TMS-CD hydro-
genation on the support led to lower activity because they might
be interfering with the adsorption of the substrate on the active
sites. Subsequently, this blockage poisons the active sites close to
the modified sites, and the hydrogenation only takes place on the
unmodified Pt nanoparticles. Consequently, the reaction on these
sites leads to a racemic mixture and thus decreasing ee values.
[12] E. Toukoniitty, B. Sˇ ev cˇ íková, P. Mäki-Arvela, J. Wärnå, T. Salmi, D.Y. Murzin, J.
Catal. 213 (2003) 7–16.
[
[
13] V.B. Shukla, P.R. Kulkarni, World J. Microbiol. Biotechnol. 16 (2000) 499.
14] A. Ghosh, R. Kumar, J. Catal. 228 (2004) 386–396.
[15] D. Ruiz, C. Mella, P. Reyes, J.L.G. Fierro, J. Chil. Chem. Soc. 57 (2012) 1394–
399.
16] D. Ruiz, M. Oportus, C. Godard, C. Claver, J.L.G. Fierro, P. Reyes, Curr. Org. Chem.
6 (2012) 2754–2762.
1
[
1
[17] C.E. Song, S.-G. Lee, Chem. Rev. 102 (2002) 3495–3524.
[18] D.D. Vos, I.F.J. Vankelecom, P.A. Jacobs, Chiral Catalyst Immobilization and
Recycling, Toronto, 2000.
[
19] M.U. Azmat, Y. Guo, Y. Guo, Y. Wang, G. Lu, J. Mol. Catal. A: Chem. 336 (2011)
42–50.
[20] C.H. Campos, M. Oportus, C. Torres, C. Urbina, J.L.G. Fierro, P. Reyes, J. Mol. Catal.
A: Chem. 348 (2010) 30–41.
[
21] C.H. Campos, C. Torres, J.L.G. Fierro, P. Reyes, Appl. Catal. A: Gen. 466 (2013)
98–207.
1
[
[
[
22] D. Ruiz, J.L.F. Fierro, P. Reyes, J. Braz. Chem. Soc. 21 (2010) 262–269.
23] D. Ruiz, P. Reyes, J. Chil. Chem. Soc. 54 (2008) 1740–1742.
24] A. Lindholm, P. Mäki-Arvela, E. Toukoniitty, T.A. Pakkanen, J.T. Hirvi, T. Salmi,
D.Y. Murzin, R. Sjöholm, R. Leino, J. Chem. Soc. Perkin 1 12 (2002) 2605.
4
. Conclusions
New catalyst systems were synthesised for the enantioselec-
tive hydrogenation of PPD. Cinchonidine was tethered directly
with prior silanisation modification over activated anatase-TiO2.
The supported metallic catalysts were obtained by the reduction
[25] E. Toukoniitty, P. Mäki-Arvela, A.K. Neyestanaki, T. Salmi, R. Sjöholm, R. Leino,
E. Laine, P.J. Kooyman, T. Ollonqvist, J. Väyrynen, Appl. Catal. A: Gen. 216 (2001)
73–83.
[26] C.D. Wagner, L.E. Davis, M.V. Zeller, J.A. Taylor, R.H. Raymond, L.H. Gale, Surf.
Interface Anal. 3 (1981) 211–225.
[
27] V. Nieminen, A. Taskinen, M. Hotokka, D.Y. Murzin, J. Catal. 245 (2007) 228–
36.
28] G. Martra, App. Catal. A: Gen. 200 (2000) 275–285.
of a metallic precursor under high H2 pressure. The characterisa-
tion data provide evidence of the immobilisation of cinchonidine.
The HR-TEM images and XPS demonstrated that the cinchonidine
content does not affect the active phase dispersion and parti-
cle size distribution. It can be deduced from the XPS results that
the Pt nanoparticles exhibited a negative charged surface, which
was attributed to the metal deposition method and the strong
interaction with the TiO2 surface. The 1% Pt/TiCD[x] catalysts dis-
played relatively high conversion and ee values (51.0%). The catalyst
with [x] = 10 presented performances similar to those previously
2
[
[29] K.S.W. Sing, Pure Appl. Chem. 54 (1982) 2201–2218.
[
[
30] J. Blümel, J. Am. Chem. Soc. 117 (1995) 2112–2113.
31] C.-H. Chu, E. Jonsson, M. Auvinen, J.J. Pesek, J.E. Sandoval, Anal. Chem. 65 (1993)
808–816.
[32] G.A. Parks, Chem. Rev. 65 (1965) 177–198.
[33] M. Muneer, M. Qamar, M. Saquib, D.W. Bahnemann, Chemosphere 61 (2005)
457–468.
[
34] L.E. Cox, D.G. Peters, E.L. Wehry, J. Inorg. Nucl. Chem. 34 (1972) 297–305.
[35] Z. Jin, Z. Chen, Q. Li, C. Xi, X. Zheng, J. Photochem. Photobiol. A: Gen. 81 (1994)
177–182.
[
36] W.A. Spieker, J. Liu, J.T. Miller, A.J. Kropf, J.R. Regalbuto, Appl. Catal. A: Gen. 232
2002) 219–235.
reported for a commercial Pt/Al O3 catalyst, where untethered
2
(
cinchonidine was used. The effect of the H2 pressure, PPD concen-
tration in the liquid phase and solvents indicated that the reaction
conditions were important in controlling the activity and enan-
tioselectivity. This is explained by the substrate adsorption on the
Pt sites. The solvent effect studies showed non-linear correlation
between enantiomeric excess and the dielectric constant of the
pure solvent. The most selective catalyst showed good reusabil-
ity under optimum conditions and deactivation by hydrogenation
of TMS-CD after the third cycle.
[37] W.A. Spieker, J. Liu, X. Hao, J.T. Miller, A.J. Kropf, J.R. Regalbuto, Appl. Catal. A:
Gen. 243 (2003) 53–66.
[
38] O.B. Bel’skaya, R.K. Karymova, D.I. Kochubey, V.K. Duplyakin, Kinet. Catal. 49
2008) 720–728.
39] B. Shelimov, J.-F. Lambert, M. Che, B. Didillon, J. Catal. 185 (1999) 462–478.
(
[
[40] S.-H. Chien, M.-C. Kuo, C.-H. Lu, K.-N. Lu, Catal. Today 97 (2004) 121–127.
[
41] B.-S. Huang, F.-Y. Chang, M.-Y. Wey, Int. J. Hydrogen Energy 35 (2010)
699–7705.
7
[
42] J. Croy, S. Mostafa, J. Liu, Y.-h. Sohn, B. Roldan Cuenya, Catal. Lett. 118 (2007)
1–7.
[
[
[
43] J.L.G. Fierro, J.M. Palacios, F. Tomas, Surf. Interface Anal. 13 (1988) 25–32.
44] S. Laursen, S. Linic, Phys. Rev. Lett. 97 (2006) 026101.
45] O. Björneholm, F. Federmann, F. Fössing, T. Möller, Phys. Rev. Lett. 74 (1995)
3017–3020.
Acknowledgements
[
[
46] M.J. Hossain, H. Tsunoyama, M. Yamauchi, N. Ichikuni, T. Tsukuda, Catal. Today
183 (2012) 101–107.
The authors thank CONICYT for the financial support (FONDE-
CYT GRANT 1061001, FONDECYT Initiation GRANT 11121631,
FONDECYT postdoctoral 3130483) and REDOC.CTA doctoral net-
work.
47] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-ray Photo-
electron Spectroscopy, Eden Prairie, 1995.
[48] J.M. Bonello, R.M. Lambert, Surf. Sci. 498 (2002) 212–228.
[
[
[
49] A.F. Carley, M.K. Rajumon, M.W. Roberts, P.B. Wells, J. Chem. Soc. Faraday Trans.
91 (1995) 2167–2172.
50] J.M. Bonello, F.J. Williams, R.M. Lambert, J. Am. Chem. Soc. 125 (2003)
2723–2729.
51] T. Evans, A.P. Woodhead, A. Gutiérrez-Sosa, G. Thornton, T.J. Hall, A.A. Davis,
N.A. Young, P.B. Wells, R.J. Oldman, O. Plashkevych, O. Vahtras, H. Ågren, V.
Carravetta, Surf. Sci. 436 (1999) L691–L696.
References
[
[
[
[
[
1] H.-U. Blaser, F. Spindler, M. Studer, Appl. Catal. A: Gen. 221 (2001) 119–143.
2] T. Katsuki, Comprehensive Asymmetric Catalysis, Berlin, 1999.
3] D. Pauluth, A.E.F. Wachter, Chirality in Industry II, 1997.
4] M. Studer, H.-U. Blaser, C. Exner, Adv. Synth. Catal. 345 (2003) 45–65.
5] W.A.H. Vermeer, A. Fulford, P. Johnston, P.B. Wells, J. Chem. Soc. Chem. Com-
mun. (1993) 1053–1054.
[
[
52] J.L. Margitfalvi, E. Tálas, App. Catal. A: Gen. 301 (2006) 187–195.
53] J.L. Margitfalvi, E. Tálas, E. Tfirst, C.V. Kumar, A. Gergely, Appl. Catal. A: Gen. 191
(
2000) 177–191.
[
[
54] J.L. Margitfalvi, E. Tfirst, J. Mol. Catal. A: Chem. 139 (1999) 81–95.
55] T. Marzialetti, M. Oportus, D. Ruiz, J.L.G. Fierro, P. Reyes, Catal. Today 133–135
[
[
[
6] N.M. Marin, G. Pecchi, P. Reyes, React. Kinet. Catal. Lett. 87 (2006) 121–128.
7] T. Marzialetti, J.L.G. Fierro, P. Reyes, Catal. Today 107–108 (2005) 235–243.
8] E. Toukoniitty, S. Franceschini, A. Vaccari, D.Y. Murzin, Appl. Catal. A: Gen. 300
(
2008) 711–719.
[
[
56] C. Urbina, C.H. Campos, P. Reyes, P. Pecchi, J. Chil. Chem. Soc. 54 (2010) 331–336.
57] B. Toukoniitty, E. Toukoniitty, P. Mäki-Arvela, J.-P. Mikkola, T. Salmi, D.Y.
Murzin, P.J. Kooyman, Ultrason. Sonochem. 13 (2006) 68–75.
(
2006) 147–154.
9] E. Toukoniitty, P. Mäki-Arvela, N. Kumar, T. Salmi, D.Y. Murzin, Catal. Today
9–80 (2003) 189–193.
10] E. Toukoniitty, P. Mäki-Arvela, J. Wärnå, T. Salmi, Catal. Today 66 (2001)
11–417.
[
[
58] E. Toukoniitty, P. Mäki-Arvela, J. Kuusisto, V. Nieminen, J. Mol. Catal. A: Chem.
7
192 (2003) 135–151.
[
[
[
[
[
59] A. Vargas, T. Bürgi, A. Baiker, J. Catal. 226 (2004) 69–82.
60] X. Luo, S.H. Goh, S.Y. Lee, Macromolecules 30 (1997) 4934–4938.
61] X. Luo, S.H. Goh, S.Y. Lee, K.L. Tan, Macromolecules 31 (1998) 3251–3254.
4
11] E. Toukoniitty, V. Nieminen, A. Taskinen, J. Päivärinta, M. Hotokka, D.Y. Murzin,
J. Catal. 224 (2004) 326–339.