School of Engineering and Science - Jacobs University

Preparation and Characterization of Pt and Pd Nanoclusters Modified with
Chiral Ligands: Examination of Catalytic Activity in Hydrogenation of
Ethyl Pyruvate
by
Alexander Kraynov
A thesis submitted in partial fulfilment
of the requirements for the degree of
Doctor of Philosophy
in Chemistry
Approved, Thesis Committee
Prof. Dr. Ryan M. Richards
Name and title of chair
Prof. Dr. Thomas Nugent
Name and title of committee member
Prof. Dr. Stefan Tautz
Name and title of committee member
Dr. N. Waldoefner
Name and title of committee member
27 September 2006
School of Engineering and Science

Table of Contents
Acknowledgment
Abstract
IV
V
Chapter 1
Introduction to the enantioselective catalysis 1
1.1 General aspects: optical activity and chirality 2
1.2 Chirality in our life 4
1.3 Obtaining pure enantiomers 4
1.3.1 Separation of enantiomers 4
1.3.2 Asymmetric synthesis 7
1.3.3 Asymmetric catalysis 7
Chapter 2
Introduction to nanoscale materials 18
2.1 General information and definitions 18
2.2 Methods of colloid and supported nanoparticles preparation 21
2.3 Preparation of nanoclusters on a heterogeneous support 24
Chapter 3
Current thesis in the EU-COST project: aims of the work 25
3.1 Introduction and goals setting 25
Chapter 4
Cinchonidine modified Pt colloidal nanoparticles: characterization and
catalytic properties 28
4.1 Introduction 28
4.2 Experimental 29
4.3 Results and discussion 31
4.3.1 General characterization 31
4.3.2 FTIR investigation of cinchonidine adsorbed on Pt 35
4.3.3 Investigation of catalytic behavior of cinchonidine
modified Pt nanoclusters 43
4.4 Summary 53
Chapter 5
Cinchonidine modified Pt colloidal nanoparticles immobilized on a
heterogeneous support 55
5.1 Introduction 55
5.2 Experimental 56
5.3 Results and discussion 57
5.3.1 Role of catalyst activation in obtaining enantiomeric
excess 57
5.3.2 Comparison of catalysts and proposed model of
catalyst activation 58
5.3.3 Catalyst stability and reuse tests 61
5.3.4 IR investigation and proposed model of activation 63
ii

5.4 Summary 66
Chapter 6
Cinchonidine modified Pd colloidal nanoparticles 67
6.1 Introduction 67
6.2 Experimental 68
6.3 Results and discussion 69
6.4 Summary 74
Chapter 7
Quiphos modified Pt and Pd colloidal nanoparticles 75
7.1 Introduction 75
7.2 Experimental 76
7.3 Results and discussion 77
7.4 Summary 85
Chapter 8
“Quiphos-spider” modified Pt colloidal nanoparticles 86
8.1 Introduction 86
8.2 Experimental 86
8.3 Results and discussion 87
8.4 Summary 89
Chapter 9
Diphos and diop modified Pt and Pd colloidal nanoparticles 90
9.1 Introduction 90
9.2 Experimental 90
9.3 Results and discussion 91
9.4 Summary 102
Chapter 10
Binap and synphos modified Pt colloidal nanoparticles 103
10.1 Introduction 103
10.2 Experimental 103
10.3 Results and discussion 104
10.4 Summary 111
References 112
iii

Acknowledgment
I extend my sincere gratitude and appreciation to many people who made this doctoral
thesis possible. Special thanks are due to my supervisor Prof. Ryan Richards for his
timely advices and cooperation.
Thanks are also due to my former and current lab mates Mr. A. Suchopar for help with
experiment organisation, Dr. K. Zhu and Dr. L. D’Souza and Dr. F. Pozgan, for many
useful advices and directing, Dr. J. C. Hu and Mrs. L. F. Chen for general help and
cooperation in lab work, Mr. L. Zhi and Mrs. Helga van Eys-Schaefer with Mrs. Kirsten
Töbe (Fraunhofer-Institut für Fertigungstechnik und Angewandte Materialforschung,
Bremen) for very important and high quality TEM photographs.
Author thanks Prof. Gerard Buono and Dr. Didier Nuel of Université Aix-Marseille III,
France for providing the quiphos and quiphos-spider ligands and for useful discussion
concerning the quiphos modifier, Dr. Virginie Ratovelomanana-Vidal (Laboratoire de
Chimie Organique, E.N.S.C.P., UMR) for providing the synphos ligand and her
colleague Dr. Véronique Michelet for useful discussion about synphos ligand.
I would like to thank Prof. U. Kortz (International University Bremen, Bremen) and Dr.
E. Kirilina (Physikalisch-Technische Bundesanstalt, Berlin) for useful discussions
about NMR spectra, Prof. Thomas Nugent (International University Bremen, Bremen)
and Prof. S. Tautz (International University Bremen, Bremen) for useful discussions
about organic chemistry and surface science, correspondingly, as well as for being my
work supervisors, Dr. T. Mallat and Prof. A. Baiker (Swiss Federal Institute of
Technology, Zurich) for useful discussions about cinchonidine on Pd and Pt systems,
Dr. Achim Gelessus (International University Bremen, Bremen) for help with work
with Gaussian 03 software, Dr. L. G. Gordeeva (Boreskov Institute of Catalysis,
Novosibirsk) for discussions about surface chemistry of alumina and silica materials,
Dr. A. G. Okunev (Boreskov Institute of Catalysis, Novosibirsk) for discussion about
kinetics of heterogeneous reactions, Dr. Serguei Soubatch (International University
Bremen, Bremen) for discussions about chemical physics of surfaces.
I am highly indebted to Dr. Yu.I.Aristov (Boreskov Institute of Catalysis, Novosibirsk),
Prof. S. A. Dzuba and Mrs. R. I. Ratushkova (Chair of Chemical and Biological
Physics, Novosibirsk State University, Novosibirsk) for introducing me to the world of
science and right directing and general support.
My gratitude goes to the International University Bremen for providing me a financial
support during my Ph.D. studies and making this work possible.
Finally, my special gratitude goes to my wife Mrs. Liliya Kulishova and my family for
their cooperation, support and encouragement.
iv

Abstract
Studies of the interactions of organic substrates with nanoscale metals and metal oxides
are of fundamental importance for both fundamental scientific understanding and the
development of commercial applications. Additionally, the development of
heterogeneous (solid/liquid or solid/gas) enantioselective catalysts is of great interest
for the pharmaceutical industry. Here, the adsorption of chiral ligands on metal surfaces
to develop heterogeneous asymmetric catalysts is presented. Functionalizing metal
nanoparticles with chiral ligands facilitates increased surface coverage and the observed
activity of these catalysts is significantly higher than their conventional phase
counterparts.
The first two chapters are an introduction to the enantioselective catalysis and
nanosized materials. As described in the chapter 3, the primary objectives of this work
are
1. to synthesize Pt and Pd nanoclusters modified by specific chiral ligands by
`bottom up' chemical synthesis,
2. to determine the geometric orientation of adsorbed modifier with respect to
the metal surface via DRIFTS (Diffuse Reflectance Infra-red Fourier
Transform Spectroscopy) and molecular modeling,
3. test enantioselective activity of the obtained samples in the hydrogenation of
ethyl pyruvate. The special effort has been given for investigation of
cinchonidine as modifier (chapter 4, 5, 6).
Among the following chiral ligands: quiphos (chapter 7), “quiphos-spider” (chapter 8),
binap and synphos (chapter 10), only diop (chapter 9) was found to be able to induce
low enantioselectivity, that was qualitatively explained through comparison with
cinchonidine/Pt system.
v

Chapter 1
Introduction to the enantioselective
catalysis
1.1 General aspects: optical activity and chirality
Any material that rotates the plane of polarized light is said to be optically active and
molecule of this materials is nonsuperimposable on its mirror image (Fig. 1-1). If a
molecule is superimposable on its mirror image, the compound does not rotate the
plane of polarized light; it is optically inactive. The property of nonsuperimposability of
an object on its mirror image is called chirality. If a molecule is not superimposable on
its mirror image, it is chiral. Inverse logic is also true, i.e. if a molecule is
superimposable on its mirror image, it is achiral. This is a necessary and sufficient
requirement.
Fig. 1-1. Superimposable (up) and nonsuperimposable (bottom) models of molecules.
If a molecule is nonsuperimposable on its mirror image, the mirror image must be a
different molecule, since superimposability is the same as identity. Both real molecule
and its mirror image are called enantiomers. Enantiomers have identical (without
considering interaction between elemental particles i.e. weak interaction [1]) physical
and chemical properties except in two important respects:
1. They rotate the plane of polarized light in opposite direction, though in equal
amounts.
2. They react at different rate with other chiral compounds. These rates may be so
close together that the distinction is practically useless, or reaction rate of one
enantiomer might be significantly far from another one. This is the reason that

many compounds are biologically active while their enantiomers are not active
or lead to different reaction results.
The mixture of enantiomers in equal amounts is called racemic mixture or racemate.
The separation of a racemic mixture into its two enantiomers is called resolution. It is
important to mention that the presence of optical activity always proves that a given
compound is chiral, but its absence does not prove that the compound is achiral. A
compound that is optically inactive may be achiral, or it may be a racemic mixture.
Optically active compound may be classified into several categories.
1. Compound with a chiral carbon atom. If there is only one carbon atom with
four different substitutes (so called chiral carbon atom), the molecule must be
optically active. The optical activity has been detected even for compounds,
where one group is hydrogen and another deuterium i.e. 1-butanol-1-d ( Fig 1-2)
[2].
C 2 H 5
H
C
D
Fig. 1-2. 1-butanol-1-d [3].
2. Compound with other quadrivalent chiral atoms. A molecule containing an
atom that has four bonds pointing to the corner of a tetrahedron will be optically
active if the four groups are different. Among atoms in this category are Si [4],
Ge [5], Sn [5] and N (in quaternary salts or N-oxides).
3. Compounds with tervalent chiral atoms. Atoms with pyramidal bonding might
be expected to be optically active if the atom is connected to three different
groups, since the unsaturated pair of electrons is analogous to a fourth group.
For example, a secondary or tertiary amine (Fig. 1-3) where X, Y and Z are
different. However due to the rapid oscillation of the lone pair from one side of
the XYZ plane to the other and following conversion of the molecule into its
enantiomers it is very difficult to isolate and even detect an enantiomer. The
inversion is less rapid in substituted ammonias [6-8] and phosphates [9-12] and
in suitable cases the inversion energetic barrier can be increased and such
enantiomers could be detected and separated at room temperature [13, 14].
OH
Z
N
X Y
Fig. 1-3. Chiral amine [3].
4. Suitably substituted adamantine. Adamantines bearing four different substitutes
at the bridgehead positions are chiral and optically active and could be resolved
[15]. This type of molecule (Fig. 1-4) is a kind of expanded tetrahedron and has
the same symmetry properties as any other tetrahedron.
CH 3
H
COOH
Br
Fig. 1-4. Example of adamantane [3].
2

5. Restricted rotation giving rise to perpendicular dissymmetric planes. Some
compound what do not have asymmetric atoms are nevertheless chiral because
their structure could be represented as two perpendicular planes neither of
which can be bisected by a plane of symmetry (Fig. 1-5). There are few
examples for illustration below. Biphenyls containing four large groups can not
freely rotate about the central bond because of steric hindrance (Fig. 1-6). In
such compound the two rings are in perpendicular planes, that fits to the
structure in Fig. 1-5.
Fig. 1-5. Two perpendicular planes [3].
There are few examples for illustration below.
Cl
O 2 N
COOH
HOOC
NO 2
Cl
O 2 N
COOH
HOOC
NO 2
A
P(Ph) 2
P(Ph) 2
(Ph) 2 P
(Ph) 2 P
Mirror
Fig. 1-6. Chiral 3'-chloro-2',6'-dinitrobiphenyl-2,6-dicarboxylic acid (A) and binap (B)
[3].
6. Chirality due to a helical shape. Some molecules existing in helical shape with
left- or right-handed orientation (Fig. 1-7). The entire molecule is usually less
that one full turn of the helix, but this does not alter the possibility of left- and
right-handedness.
B
3

Fig. 1-7. Hexahelicene [3].
7. Chirality caused by restriction rotation of other types. In this case (e.g. Fig. 1-8.)
chirality results because the benzene ring can not rotate in such a way that the
carboxyl group goes through the alicyclic ring.
HOOC
(CH 2 ) 10
Fig. 1-8. Example of chiral molecule with restricted rotation [3].
1.2 Chirality in our life
Chemical compound like drugs, fragrances, agrochemicals, etc. are needed in
enantiopure form, because the all living organisms are built from chiral molecules, e.g.
DNA. The enantiopure drugs, for instance, thus can work more efficiently than in
racemate form because they influence on specific biochemical reactions by a lock-key
mechanism. When one enantiomer is needed as an active compound, the other one is
ballast, in the best case, or can induce a poison effect, like the (S)-thalidomide (where it
was in racemic mixture with (R)- thalidomide), which caused birth defects in the
beginning of 1960s. In present days drugs manufacturers need enantiopure chemical
compound for producing enantiopure drugs. Thus the market related to the singleenantiomer
drug business reached 40 % of total drug market in 2000. Whereas it
covered 4 % only in 1985.
1.3 Obtaining pure enantiomers
Since huge number of natural compound are (e.g. cinchonidine etc.) chiral thus one of
the simplest way for obtaining chiral compounds is through extraction from nature i.e.
from plants, living organisms etc. However it is clear that not all needed chiral
compounds are available from the natural “chiral pool”. At the same time difficulties in
obtaining chiral compound in sufficient amounts might lead to increases of products
cost. Therefore methods for obtaining required enantiopure compounds have to be
developed.
1.3.1 Separation of enantiomers
The resolution of the racemic mixture could be performed through removal or
decomposition partially or completely of one enantiomer.
Mechanical separation
In 1848 L. Pasteur has shown an opportunity of mechanical separation of crystals of Na
salt of racemic tartaric acid. In this example crystals of enantiomers form separately
and visually look different, like mirror reflections of each other (Fig. 1-9), obtaining
4

crystals are called enantiomorphic crystals. This method is known to as spontaneous
resolution by crystallization.
Fig. 1-9. Enantiomorphic crystals. Abstract model [16].
Preferential crystallization by inoculation
The selective crystallization of a single enantiomer is possible to perform in the
presence of traces of the same enantiomer that is to be separated. In this case an
oversaturated racemic solution is "seeded" in small amount of pure enantiomer and the
resulting precipitate will have domination of the enantiomer in the seed. However, with
the exception of the separation of glutamic acid and of a copper complex of DLaspartic
acid, this method has been found to be impractical or resulted in portion
resolution only.
Biochemical processes
The chiral compounds that react at different rates with the two enantiomers may be
present in a living organism. For example, in 1858 Pasteur discovered that aqueous
solution of racemic tartaric acid after contact with the mould fungus (Penicillium
glaucum) slowly became enantioenriched. The mould was preferentially metabolizing
the right rotatory enantiomer. This method is limited, since it is necessary to find the
proper organism and since one of the enantiomer is destroyed in the process. However,
when the proper organism is found, the method leads to a high extent of resolution
since biological processes are usually very stereoselective. Another disadvantage of this
method is that only dilute solutions can be used.
Conversion in diastereomers
If the racemic mixture to be resolved contains a carboxyl group (and no strongly basic
group), it is possible to form a salt with an optically active base. Fig. 1-10 illustrates
reaction of S-brucine with racemic hydroxypropanoic acid produced a mixture of two
salts having the configuration SS and RS. Although the acids are enantiomers, the salts
are diastereomers and have different properties e.g. solubility. The mixture of
diastereometric salts is allowed to crystallize from a suitable solvent.
5

R
H
COOH
C OH
CH 3
H
COO - Bricine-H +
C OH
CH 3
S-brucine
R
S
S
COO - Bricine-H +
COOH
HO C H
HO C H
CH 3
CH 3
S S
Fig. 1-10 formation of diastereometric salts [3].
Since the solubilities are different, the initial crystals formed will be richer in one
diastereomer. Once the two diastereomers have been separated, it is easy to convert the
salt back to the free acids and the recovered base can be used again. Advantages of this
approach are that naturally occurring optically active bases (mostly alkaloids: brucine,
ephedrine, morphine, etc.) are readily available. However, unfortunately, the difference
in solubilities is rarely if ever great enough to effect total separation with one
crystallization.
Differential absorption
When a racemic mixture is placed on a chromatographic column consisting of chiral
phase, then, in principal the enantiomers should move along the column at different
rates and should be separated. This technique is widely in use in different
chromatography applications: GC, HPLS.
Chiral recognition
In order to separate enantiomers from racemic mixture the use of chiral host to form
diastereometric inclusion is possible. Some hosts can form an inclusion with one
enantiomer only or this formation is faster then with another enantiomer. For example,
chiral host (Fig. 1-11-A) can partially resolve the racemic amine salt [17] (Fig. 1-11-B).
O
O
O
O
O
O
Ph
CH 3
C
+
NH 3
-
PF 6
H
A
B
Fig. 1-11. Chiral host-A and guest molecule –B [3].
Kinetic resolution
Since enantiomers react with chiral compounds at different rates, it is sometimes
possible to effect a partial separation by stopping the reaction before completion. An
example is the resolution of allylic alcohol [18] (Fig. 1-12-A) with one enantiomer of a
chiral epoxidizing agent, in this case the discrimination was extreme. One enantiomer
was converted to the epoxide and the other was not, the ratio being more than 100.
Reactions catalyzed by enzymes can be utilized for this kind of resolution.
6

Bu
O
Bu
Bu
OH
OH
OH
racemic (R)-enantiomer (S)-enantiomer
A B
Fig. 1-12. Kinetic resolution of racemic allylic alcohol –A, yielding its (S) enantiomer
(B) [3].
1.3.2 Asymmetric synthesis
There are two basic ways to synthesize a chiral compound. The first way is to begin
with single stereoisomer and to use a synthesis that does not affect the chiral centre.
The other basic method is called asymmetric synthesis. The definition of asymmetric
synthesis is as follows: "Asymmetric synthesis is a reaction in which an achiral unit in
an ensemble of substrate molecules is converted by a reactant into a chiral unit in such
a manner that the stereoisomeric products are formed in unequal amounts" [19].
1.3.3 Asymmetric catalysis
As was mentioned above in order to synthesize a chiral compound another chiral
substance is needed. Before such a chiral substance was considered as a reactant,
however it can also be a chiral catalyst in asymmetrical synthesis. According to the
definition catalyst influences on reaction but keeps its structure unchanged in the end of
reaction. Thus, a small amount of specific chiral catalyst can induce synthesis of chiral
product in large scales. This strategy attracts much attention from industry and science,
in fact, the number of publication on asymmetric syntheses has been increasing
exponentially every year since the 80s [20].
Willian S. Knowles, Ryoji Noyori and Barry Sharpless were awarded the Nobel Prize
for chemistry in 2001 for their contribution in asymmetrical catalysis, sealing the
importance of this field in contemporary chemistry and chemical technology.
Asymmetrical reactions catalyzed with natural and man-made chiral catalysts becomes
main-stream in different fields of industry (chemical, pharmaceutical, etc.).
Asymmetrical catalysis can be divided into two main areas: homogeneous and
heterogeneous. It is clear from terms that in the case of homogeneous catalysis the
important reactions are occurring in one phase, e.g. molecules in a solution. Whereas,
in the case of heterogeneous catalysis reactions are occurring on the interface of two
phases e.g. solid-liquid, solid-gas or in other words, on the surface of solid bodies.
It is important to mention here briefly the important requirements [21] for
enantioselective catalysts, since they are applied either for heterogeneous and
homogeneous catalysts.
Enantioselectivity, expressed as enantiopurity of a system in terms of enantiomeric
excess (ee, %). A catalyst has to be able to induce enantioselectivity in the range of 99
% for further pharmaceutical applications. However ee’s > 80 % are acceptable also if
further enantiopurification could be easily performed, e.g. via recrystallization.
Catalyst productivity, characterized in terms of turnover number (TON). For example,
in case of homogeneous enantioselective hydrogenation reactions the following values
of TON can be used for the catalyst productivity classification. TON’s > 1000 for
small-scale, high-value products and > 50000 for large-scale or less expensive products.
7

Catalyst activity, expressed in the turnover frequency (TOF). For large-scale synthesis
TOF > 10000 1/h are acceptable (under high pressure).
Separation, should be able to perform by a simple low cost operation e.g. settling,
filtration, distillation and at least 95 % of catalyst should be recovered.
Stability, catalyst should be chemically and mechanically stable without loss of activity
and enantioselectivity.
Price of catalyst, consisting of price of expensive chiral ligand (homogeneous
catalysis) and depending on type of catalyst in heterogeneous catalysis.
Availability of the catalyst at the right time in an appropriate quantity.
These criteria determine the choice of chiral catalyst for the specific reaction, as was
illustrated in the synthesis of HPB (Ethyl (R)-2-Hydroxy-4-Phenylbutyrate) ester [22].
In the following paragraphs the description of the most important aspects and ideas is
summarized, however more detailed reviews can be found in the references that are
given below.
Homogeneous asymmetric catalysis
The homogeneous enantioselective catalysis is widely used in industry covers the
following asymmetric reaction [20]: hydrogenation, hydrosilylation (and related
reaction), isomerisation of allylamines, carbometallation, oxidation (and related),
carbonylation, polymerization and other that are not mentioned here.
The origin of homogeneous asymmetric catalysis is in chiral organometallic complex
catalyzing enantioselectivly specific types of reactions performed in one phase, e.g. in
solutions. The electronic properties of the transition metals promote the formation of
organometallic complex (e.g. binap-Ru Fig. 1-13-A) with lower activation barrier for
the reaction, whereas steric properties of used ligand create a chiral environment around
the metal, leading to the specific reaction pathway and ending in enantioselectivity. The
model of the reaction pathway for the C=O bond hydrogenation in the original work of
Noyori [23], here we only mention that formation of the transition state Fig 1-13-B
which is responsible for enantioselectivity results in excellent, up to 98-99 % of ee [23].
H
Ar 2 R
H 1 O
P Cl 2
CH H H
N R 2
H
Ru
N
P Cl
N R 3
X(R 3 P) 2 Ru
H 2
Ar 2
R 4
H 2 N
Ar=C 6 H 5 ; R 1 =R 4 =H; R 2 =R 3 =C 6 H 5
X=H, OR, etc
A
B
Fig. 1-13 Chiral binap-Ru complex (A) and based on it transition state (B) in
mechanism of asymmetrical hydrogenation of C=O bond [23].
Binap-Ru complex is not only one example of chiral homogeneous catalysis, in general,
the suitable combination of a metal species and chiral ligand might be able to induce
enantioselectivity. Here (Fig. 1-14) we give some examples of ligands: quiphos [9-12],
diop [24], BPPFA [25], synphos [26-35], PPCP [36], BPPM [37] which were
successfully applied in corresponding asymmetrical reactions. Some of these ligands
were used as chiral modifiers of surface of Pt and Pd nanoparticles in the current thesis,
please see corresponding chapters below.
8

N
N
P
Fe P(C 6H 5 ) 2
O
P
P
N
O O
P(C 6 H 5 ) 2
Quiphos
BPPFA: X= (CH 3 ) 2 N
Diop
O
O PPh 2
(C 6 H 5 ) 2 P
P(C 6 H 5 ) 2
O PPh 2
P(C 6 H 5 ) 2
P(C 6 H 5 ) N
2
O
X
BPPM: X=(CH 3 ) 3 COCO
PPCP
Synphos
Fig. 1-14. Some chiral ligands used in homogeneous asymmetrical catalysis.
In spite of the fact that chiral homogeneous catalysts demonstrate excellent
enantioselectivity, due to the separation criterion mentioned above, the disadvantage
this type of catalysts appears. In fact, these soluble catalyst are more difficult to
separate and handle that the heterogeneous ones. One promising strategy to combine
the best properties of the two catalyst types in the heterogenization or immobilization of
active metal complexes on supports or carriers, which may be separated by filtration or
precipitation. However we leave this approach out of the current thesis, but keeping it
in EU-COST project, please see below. Consequently, emphasis has to be given to the
design of stable heterogeneous catalysts that are capable of high enantioselectivity.
Heterogeneous asymmetric catalysis
To date there have been numerous approaches to the design of heterogeneous
asymmetric catalysts, since Schwab and coworkers first demonstrated that Cu and Ni
could be supported on chiral silica surfaces [38, 39] and that the resulting catalysts
could give low enantioselection in the dehydration of butan-2-ol.
There are several strategies for the design of chiral heterogeneous catalysts, the three
mentioned below, lead to stable catalysts that allow high enantioselectivity.
• Modification of metal interfaces with chiral molecules;
• Tethering of active homogeneous catalysts to a three-dimensional structure;
• Electrostatic interaction as a means of immobilization.
An overview of the about three strategies can be found in the work of Hutchings [40],
whereas the current thesis focuses more on modification of the surface of the metal
nanoclusters by chiral molecules.
In contrast to homogeneous enantioselective catalysis the heterogeneous one is rather
young and has a limited success in applications. Several factors contributed to this
situation: a) the more complex structure of the heterogeneous catalyst surface on which
coexist centers with different catalytic activity and selectivity, which can lead to
undesired secondary reactions, b) an increased difficulty to create an effective
asymmetric environment and to accommodate it with the multitude of reactions that are
interesting to be carried out under enantioselective restrictions.
X
9

In fact, there are only two heterogeneous catalysts that reliably give high
enantioselectivities (90 % e.e. or above). These are Raney nickel (or Ni/SiO 2 ) system
modified with tartaric acid or alanine for hydrogenation of β-ketoesters [41] and β-
diketones [41], methyl ketones [42, 43] and others summarized in the reviews of Blaser
et al. [44] and platinum or platinum-on-support modified with cinchona alkaloids first
reported by Orito [45-47] for the hydrogenation of α-ketoesters (Orito reaction), α-
ketolactones [48-52], α-diketones [53-57], α-keto acetals [58, 59], α- α- α-
trifluoroketones [60-63] and linear and cyclic α-ketoamides [64-66]. Cinchonidine
modification of Pd surface was found to be successful in enantioselective
hydrogenation of C=C bonds [67].
Below we give basic aspects and models concerning behavior of cinchonidine based
heterogeneous catalysts and origin of its enantioselectivity, since significant part of the
current thesis is devoted to cinchonidine modified Pt and Pd systems, more detailed
reviews can be found in work of Baiker [50, 68], Blazer [44], Murzin [69], Hutchings
[40] and others [49].
Conformations of cinchonidine molecule
Several geometrical conformations were found for the cinchonidine molecule. The most
important ones are in the relative orientation of quinoline and quinuclidine moieties
(Fig. 1-15) via changing torsion angles τ 3’,4’,9,8 and τ 4’,9,8,N . The two types of major
conformers are called closed (Fig. 1-16-A) and open (Fig. 1-16-B). The terms “open”
and “closed” indicate whether the lone pair of quinuclidine nitrogen point away or
towards the quinoline ring, respectively.
Chiral center
8
HO
9
4'
3'
N
N
Quinuclidine
Quinoline
Fig. 1-15. Scheme of cinchonidine molecule
It was shown by NMR spectroscopy [70], that cinchonidine molecule can exist in open
and closed conformations at room temperature in different solvents. It has to be noted
that there are other conformations (e.g. closed 1, 2, 4 they are not shown here) of
cinchonidine which belong to the open or closed types.
10

Fig. 1-16. Closed –A and open –B types of cinchonidine conformers. Hydrogens are
not shown for clarity.
The relative population of these conformers depends on polarity of the solvent and
possible chemical interaction (cinchonidine-solvent), in fact as was show by
calculations and NMR spectroscopy the open 3 conformer is most abundant in the
solvent with low dielectric constant.
Fig. 1-17. Dependence of relative abundance of conformer (Open 3) on solvent polarity
determined by NMR spectroscopy and calculated using reaction field model [41].
Solvents: (1) benzene, (2) toluene, (e) ethyl ether, (4) tetrahydroguran, (5) aceton, (6)
dimethylformamide, (7) dimethyl sulfoxide, (8) water, (9) ethanol. The calculated
model, does not take into account specific interaction such as the hydrogen boning
between cinchonidine and ethanol. The figure is adapted from review of Baiker [50].
It is important to note, that in the hydrogenation of ketopantolactone over cinchonidinemodified
Pt ee decreases with increasing the dielectric constant of a solvent in the very
11

similar way as relative abundance of the Open 3 conformer. Such relation is in
agreement with suggestion that the Open 3 plays a crucial role in the mechanism of the
enantioselectivity, see below.
Fig. 1-18. Combined plot of the enantiomeric excess achieved in the hydrogenation of
ketopantolactone over cinchonidine-modified Pt [52, 71] (left axis) and the population
of conformer Open(3) as calculated by density functional theory in combination with a
reaction field model (P Open(3) , right axis) versus the dielectric constant of the solvent.
The axis scale is arbitrarily chosen. The solvents are (1) cyclohexane, (2) hexane, (3)
toluene, (4) diethyl ether, (5) tetrahydrofurane, (6) acetic acid, (7) ethanol, (8) water,
(9) formamide. The picture is adapted from [70].
Adsorption modes on Pt and Pd macro support
Adsorption of quinoline (part of cinchonidine) and dihydrocinchonidine on the surface
of Pt (single or polycrystal) was studied by XPS [72, 73], LEED [72], NEXAFS [74],
H/D exchange experiments [75] in vacuum. It was found that adsorption of these
compounds takes place preferentially in the flat mode through the π-aromatic system at
room and lower temperatures, whereas even at 50 °C on average the quinoline ring was
considerably tilted with respect to the surface [74]. Since experimental conditions there
were far from used in the reactions, Ferri and Bürgi [76] performed ATR studies of the
cinchonidine adsorption on solid-liquid interface, where 100 nm layer of Pt was
deposited on alumina and cinchonidine in different concentration in dichloromethane
was used. It was found that at low concentration (10 -6 M) a predominantly flat
adsorption mode prevails (Fig. 1-19-1), at increasing coverage two different tilted
species, α-H abstracted (Fig. 1-19-2) and N lone pair bonded (Fig. 1-19-3) cinchonidine
were observed, at higher (10 -3 – 10 -4 M) concentrations all three species coexist on the
Pt surface, this was also confirmed by the RAIRS experiments of Zaera and et al. [77].
Adsorption of cinchonidine on Pt (111) was also studied by STM in vacuum and with
presence of hydrogen, where mobility of cinchonidine was observed at high hydrogen
pressure.
12

Fig. 1-19. Three schematically favored adsorbed orientations of cinchonidine on wide
Pt support [76]. Species 1: “π-bonded”, 2: “α-H abstracted”, 3: “N lone pair bonded”.
Hydrogen atoms are omitted.
It is interesting to note, that maximum observed ee values were found at low
cinchonidine concentration, for example, in the works of LeBlond [78] and Bartok [51,
79].
The Fig. 1-20 illustrates dependence of ee on concentration of cinchonidine above the
Pt/Al 2 O 3 catalyst.
Fig. 1-20. Effect of cinchonidine concentration of the ee and the initial rate of
hydrogenation. Filled squares - ee , filled cicles - initial rate in ethyl pyruvate
hydrogenation, open squares - ee and open cycles initial rate in bute-2,3-dione
hydrogenation. Reaction conditions: dichlormethane (12.5 ml), reactant (66 mmol),
catalyst 0.25 g, 50 bar H 2 , 20 C, 1200 rpm stirring. The picture is adapted from [80].
The observed dependence of ee on cinchonidine concentration can be explained in the
following way, in fact, at low concentration of modifier and small part of Pt surface is
modified, whereas on unmodified sites reaction goes racemically (ee =0). When the
13

surface concentration of cinchonidine becomes optimum ([cinchonidine]/[Pt sur ] ≈ 1/20
[78, 81]) the maximum of ee is reached. Adsorption of cinchonidine in tilted modes
becomes dominated at further increase of cinchonidine’s concentration in solution, that
leads to decrease of relative concentration of π-bonded cinchonidine (active chiral site,
see below) and thus decrease of ee.
With respect to the adsorption on Pt, cinchonidine demonstrates differences and
similarities in adsorption on the Pd/Al 2 O 3 (as a thin film) surface [82]. In fact,
cinchonidine was found in adsorption mode with the quinoline moiety nearly parallel to
the Pd surface, likely through the π-system (flat adsorption mode) and in the mode with
interaction with Pd through the lone pair of the quinoline N. Compared to Pt, the
relative abundance of the two species under similar conditions was found to be different
on the two surfaces, such that the π-bonded species are less stable on Pd than on Pt.
Based on the comparison with adsorbed pyridine on Pd and Pt, the most obvious
difference in the adsorption of cinchonidine on the two metals is that the : “α-H
abstracted” species observed on Pt are absent on Pd. The difference in the diffuseness
of the d orbital of the metal was proposed to be the origin of the different behavior with
respect to the adsorption of cinchonidine.
Origin of enantioselectivity and rate enhancement and two-cycle model
Through the systematic variation of the cinchona alkaloid structure it was found that
changing the absolute configuration at C-8 (S) and C-9 (R) of cinchonidine, i.e.,
substituting cinchonidine by the diastereomer cinchonine, alters the chirality of the
product from (R)- to (S)-lactate. The most important is the finding that the
enantioselectivity is completely lost upon alkylation of the quinuclidine nitrogen atom
which indicates that this center plays a crucial role in the mechanism of
enantioselection [83].
According to the modern model, based on theoretical studies using quantum chemistry
techniques, at both ab initio and semiemperical levels, and molecular mechanics [84-
86] enantioselectivity is assumed to be thermodynamically controlled, i.e. by the
difference in the energy of adsorption and complex formation ΔG 0 ad, of the pro-(R),
[EP-CD] R ad and pro-(S), [EP-CD] S ad (Fig. 1-21). When adsorption equilibrium is
established the relative surface coverage of [EP-CD] R ad and [EP-CD] S ad can be
0
ΔΔGad
θ * −(
)
R
RT
expressed as = e . If no kinetic factor controls, i.e. if activation energies and
θ *
S
pre-exponential factors of the subsequent hydrogen addition are the same for pro-(R)
[ R]
[ θ * ]
R
and pro-(S) complexes, the ratio of surface coverage = . Thus, under these
[ S]
[ θ * ]
S
conditions the origin of enantioselectivity should be determined by stability of the
intermediate pro-(R) and pro-(S) transition states. The question regarding whether the
mechanism is kinetically or thermodynamically controlled, or whether the
thermodynamic and kinetic factors are working in concert with each other is not yet
defined.
Enantioselectivity and rate enhancement (in factor of 10-100 compared with racemic
EP hydrogenation) occur always together in this system. The H-bonding (in Pro-R and
Pro-S state, Fig. 1-21) would stabilize the half-hydrogenated state, increasing its life
time and, hence, its concentration, thereby increasing the observed reaction rate [87].
14

Fig. 1-21. The optimized structures of the proposed transition state formed on
interaction of cinchonidine’s protonated nitrogen and the pyruvate ester [88].
It has to be noted, that in the transition states the α-ketoesters are stabilized in their
half-hydrogenated form [88] and cinchonidine (in the flat adsorption mode) is
represented in open 3 conformation in these transition states as was suggested from the
investigation of obtained ee with conformation analysis.
Today it is practically accepted that enantioselective hydrogenation proceeds through
the “two-cycle” mechanism as presented in Fig. 1-22. It is assumed, that the active
chiral sites associated with adsorbed cinchonidine [CD] ad on Pt surface. Ethyl pyruvate
from the liquid phase adsorb reversibly on these sites in its two enantiofacial
configurations forming the diastereomeric intermediate complexes [EP-CD] R ad and
[EP-CD] S ad, which upon hydrogenation afford the (R)- and (S)- ethyl lactate,
respectively. It has been suggested that the adsorbed cinchonidine interacts with the
adsorbed ethyl pyruvate via hydrogen bonding between quinuclidine N and O atom of
the α-carbonyl moiety [73, 84]. Blaser et al. [89] suggested that for hydrogenation on
unmodified sites, leading to the racemic products, addition of the first hydrogen is ratedetermining
whereas for hydrogenation on modified sites, leading to the major
enantiomer, the rate determining step is the addition of the second hydrogen.
Fig. 1-22. Two-cycle mechanism suggested for enantiomeric hydrogenation of ethyl
pyruvate [EP] over Pt modified by cinchonidine. The [CD] ad present an active chiral
15

site on Pt surface formed by a adsorbed cinchonidine, [EP-CD] R ad and [EP-CD] S ad,
represent half-hydrogenated complexes formed by interaction of chiral sites with EP,
affording (R)- and (S)-ethyl lactate (EL). The picture is adapted from [50].
It has to be noted, that the alternative models (e.g. the Chemical Shielding Model [90-
93]) explaining the origin of enantioselectivity were developed, but it was found to be
less successful with respect to the model mentioned above. The detailed comparison of
existing models with underlining their strong and weak point can be found in the work
of Vargas [16].
Several features with cinchonidine based systems
Here we briefly mention some features and parameters of cinchona alkaloid based
heterogeneous catalyst and give corresponding references for detailed examination,
whereas some of the below mentioned features and reaction parameters (metal particle
size, type of metal, pressure, conversion dependence on ee and catalyst reuse) and their
role in enantioselectivity will be under investigation and discussion in the following
paragraphs and chapters.
Instead of Pt, Ir [94, 95] is also a suitable metal for enantioselective hydrogenation of
α-ketoesters. The Rh/Al 2 O 3 [94] gave 30 % lower ee than Pt/Al 2 O 3 in the hydrogenation
of α-ketoesters. The Ru/Al 2 O 3 [96] afforded no enantioselectivity with cinchonidine in
the hydrogenation of methyl pyruvate. The use of Pd will be discussed in a later
chapter.
The most convenient support material, like Al 2 O 3 , SiO 2 and carbon, have been reported
to be suitable for ethyl pyruvate hydrogenation and have no particular effect of
enantioselection [97]. However an ee of 72 % was reported in the hydrogenation of E-
α-phenylcinnamic support texture and a large effect on enantioselectivity [98, 99].
The hydrogen pressure in the range of 1-100 bar was found to effect on reaction rate as
the first order dependence, whereas ee increases with plateau [44].
Hydrogenation of ethyl pyruvate started to decrease at temperature of 70 °C [89],
whereas the drop in ee has been observed above 50 °C [100]. Moreover the adsorption
mode of the quinoline ring is changed above 50-60 °C towards Pt (111) surface [89].
Effect of concentration of ethyl pyruvate induces changes in the reaction order 1 → 0
(saturation) or sometimes maximum [44, 89, 101].
The increase of ee in the beginning of the reaction is the so called initial transient
period. The nature of the effect is still not well understood [40], whereas its
understanding might be useful in developing a catalyst which induces constant high
enantiomeric excess during reaction, at low hydrogen pressure. This effect will be
considered in details in the chapter 4.
Additivies can effect on enantioselective hydrogenation in different ways, e.g., act as
poison, interact with the products, change the chemical nature of modifier or interact
with the modifier. All this effects are summarized in the review of Murzin [69].
The very interesting work has been done by Baiker and co-workers through variation of
the modifier structure, briefly the new alternative to quinuclidine chiral moisture
attached to the anchor part was designed [50, 102-104], the new obtained modifier
results in high ee in the hydrogenation of α-ketoesters. At the same time it was show
that the anthracene anchor of the new chiral ligand results in ~ 10 % higher ee with
respect to naphthalene or quinoline anchors.
Quasi-homogeneous catalysis
16

The term quasi-homogeneous (it is also known as colloidal or soluble heterogeneous or
soluble hybrid heterogeneous-homogeneous) was introduced first by Schmid [105]. It
means that the catalyst forms the same (homogeneous) phase with the reaction mixture
i.e. it is soluble in the used solvent however catalyst is not consisting of molecules or
complexes like e.g. mentioned above binap-Ru catalyst is. The quasi-homogeneous
catalysis consists of small particles (generally few nm in size) which are soluble since
their surface is modified by a specific ligand.
Colloidal catalysts are of interest for two reasons. Firstly the support effect is
eliminated and secondly, it is possible to better control the morphology (size and shape)
of the metal particles compared to classical supported catalysts. Among several studies,
where Pt and other colloids were used for the cinchonidine-modified hydrogenation of
pyruvates, diketones [106-112] and trifluoroacetophenone [113] we would like to mark
the works of Bönneman [114, 115] since in his work dihydrocinchonidine was used as
chiral modifier and stabilizer, simultaneously, that has allowed to investigate this
system and compare with conventional supported Pt catalysts. Others interesting works
[106, 116, 117] where Pt colloidal catalyst stabilized by PVP (poly-N-vinyl-2-
pyrrolidone) (about stabilization see the next chapter) demonstrated the maximum
obtained 98 % ee in methyl pyruvate hydrogenation. Very recently Pd colloids
modified with specific chiral ligands demonstrated up to 98 % of ee in allylic alkylation
reaction [118].
Similar rate acceleration [109] and solvent effect [106] which are known for
conventional Pt/Al 2 O 3 were observed for colloidal systems.
The very interesting questions arise, wherever the size of Pt nanoparticles influence on
ee or not and how adsorption of cinchonidine on small ~1-2 nm nanoparticles is
different from adsorption of big macroscopic Pt support (plane). Thus, in the current
thesis these questions will be under focus of our attention.
17

Chapter 2
Introduction to the nanoscale materials
2.1 General information and definitions
Introduction
Colloidal metal (originally called "sols") particles are known at lease since work [119]
of Hans Heinrich Helcher (1718) about gold colloids. In the Middle Ages glass-blower
workers used gold and other metals in colloidal form in order to make coloured glass
(Fig. 2-1) which were used e.g. as stained-glass window in churches or as service
dishes. In 1857, Faraday reported the formation of deep red solutions of colloidal Au by
reduction of an aqueous solution of AuCl 4 using phosphorus in CS 2 (a two-phase
system) in a well-known work [120].
Fig. 2-1. Red coloured glass based on gold colloids. Both pictures were found via
google.com
Now the field of nanomaterials is booming with scientific investigations and
application of the different areas of technology. In fact, for example the catalytic
properties and the electronic structure of the nanomaterials can be considered from
changing the cluster size, composition and structure [121, 122]. Herein, we will provide
an overview of colloidal methods and properties as well as the approaches being used
to bridge the fields of nanotechnology and catalysis. In our opinion, it is better to start
acquaintance with nanomaterials from definitions developed by Schmid (Table 2-1)
[123] as well standardized by IUPAC [124].
Classification
Nanoparticles are constituted of several tens or hundreds of atoms or molecules and can
have a variety of sizes and morphologies (amorphous, crystalline, spherical, needles,
etc.). According to the IUPAC definition the term colloidal refers to a state of
subdivision, implying that the molecules or polymolecular particles dispersed in a
medium have at least in one direction a dimension, roughly between 1 nm and 1µm, or
that in a system discontinuities are found at distances of that order [124].
Table 2-1. Classification of nanoscale materials.
18

Discrete, single Metal clusters, Traditional metal Bulk metal
metal a complexes including
nanoclusters
colloids
1-10 nm > 10 nm
Comments: A - the classification can be used not only for metal materials.
Another useful classification of variety form of colloids is illustrated in the Table 2-2,
adapted form online encyclopaedia (Wikipedia) [125].
Table 2-2. Examples of variety of colloids.
Dispersed medium
Gas Liquid Solid
Continu Gas None (all gases are Liquid Aerosol Solid Aerosol
ous
soluble)
(fog, mist) (smoke, dust)
media Liquid Foam (cream) Emulsion (milk, Sol (paint, ink)
blood)
Solid Solid foam (aerogel, Gel (gelatine, Solid sol (ruby glass)
pumice)
cheese)
Full shell clusters model
According to the Schmid [123], metal clusters which have a complete, regular outer
geometry are designated as full-shell, or “magic number”, clusters (Table 2-3).
It is believed, that full-shell clusters have added stability, as their densely-packed
structures provide the maximum number of metal-metal bonds [126, 127].
Full-shell clusters are constructed by successively packing layers – or shells- of metal
atoms around a single metal atom. The total number of metal atoms, y, per n th shell is
given by the equation y 2 n
= A ⋅ n + 2 (n>0). In this formula parameter A depends on
geometry of the current crystal, the values of parameter A are given in the Table 2-4 for
several types of crystals.
Table 2-3. Full shell model on the example of cuboctahedrone crystal. The table is
adapted from [128].
Table 2-4. Full shell model applied for different polyhedrons [129, 130].
Polyhedron
Illustrated A values
index a
Truncated tetrahedron 1 14
Cuboctahedron or twinned cuboctahedron 2 10
Truncated octahedron 3 30
Truncated cube 4 46
19

Triangular prism 5 4
Comments: A – The geometry of the corresponding polyhedron can be found in the Fig.
2-3.
Fig. 2-3. Geometrical models of some polyhedrons. Numeration corresponds to
illustration index from Table 2-4.
The full shell model is very useful in calculation of concentration of surface atoms for a
nanocluster with known size, that can be considered as number of active catalytic sites
e.g. for estimation of initial rate values normalized on number of active sites.
Stabilization of nanoparticles
Colloidal particles generally are not thermodynamically stable, because of high free
surface energy. As any physical system, nanoparticles always have a tendency to
minimize their full energy that leads to their agglomeration and losing nanoscale size, it
should be noted that Van der Waals attraction (see below) can also take part in inducing
an agglomeration. In order to prevent agglomeration particles have to be stabilized.
However before considering stabilization principals it makes sense to mention different
type of interactions between colloidal particles [125], they are summarized below.
Excluded Volume Repulsion: This refers to the impossibility of any overlap between
hard particles.
Electrostatic interaction: Colloidal particles often carry an electrical charge and
therefore attract or repel each other. The charge of both the continuous and the
dispersed phase, as well as the mobility of the phases are factors affecting this
interaction.
Van der Waals forces: This interaction is due to induced dipole-dipole interaction. Even
if the particles don't have a permanent dipole, fluctuations of the electron gas give rise
to a temporary dipole, meaning that Van der Waals forces are always present, although
possibly at a much lower magnitude than others.
Entropic forces: According to the second law of thermodynamics, a system progresses
to a state in which entropy is maximized. This can result in effective forces even
between hard spheres.
Steric forces between polymer-covered surfaces or in solutions containing nonadsorbing
polymer can modulate interparticle forces, producing an additional repulsive
steric stabilization force or attractive depletion force between them.
Steric stabilization and electrostatic stabilization are the two main mechanisms for
colloid stabilization. Electrostatic stabilization is based on the mutual repulsion of like
20

electrical charges; in fact the total electric potential is the sum of electrostatic and Van
der Waals potentials and results in energy barrier. Electrostatic repulsion (Fig. 2-4 left)
is generally caused by an electrical double (really multi) layer [131] formed by ions
adsorbed on the particle surface (e.g. halides) and the corresponding counter ions (e.g
tetraalkylammonium). Small particle sizes lead to enormous surface areas, and this
effect is greatly amplified in colloids. In a stable colloid, the mass of a dispersed phase
is so low that its buoyancy or kinetic energy is too little to overcome the electrostatic
repulsion between charged layers of the dispersing phase. The charge on the dispersed
particles can be observed by applying an electric field: all particles migrate to the same
electrode and therefore must all have the opposite sign charge with respect to the
electrode’s charge.
Fig. 2-4. Schematic image of two electrostatically-stabilized nanoparticles (top-left) and
plot of electrical potential space distribution (bottom-left). Two sterically-stabilized
nanoparticles (top-right) and caused a large energy barrier (bottom-right) against
particle interaction. Pictures are adapted from [128, 132].
Steric stabilization (Fig. 2-4-right) is achieved by the coordination of sterically
demanded molecules that act as protective shields on the metallic surface. In this way,
nanometallic cores are separated from each other and agglomeration is prevented.
2.2 Methods of colloid and supported nanoparticles
preparation
Nanoparticles can be obtained by two general approaches: “top down” and “bottom up”.
In “top down” method bulk materials are mechanically ground to the nanosized scale
and stabilized by a suitable molecule [133, 134]. The problem with this method is
difficulty in achieving the narrow size distribution and control both shape and average
size of the particles.
Moreover, bimetallic nanoparticles with core shell structures cannot be obtained by this
method. In “bottom up” methods, nanoparticles are obtained by starting with molecular
21

precursors and building-up. This method allows control of size and shape of
nanoparticles and their size distribution. Taking into account these reasons we will
focus on “bottom up” methods only.
The “bottom up” method can be classified conditionally as methods occurring in
condensed phase and gas (or vacuum) phase. In condensed phase synthesis,
nanoparticles are prepared by means of chemical synthesis; this method is also called as
wet chemical preparation. For example, in order to prepare metal nanoparticles, metal
salts are usually reduced in the presence of suitable capping agents (stabilizer) that
stabilize the nanoparticles from agglomeration. In the gas phase synthesis, metal is
vaporized and then vaporized metal atoms are condensed. The same idea can be
realized in vacuum where the metal of interest is vaporized with high energy ions or by
heating or by laser beam (the latter method can also be used in liquid media) and thus
generated metal vapour is deposited on a support. Below we describe methods for
preparation metal (due to their particular interest in the current thesis) nanoparticles in
condensed and briefly in gas phase.
Wet chemical preparation method
Wet chemical reduction procedures have been widely used to stabilize the different
types of metal nanoparticles with different kinds of organic as well as inorganic
stabilizers.
As was mentioned above, metal precursors are usually reduced in the presence of a
stabilizer in order to prevent the nanoparticles from agglomeration. The choice of metal
precursor, stabilizer, reducing agent and reaction conditions (relative
precursor/stabilizer concentration, solvent, temperature and even stirring rate) strongly
influence in properties of obtained nanoparticles, i.e. average particle size.
The main classes of stabilizer predominant in the literature are macromolecules
containing P, N, and S (e.g. phosphanes, amines, thioethers) atoms [105, 135-145].
Such solvents as THF [135, 146] or THF/MeOH [147] also can act as stabilizers.
Molecules like ionic surfactants stabilize the nanoparticles by means of both
electrostatic repulsion and steric hindrance. Their polar group forms an electrostatic
layer around them and the long organic moiety gives steric hindrance. In general,
lipophilic protecting agents give metal colloids that are soluble in organic media
(“organosols”), while hydrophilic agents yield water-soluble colloids (“hydrosols”).
Reducing agents play an important role in directing the size and shape of particles.
Among different kinds of reducers (NaBH 4 , R-CHO, etc.) we would like to underline
hydrogen. In fact, it has been applied in the synthesis of various metal clusters. The
advantage of using hydrogen lies in the avoidance of by-products, the disadvantage is a
weak reducing agent that is unable to reduce many metal complexes and salts.
However, the use of hydrogen makes possible to perform reduction under mild
condition that might be important for sensitive stabilizers.
It is interesting to note, that reduction of metal salts e.g. Na 2 PtCl 4 by microwave
radiation, thus in presence of stabilizer gives 20-30 Pt nanoparticles [148].
The decomposition of organometallic complexes with null valence metal atom(s) in
presence of stabilizer can also be applied as a method of preparation of nanoparticles.
This decomposition is typically initiated by an external force e.g. heat, ultrasonic or by
a chemical agent capable to induce decomposition e.g. H 2 or CO. Classical example is
in thermal decomposition of iron Fe(CO) 6 [149] or cobalt complex Co 2 (CO) 8 [150] at
130-170 °C in presence of polymer yielding 45 nm Co nanoparticles [151].
22

The advantage of use of organometallic route in preparation of nanoclusters is obvious
when it is needed to stabilize (modify) nanoparticles by hydrophobic surfactant or in
other words, when required stabilizer is water insoluble.
An interesting method of preparation of metal nanoparticles is through electrochemical
approach established by Reetz [152]. In this method, an anode is made up of the
specific metal (required for nanoparticles) and a cathode is of any other metal. Under
the suitable applied current density the anode sacrificially dissolves in the electrolyte,
the metal ions migrate towards the cathode and reduction occurs. The nucleation and
growth of reduced metal atoms occurs at the electrode surface. The stabilizing agent in
the reaction vessel arrests the growth of the nanoparticles and facilitates the formation
of stable nanostructures. The final step is dissolution of nanoparticles from the
electrode surface into the bulk of the solution. The advantages of the electrochemical
pathway are that the contamination with by-products resulting from chemical reducing
agents is avoided and that the products are easily isolated from the precipitate. Further,
the electrochemical preparation allows for size-selective particle formation. The
particle size obtained by the electrochemical route depends on many factors, the
distance between the electrodes, reaction time, temperature, and polarity of the solvent
contribute to the particle size. Experiments have also shown that the applied current
density also has a major influence on the particle size.
Vapor phase synthesis
As was mentioned above metal atoms can be obtained not only through reduction of
metal salts, but also through solid to gas phase transformation of metal. The Solvated
metal atom dispersion (SMAD) techniques is based on this method. Typically, the metal
of interest is heated in a crucible at elevated temperature evaporated under vacuum and
then co-condensed with a specified amount of liquid ligand substrate on the liquid
nitrogen cooled walls of the reactor vessel. The reactor is removed from the liquid
nitrogen and the reactor is allowed to warm slowly. After the warming stage, particles
can be stabilized either sterically (by solvation) or electrostatically (by incorporation of
negative charge). The SMAD techniques has been successfully applied e.g. in
preparation of thiol stabilized gold colloid [153]. The main advantage of this “clean”
method of preparing nanoclusters in gram scale is in fact, that it does not involve
starting materials containing counter ions like nitrate, chlorides, sulfates which are
common in reduction synthetic methods. However, the use of SMAD method is limited
because of difficulties in the operation of the apparatus and it is difficult to obtain
narrow particle size distributions.
Another interesting method is through the laser ablation of bulk metal target crystal
being in liquid phase. The recent example is in preparation of Pt nanoparticles by laser
ablation of Pt crystal in water [154-156]. Due to the laser induced evaporation of Pt
atoms into the liquid phase with formation of 1-15 nm nanoparticles, it was found that
the yield of nanoparticles depends on the wavelength of the used laser beam whereas it
has no significant influence on the average size of produced nanoparticles. The laser
ablation method has the same important advantages as SMAD method demonstrates,
moreover there is not need to work with very high temperatures (for metal evaporation)
and keep system under high vacuum. However with respect to the chemical methods
the laser ablation has the following disadvantages. There is more insight about the
reaction at plasma and liquid interface and it is relatively more expensive to obtain the
large amount of products, in fact the rate of nanoparticles production was about 4.4
mg/h.
23

2.3 Preparation of nanoclusters on a heterogeneous
support
Metal clusters especially in nanoscale size are widely used in heterogeneous catalysis.
Heterogeneous catalysts are traditionally obtained by impregnation methods consisting
of immersing a solid support like Al 2 O 3 , SiO 2 or ZrO 2 in a solution of metal salts of
interest and inducing deposition. After some specified time, the solid substance is
filtered and dried. The dried product is calcinated at high temperature in the presence of
gaseous reducing agents such as H 2 or CO to yield the final catalytic material. The
problem with impregnation methods is the difficulty in controlling the particle size and
that the metal particle formed after the reduction may be located where it is not
accessible for further application.
An alternative method for the preparation of heterogeneous catalysis is the “precursor
concept“ which was originally developed by Turkevich in 1970 [157] and was further
developed in the 1990's by many groups [135, 158-161].
This method consists of depositing the pre-prepared metal colloid on a solid support; by
applying this method it is possible to control the size and shape of the nanoparticles.
Application of the precursor concept also allows full pre-characterization of the catalyst
components (support and colloid) prior to deposition, thus facilitating the possibilities
for comprehensive study of heterogeneous catalyst support/metal effects. Moreover,
almost all of the particles on the inert support are accessible because the preformed
colloids generally do not “fill” cracks and other defects of the support’s surface as
molecular precursors often do.
Metal nanoparticles have a high active metal surface area (50-100 m 2 /g for 6-3 nm
particles) that facilitates proper surface modification (roughly, due to the high activity
of “freshly” prepared nanoclusters) and sometimes unique size dependence properties.
The catalytic properties of supported and not supported metal nanoclusters were found
in many important reactions such as oxidation (Co, Pt, Pd, Ag and Au), hydrogenation
(Pt, Pd, Pt/Pd, Au/Pt, Au/Pd and Rh) and others. The detailed overview can be found in
the work of D’Souza [162].
24

Chapter 3
Current thesis in the EU-COST project:
aims of the work
3.1 Introduction and goals setting
The idea to use metal nanoclusters modified with chiral ligands is a much promising
strategy in research and development of enantioselective catalysts. In fact, making
nanoparticles stabilized with chiral ligands it is possible to reach maximum chiral
coverage of the metal surface, that is very important for obtaining high
enantioselectivity. Since Pt and Pd (generally alumina or silica supported) were found
to be the best metals for heterogeneous enantioselective hydrogenation, in this thesis,
we focused on the chiral modification of the surface of Pt and Pd nanoclusters.
The current thesis is a part of the “EU COST D24 WORKING GROUP 7
D24/0007/02” working project, which besides International University Bremen (Prof.
R. Richards) also involves research groups from the Marseille University (Prof. G.
Buono), Ecole Nationale Supérieure de Chimie de Paris Institute (Prof. J.-P. Genet),
University of Ljubljana (Prof. M. Kocevar) and University of Bucharest (Prof. V. I.
Parvulescu).
The working scheme (Fig. 3-1) shows that the groups from Paris and Marseille provide
us with specific chiral ligands (synphos and quiphos with quiphos-spider, respectively).
At the International University Bremen we have prepared Pt and Pd nanoclusters
modified with obtained from our colleagues (and purchased from Fluka and Aldrich)
chiral ligands for chiral modification of them on the surface of metal nanoclusters. The
groups in Paris, Ljubljana and Bucharest study catalytic activity (as well as
enantioselectivity) over obtained samples in corresponding reactions. The details about
their work can be found in the description of the EU-COST project [163] and from the
corresponding research groups.
25

Paris
Marseille
Ljubljana
Synphos
Quiphos
Quiphos - spider
IUB catalysts
preparation and
characterization
Bucharest
Aldrich
Fluka
Cinchonidine,
binap, diop, diphos
Fig. 3-1. The working scheme of EU-COST project.
The all used in the current work ligand for the surface modification are summarized in
the Fig. 3-2.
O
HO
N
N
N
P
N
O
PPh 2
O
PPh 2
PPh 2
O PPh 2
N
Cinchondine
Quiphos
(2R, 5S)-3-Phenyl-2-(8-quinolinoxy)
-1,3-diaza-2-phosphabicyclo-octane
Binap
2,2'-Bis(diphenylphosphino)-
1,1'-binaphthyl
O
Synphos
(2,3,2',3'-tetrahydro-5,5'-bi(1,4-
benzodioxin)
-6,6'-diyl)-bis(diphenylphosphane)
P
P
P
P O
P
O
O
Diphos
Bis(diphenylphosphino)ethane
Diop
4,5-Bis(diphenylphosphinomethyl)-
2,2-dimethyl -1,3-dioxolane
Quiphos-Spider
4-isopropyl-2 methyl cyclohexyl-1methylphosphonite
-2-2'-1,5 binaphtyl
Fig. 3-2. Ligands used for modification of surface of Pt and Pd nanoclusters.
26

Besides preparation of chiral ligand modified Pt and Pd nanoclusters another aim of
the current thesis is in determination the geometrical models of adsorption of chiral
ligands (Fig. 3-2) on the surface of Pt and Pd nanoclusters via DRIFTS (Diffuse
Reflectance Infra-red Fourier Transform Spectroscopy) and molecular modeling. Since
the ability of the adsorbed cinchonidine to induce enantioselectivity e.g. in
hydrogenation of α-ketoesters is well known in heterogeneous enantioselective
catalysis we placed high emphasis on the examination of cinchonidine modified Pt and
Pd nanoclusters in ethyl pyruvate hydrogenation. The work is also focused on
obtaining active, highly enantioselective stable catalyst for the performing
hydrogenation under mild conditions that is very important from application point of
view. Furthermore we investigated catalytic and enantioselective activity of all
synthesized samples in typical Orito reaction.
27

Chapter 4
Cinchonidine modified Pt colloidal
nanoparticles: characterization and
catalytic properties
4.1 Introduction
Cinchonidine modified transition metal surfaces (primarily Pt) represent one approach
towards asymmetric heterogeneous catalysis. Since the first report by Orito [45, 46],
numerous experimental and theoretical investigations into the cinchona on Pt and Pd
systems have revealed a great deal of insight into the system and its applicability. As
was mentioned above, adsorption geometry of the cinchonidine molecule on metal
surface has a significant role in the enantioselectivity. It is known from the work of
Ferri [76], Zaera [77] and Wells [74], that cinchonidine can adsorb on the surface of
macroscopic Pt (plane, single crystal with sizes are much bigger than the size of the
cinchonidine molecule) in three different orientations (Fig. 1-19): “π-bonded”, “α-H
abstracted” and “N lone pair bonded”. It is also considers, that only cinchonidine in the
flat “π-bonded” orientation is able to induce enantioselective hydrogenation. However
adsorption mode of cinchonidine on the surface of nanosized Pt crystals has never been
investigated and it attracts special attention, for instance, because the maximum
reported enantiomeric excess (98 %, in the hydrogenation of methyl pyruvate) was
obtained on small (1.2 nm) Pt colloids. Thus the beginning of this chapter is dedicated
to the determination of adsorption modes of cinchonidine on Pt nanoclusters.
The initial transient period (ITP) is an interesting phenomenon which consists of an
increase of enantiomeric excess in the beginning of e.g. ethyl pyruvate hydrogenation.
The ITP effect was observed in the examination of enantioselectivity of cinchonidine
modified Pt nanoclusters and discussed in the second part of the current chapter. The
initial transient period (ITP) for the Pt/Al 2 O 3 system has been reported and investigated
by numerous groups and has been the subject of intense discussions [164, 165]. ITP
was observed by Baiker and co-workers [166] and was attributed to an impurity effect
(from destructive adsorption of EP and alcoholic solvents on Pt) which could be
eliminated by pre-activation with hydrogen. Blackmond and co-workers proposed the
“reaction-driven equilibration" explanation [167] and demonstrated the significant
influence of mass transport (gas-liquid diffusion of hydrogen) limitation on the
enantioselectivity [168]. Recently, it was shown that ITP is not only the effect of
impurities (lactates in ethyl pyruvate) but could also be a result of competitive
adsorption of reactant, modifier and solvent [169]. From the work of Hutchings et al.
[170], we know, that the ITP effect was found with cinchonidine premodified
conventional Pt/Al 2 O 3 in ethanol and dichloromethane solution and explanation through
“removal of adsorbed oxygen species from the surface” was proposed.
Here, we investigate the behavior of enantiomeric excess (ee) during hydrogenation of
ethyl pyruvate with conventional Pt/Al 2 O 3 and with cinchonidine Pt nanoclusters
prepared by two different methods, an aqueous (first reported by Bönnemann [114])
and a novel organometallic method through decomposition of Pt 2 (DBA) 3 (DBA is bisdibenzylidene
acetone) complex. The nature of the increase in ee at initial reaction
time, further decreases and differences in the behavior of conventional Pt/Al 2 O 3 and
28

“quasi-homogeneous” cinchonidine stabilized Pt nanoclusters are discussed below.
Further, the experimental conditions selected for these experiments were chosen to be
mild: room temperature and low hydrogen pressure (2-10 bar) since these are likely to
be the most relevant conditions for hydrogenation of α-ketoesters on an industrial scale.
4.2 Experimental
Materials
Unless otherwise mentioned the following materials: Pt on alumina (Aldrich 5% Pt),
cinchonidine (> 98%, Fluka), acetic acid (99.8% Fluka), ethyl pyruvate (98%, Aldrich),
THF (Applichem, 99.8%), formic acid (Aldrich, 99%), bis-dibenzylidene acetone
(Fluka, ≥96.0%), K 2 PtCl 4 (Acros Organics) and ethanol (Applichem, 99.9%) were used
as received in all chapters.
The 5 % Pt on alumina (E4759) catalyst was kindly provided by Engelhard company by
request.
Sample preparation
In order to activate the conventional Pt/Al 2 O 3, it was heated at 300 °C under vacuum
(10 -3 mbar) for 1 hour, then kept for 3 hours under flowing hydrogen (100 ml/min) and
cooled to room temperature. Then the required amount was weighed and transferred
into a glass reactor with acetic acid, ethyl pyruvate and cinchonidine.
Chiral modification of conventional Pt/Al 2 O 3 . 40 mg of conventional Pt/Al 2 O 3 was
heated to 400 ºC under vacuum (10 -3 mbar) for 1 hour, then the mixture of N 2 (93%)
and H 2 (7%) gases were passed (100 ml min -1 ) over the sample for 3 hours. Then, under
flowing gases (H 2 and N 2 ), the support was cooled to room temperature and the solution
of cinchonidine in CHCl 3 (6.0 ml, 3 mM) was added by syringe through a stopper. The
system was kept at 10-15 ºC for 24 hours, then washed 3 times with CHCl 3 and finally
dried under vacuum at room temperature.
Preparation of 10.11-dihydrocinchonidine (DHCIN) has been done according to the
protocol from work of Morawsky [171]. 10 g cinchonidine (33 mmol) was dissolved in
100 ml 0.5 M H 2 SO 4 and placed in a metal reactor. 0.5 g Pd/C (5 % wt, Degussa
E101/0/W) was added to the solution. The H 2 pressure was set to 6 bars, reaction
started by stirring. After 1.5 h reaction was finished. The catalyst was removed via a G4
filter, 50 ml 2 M NaOH was added to precipitate the hydrated alkaloid. The white
precipitate was washed with 500 ml H 2 O and recrystallized in ethanol. It has to be
noted that resulted product had some amount of cinchonidine, as was detected as
presence of attenuated peak at 1636 cm -1 in IR spectrum of the product, which
corresponds to the stretching of C=C bond.
The preparation of Pt 2 (DBA) 3 organometallic complex has been done according to the
literary protocol [172, 173]. Typically, to a solution of 2.8 g sodium acetate
and 2.36 g of bis-dibenzylidene acetone (dba) in 60 mL of ethanol at 50 °C was added a
solution of 1.40 g K 2 PtCl 4 in 12 mL of freshly distilled water. The initial pale yellow
suspension redissolved while the mixture was heated to the refluxing temperature (90
°C). After refluxing for 1 h a dark violet precipitate formed. The mixture was allowed
to settle overnight. The solution was eliminated by filtration and the solid washed three
times with 200 mL of water, dried overnight under vacuum, further washed three times
with 200 mL of pentane and finally dried under vacuum overnight.
A water soluble Pt nanoclusters modified with DHCIN was prepared based on the
protocol published by Bönnemann [114]. Typically, 0.208 g (0.6 mmol) PtCl 4 in 160 ml
distilled water was reduced with 0.1 M aqueous formic acid (15 ml) with dissolved
29

dihydrocinchonidine (355 mg or 1.2 mmol) under reflux, at ~95 ºC. After 10-15
minutes the product was precipitated into 200 ml of aqueous saturated NaHCO 3
solution, then the precipitate was filtered with a G4 fritted glass filter and washed first
with a half-saturated NaHCO 3 (400 ml) solution followed by water. Finally, the black
powder was taken up from the filter using an acetic acid solution. The acetic acid – Ptcinchonidine
system was dried at room temperature under vacuum. This catalyst was
found to be redispersible in water, acetic acid and toluene. This catalyst will be referred
to as B – henceforth in this chapter. Physical state: black powder.
The cinchonidine modified Pt nanosized catalyst was prepared by decomposition of the
organometallic complex Pt 2 (DBA) 3 under hydrogen in the presence of cinchonidine.
Pt 2 (DBA) 3 (402 mg or 0.37 mmol) and an excess of cinchonidine (4.5 g or 15 mmol)
were dissolved in 160 ml THF and flushed with Ar for 20 min. The mixture was kept in
the glass reactor under 2 bar of H 2 with constant stirring at 300 rpm. After 18 hours the
black precipitate was collected and washed first with 50 ml pentane, then with a
pentane/THF mixture (5:1) and chloroform until all hydrogenated DBA and excess
cinchonidine was removed as monitored by IR and XRD spectroscopy. Thus prepared
catalysts were found to be redispersible in acetic acid and THF and will be referred as
D – henceforth in this chapter. Physical state: black powder.
In order to purify ethyl pyruvate, 20 ml of it was added to 80 ml of water/n-pentane
(1/1 by volume) mixture and kept for two days in a separation funnel. The n-
pentane/ethyl pyruvate fraction was separated and n-pentane was evaporated in a rotary
evaporator. The procedure was repeated two times and the collected purified ethyl
pyruvate was found to contain 0.05 % R, S ethyl lactate (from the initial 0.5%) by GC.
Characterization and experiments
Hydrogen chemisorption on conventional Pt/Al 2 O 3 (Aldrich) catalyst was performed on
the Autosorb instruments supplied by Quantachrome, Germany. Typically the sample
was preheated at 400 °C for 1 h under vacuum (10 -2 mm. Hg.) and for 3 hour under
flow of hydrogen then chemisorption has started at 40 °C in the 0-586 mm. Hg.
hydrogen pressure interval.
Transmission electron microscopy (TEM) micrographs were obtained by using a
TECNAI F20 instrument in the IFAM Bremen. Specimens were prepared by placing a
drop of the colloidal onto a copper grid with a perforated carbon film and then allowing
the solvent to evaporate. Around 100-150 particles were considered for size distribution
calculations.
X-ray diffraction (XRD): Powder XRD analyses were performed using a Siemens
D5000 instrument with Cu Kα radiation (wavelength 1.54·10 -10 m) operated at 40 kV
and 40 mA.
Gas chromatography: (GC) Measurements were performed at 88 °C (isothermal) on a
Varian GC 3900, FID with a Lipodex-E chiral column.
Catalytic Reactor: The catalytic reactor was supplied by Parr Instrument (Germany)
GmbH. The apparatus consists of five major components; reactor, H 2 reservoir, oil
reservoir, a temperature and pressure controller and computer. The batch type reactor
(300 ml) is made of double walled borosilicate glass with a maximum pressure rating of
10 bars. A regulator situated between the reactor and reservoir maintains a constant
hydrogen pressure inside the reactor. The reactor chamber also consists of a water
cooler, an electric stirrer and thermocouple.
Hydrogenation of ethyl pyruvate was carried out in the described above 300 ml glass
reactor under a pressure of either 2 or 10 bars of hydrogen at room temperature (22±1
30

°C). The stirring speed was 300 rpm. The ethyl pyruvate was added to 70 ml of acetic
acid, then the mixture was sonicated for 10 min, Ar was then flushed through the
solution for 10 min to remove dissolved air. The required amount of catalyst was
added, the mixture was sonicated for 5 min, transferred into the reactor, flushed with Ar
for 10 min and finally after the pressure of hydrogen had been stabilized for 10-15 sec
the initial time (t=0) was set and the reaction kinetics were monitored. Around one ml
aliquots of the reaction mixture were taken at certain time intervals, filtrated with a G4
filter and 1 μl was injected into the gas chromatograph (GC).
FTIR spectra of D type of catalyst were measured with a “Thermo Nicolet Avatar”
spectrometer in transmission mode, resolution 4 cm -1 , number of scans 500. IR pellets
were prepared by mixing 100 mg KBr and 1.5 mg sample then pressing into a pellet
under a pressure of 10 tons. For IR investigation of the B type of catalyst a “Thermo
Nicolet 4700 FT-IR” spectrometer with liquid nitrogen cooled detector and DRIFT
accessory was used with the following parameters: 3000 scans, 600-2000 cm -1 scan
range, 4 cm -1 resolution. The DRIFT spectra were chosen for analysis because it is the
more applicable technique for observing adsorbed species on finely divided metal
powder [174].
The carbon and nitrogen elemental analysis measurements were done on a Carlo Erba
NA2500 C/N analyzer in the School of GeoSciences at Grant Institute in the University
of Edinburgh. Measurements were repeated several times with an error ~1%. The metal
content was found from thermogravimetric Analysis (TGA) in 30-1200 °C temperature
interval under nitrogen 100 ml/min with an SDTQ 600 instrument from TA
Instruments.
The cumulative enantiomeric excess (ee) was calculated according to the formula
[ R]
− [ S]
ee = ⋅100%
.
[ S]
+ [ R]
The actual enantiomeric excess (ee*) was calculated by the previously established
[169] formula
ee y − ee y
ee
*
i+
1 i+
1 i i
i+
1 / 2
=
,
yi+
1
− yi
where y i - mol of both ethyl lactates, ee –cumulative (observed) enantiomeric excess.
Computational methods: Geometry optimization and calculation of IR spectra of
cinchonidine and quinoline were carried out using the Gaussian 03 program [175].
Density functional theory (DFT) calculations were performed by using the B3LYP/6-
31G method and basis set. The vibrations and orientations of dipole moment were
visualized with GaussView and MOLDEN programs.
4.3 Results and discussion
4.3.1 General characterization
Choice of conventional catalyst
The Engelhard 4759 (5% Pt/Al 2 O 3 ) is known from 90 th to be one of the most studied
catalyst for Orito reaction, however the Engelhard company does not produce one any
more and it is not widely available. Due to this reason and the fact that under the same
reaction conditions (Table 4-1) catalyst from Aldrich demonstrates higher
enantioselectivity and reaction rate, we chose it to be the basic conventional catalyst for
further comparisons.
31

Table 4-1. Comparison between 5 % Pt/Al 2 O 3 catalysts from Aldrich and Engelhard
under the same reaction conditions A .
Supplier Initial rate, Conversion, % after Ee, % (dominated
mmol·(min·g) -1 (time, min) enantiomer)
Aldrich 28 80 (60) B 91 (R)
Engelhard 2 34 (250) B 70 (R)
Comments: A - To 4 ml ethyl pyruvate (36 mmol) and 20 mg cinchonidine (1 mM) 20
mg thermally activated catalyst (please find the activation procedure in the chapter 5)
was added, the mixture purged with Ar for 10 min and hydrogen pressure of 10 bar was
set under the constant stirring (1000 rpm), B - according to the curve profile
(conversion vs. time) reaction can go further.
From the hydrogen chemisorption experiment (Fig. 4-1) with conventional 5% Pt/Al 2 O 3
(here and further Aldrich catalyst is mentioned as a conventional) it was determined
that the metal dispersion is 29 %, active metal surface area is 3.6 m 2 g -1 and
concentration of active sites is 37 μmol g -1 , average crystal size is 3.8 nm.
Fig. 4-1. Combined, weak and their difference (strong) curves of hydrogen
chemisorption data on conventional 5% Pt/Al 2 O 3 .
X-ray diffraction phase analysis of the cinchonidine modified Pt nanoclusters was
carried out using powder XRD to ascertain the polycrystalline nature of the particles.
32

800
111
Intensity, a.u.
400
200
220 311
0
2Θ
40 60 80
Fig. 4-2. XRD spectrum of B sample.
600
111
Intensity, a.u.
400
200
200
220
0
40 50 60 70
2Θ
Fig. 4-3. XRD spectrum of D sample.
As can be seen from XRD spectra of sample B (Fig. 4-2) and D (Fig. 4-3) they both
have dominant peak at ≈ 39 °, which corresponds to the (111) crystal face, as well as
peak at 66 °, which corresponds to (220) crystal phase. The peak at 46 ° (corresponding
to the (220) crystal phase) is clearly seen in the spectra of D sample, however it is
present as poorly observable shoulder in the spectra of B sample, probable due to the
widening of the dominant peak at 39 ° and their overlapping. The assignment of the
peaks was based on the work of Davey [176]. The quantitative value of a peak
33

widening can be used in estimation of the average crystal size, according to the
Scherrer equation [177]. The obtained values can be found in the Table 4-1 below.
K ⋅λ
d hkl
= , where K is a constant whose value depends on the shape of the
Δ ⋅cos(Θ)
particle and the method used to measure the peak width, in this work K was taken to be
0.94, according to the work of Mukherjee [148], Δ – Peak width, taken on a half height.
TEM images of the B and D samples (Fig. 4-4) and their corresponding size distribution
histograms (Fig. 4-5) clearly show that D type of sample contains Pt nanoparticles with
average size of 2.3 nm, whereas average size of nanoparticles of B sample is 1.4 nm.
Fig. 4-4. TEM of B and D-sample.
The very similar values of the average crystal size were estimated from the Scherrer
equation based on XRD data are summarized in the Table 4-2 together with data
obtained from TEM.
Table 4-2. The average size of cinchonidine modified Pt nanoclusters obtained by
aqueous (B) and organometallic (D) methods determined via TEM and XRD
measurements.
Type of sample (TEM), nm. (XRD), nm.
B 1.4 1.2
D 2.3 2.3
34

Number of particles
30 Types of sample
D
B
25
20
15
10
5
0
1.0 1.5 2.0 2.5 3.0 3.5
Particle size, nm
Fig. 4-5. Particles size distribution for B and D types of sample.
From the elemental analysis it was found, that the B-series sample contains 43.5 % C,
4.3 % N or 45 % cinchonidine, based on nitrogen content. From the TGA, the Pt
content was found to be 25 % thus differing from the 40 % metal content in the original
work [11] for 1.5 nm sized Pt cluster, probably because of the presence of
moisture/CO/CO2 and possibly acetic acid (30 %).
Elemental analysis of the D-catalysts before reaction (please see below) showed 16.6 %
C and 1.8 % N or 20 % cinchonidine, by making average between C and N. 63 % Pt
was found from TGA. The remaining content was attributed to moisture/CO/CO 2 . The
content of C and N on the D catalyst after 50 min of reaction was found to be 5.3 % and
0.8 %, correspondingly or 7 % of cinchonidine.
4.3.2 FTIR investigation of cinchonidine adsorbed on Pt
FTIR investigation of cinchonidine adsorbed on Pt was performed with cinchonidine
modified conventional Pt/Al 2 O 3 (Fig. 4-6) and cinchonidine modified Pt nanoclusters B
(Fig. 4-7) and D (Fig. 4-8) samples. The use of different Pt sample allows comparison
of adsorbed modes of the cinchonidine molecule of Pt crystals with different sizes,
starting from 1.2 (B-sample), passing 2.3 nm (D-sample) with 3.8 nm (conventional
alumina supported Pt) and ending with comparison of adsorption mode of cinchonidine
on microscopic Pt plane from the work of Ferri [76] and Zaera [77, 178]. In the Fig. 4-6
the region below 1300 cm -1 is not accessible of IR investigation, due to the contribution
of alumina background, whose FTIR spectrum can be found as an inset in the Fig. 5-5
from the chapter 5.
35

1
Absorbance
0.1
2
1600 1400 1200 1000 800
Wavenumber, cm -1
Fig. 4-6. DRIFT spectra of the cinchonidine modified conventional Pt/Al 2 O 3 (1) and
free cinchonidine (2), intensity of that spectrum is decreased for better presentation.
Absorbance
0.13
1
2
1600 1400 1200 1000 800
Wavenumber cm -1
Fig. 4-7. DRIFT spectra of the cinchonidine modified Pt nanoclusters (1, sample B) and
free cinchonidine (2), intensity of that spectrum is decreased for better presentation.
It is important to note, that the absence of Al 2 O 3 (normally used as a support for Pt
[76]) allows investigation down to the range of 700-900 cm -1 thus facilitating
assignment of the important out of plane vibrations of aromatic rings.
36

Absorbance
0.05
1
2
1600 1400 1200 1000
Wavenumber, cm -1
Fig. 4-8. FTIR spectra of the cinchonidine modified Pt nanoclusters (1, sample D) and
free cinchonidine (2), intensity of that spectrum is decreased for better presentation.
The spectra measured in transmission mode with KBr pressed pellet.
The 3-9 cm -1 blue and red shifts in the wavenumber of the corresponding peaks in the
spectra of free and adsorbed modifier clearly indicate on the fact that cinchonidine is
adsorbed on the Pt cluster. The quality of shown spectra increases significantly with
absence of a support due of the mentioned background contribution and because of
higher cinchonidine concentration of surface of Pt for B and D samples with respect to
modified conventional Pt/Al 2 O 3 . However, despite of the fact that not all peaks in the
spectrum of cinchonidine modified conventional Pt/Al 2 O 3 (Fig. 4-6) are resolved well
(due to the alumina background effect) and some corresponding peaks from
cinchonidine modified Pt nanoclusters (Fig. 4-7, 4-8) presence in the Fig. 4-6 as
shoulders, it is possible to conclude, that the spectra of adsorbed cinchonidine on all
three samples have no principal difference. And only difference is in quality of the
spectra. Through comparison of all three spectra is also possible to conclude that the
cinchonidine molecule is still on the Pt surface and did not decompose after samples
preparation, that is especially important to note for B and D sample. Since the spectrum
quality of B type of nanosized sample is better that the quality of modified conventional
Pt/Al 2 O 3 and D sample, this spectrum had been chosen for further detailed analysis,
where as brief analysis of two left spectra can be found below in this chapter (sample
B) and in the chapter 5 (cinchonidine modified conventional Pt/Al 2 O 3 ).
The DRIFT spectrum of sample B (Fig. 4-7) has many peaks and shoulders, thus it
gives difficulties to find corresponding peaks in the spectrum of free cinchonidine
unambiguously for all of them. This is due to the adsorption selection rule [179] by
which intensities of some vibration bonds of adsorbed molecules are changed. That is
why only some of the peaks and shoulders (which are well separated and/or with high
intensity) were chosen for further analysis, they are summarized in Table 4-3. Some of
the peaks from Table 4-3 could be easily assigned to the ‘flat’ or/and ‘tilted’ orientation
mode of cinchonidine by using data from the works of Ferri [76] and Zaera [77].
However, not all vibrations from Table 4-3 could be correlated with prior work due to
37

the focus of these prior experiments on selective regions of the spectra. In the work of
Ferri and co-workers the interval 1321-1635 cm -1 is particularly covered and in the
work of Zaera and co-workers, the description of many peaks is given, but some of
them, which were observed in our experiments, are not mentioned. This why the
assignment and interpretation of summarized in the Table 4-3 peaks of adsorbed
cinchonidine has been performed with help of modelling of the IR vibrations of the
molecule and metal adsorption selection rule [179]. The metal adsorption selection rule
[179] states, if a vibration of a molecule adsorbed on a metal surface produces a dipole
moment orientated perpendicularly to the surface, the intensity of this vibration is
enhanced (or can be observed) and vice versa; i.e. if the dipole moment is orientated
parallel to the metal surface, then intensity of the corresponding vibration is reduced (or
not present). It means that dipole moments corresponding to all vibrations from Table
4-3 are orientated mostly perpendicularly to the metal surface.
Fig. 4-9. Illustration of the metal adsorption selection rule. Image charges are drawn
dotted.
It is important to note, that intensity of the IR radiation adsorbed by a molecule is
proportional to the first time derivative of dipole moment in power of two, or, roughly,
it is proportional to the squared dipole moment. At the same time the absolute value of
electric field induced by any dipole moment is inversely proportional to the distance in
power of three. Thus contribution of the electric field from dipole moment of the
molecule adsorbed on the neighbor nanoparticle is significantly less (approximately in
the factor of 10 3 ) that electric field from dipole moment of the molecule adsorbed on
the current nanoparticle. Here we took into account, that size of the molecule is ~ 10-20
% of the size of nanocluster and typical distance between nanoparticles (in the solid
state) can be estimated as a size of nanoparticle. Since, an intensity of any
electromagnetic field is proportional to the squared amplitude of electric field, we can
conclude, that observed IR peaks belong to the molecule adsorbed on the current metal
nanoparticle only.
In the case of cinchonidine molecule, its optimized geometry was chosen in the open 3
conformation [77] and vibration modes with corresponding dipole moments orientation
were found from DFT calculation (with correlation coefficient dω theory /dω exp = 1.02).
Then every observed vibrational frequency from Table 4-3 was analyzed with respect
to the metal adsorption selection rule. The analysis results are also summarized in the
38

Table 4-3 and the ‘flat’ and ‘tilted’ modes of adsorbed cinchonidine are indicated by a
vibration band e.g. at 768 cm -1 and at 1514 cm -1 are illustrated in the Fig. 4-10 and Fig.
4-11, respectively.
Fig. 4-10. Flat adsorbed cinchonidine demonstrates vibration mode at 768 cm -1 with
dipole moment (big arrow) orientated mostly perpendicularly to the metal surface. The
most intensive displacement of atoms is shown by small arrows.
39

Fig. 4-11. Tilted adsorbed cinchonidine demonstrates vibration mode at 1514 cm -1 with
dipole moment (big arrow) orientated mostly perpendicularly to the metal surface. The
most intensive displacement of atoms is shown by small arrows.
It should be noted, that observed “blue and red shifts” in frequencies (see Table 4-3)
indicate the changes in the bond force constants and thus changes in the skeletal
geometry of an adsorbed molecule with respect to the free species. However, for the
purpose of this work, we will leave these considerations aside, since they do not hinder
determination of the orientation of adsorbed cinchonidine and other used molecules
(see below). Calculated (correlation coefficient 1.06) and measured IR spectra of free
quinoline (mostly medium and strong peaks were chosen) were used as a particular
reference for facilitating assignment of the spectra of the cinchonidine (as well as
quiphos with aniline from the next chapter), Fig. 4-12.
3
A
Relative units
2
B
1
C
0
D
1600 1400 1200 1000 800 600
Wavenumber cm -1
Fig. 4-12 Comparison of IR peaks between free cinchonidine (A), quinoline (B),
quiphos (C) and aniline (D).
Table 4-3. Assignment of some frequencies (cm -1 ) from IR spectrum of adsorbed and
free cinchonidine and free quinoline.
Cinchonidine
on Pt
Free
cinchonidine
Quinoline Cinchonidine
theoretical
Vibration
description
Orientation
of adsorbed
cinchonidine
1634 (sh) 1636 (med) - 1709
19,20 C strch. Flat
1613 (sh) 1619 (wk3) 1620
(med)
1594 (med) 1590 (str4) 1596
(med)
1670 C-Qu.
strch., H-
Qu. b.
1643 C-Qu.
strch., H-
Qu. b.
Tilted
Tilted
40

1576 (sh) 1567 (med) 1571
(med)
1622 C-Qu.
strch., H-
Qu. b.
1514 (str) 1507 (str) 1502 (str) 1556 C-Qu.
strch., H-
Qu. b.
1463 (str) 1452 (str) - 1494
1427 (sh) 1420 (med) 1431 (wk) 1445
15,17,22 CH 2 -
QL. sci.
2,3 C strch.,
C-H-Qu.,
ip. def.,
15 CH 2 ,
HO 2 C sci.,
H 2 C b.
1208 (med) 1207 (med) - 1240 C-H-Qu.
ip. b., 2,3 C
strch., HO,
2 CH b.,
QL. comp.
1179 (wk) 1180 (wk) - 1223
1168 (wk) 1163 (wk) - 1189
30 H-Qu.,
23,24 H b.,
H- 17 C-Ql.
comp.
25,26,30 H-
Qu. ip. b.,
23,24 H b.,
QL. comp.
1145 (wk) 1133 (wk) - 1143 H-Qu. ip.
b.,
23,24 H
b., Ql.
comp.
1118 (med) 1099 (wk) - 1128 H-Qu. ip.
b., O 2 C
strch., QL.
comp.
856 (med) 862 (wk) - 873
29,10 H-Qu.
oop. wag.,
QL. comp.
831 (med) 823 (med) 807 (str) 843 H-Qu. oop.
wag, H 20 2 C
sci., QL.
Comp.
809 (med) 803 (med) 787 (str) 824 H-Qu. oop.
Wag.,
40,41 H 20 C
sci., QL.
comp.
782 (str) 778 (wk) 761 (wk) 791 H-Qu. oop.
wag.,
40,41 H 20 C
Tilted , flat
Tilted
Tilted
Flat
Tilted
Flat
Flat, tilted
Flat
Flat
Flat
Flat
Flat
Flat
41

sci., QL.
comp.
768 (str) 756 (str) 740 (str) 775
25,26,27,28 H- Flat
Qu. oop.
wag.
Comments: sh. – shoulder, med. – medium, wk – weak, st. – strong., 15C –carbon atom
number 15. Qu – quinoline, QL – quinuclidine. C-Qu. – only carbons in quinoline, H-
Qu. – only hydrogens in quinoline, skel. – skeleton, comp. – complex: more then one
type of vibrations (stretching, wagging, rocking, and scissoring) are present, def. –
deformation, C-H-QL. – carbons and hydrogens in quinuclidine, b. – bending, strch. –
stretching, sci. – scissoring, wag. – wagging, ip.- in plane, oop. – out of plane vibration.
As it is seen from the Table 4-3, cinchonidine on Pt nanoclusters was found in both the
‘flat’ and ‘tilted’ adsorption modes. Here, we did not differentiate between the two
types of tilted adsorption modes, the so called “N-lone pair bonded” and “α-H
abstracted" because we found assignment of the corresponding species is somewhat
ambiguous and since the flat adsorption mode is more determinant in enantioselective
catalysis.
It is very important to compare the adsorbed modes of cinchonidine on macroscopic
and nanocluster surfaces of Pt. In fact, through comparison of IR investigations of
cinchonidine adsorbed on bulk Pt crystals (ATR, RAIRS) and on Pt nanoclusters
(DRIFTS) we found no principal difference in the spectra as well in the spectra of
sample D and cinchonidine modified conventional Pt/Al 2 O 3 (the only difference we
found was in the quality of the spectra). However, in the spectra of cinchonidine on Pt
nanoclusters small shifts of the spectra from free cinchonidine were found and between
the spectra of cinchonidine adsorbed on Pt. For example: in case of free cinchonidine,
the peak at 1593 cm -1 in work of Ferri [76] was observed at 1590 cm -1 in the current
thesis and 1591 cm -1 in the work of Zaera [77]; in case of cinchonidine on Pt, the peak
at 1458 cm -1 in work of Ferri was found at 1463 cm -1 in the current work and at 1463
cm -1 in the work of Zaera. It seems most likely that these small shifts are due to the role
of solvents/purity of cinchonidine used and other secondary effects. Another difference
is that the spectra of the ligands on nanoclusters contains less noise and, mostly, have
better quality with respect to the spectra of ligands on bulk metal. The reason for this
may be due to the high surface area of nanoclusters and thus higher concentration of
adsorbed ligand in the IR beam. If size of the cinchonidine supported Pt nanoclusters
(2.3 nm sample D, or 4 nm conventional Pt/Al 2 O 3 ) is bigger than typical size of
molecule (about 0.5 nm, the length of the quinoline anchor), the support can be
considered as a plane and molecule behaves similarly as to being adsorbed on a
continuous surface of a (bulk) metal crystal.
However, if the size of the clusters becomes comparable with the size of the molecule,
as we have in case of cinchonidine on Pt (1.2 nm in sample B) the ligand might behave
differently (especially in terms of enantioselectivity) and would be expected to
experience less steric hindrance due to the curvature of the surface.
In fact, Yang and co-authors [106] reported enantiomeric excess in hydrogenation of
methyl pyruvate to be 90 % over 1.2 nm Pt cluster and 85 % over 3.4 nm under 40 bar
of H 2 pressure. They explained this difference as a size effect resulting from
cinchonidine and methyl pyruvate forming a half-hydrogenated complex which is
adsorbed on different crystal faces. However, these authors did not focus on stability of
Pt nanoclusters during hydrogenation and changes in size, especially after removal of
42

PVP stabilizer. In the work of Bönnemann [114, 115] , no such dependence was found
(ee = 76% over 1.5 nm Pt cluster and 78 % over 3.9 nm, under 1 bar of H 2 pressure).
In our opinion, one of the primary effects causing enhanced ee is a higher concentration
of active chiral sites (“π-bonded” cinchonidine) on the catalyst surface. In the work of
Yang and co-authors cinchonidine adsorbs on the Pt surface from solution (optimum
concentration of cinchonidine [80, 102, 180]) whereas high hydrogen pressure (40 bar)
promotes a surface cleaning (from PVP, AcOH and etc.) for obtaining more space for
flat adsorbed cinchonidine ( cinchonidine in any tilted mode needs, definitely, less free
space for the adsorption). The catalyst prepared according to Bönnemann’s procedure
(sample B) possesses adsorbed cinchonidine in flat and in two tilted modes as
demonstrated in this manuscript and a high pressure of hydrogen was not applied.
Moreover ee=94-95 % was obtained under high (100 bar) H 2 pressure with
conventional 5R94 [180] and Engelhard 4759 [181]- Pt/Al 2 O 3 catalysts (average
particles size 3.6 nm [182]) where the above mentioned size effect seems to be less
probable.
Based on these facts, we would like to conclude here, that it was found that the size of
Pt clusters (in range 1.4- ∞ nm) has no observed influence on the adsorption mode of
the cinchonidine (at least in case of sample B (1.4 nm), sample D (2.3 nm),
conventional Pt/Al 2 O 3 (4 nm) and macroscopic samples from work of Zaera [77] and
Ferri [76]). This is important because it has been reported in the literature that cluster
size does influence both catalytic activity and the observed ee [106].
Thus, it was found that both ‘flat’ and ‘tilted’ modes of adsorbed cinchonidine are
present on Pt nanoclusters and could be an explanation of the fact that this type of
catalyst does not provide the maximum possible enantiomeric excess (99 %).
4.3.3 Investigation of catalytic behaviour of cinchonidine
modified Pt nanoclusters
Investigation of catalytic behavior of cinchonidine modified Pt nanoclusters was
studied with B, D and conventional modified Pt/Al 2 O 3 in the hydrogenation of ethyl
pyruvate.
Initial transient period
Under all reaction conditions, summarized in Table 4-4, enantiomeric excess was found
to depend on the reaction time in a manner similar to that shown in the Fig. 4-13.
During the first moments of hydrogenation (the period of time termed the initial
transient period) enantiomeric excess increases to a maximum value (ee max ) and then
decreases until the entire amount of ethyl pyruvate is converted in the reaction mixture.
43

Conversion, %
100
80
60
40
ITP time
ee max
80
70
Actual enantiomeric excess, %
20
0
Y =83.7-0.19 X
60
0 25 50 Time, min75 100 125
Fig. 4-13. Typical catalytic behaviour of cinchonidine modified Pt nanoclusters during
ethyl pyruvate hydrogenation.
Effect of hydrogen pressure
The effect (Exp. 1 and 2, see Table 4-4) of changing the hydrogen pressure in the
reaction system is shown in the Fig. 4-14 (other graphs are similar, thus they are not
shown in further, important features are mentioned in the text, in the corresponding
paragraphs and in the Table 4-4). With an increase of hydrogen pressure from 2 to 10
bar the initial rate increases by a factor of ten from 14 min -1 to 140 min -1 An increase of
initial rate could be qualitatively explained as an increase of hydrogen migration from
the gas to the liquid phase, but the fact that the maximum enantiomeric excess reached
significantly increases from 36 % to 78 % seems to indicate a specific hydrogen –
catalyst interaction, as a consequence of this, the catalyst demonstrates increasing
enantioselectivity. A decrease by 4 fold in the time required to reach ee max is observed
by changing the pressure from 2 (110 min) to 10 (30 min) bar.
44

80
Conversion, %
100
80
60
40
Y =83 - 0.19 X
Hydrogen pressure
10 bar
2 bar
70
60
50
40
Actual enantiomeric excess, %
20
30
0
20
0 50 100 150 200 250 300
Time, min
Y =44 - 0.07 X
Figure 4-14. Effect of hydrogen pressure.
Experiment 1: 5 mg (13 μmol Pt) D- type of catalyst, [EP 0 ] =2.571 mol l -1 , no free
cinchonidine, P H2 = 2 bar. Full squares (left axis) – ethyl pyruvate conversion, crossed
squares (right axis) – actual EE, line – linear approximation of actual ee decrease.
Experiment 2: 4 mg (13 μmol Pt) D- type of catalyst, [EP 0 ] = 2.571 mol l -1 , no free
cinchonidine, P H2 = 10 bar. Full triangles (left axis) – ethyl pyruvate conversion,
crossed triangles (right axis) – actual ee, line – linear approximation of actual EE
decrease.
Type of
[EP ],
Table 4-4. Details of catalysts and reaction parameters.
Exp. P H2 , Pt surf.
ee * max, Cinchonidine, Time slope Initial
0
N. bar Atoms, catalyst % concentration, ee * max, min -1 mol l - rate b ,
μmol a and
mmol l -1
1
min
min -1
loading,
mg
1 2 5.8 D, 4 36 0 110 0.07 2.571 14
2 10 5.8 D, 4 78 0 30 0.07 2.571 140
3 c 10 5.8 D, 4 76 0 20 0.19 0.514 99
4 c 10 29 D, 20 65 0 9 0.22 0.514 37
5A c 10 1.11 Pt/Al 2 O 3 , 76 1 - 0.24 0.257 128
20
5B c 10 0.56 Pt/Al 2 O 3 ,
10
76 1 - 0.21 0.257 132
45

6A c 10 19.5 B, 20 61 0 3 0.18 0.514 24
6B c 10 3.9 B, 4 76 0 - - 0.514 -
7A d 10 2.2 D, 1.5 79 1 30 - 0.514 e 130
7B 10 2.2 D, 1.5 78 1 30 - 0.514 120
8 c 10 5.8 D, 4 76 1 55 0.04 0.514 121
9 c 10 3.9 B, 4 74 1 34 0.07 0.514 42
Comments: a – The number of Pt surface atoms was estimated from the medium
particles size according to “full shell” model of Pt nanoclusters and from chemisorption
on conventional Pt/Al 2 O 3 , b – Initial rate values [min -1 ] were calculated as
normalization a reaction rate [mmol min -1 ] on number of Pt surface atoms [mmol] in
the reaction mixture, c – The graph is not shown; important parameters of reaction are
presented, d – The purified EP was used.
Effect of catalyst loading
The decrease of initial rate from 99 min -1 to 37 min -1 with an increase of catalyst
loading from 4 mg (13 μmol Pt) to 20 mg (65 μmol Pt) was found in Exp. 3 and 4.
Also, the maximum of actual enantiomeric excess (ee* max ) has also decreased from 76
% to 65 %. A possible explanation for this is that the time necessary to reach the
maximum enantioselective activation level may not be the same for various catalyst
loadings and the full conversion of ethyl pyruvate has already been accomplished. In
other words, the number of chiral sites (π-bonded cinchonidine) did not reach the
maximum possible value under the 10 bars of hydrogen pressure. It should be noted,
that this effect can not be explained as mass transport phenomena in liquid phase, since
according to Minder, Baiker and co-workers [183] the kinetically controlled region lays
below 15 g l -1 of catalyst concentration while in the Exp. 3 and 4 the maximum catalyst
concentration is 0.28 g l -1 . In fact, we performed the reactions with 20 mg (Exp. 5A,
ee max = 76 %, initial rate = 128 min -1 ) and 10 mg (Exp. 5B, ee max = 76 %, initial rate =
132 min -1 ) with conventional Pt/Al 2 O 3 and concluded that in our experiments mass
transport effect in liquid phase does not limit the reaction. The rate of migration of
hydrogen from gas to the liquid phases was the same, since both experiments were
performed under the same hydrogen pressure (10 bar).
A similar decrease of ee max with an increase of catalyst loading was found for the B
type catalyst (see Exp. 6A and 6B in Table 4-4).
By comparing these results with those reviewed by Studer, Blaser and Exner in 2003
[44] we can see that there are intrinsic differences between the colloids stabilized by
chiral ligands and the conventional Pt/Al 2 O 3 heterogeneous catalysts which are
unmodified prior to introduction to the reaction vessel. In fact, decrease of initial rate
and maximum enantiomeric excess with an increase of catalyst loading are in
contradiction to the results observed for the conventional Pt/Al 2 O 3 system which may
be attributed to the fact, that our catalysts have adsorbed cinchonidine already on them,
whereas the catalyst systems reviewed in the literature are exposed to cinchonidine
during the reaction (in situ).
Role of the ethyl pyruvate in nature of ITP
The ITP time was found to be almost independent of the initial concentration of ethyl
pyruvate (Exp. 2 and 3). The ITP was found to be 20-30 min in Exp. 2 (20 ml EP) and
20-25 min in the case of Exp. 3 (4 ml EP). Ee* max is almost the same value (76-78 %)
in both experiments, but initial rate increased by ≈ 30 % with an increase of the initial
concentration of ethyl pyruvate.
46

In order to clarify the role in ee of the initial presence of ethyl lactate (0.5 % each S and
R isomers) as an impurity in EP, the catalytic hydrogenation reaction with the D type of
catalyst was performed with both the unpurified and purified EP. Fig. 4-15 (Exp. 7A,
7B) shows that in this experiment the ee observed is similar for both the purified and
unpurified experiments with respect to the reaction time and both demonstrate the same
ITP = 30 min and initial rate = 120-130 min -1 .
80
Conv. 47 %
Enantiomeric excess, %
76
72
Purified EP
Normal EP
Conv. 4 %
68
0 20 40 60 80
Time, min
Fig. 4-15. Role of the impurities in the ethyl pyruvate on EE.
Experiment 7A: 1.5 mg (4.8 μmol Pt) D type of catalyst, [EP 0 ] = 0.514 mol l -1
(purified), [Cinchonidine 0 ] =1 mmol l -1 , P H2 = 10 bar.
Experiments 7B: 1.5 mg (4.8 μmol Pt) D type of catalyst, [EP 0 ] = 0.514 mol l -1 ,
[Cinchonidine 0 ] =1 mmol l -1 , P H2 = 10 bar.
However, the initial presence of R, S ethyl lactates gives a systematic shift in ee. This
shift has a maximum value (≈ 3 %) in initial times (at low conversion) and decreases
with time (with increase of conversion). Taking into account the initial concentration of
both ethyl lactates, the values of the shift lie within the estimated limit.
In order to show that the presence of EP does not play a significant role in increasing
ee, for example as a reagent in side reactions (aldol condensation) reaction [184] we
performed a blank experiment in the absence of EP with free cinchonidine under
reaction conditions (acetic acid, 10 bar of H 2 ). However, both B and D samples are
unstable with prolonged hydrogen exposure especially at high pressure. Thus, to
improve stability, we prepared cinchonidine modified Pt nanoclusters deposited on
alumina. This catalyst was found to be stable under the reaction conditions for a long
time and demonstrates (P H2 = 10 bar, [EP 0 ] = 0.514 mol l -1 , [cinchonidine] = 1 mmol l -
1 ) ee = 79 %, however, after pretreating the catalyst for 16 hours in acetic acid it
demonstrates ee = 84 %. This experiment obviously shows that the ethyl pyruvate does
not influence significantly to the increase of ee after exposing catalyst under reaction
condition. Another interesting fact is that under optimized conditions (hydrogen
47

pressure, exposing time, cinchonidine concentration, etc.) the maximum obtained ee
was found to be 88 %.
In order to keep clarity in description of the samples and their characterization, it is
recommended to the reader to see Chapter 5 for further information about cinchonidine
modified Pt nanoclusters deposited on alumina, where we give detailed models of effect
of ee increasing, whereas here we would like to continue description of observed ITP
effect.
Role of free cinchonidine in ITP
The addition of 20 mg (0.068 mmol) of free cinchonidine into the reaction mixture
(Exp. 3 and 8) increases ITP by a factor of almost three (from 20 min to 55 min) while
ee* max ≈ 76 % remained nearly constant and the initial rate increases only slightly. The
stability in ee with addition of free cinchonidine was also found for B type of catalyst in
out experiments and in original work of Bönnemann [114]. This fact might be
explained by competitive formation and disappearance of new chiral sites, as is
discussed below.
Post ITP decrease of enantiomeric excess
The decrease of ee during the hydrogenation of ethyl pyruvate has been previously
observed both at high (40-100 bar) [50, 185] and low (~1 bar) [186] pressures of
hydrogen and has been shown to result from the partial hydrogenation of the quinoline
(anchor) portion of the cinchonidine. The partially hydrogenated quinoline ring is
unable to function as an anchor any longer due to the loss of the π-aromaticity and as a
result, the cinchonidine desorbs from the surface resulting in an unmodified catalyst
which is not capable of inducing enantioselectivity. Our results support the possibility
that this effect also takes place with ‘quasi-homogeneous’ catalysts in the 2-10 bar
range.
To the best of our understanding and characterization of the effect, the decay of the
actual ee after ITP was linearly approximated and the absolute value of the line’s slope
was used to determine the stability of catalyst enantioselectivity with respect to the
catalyst type and reaction conditions. Thus, in experiments with the same initial
concentration of ethyl pyruvate we can easily see where no additional cinchonidine was
added (Exp. 3, 4, slope ~ 0.18-0.22 min -1 ) the rate of decrease of actual enantiomeric
excess was observed to be 3-4.5 times greater than with (Exp. 8, 9) free (20 mg or 1
mmol l -1 ) cinchonidine (slope ~ 0.04-0.07 min -1 ).
Qualitatively, this effect can be explained by maintaining a more or less constant π-
bonded cinchonidine concentration on Pt surface due to the replacement of “old”
molecules (with partially hydrogenated quinoline anchor) by “new” molecules of
cinchonidine from solution.
Comparison of D, B ‘quasi-homogeneous’ and conventional
heterogeneous catalysts
Comparison of catalytic behavior
The important principal difference in catalytic behavior of D, B and conventional
Pt/Al 2 O 3 catalyst systems is the negligibly small increase of actual (non-cumulative)
enantiomeric excess during the reaction in the case of the previously non-modified
48

Pt/Al 2 O 3 (Exp. 5A). This increase of enantiomeric excess was found to be significantly
smaller than in the series of experiments with pre-modified catalysts. This indicates the
ITP effect is nearly non-existent for the systems which are modified in situ under our
reaction conditions. That the ITP effect is hardly observed in acetic acid for catalysts
without cinchonidine pre-modification is in agreement with the findings of Balazsik
and Bartok [169]. In our system, the conventional Pt/Al 2 O 3 shows ee* max = 76 % which
is the same value as was found in Exp. 3 and 8 with the D type of catalysts and also
demonstrated similar values of initial rate. However, the rate of decrease of the actual
enantiomeric excess is 3-6 times higher for the conventional catalyst than for the B and
D types of catalyst with the presence of the same amount of free cinchonidine in the
reaction mixture (Exp. 8 and 9). Also, the D type of catalyst gave slightly higher
maximum enantiomeric excess and initial rate than the B type under the same reaction
conditions (compare Exp. 3 with 6B, Exp. 4 with 6A and Exp. 8 with 9) possibly due to
the different preparation methods (used solvents, washing procedure, preparation
temperature, etc.).
The two-cycle mechanism proposed by Blaser, Garland and Jallet [81, 180], based on
the assumption that the reaction goes with [R]/[S] = 1 on unmodified sites and with
[R]/([R]+[S]) = s on modified sites (near flat adsorbed cinchonidine), explains the
observed correlation between ee max and initial rate (R) as ee max = A-B*R -1 , where
A=(2s-1)k m /(k m -k u ), B=A*k u , k u and k m are pseudo-first-order rate constants of
unmodified and modified cycles. In fact, the graph ee max vs. R -1 (Fig. 4-16) shows a
good linear correlation for all our experiments. This means that on modified sites the
ratio [R]/[S]=(1/s)-1 is the same for all catalysts and does not depend (within the
margin of error) on reaction conditions, this seems reasonable because the [R]/[S] ratio
is determined by the nature of the interactions between molecules of cinchonidine, ethyl
pyruvate and hydrogen on Pt surface. However the ratio of modified/unmodified sites
might strongly depend on experimental conditions, hydrogen pressure (cleanliness of
the surface), type of catalyst, concentration of modifier, etc.
80
ee max
, %
60
ee max
= A - B*R -1
40
0.00 0.02 0.04 0.06 0.08
Rate -1 , min.
49

Fig. 4-16. Linear relationship between maximum reached enantiomeric excess and
inversed initial rate for Exp.1-9.
Comparison of the stability of samples
It is obvious that for the high performance, long activity and reuse of the catalyst the
stability of its structure is very important aspect to focus on. Here we qualitatively
compare stability of B type of colloidal sample before (Fig. 4-16-1) and after (Fig. 4-
16-2) ethyl pyruvate hydrogenation. Precipitation of Pt particles due to the
agglomeration says about low stability of the sample under exposure to hydrogen for a
long time (approximately 2-3 hours, 10 bars of H 2 , no free cinchonidine). It should be
notes, that sample can be kept in colloidal form in acetic acid solution for long (~1
year) time if no hydrogen is applied.
Fig. 4-17. Soluble (stable) cinchonidine modifier colloidal Pt nanoparticles in acetic
acid before hydrogenation reaction (1) and precipitated (2) after exposure under 10 bars
of hydrogen (in AcOH) for 2 hours.
IR investigation
The orientation of the cinchonidine species on Pt during the reaction was investigated
by measuring the IR spectra of D type catalyst prior to and after 50 min and 24 hours
under reaction conditions, at 10 bar. After 50 min, the reaction in Exp. 4 (P H2 = 10 bar,
[cinchonidine] = 0 ml l -1 , [EP] = 0.514 mmol l -1 ) was terminated, the hydrogen pressure
was released and the reactor opened to ambient conditions. After 1 hour the colloidal
catalyst precipitated and a black powder was collected, washed three times with acetic
acid, dried at 30 ºC under vacuum and finally was used for IR investigation. To insure
that the washing procedure does not influence results, a control experiment was
conducted both before and after washing B and D type catalysts (without exposure to
the catalytic reaction conditions). Because the spectra revealed the presence of both the
tilted and flat orientations of cinchonidine for all samples it was determined that the
washing procedure does not influence the results.
The reaction in Exp. 4 was repeated and when finished, the reaction mixture was kept
in the reactor under 10 bar of hydrogen pressure for 24 hours after which the colloid
solution formed a black precipitate that was collected, washed three times with acetic
acid and dried at 30 ºC under vacuum and finally was used for IR investigation.
For better visualization, intensity of the spectra (Fig. 4-18) of free cinchonidine was
decreased. Intensities of the spectra of the D type catalyst are presented as measured.
To aid in spectral analysis, the observed vibration frequencies are summarized in Table
4-4.
50

0.16
0.12
1
1 Free cinchondine
2 Before reaction
3 After 1 h
4 After 24 h
Absorbance
0.08
2
0.04
3
4
0.00
1700 1600 1500 1400 1300 1200 1100 1000
Wavenumber, cm -1
Fig. 4-18. FTIR spectra of the D type of catalyst at different reaction time and spectra
of free cinchonidine.
Table 4-4. Vibration band assignment (cm -1 ) of the cinchonidine spectra of the free
cinchonidine, and D type of catalyst before reaction, at 50 min and after 24 h under
reaction condition.
Free
Before reaction At 50 min At 24 h Assignment a
cinchonidine
1162 1152 1152 1152 F: 1
1207 1209 1209 1209 T: 2, 3
1383 1383 - 1382 F: 1, T: 2, 3,
acetic acid
1420 1402 1402 - F: 1, acetic acid
1451 1456 - - T: 2, 3
1507 1507 - - T: 3
1569 1570 1567 - F: 1, T: 2, 3
1590 1590 - - T: 3
Comments: a – F: 1 – flat adsorbed cinchonidine, T: 2 and T: 3 – cinchonidine adsorbed
in tilted mode, according to the Fig. 1-19. Assignment of peaks had been done
according to the work of Ferri and Burgi [76], Zaera [77, 178, 187] and Table 4-3 .
Analysis of the IR spectra of the D system shows that prior to reaction, cinchonidine is
adsorbed in the flat (1) and in two tilted modes: α-H abstracted (2) and in N-lone pair
bonded (3). It was found that after 50 min under 10 bar of hydrogen the peaks at 1456
cm -1 and 1383 cm -1 corresponding to the tilted (2 and/or 3) and peaks at 1407 cm -1 and
1590 cm -1 tilted (3) form of cinchonidine are absent and only those IR peaks at 1152
cm -1 , 1209 cm -1 , 1402 cm -1 and 1567 cm -1 and indicating the flat-mode (1) and tilted
mode (2) are observed. After 24 hours under 10 bar of hydrogen only peaks at 1152 cm -
51

1 , 1209 cm -1 and 1382 cm -1 remain. The peaks at 1152 cm -1 and 1209 cm -1 could be
assigned to flat (1) and tilted (2) modes, whereas the peak at 1382 cm -1 corresponds
here to the acetic acid.
The absence of the tilted (3) mode of the adsorbed cinchonidine (absence of the peaks
at 1590 cm -1 , 1507 cm -1 and possibly 1456 cm -1 ) after approximately one hour of
hydrogen exposure indicates that N-lone pair bonded cinchonidine on Pt nanoclusters
has a weak adsorption strength with respect to the flat (π-bonded) and α-H abstracted
species. This fact has been demonstrated previously in the work of Ferri and Burgi [76]
for a thin film of Pt. Further, we have recently demonstrated above that the adsorption
modes for cinchonidine for both the thin film and nanocluster systems are similar
(please the beginning of this chapter).
Collectively, these experimental data of the system have led us to propose the model
illustrated in the Fig. 4-19 for the observed phenomena. During ITP for our Pt
nanocluster systems, desoprtion of weakly bound cinchonidine molecules (tilted
orientation) is occurring as a result of interaction with hydrogen and possibly solvent,
making free space for the formation of new chiral sites via adsorption of cinchonidine
in the ‘flat’ conformation thus leading to an increase of ee. When cinchonidine
coverage of the Pt surface reaches some minimum (due to hydrogenation and/or
desorption) the nanoclusters become unstable and agglomeration takes place as was
observed with B and D catalysts.
cinchonidine
in solution
H 2
1507, 1590 cm -1
N
N
HO
H
N
OH
H
H
H
N
N
H H CO/CO 2 /H 2 O
OH
H
H
N
1 2 3
Pt surface
Fig. 4-19. Proposed model of behaviour of cinchonidine on Pt under hydrogen
exposure. 1-“π-bonded”, 2- “α-H abstracted”, 3- “N lone pair bonded”.
The addition of free cinchonidine leads to an increase of enantioselective stability of
these catalysts and increase of ITP likely due to the replacement “old” to “new”
cinchonidine molecules on Pt.
It is interesting to compare adsorption energies of hydrogen and cinchonidine on Pt
surface in order to check a principal ability of hydrogen to replace cinchonidine from
the surface. It is well known that hydrogen adsorb on Pt surface with further
dissociation to two hydrogen atoms (here we consider strong hydrogen adsorption
only), however adsorption energy of hydrogen atom on Pt depends on many factors
such as type of Pt crystal plane, adsorption site (atop, bridge, etc.), size of the Pt cluster,
nature of a support (Al 2 O 3 etc.), presence of electric charge, etc. The experimental
values for hydrogen adsorption on Pt (111) range from 29 to 90 kJ mol -1 [188-190] .
52

We do not consider an extreme values, and would like to take the average 45 kJ mol -1
values for rough comparison, since it is considered for the uncharged, unsupported Pt
(111) plane [191]. The adsorption energy of cinchonidine on Pt continuous plane
deposited on Al 2 O 3 support in CH 2 Cl 2 was estimated to be 31 kJ mol -1 [76], note that
this is average value assigned for all three species of cinchonidine adsorption. Thus, we
can conclude, that hydrogen atoms, in principal, can remove all adsorbed species of
cinchonidine from Pt surface, however there is a high chance, that weakly adsorbed
mode of cinchonidine (N-lone pair bonding) will be removed (desorbed) faster that
others.
4.4 Summary
The orientation of the chiral modifier cinchonidine on the surfaces of Pt nanoclusters
has been investigated for the first time using a combination of DRIFTS and molecular
modeling methods. It was found that the size of Pt clusters (in range 1.2- ∞ nm) has no
observed influence on the adsorption mode of the cinchonidine. Thus, it was found that
both ‘flat’ and ‘tilted’ modes of adsorbed cinchonidine are present on Pt nanoclusters
and could be an explanation of the fact that this type of catalyst does not provide the
maximum possible enantiomeric excess (98 %).
For the first time, the ITP was observed and studied for cinchonidine modified Pt
‘quasi-homogeneous’ systems. The ITP was observed at room temperature (22 ± 1 ºC),
in acetic acid media with catalysts prepared by two different methods (D and B type),
with different catalyst loadings (1.5 - 20 mg), with different ethyl pyruvate
concentration (0.257, 0.514 and 2.57 mol l -1 ) under 2 and 10 bar of hydrogen pressure
both with (1 mmol l -1 ) and without addition of free cinchonidine into the reaction
mixture.
The ITP was found to be dependent on hydrogen pressure and catalyst loading,
increasing from 30 min to 100 min with a decrease of hydrogen pressure from 10 bar to
2 bar. Further, ITP decreases from 30 min to 9 min with an increase of catalyst (D-type)
loading from 4 to 20 mg. ITP was also found to be independent of the initial
concentration of ethyl pyruvate in the range of 0.257-2.57 mol l -1 . It was also found that
cinchonidine modified Pt nanoclusters on alumina catalyst demonstrates (5-10%)
higher ee after treatment under reaction condition without EP.
Addition of free cinchonidine (1 mmol l -1 ) shows almost no detectable influence on the
maximum value of ee and slight increase of initial rate, but increases ITP by a factor of
3 and enantioselective stability of catalyst by nearly 5 fold. The highest ee was also
found to be independent on the initial concentration of the ethyl pyruvate and decreases
with an increase of the catalyst loading from 78 % to 65 % in the range of 4 mg to 20
mg, probably due to the incomplete chiral activation of the catalyst.
The racemic ethyl lactates as an impurity in EP does not play a significant role in
behavior of enantiomeric excess, however they make systematical down shift in ee,
which decreases with conversion.
From IR investigation, it was found that after approximately one hour under 10 bar of
hydrogen pressure in acetic acid, the weakly bound species which correspond to
cinchonidine adsorption in the tilted (“N lone pair bonded”) mode are not observed and
only the strongly bound species corresponding to cinchonidine adsorption in the flat
and in tilted (“α-H abstracted”) mode remain. In other words, “N lone pair bonded”
species of the adsorbed cinchonidine are desorbed due to the influence of hydrogen and
possibly solvent.
53

Our results are in line with previously published work of Balazsik and Bartok [169] in
which the authors assigned the nature of the ITP to competitive adsorption of reactants,
modifier and solvent and supplement it in desorption of weakly bonded cinchonidine
adsorbed in tilted “N lone pair bonded” mode. It also has to be noted, that the hydrogen
effect of cleaning Pt surface from previously adsorbed impurities: H 2 O/CO/CO 2 and
oxygen plays a major role in ITP [166].
54

Chapter 5
Cinchonidine modified Pt colloidal
nanoparticles immobilized on a
heterogeneous support
5.1 Introduction
Orito [45, 46] first reported the system of cinchonidine on Pt/Al 2 O 3 as a catalyst for
enantioselective hydrogenation of α-ketoesters. The ability of Pt modified with
cinchonidine to induce enantioselectivity has great fundamental and practical interest
[21], for example in the synthesis of HPB ester (Ethyl (R)-2-hydroxy-4-phenylbutyrate)
[22]. In order to obtain high enantiomeric excess, the reaction conditions were
optimized by Orito and Blaser and summarized in a review of Studer, Blaser and Exner
in 2003 [44]. As can be seen from this review, that high (≥ 40 bar) hydrogen pressure
and thermal pretreatment (generally at 400 °C under flow of H 2 ) of the conventional
Pt/Al 2 O 3 are critical requirements to obtain high ee. However, the use of high hydrogen
pressure and thermal pretreatment of the Pt/Al 2 O 3 catalyst can induce technical
difficulties, safety issues and increase production cost especially on an industrial scale.
The available literature dealing with mild conditions (no thermal treatment, low (≤ 10
bar) hydrogen pressure) are limited [51, 78, 192, 193].
In fact, ee of 94 % was obtained in ethyl pyruvate hydrogenation by Sun and coworkers
[78], under 5.8 bar H 2 , 17 °C, however, this system required maintaining low
dihydrocinchonidine concentration through dosing of modifier (otherwise ee
significantly decreases) throughout the reaction and a very specific reaction procedure
which included “add catalyst and AcOH to reactor, hydrogenate for 1 h under the
hydrogenation condition (e.g. 17 °C, 5.8 bar). Evacuate H 2 and add the modifier to the
reactor under a N 2 purge. Apply H 2 to desired pressure and quickly add pyruvate to
reactor under H 2 pressure using a syringe pump”. Despite the fact that the authors
obtained 94 % ee without thermal pretreatment of the catalyst, such a reaction
procedure seems to present its own set of difficulties.
In another interesting work of Bartok and coworkers [51], 97 % ee was obtained in
ethyl pyruvate hydrogenation under 10 bar of H 2 with ultrasonic activation (20 kHz, 30
W) in a closed hydrogen atmosphere, the activated catalyst thus requires handling
without exposure to air. Further, using this ultrasonic activation it was reported that
changes in the size and morphology of Pt nanoclusters on alumina support occurred.
It is also should be noted, that the “CatASium F214” [192] catalyst developed by
Degussa and Solvias does not require thermal activation and can induce e.g. 94 % ee in
ethyl pyruvate hydrogenation under 60 bar of H 2 . “Alternatively, the reaction can also
be carried out at 1.1 bar hydrogen pressure in a glass shaker and the hydrogen can be
supplied from a balloon. However, activity and selectivity don't reach an optimum”
[193].
Herein, we present the detailed preparation of a novel heterogeneous catalyst:
cinchonidine modified Pt nanoclusters deposited on non-porous alumina which
demonstrates 80-88 % ee in hydrogenation of ethyl pyruvate under relatively low
hydrogen pressure (2.5-10 bar). The reaction procedure does not require thermal or
ultrasonic activation of the catalyst and allows storage of the catalyst in air before
55

eaction. The stability of the catalyst structure under reaction and preparation
conditions is considered as well as catalyst recycling and reuse. The influence of the
catalyst nature towards obtaining high enantiomeric excess in this pressure range is also
considered.
5.2 Experimental
Materials
5% Pt on alumina (Aldrich, dispersion 29 %, Pt particles size 3.9 nm), cinchonidine
(98%, Fluka), acetic acid (99.8% Fluka), ethyl pyruvate (98%, Aldrich), formic acid
(Aldrich, 99%), acetone and aluminum oxide (4 m 2 g -1 ) 90 active neutral (0.063-0.2
mm) by Applichem were used as received.
Samples preparation
Adsorption of PtCl 4 on alumina was performed according to the following procedure. 4
g Al 2 O 3 was heated at 180 °C under vacuum (10 -3 mbar) for 16 hours and then was
cooled to room temperature under vacuum. To this alumina, 16 mg PtCl 4 dissolved in
40 ml of acetone solution was added by syringe. After 5-10 minutes stirring, the
initially dark supernatant acetone solution became colorless and an additional 10 ml
aliquot of acetone (with 30 mg PtCl 4 dissolved) was added by syringe. This procedure
was repeated several times, until 480 mg PtCl 4 was added, at which point the
supernatant no longer became colorless. The mixture was then refluxed for 12 hours
and the supernatant was evaporated under vacuum. Thus 480 mg PtCl 4 on 4000 mg
alumina (~11 % PtCl 4 ) was obtained.
In order to reduce Pt The 600 mg of alumina with adsorbed PtCl 4 was transferred into a
round bottom flask with distilled water (100 ml). The mixture was refluxed with
constant stirring. 160 mg dihydrocinchonidine dissolved in 25 ml aqueous solution of
formic acid 0.1 mol·L -1 was added rapidly by syringe and Pt was reduced. The heat was
removed after 10 min. The mixture was allowed to stir for 24 hours, the black powder
was isolated by filtration and washed with water, then with acetic acid and finally dried
in air at room temperature. Thus prepared catalysts will be referred to as A1 henceforth
in this manuscript.
For catalyst activation140 mg of the A1 catalyst was transferred into a metal reactor
with 70 ml acetic acid and 20 mg cinchonidine. The mixture was maintained under
hydrogen pressure (10 bar) for 1 hour with constant stirring (700 rpm). The pressure
was then released and the black powder was collected, washed 3 times with acetic acid
(100 ml), and then dried in air at room temperature. Thus prepared catalysts will be
referred as A2 henceforth in this manuscript.
Unless otherwise mentioned, the conventional Pt/Al 2 O 3 was heated to 400 °C under
vacuum (10 -3 mbar) for 1 hour, then kept for 3 hours under a flowing mixture of N 2
(93%) and H 2 (7%) gases (100 ml min -1 ) and subsequently cooled to the room
temperature. The required amount was then weighed (in open atmosphere) and
immediately transferred into a glass reactor with acetic acid, cinchonidine and ethyl
pyruvate.
Modification of conventional Pt/Al 2 O 3 has been done, according to the following
procedure. 40 mg of conventional Pt/Al 2 O 3 was heated to 400 ºC under vacuum (10 -3
mbar) for 1 hour, then the mixture of N 2 (93%) and H 2 (7%) gases were passed (100 ml
min -1 ) over the sample for 3 hours. Then, under flowing gases, the support was cooled
to room temperature and the solution of cinchonidine in CHCl 3 (6.0 ml, 3 mM) was
56

added by syringe through a stopper. The system was kept at 10-15 ºC for 24 hours, then
washed 3 times with CHCl 3 and finally dried under vacuum at room temperature.
Characterization and experiments
Transmission electron microscopy (TEM) micrographs were taken using a TECNAI
F20 instrument operating at 200 kV. Specimens were prepared by placing a drop of an
ethanolic sample suspension onto a copper grid with a perforated carbon film and then
allowing the ethanol to evaporate.
For IR investigation a “Nicolet 4700 FT-IR” spectrometer with liquid nitrogen cooled
detector and nitrogen purged DRIFT accessory was used with the following parameters:
200 scans, 600-4000 cm -1 scan range, 4 cm -1 resolution. KBr background was acquired
before spectra measurements. The DRIFT spectra were chosen for analysis because it is
the more applicable technique for observing adsorbed species on finely divided metal
powders [174].
Hydrogenation of ethyl pyruvate
Hydrogenation of ethyl pyruvate was carried out in a metal reactor (Parr Instrument
GmbH) under a pressure of 2.5-10 bar of hydrogen at room temperature (22±1 °C). The
stirring speed was 1000 rpm.
Typical conditions were such that 4 ml (36 mmol) ethyl pyruvate was added to 20 mg
(68 µmol) cinchonidine in acetic acid (70 ml) solution. 20 mg of catalyst was added
into the mixture and then transferred into the reactor, flushed with Ar for 10 min and
finally, after the pressure of hydrogen had been stabilized for 10-15 sec, the initial time
(t=0) was set and the reaction kinetics were monitored. About one ml aliquots of the
reaction mixture were taken at certain time intervals, filtrated with a G3/G4 glass filter
and 1 μl was injected into the gas chromatograph (GC). GC measurements were taken
using a Varian 3900 instrument with FID detector and Lipodex-E column.
Catalyst stability and reuse tests
In order to test the stability of the catalysts under the typical reaction conditions, 30 mg
A1 was transferred into the reactor with cinchonidine (250 mg) in 70 ml acetic acid
solution and left with continuous stirring under 10 bar of hydrogen pressure for 16
hours. The pressure was then released and samples were collected for TEM analysis.
In the case of catalyst reuse experiments, after the reaction is finished (conversion
reached 100 %), hydrogen pressure was released and the reactor was unplugged. The
black powder was collected through centrifugation and washed three times with acetic
acid, then was dried in air at room temperature; the catalyst was then used again in the
identical manner to that described above.
5.3 Results and discussion
5.3.1 Role of catalyst activation in obtaining enantiomeric
excess
The A1 catalyst (Table 5-1) demonstrates 79-80 % of ee which is similar to the data
reported by Bönnemann [114, 115] for alumina supported cinchonidine modified Pt
nanoclusters. The same catalyst after activation (A2) demonstrates 85 % of ee and an
increase of reaction rate by a factor of ~9. The A2 catalyst demonstrates stable ee and
100 % of conversion after 20-25 min (Fig. 5-1).
57

Table 5-1. Ee and reaction rate induced by cinchonidine modified catalysts before and
after activation.
Catalyst A Ee, % Reaction rate, mmol
(min·g) -1
A1- Before activation 79-80 3
A2 -After activation 85 28
100
100
Conversion, %
80
60
40
20
80
60
40
20
Enantiomeric excess, %
0
0 5 10 15 20
Time, min
0
Fig. 5-1. Evolution of conversion (squares) and enantiomeric excess (triangles) for A2
catalyst under 10 bar of hydrogen pressure.
A similar effect of increased ee (10-20%) after catalyst exposure to reaction condition
was found by Bartok and Balàzsik [194] for conventional Pt/Al 2 O 3 (E4759) in toluene
but was not found in acetic acid, probably, due to the differences in the catalysts. The
proposed origin of ee and rate enhance for our catalyst is discussed below.
5.3.2 Comparison of catalysts and proposed model of
catalyst activation
In order to compare cinchonidine modified catalysts A1 and A2 with modified and
unmodified conventional catalyst (Aldrich) (the latter is modified with cinchonidine in
situ) we tested them in ethyl pyruvate hydrogenation in the range of 2.5-10 bar of
hydrogen pressure.
The major difference in selectivity between cinchonidine modified conventional
Pt/Al 2 O 3 (see modification procedure in experimental part) and A1 catalyst was
observed in the ethyl pyruvate hydrogenation without addition of free cinchonidine. In
fact, the A1 catalyst induces 71 % of ee, whereas cinchonidine modified conventional
Pt/Al 2 O 3 demonstrates only 40 % ee.
It should be noted that without free cinchonidine, the A1 catalyst shows 71 % of ee and
ee decreases with reaction time (Fig. 5-2), whereas in the experiment with the addition
58

of free cinchonidine, the catalyst demonstrates 80 % of ee and no decrease of ee was
found during reaction. Another important fact is that just 10 % of conversion after 80
min (Fig. 5-2) was reached in the ethyl pyruvate hydrogenation without free
cinchonidine, whereas conversion of 70 % after 80 min (not shown) was reached with
free cinchonidine with the same catalyst and under the same conditions. This
phenomena can be explained as sorption/desorption of cinchonidine between Pt and
solution. Desorption of cinchonidine from Pt surface has been already observed even
under 1 bar of hydrogen and explained as partial hydrogenation of quinoline anchor
[169].
10
72
Conversion, %
8
6
4
70
Enantiomeric excess, %
2
68
0
0 20 40 60 80
Time, min
Fig. 5-2. Evolution of conversion (squares) and enantiomeric excess (triangles) for A1
catalyst without free cinchonidine, hydrogen pressure 10 bar.
Based on IR investigations, we would like to suggest here that A1 (and A2) possess a
higher concentration of adsorbed cinchonidine with respect to modified conventional
Pt/Al 2 O 3 . Despite the fact that both catalysts have Pt particles of similar size (3 (TEM)
and 3.9 nm (chemisorption)) the A1 catalyst (made by reduction of the Pt salt) is more
active than conventional Pt/Al 2 O 3 , most probably due to the fact the latter has a lower
level of cleanliness of the Pt surface (even after treatment with H 2 at 400 °C) with
respect to freshly reduced and modified Pt.
For ethyl pyruvate hydrogenation with the addition of free cinchonidine (Fig. 5-3) it
was found that in the range of 2.5- 10 bar of hydrogen pressure, A2 demonstrates
approximately 7-8 % higher enantiomeric excess and a higher reaction rate by a factor
of 1.7-2.3 than the conventional (non modified in prior) Pt/Al 2 O 3 at the same
conditions.
59

Enantiomeric excess, %
85
80
75
A2
Conventional Pt/Al 2
O 3
40
30
20
Reaction rate, mmol/min*g cat
10
70
2 4 6 8 10
Hydrogen Pressure, bar
Fig. 5-3. Influence of hydrogen pressure on enantiomeric excess (filled symbols) and
reaction rate (open symbols) for conventional Pt/Al 2 O 3 (triangles) and A2 catalysts
(squares).
Our proposal that this is probably due to higher surface concentration of active chiral
sites for the A2 catalyst is based on the fact that the A2 catalyst has already adsorbed
cinchonidine (in flat and in tilted adsorption mode, see section 5.3.4 for IR
investigation) and adsorbs additional cinchonidine (in both modes) from solution and at
the same time the conventional Pt/Al 2 O 3 (which is previously unmodified) adsorbs
cinchonidine only from solution (also in both modes) during reaction (in situ).
An increase of ee with increased hydrogen pressure was found for both A2 and the
conventional catalyst. This increase could be explained through the hydrogen cleaning
effect. In fact, it was shown that hydrogen cleans the Pt surface of different impurities
from solvent, ethyl pyruvate, and air (O 2 , CO/CO 2 ) [166, 195]. It is known, that the
likelihood of the occurrence for flat adsorption mode of the cinchonidine (active chiral
site) depends on the concentration of modifier in the solution [76, 187], however it is
clear, that cinchonidine adsorbed in flat mode requires more free space on the Pt
surface with respect to any tilted adsorption mode. Therefore, we suppose that with an
increase of concentration of free sites (due the hydrogen cleaning effect) the chance of
flat adsorption increases.
A1 after activation (see experimental part) is referred to as A2 was found to have a
lower concentration of CO contamination on the Pt surface (see IR investigation section
5.3.4) with respect to its state before activation, and demonstrates a plateau in the plot
of ee vs P H2 (Fig. 5-3). The presence of the plateau for the A2 catalyst and its absence
for conventional catalysts indicates differences in the processes associated with each
system.
It should be noted that, since all catalysts were kept in air and thermal treatment at 400
°C is not possible for those catalysts modified during preparation, these catalysts as
well as the conventional Pt/Al 2 O 3 catalyst were used without treatment at high
60

temperature in all of the above mentioned experiments. However, thermally treated
(with hydrogen according to procedure described above) conventional Pt/Al 2 O 3
demonstrates 91 % ee (10 bar H 2 ) which is 3 % higher than maximum obtained with A2
catalyst under the same reaction conditions. This significant (~14 %) increase in ee
after catalyst treatment clearly demonstrates importance of surface cleaning.
It is known, that platinum can be oxidized by O 2 and form PtO or PtO x [196] under red
heat and Pt is able to dissociatively chemisorb oxygen at room temperature, other
proposed models of oxygen adsorption on Pt can be found in the works of Smith [197]
and Matsushima [198] and others [199-201]. In our experiments with and without
thermal pretreatment (catalyst was transferred to reactor in air at room temperature) of
the Pt catalyst, chemisorption of oxygen on Pt likely occurs. However, it does not
prevent cinchonidine and ethyl pyruvate adsorption with the presence of hydrogen in a
solvent (in situ) and min. 77 % of ee (conventional Pt/Al 2 O 3 ) was observed due to the
surface cleaning effect, induced by hydrogen [76, 82, 195, 202]. The significant
increase of ee (up to 91 %, in our work) with thermally treated Pt/Al 2 O 3 has been
known since report of Orito et al. [45] and was explained as further surface cleaning
[203] e.g. from organic species, H 2 O, CO, CO 2 , etc.
The global role of oxygen in the reaction is rather ambiguous [69], for example, it was
shown that oxygen induces a cleaning effect through CO removal [195, 204] and
oxygen was to reduce cinchonidine coverage on Pt in such a way that a larger fraction
of the Pt surface would be available for ethyl pyruvate hydrogenation [205].
5.3.3 Catalyst stability and reuse tests
The TEM investigation of sample A1 before the stability test (Fig. 5-4) shows that
alumina globules are fully covered with Pt nanoclusters with an average size about 2.7
nm. The small increase of average particles size (maximum of 0.5 nm) was observed
from the sample following the stability test; in fact the average size of Pt nanoclusters
after the stability test was found to be about 3.2 nm. It should be noted that the surface
of an alumina globule after the stability test is still fully covered with Pt nanoclusters
and ultrasonic treatment (at least for half an hour) does not influence the size and
stability of alumina supported Pt nanoclusters.
61

35
30
Before
After
Percent of particles, %
25
20
15
10
5
0
1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
Size, nm
Fig. 5-4. TEM images of fully covered alumina globule before (above) and after
(below) stability test and a histogram representing the Pt particle size distribution
before and after catalyst stability test.
Since, there is no significant change in the size of Pt nanoclusters after the reaction time
was exceeded; the possibility of recycling was explored. The A2 catalyst (see Table 5-
2) was found to be fully reusable (at least four times) without loss of activity and
enantioselectivity. Moreover, a slight increase of ee was observed, which could be
explained as the result of additional activation after first reaction.
Table 5-2: A2 catalyst reuse data under 10 bar of hydrogen pressure.
Number of use Ee, % Reaction rate, mmol
(min·g) -1
1 85 28
2 87 25
3 86 29
4 88 30
It is known from the work of Reschetilowski and co-workers [206] (home-made
Pt/zeolite) and work of Bartok and co-workers [194] (conventional Pt/Al 2 O 3 ), that ee
could be constant (Reschetilowski) or decrease of 5 % in catalyst reuse experiments,
where “fresh” (dihydro)cinchonidine was added in every reuse. However, a significant
decrease of reaction rate was detected in both works, which might indicate a decrease
of catalyst activity, whereas the A2 catalyst does not demonstrate a decrease of ee or
62

eaction rate, moreover, the ee and reaction rate were slightly increased, probably due
to the further activation.
5.3.4 IR investigation and proposed model of activation
As we have shown already (chapter 4) cinchonidine demonstrates similarity in
adsorption on Pt nanoclusters (1.2 nm size) with respect to the macroscopic Pt support.
In fact, cinchonidine was found in both the tilted and flat adsorption modes in both
cases. By using previously obtained data from modeling of IR vibrations of the
cinchonidine molecule in open 3 conformation (chapter 4) and data from the work of
Ferri [76] and Zaera [77] the assignment of the observed peaks has been accomplished.
Fig. 5-5 shows the presence of peaks and shoulders that are in correlation with the
spectrum of free cinchonidine. Here, modification of both: conventional catalysts and
A1 prepared as described above is clearly established. It should be noted that, in the
case of A1, peaks corresponding to the adsorbed cinchonidine are better resolved and
have higher intensity with respect to cinchonidine modified conventional Pt/Al 2 O 3 due
to the higher concentration of adsorbed cinchonidine on the Pt surface.
1640
0.5
D
F
A - 71 % ee
1500 1000
Adsorbance
0.25
B - 40 % ee
C
1700 1600 1500 1400
Wavenumber,
1300
cm
1200 1100 1000
Fig. 5-5. DRIFTS of cinchonidine modified A1 (A), conventional Pt/Al 2 O 3 (B) and free
cinchonidine (C). The unmodified Pt/Al 2 O 3 (Aldrich) and alumina presented in inset as
F and D, correspondingly. Numbers near symbols A and B show maximum obtained
enantiomeric excess with respective catalyst without addition of free cinchonidine
under 10 bar of H 2 . The intensity of spectra of free cinchonidine was decreased for
better comparison.
A DRIFTS investigation of cinchonidine modified Pt nanoclusters on alumina support
was performed before (A1) and after catalyst activation (A2). The practical absence of
the peaks at ~2100 cm -1 (Fig. 5-6) which corresponds to adsorbed CO [202, 207] after
63

catalyst activation seems to indicate a Pt surface cleaning effect. This cleaning effect
was also observed by Baiker and co-authors for similar systems [166, 202].
Absorbance
0.004
A1
A2
2300 2200 2100 2000 1900
Wavenumber, cm -1
Fig.5-6. DRIFT spectra of CO adsorbed on A1 and A2 catalysts.
It should be noted, that the spectral range below 1200 cm -1 is not accessible for
investigation, due to the alumina background (Fig. 5-5, 5-7) which covers the frequency
range below 1200 cm -1 where the most useful (for unambiguous assignment of the flat
adsorbed species of the cinchonidine) out of plane vibrations of quinoline ring are
located. Further, the presence of alumina “dilutes” the cinchonidine in the IR beam and
lowers peak resolution (especially around 1580 cm -1 ) with respect to cinchonidine
stabilized Pt nanoclusters (support free) (please see chapter 4). Both of these effects
hinder unambiguous interpretation of the change in spectra. However, some significant
changes are clearly observed, they are summarized in the Table 5-4 and we provide the
following interpretation.
64

A2 -88 % ee
Adsorbance
1709
1613
1583
1572
1470
1463
1422
1422
1371
1340 1336
1280
A1 -80 % ee
C
0.5
1700 1600 1500 1400 1300 1200 1100 1000
Wavenumber, cm -1
Fig. 5-7. DRIFT spectra of A1 and A2 catalysts and free cinchonidine (C). Numbers
near symbols A and B show enantiomeric excess obtained with respective catalyst with
addition of free cinchonidine (1 mol L -1 ) under 10 bar of H 2 . The intensity of spectra of
free cinchonidine was decreased for better performance.
The appearance of the shoulder at 1613 cm -1 after activation could be explained as an
increase of concentration of cinchonidine in the “N lone pair bonded” adsorption mode.
The peak at 1572 cm -1 before and 1583 cm -1 after activation are super positions of
peaks/shoulders (which could be seen separately in DRIFT spectra of unsupported
cinchonidine on Pt nanocluster samples (chapter 4)) corresponding to 1619, 1591 and
1568 cm -1 peaks in the spectra of free cinchonidine. Such a shift as that from 1572
before to 1583 cm -1 after activation can be explained as partial interconversion of one
or more adsorption mode (flat 1, tilted 2, tilted 3) adsorption mode to tilted 3. Also, an
increase of the relative intensity and blue shift in wavenumber from 1463 (wk) to 1470
(med) cm -1 allows interpretation as an interconversion of adsorption modes (flat 1,
tilted 2, tilted 3) between each other and taking into the account the assumption that a
larger shift in wavelength results from stronger adsorption, thus it is possible to propose
that the concentration of tilted 2 mode decreased with respect to flat 1 or/and tilted 3
modes. No significant change in intensity or in wavenumber was found for the peak at
1422 cm -1 , corresponding to the flat 1 adsorption mode, whereas an increase of
intensity at 1371 cm -1 led to the appearance of the shoulder which corresponds to the
flat 1 adsorption mode. The decrease of intensity at 1340 cm -1 before activation and red
shift to 1336 cm -1 after activation is difficult to unambiguously explain since there is
not a clear assignment of this vibration band with an adsorption mode. The situation is
similar with the peak at 1264 cm -1 before activation and the weak shoulder at 1264 cm -1
after activation. The latter is almost covered by the broad peak at 1275 cm -1 which can
be assigned to acetic acid.
65

Table 5-3. Assignment of some vibrations (cm-1) of cinchonidine adsorbed on Pt
before (A1) and after activation (A2).
A1 A2 Free Assignment and orientation of
cinchonidine. adsorbed cinchonidine
- 1709 (st. wd.) - Adsorbed acetic acid
- 1613 (sh) 1619 (wk) Tilted 3
1572 (st) 1583 (st) 1591 (st), Tilted 3
1568 (med) Flat 1, tilted 2 and tilted 3
1463 (wk) 1470 (med) 1452 (str) Flat 1, tilted 2 and tilted 3
1422 (med) 1422 (med) 1422 (med) Flat 1
- 1371 (sh) 1355 (med) Flat 1
1340 (med) 1336 (sh) 1327 (med), -
1337 (med) Flat 1
- 1275 (wd) - Adsorbed acetic acid
1264 (med) 1264 (wk, sh.) - -
Comments: st. – strong, wd. – wide, sh. – shoulder, wk. – weak, med. – medium, flat: -1
“π-bonded” adsorption mode, tilted: 2 - “α-H abstracted” adsorption mode, tilted: 3 -
“N lone pair bonded” adsorption.
In order to assess the ability of the cinchonidine to adsorb on used alumina (pure
alumina passed all preparation and activation steps) the control experiments were
performed and it was found from DRIFT spectra that no cinchonidine peaks were found
on alumina after washing with acetic acid. Further, no conversion of ethyl pyruvate was
observed by using cinchonidine and alumina without Pt.
Based on the data described above we would like to propose a model for catalyst
activation. Thus, activation results from a decrease of CO coverage of Pt surface and an
increase of the concentration of flatly bound cinchonidine (active chiral site), probably
through reorganization between adsorbed cinchonidine species (partial desorption of
weakly bound cinchonidine in tilted mode to solution) and additional adsorption of
cinchonidine from solution onto the produced free sites as well of adsorption of acetic
acid, which could be competitive.
5.4 Summary
High enantiomeric excess in ethyl pyruvate hydrogenation was obtained at low – 2.5-10
bar hydrogen pressure with conventional Pt/Al 2 O 3 - 91 % ee (after thermal pretreatment
with hydrogen) and 80-88 % ee (without thermal pretreatment) with cinchonidine
modified Pt nanoclusters deposited on nonporous alumina (after activation with
hydrogen and cinchonidine in acetic acid at room temperature) without requirement of
avoiding catalyst contact with air. The stability of the catalyst under reaction conditions
and recycling and reuse of the catalyst was successfully demonstrated.
The positive effect of activation under reaction conditions was demonstrated and
explained as removal of contaminants from the Pt surface and reorganization of the
adsorbed cinchonidine species and adsorption of a new portion of ligand from solution.
These effects led to an increase of the concentration of active chiral sites on Pt surface
and thus obtained ee and rate enhancement.
66

Chapter 6
Cinchonidine modified Pd colloidal
nanoparticles
6.1 Introduction
It is known, that Pd is a suitable catalyst for hydrogenation of double bonds.
Cinchonidine modified Pd system was found to be able to induce enantioselectivity in
hydrogenation of substituted alkenes. In 1985 Perez and co-workers reported the
hydrogenation of (E)-α-phenyl-cinnamic acid to 2,3-diphenylpropionic acid over
cinchonidine-modified Pd/C; the enantiomeric excess was 30 % in favour of the S-
product [208]. The role of a support for Pd, metal loading, dielectric constant of the
solvent, concentration of cinchonidine and substrate, hydrogen pressure and
temperature has been studied by Nitta and co-workers [98, 209-213]. Enantioselective
hydrogenation of the carbon-nitrogen double bond on an example of Ph(C=NOH)Me
(ee 18 % ) is known from reports of Blaser [214] and Nakamura [215]. More examples
for enantioselective hydrogenations of double carbon and carbon-nitrogen bond can be
found from the review of Wells [87].
In this chapter we focus on hydrogenation activated ketones over cinchonidine
modified Pd catalysts. Hydrogenation of ethyl (or methyl) pyruvate on cinchonidine
modified Pd catalyst leads to the interesting phenomena, so called inversed
enantioselectivity. In fact, in 1988 Blaser and co-workers reported that cinchonidinemodified
Pd/carbon was mildly enantioselective for the hydrogenation of ethyl
pyruvate, the reaction giving an enantiomeric excess of 4% in favour of the S-
enantiomer [94], where as hydrogenation over cinchonidine modified Pt catalyst results
in enantiomeric excess in favour of R-ethyl lactate (see chapters 4 and 5). Closer
examination showed that reaction over conventionally supported Pd differed very
significantly from the corresponding Pt system [49, 216]. In fact, rate enhancement was
absent over cinchonidine modified Pd, enantioselective reaction was half order in
hydrogen whereas it was first-order for Pt, reactions over Pd were only achieved in
certain solvents and Pd was enantioselective only if reduced at low temperature.
It is known, that in the case of Pt catalyst the use of cinchonidine modifier results in
excess of R enantiomer and cinchonine gives S lactate. However very interesting
phenomenon takes place with use of Pd based catalyst, in fact, the use of cinchonidine
over Pd catalysts leads to positive (excess of R enantiomers) or negative (excess of S
enantiomers) ee, at the same time cinchonine over Pd catalysts is able to induce
negative and positive ee [49]. Wells and Whyman [217] first reported this phenomenon,
they found, that the direction of the enantioselectivity being solvent and/or substrate
(methyl and ethyl pyruvate) dependent.
In this chapter we define the following aims: to investigate orientation of the
cinchonidine molecule being adsorbed on Pd nanoclusters (~1.4 nm, see below) via
DRIFTS similarly to the Pt nanoclusters (please see the chapter 4) and to show
differences in the direction of enantioselectivity between modified with cinchonidine in
situ conventional Pd/Al 2 O 3 catalyst and cinchonidine modified Pd nanoclusters in ethyl
pyruvate and ethyl benzoyformate hydrogenation reactions. The choice of ethyl
benzoyformate was caused because it can not, in principal, form an enolate tautomer
state (please find details below).
67

6.2 Experimental
Materials
Cinchonidine (>98% Fluka), 5 % Pd/Al 2 O 3 (Fluka), Pd(DBA) 2 (Aldrich), THF (> 99.5
% Fluka), cyclohexene (> 99% Fluka) and ethyl benzoylformate (95 % Aldrich) were
used as received.
Samples preparation
Cinchonidine modified Pd nanoclusters were prepared according to the following
procedure. 1084 mg Pd(DBA) 2 (1.88 mmol Pd) and 1658 mg cinchonidine (5.64 mm)
in 70 ml THF were transferred into the metal reactor. After purging with Ar for 10 min,
3 bar of hydrogen pressure was applied under constant stirring. After 3 hours pressure
was released and the mixture was transferred into 500 ml cyclohexene and stirred for
24 h, then stirring was stopped and after 10 hours a precipitate was collected and
washed with cyclohexene (200 ml), then dried in air at room temperature. The washing
procedure was repeated one more time using 50 ml THF only, for complete removal of
side products. Obtained black precipitate was dried in air and transferred into 40 ml
AcOH (sample was completely dissolved). After stirring (30 min), the mixture was
transferred (Bönnemann’s procedure) into the 500 ml saturated NaHCO 3 water
solution, after 2 h filtrated with G4 glass filter, then washed with 200 ml half saturated
water solution of NaHCO 3 , then with water. Then it was picked up by AcOH (10-15
ml) (with few drops of water) and left for drying on the glass plate in air at room
temperature. Finally a black powder was scratched from the glass.
The thermal activation of conventional Pd/Al 2 O 3 was performed at 400 °C under
vacuum (10 -3 mbar) for 1 hour, then kept for 3 hours under a flowing mixture of N 2
(93%) and H 2 (7%) gases (100 ml min -1 ) and subsequently cooled to the room
temperature. The required amount was then weighed (in open atmosphere) and
immediately transferred into a glass reactor with acetic acid, cinchonidine and ethyl
pyruvate.
Characterization
Molecular modelling, GC, XRD spectra measuring and DRIFTS investigations have
been done as it is described e.g. in the chapter 4 and 5.
Hydrogenation of ethyl pyruvate (EP) and ethyl benzoylformate (EB) was conducted in
the metal reactor under 10 bars of hydrogen. Typically, to 70 ml AcOH with 36 mmol
substrate and (unless otherwise mentioned) 20 mg cinchonidine (1 mM) and 20 mg
catalyst was added. In the case of colloidal catalyst the mixture was sonicated for 10
min. Finally it was transferred into the metal reactor, purged with Ar for 10 min and
hydrogen pressure (10 bar) has been set, the stirring frequency was 1000 rpm. In the
case of the conventional Pt/Al 2 O 3 (was used for comparison) and Pd/Al 2 O 3 the catalyst
was thermally activated under hydrogen (as described in details in the e.g. chapter 5)
before use.
68

O
OH
OH
O
O
O
Catalyst, AcOH
O
P H2 = 10 bar
O
O
O
OH
OH
O
O
O
Ph
Ph
Ph
O
O
O
Fig. 6-1. Scheme of ethyl pyruvate and ethyl benzoylformate hydrogenation.
6.3 Results and discussion
XRD spectrum of cinchonidine modified Pd nanoclusters (Fig. 6-2) shows strong peaks
at ~40° which corresponds to the (111) crystal face and small peak at 67° (220) face,
according to work of Haglund [218].
400
300
200
100
0
-100
20 30 40 50 60 70 80
2Θ
Fig. 6-2. XRD spectrum of cinchonidine modified Pd nanoclusters.
The average crystal size has been estimated to be 1.6 nm based on the Scherrer’s model
[177] from the width on a half height of the strongest peaks (2Θ=40°). The TEM image
(Fig. 6-3) shows particles with 1.4 and 2.9 nm size. It is important to note, that at close
examination it is seen that “big” particles consist of at least two “small” particles. The
average size of “small” particles is very similar to the size obtained from XRD analysis
and in further we consider 1.4 nm as the average size of basic Pd nanoclusters.
69

Fig. 6-3. TEM image of cinchonidine modified Pd nanoclusters.
Determination of adsorption geometry of cinchonidine on Pd nanoclusters via DRIFTS
It is convenient to perform FTIR investigations of cinchonidine modified Pd
nanoclusters through direct comparison with the spectrum of cinchonidine modified Pt
nanocluster (considered in the chapter 4). Fig. 6-4 shows that DRIFT spectra of
cinchonidine modified Pt and Pd nanoclusters are very similar, however some
distinctions are present and they are discussed below. The most important peaks and
shoulders (from both DRIFT spectra) for determination the adsorption modes of
cinchonidine are summarised in the Table 6-1.
Table 6-1. Assignment of some frequencies (cm -1 ) from DRIFT spectra of cinchonidine
modified Pt and Pd nanoclusters.
Cinchonidine on Pd Cinchonidine on Pt Free cinchonidine Direction of dipole
moment A .
n.o. 1637 (sh) 1636 (med) (0, 0, 1)
1613 (sh) 1613 (sh) 1619 (wk) (-1, 1, 0)
1586 (med) 1594 (med) 1590 (st) (1, 1, 0)
1570 (sh) 1576 (sh) 1567 (med) (1, -1, 0)
1515 (st) 1514 (st) 1507 (st) (-1, 1, 0)
1462 (med) 1463 (st) 1452 (st) (-1, 1, 0)
811 (wk) 812 (med) 806 (med) (1, -1, -1)
768 (med) 766 (str) 758 (st) (0, 0, 1)
687 (st) 697 (med) 665 (wk) (-1, 1, 0)
652 (sh) 655 (st) 648 (wk) or
638 (med)
(1, 1, -1) or
(1, 0, -1)
Comments: A – Orientations of dipole moments were characterized as directions of the
unit vector (X, Y, Z) in the coordinate system having X and Y axes along the long and
short axis of the quinoline and Z axis along the normal of the quinoline plane towards
to quinuclidine.
70

The presence of in plane vibrations of cinchonidine on Pd nanoclusters e.g. at 1515 cm -
1 clearly indicates the tilted adsorption mode. The out of plane vibrations of quinoline
moiety e.g. 811 cm -1 were also observed on Pd that confirms the flat adsorption mode.
The proposed adsorption modes (schematically shown in Fig. 4-10 and 4-11) of the
cinchonidine on Pd nanoclusters (1.4 nm) were found to be very similar to the
adsorption modes determined from the work of Ferri, Burgi and Baiker [82] over
macroscopic film of Pd on Al 2 O 3 .
1609
1640
3
2
1
1600 1400 1200 1000 800
Wavenumber, cm -1
Fig. 6-4. DRIFT spectra of free (1) and adsorbed on Pd (2) and Pt (3) nanoclusters
cinchonidine. Intensity of spectrum of free cinchonidine was decreased for better
performance.
Comparison of adsorption modes of cinchonidine on Pt and Pd nanoclusters
The relative intensity of the peaks (811 cm -1 and 768 cm -1 ) corresponding to the out of
plane vibrations of quinoline ring is significantly lower with respect to the cinchonidine
on Pt system, probably due to the lower concentration of the π-bonded cinchonidine
with respect to the tilted mode. In fact, in the preparation of the samples the used ratio
cinchonidine/Pd was 3, whereas in the case of Pt based sample the ratio cinchonidine/Pt
was 2, at the same time it is known that with increase of concentration of cinchonidine
the chance of flat adsorption decreases, as was shown for Pt [76] and Pd [82]. It is also
can be expressed as a favour of tilted adsorption on Pd with respect to Pt, due to the
nature of the adsorption of cinchonidine on metal.
It is interesting to note the absence of the peaks (or shoulder) at 1640 cm -1 which
corresponds to the stretching of C=C bond, that might indicate to its hydrogenation.
However the parallel to the Pd surface orientation of this bond due to the specific
adsorption mode of cinchonidine can not be fully excluded, especially at high
concentration of modifier, whose adsorption modes are less studied in high
concentration range.
Hydrogenation of α-ketones
71

From the results of EP and EB hydrogenations which are summarized in the Table 6-2,
it is seen that presence of the cinchonidine (entry 2) results in 14 % of conversion (in
contrast with 0 % without cinchonidine, entry 1) and in excess of R lactate (37 % ee). It
is clear that cinchonidine induces a reaction rate and positive ee (excess of R
enantiomer) similarly to the Pt catalyst, but reaction is significantly less intensive over
Pd/Al 2 O 3 , In fact only 14 % of conversion has been achieved after 19 hours (entries 1
and 2), whereas under the same conditions Pt/Al 2 O 3 results 91 % ee (R lactate) and
80% conversion after ~60 min (chapter 5). The different situation was found with
cinchonidine modified Pd nanoclusters (entry 3), this catalyst without presence of free
cinchonidine demonstrates 11 % enantiomeric excess in favour of S lactate and high
conversion after 60 min. Surprisingly the normalized on number of metal surface atoms
reaction rates were found to be almost the same as for conventional Pd/Al 2 O 3 , however
the direct comparison between colloidal and thermally activated catalyst is laboured
due to the presence of impurities blocking active sites.
Conv. B , %
EP ≈0 ≈0 0 D (75) -
EP ≈0.2 ≈0.1 1.3 D (90) -
EP ≈0 ≈0 0 D (60) -
Table 6-2. Results of ethyl pyruvate and ethyl benzoylformate hydrogenation over Pd
based catalysts.
Run Catalyst Substra Initial rate, Initial
Ee, %
(min) -1 (time, enantiomer)
te mmol·(min·g) -1 rate A after (dominated
min)
1
C
Pd/Al 2 O 3 EP ≈0 ≈0 ~0 (16 h) -
2 Pd/Al 2 O 3 EP 0.36 9.8 14 D (19 h) 37 (R)
3 Cinchonidine EP 13 9.6 59 E (60) 11 (S)
on Pd coll. C
4 Cinchonidine EP 9 6.7 71 E (120) 19 (S)
on Pd coll.
5 Diphos on
Pd coll. C,F
6 Diop on Pd
coll. C,F
7 Quiphos on
Pd coll. C,F
8 Cinchonidine EB 3.9 2.9 61 E (280) 0
on Pd coll.
9 Pt/Al 2 O 3 EB ≈0 ≈0 0 D (19 h) -
Comments: A – Values were estimated based on TEM, metal content of the samples
and chemisorption data, B - initial amount of racemic ethyl lactate (as impurity in
commercial ethyl pyruvate) was taken into account, C – no free cinchonidine was
added, D – according to the curve profile (conversion vs. time) reaction can not go
further, E - according to the curve profile (conversion vs. time) reaction can go further,
F – the sample was show for comparison, detailed information about it can be found in
the corresponding chapter.
Addition of free cinchonidine (entry 4, 1 mM) increases the enantiomeric excess to the
maximum reported value of 19 % with respect to the S enantiomer, but practically does
not influence on the reaction rate. The change of direction of enantioselectivity towards
to the excess of S enantiomer over cinchonidine modified Pd nanoclusters in particular
(entries 3 and 4) and over different types of Pd catalysts in general [217] can be
explained through formation of enol intermediate state in a rate-determining step (Fig.
72

6-5) [87], the latter was verified after reaction with deuterium yielded the product
exchanged at the methyl group near α-keto group, i.e. giving the product
CX 3 CX(OX)COOCH 3 (where X = H or D) [219]. It is assumed that due to the
intermolecular repulsion between the >C=C< and >C=O moiety of the adsorbed state 3
(Fig. 6-5) can be released by rotation around the C-C bond to give species 3´ which
yields S enantiomer upon hydrogenation. According to this mechanism, these is not any
supposed enhancement of the reaction rate (it is assumed that intermediate state 2
which is involved in the rate=determining step 2-3, can not participate in hydrogen
bonding making semi hydrogenated complex with cinchonidine). However Pd colloids
modified with other chiral and non chiral ligands (entries 5, 6 and 7) demonstrate very
low (practically zero) reaction rate. The similar effect of enhancement of reaction rate
was observed by Wells and co-workers [217] over Pd colloids but was not explained. In
the current thesis we also do not give exact explanation, because this problem requires
addition investigation that can not be covered by current PhD thesis having other aims,
we give only an idea of interpretation of the effect which can be found below.
O OEt
1
*
H 3 C O
*
+H -H
2
O
*
H 2 C
OEt
* O
+H slow
3
HO OEt
H 2 C * * O
+2H
fast
HO
H
H 3 C
OEt
O
*
(R)
fast
H 2 C OEt
+2H
H OEt
HO
(S)
3' *
HO O
*
fast
H 3 C O
Fig. 6-5. Mechanism for the enantioselective hydrogenation of the EP over Pd, adapted
from [87].
It is very interesting to note, that in our experiments (entries 2 and 4) under the same
reaction conditions enantiomeric excess in favour of R enantiomer (37 %) and in favour
of S enantiomer (19 %) was found for unmodified in advance (modified in situ)
Pd/Al 2 O 3 and already modified with cinchonidine Pd nanoclusters, respectively. The
similar inverse of ee with cinchonidine Pd system has been observed for different type
of Pd catalysts, solvents and substrates [217]. This clearly indicates on the fact that
there are competing mechanisms of hydrogenations e.g. the presence of both enols and
keto forms on the Pd surface. The keto- mechanism results in excess of R product (in
the similar way to the cinchonidine-Pt system, probably with rate enhancement) and
enol- mechanism gives excess of S enantiomer as it is shown in Fig. 6-5 without rate
73

enhancement. However the exact factor (nature of solvent, Pd catalyst (e.g. presence of
Pd δ+ ), substrate, hydrogen pressure or present of chiral or non chiral impurities) playing
a major role in domination of one mechanism with respect to another one is not known
still. As was mentioned above the origin of the rate enhancement is also unknown and it
can not be explained through domination of the keto – mechanism because it will give
an excess of R lactate. Our idea is based on another adsorption mode of cinchonidine
on Pd surface i.e. via its double carbon bond (Fig. 6-6-right).
Fig. 6-6. Proposed adsorption modes: π-bonded (left) and via C=C bond (right) of
cinchonidine on Pd surface results in different asymmetrical environments caused by H
and OH groups.
In fact, in this adsorption mode (opposite to adsorption via quinoline moiety)
cinchonidine might behave in the similar manner to cinchonine, i.e. having replaced
OH and H groups, thus induces excess of S enantiomer and rate enhancement (third
mechanism). The proposed adsorption mode probably is not stable on the metal surface
due the C=C bond hydrogenation. Probably because of the addition of free cinchonidine
(entry 4) ee has increased from 11 to 19 % in favour of S enantiomer.
Interestingly that hydrogenation of ethyl benzoylformate (where formation of enole
state is not possible in principal, thus enol- mechanism is forbidden) with the same
cinchonidine modified Pd nanoclusters results in zero enantiomeric excess. This means
that the reaction was performed racemically (on unmodified sites) or equal contribution
from keto- mechanism (yielding excess of R lactate) and the third mechanism (excess
of S lactate). At the same time, conventional activated Pt/Al 2 O 3 (entry 9) with presence
of free cinchonidine (1 mM) showed no conversion of EB in contrast with cinchonidine
modified Pd colloids under 10 bars of hydrogen.
The observations show that the Pd catalyzed enantioselective hydrogenation of
pyruvate esters is much more complicated system than the corresponding Pt catalyzed
reaction and co adsorption of modifier, solvent and substrate [217] and possible
impurities on Pd surface might be critical in determining the direction of
enantioselectivity. The Pd cinchona system requires more investigation for concluding
reaction mechanisms especially in the light of recent work of Garland et al. [220] where
74

inverse of ee in hydrogenation of EB and EP was observed even over Pt/Al 2 O 3 and Pt/C
catalysts.
6.4 Summary
The orientation of the cinchonidine on the surfaces of 1.4 nm Pd nanoclusters has been
investigated for the first time using a combination of DRIFTS and molecular modeling
methods. It was found that both ‘flat’ and ‘tilted’ modes of adsorbed cinchonidine are
present on Pd nanoclusters having reduced relative concentration of flat bonded
cinchonidine with respect to the cinchonidine modified Pt nanoclusters having the
similar size (1.4 nm).
The highest known enantiomeric excess of 19 % with respect to the S ethyl lactate was
reported over cinchonidine modified Pd nanoclusters whereas under the same reaction
conditions enantiomeric excess induced by conventional Pd/Al 2 O 3 modified with
cinchonidine in situ was found to be 37 % with excess of R ethyl lactate.
The inverse of enantioselectivity with respect to the Pt catalyzed hydrogenation was
explained with literature model through the enol- mechanism, whereas rate
enhancement effect and dependence of direction of enantiomeric excess on the nature
of Pd catalyst and reaction conditions require additional investigation.
Chapter 7
Quiphos modified Pt and Pd colloidal
nanoparticles
7.1 Introduction
The Quinoline diazaphospholidine ‘quiphos’ (Fig. 7-1) family of ligands represents a
possible alternative to the cinchonidine class of ligands for application in
enantioselective catalysis. Quiphos is a chiral molecule because of the chiral
phosphorus atom due to the inversion high barrier at room temperature. In this work we
used enantiopure quiphos that was confirmed from synthesis and
31 P-NMR
spectroscopy, for the further information about its synthesis we recommend to contact
to Dr. Didier Nuel of Université Aix-Marseille III, France.
These ligands have attracted attention as a result of their ease of synthesis and stability
to air and moisture as well as possessing the same quinoline anchor as the cinchonidine
systems and thus may be able to assume a flat “π-bonded” geometrical orientation
which is known to be necessary for inducing enantioselectivity in the cinchonidine
system.
N
N
N
P
O
Fig. 7-1. Quiphos molecule.
75

Further, this family of man-made ligands offers numerous possibilities for structural
modification which may enable ‘tuning’ of the resulting system and exploration of a
wide variety of derivitization. This family of ligands has also been successfully applied
to a variety of transition metal catalysts for homogeneous enantioselective synthesis [9,
11, 12].
7.2 Experimental
Materials
Quiphos ligand was provided by Prof. Gerard Buono and Dr. Didier Nuel of Université
Aix-Marseille III, France. Other chemicals were purchased and used as received.
Samples preparation
The modification of conventional Pt/Al 2 O 3 with quiphos was performed in the similar
way, how it was described for cinchonidine in the chapter 4. Typically, 40 mg Pt/Al 2 O 3
was heated at 200-220 ºC under vacuum (~10 -2 mm) for 2-3 hours, then the gases N 2
(93%) and H 2 (7%) were passed (50 ml/min) over the support for 2-3 hours without
contact with air at the same temperature. While still under flowing gases, the support
was cooled to room temperature and ligand was added as a solution in CHCl 3 (6.0 ml, 3
mM). Then the system was kept at 10-15 ºC for one day, washed 3 times with CHCl 3
and finally dried under vacuum at room temperature.
For the preparation of quiphos stabilized Pt nanoclusters 70 ml of chloroform was
flushed with nitrogen for 10 minutes, then 405 mg of Pt 2 (DBA) 3 (0.74 mmol of Pt) was
added together with 775 mg of quiphos (2.22 mmol). The solution was flushed with Ar
for 15 min under stirring. Then the contents were transferred into the glass reactor and
3 bar of hydrogen pressure was applied under constant stirring. After 18 hours a black
reaction mixture was precipitated by adding the reaction mixture to 230 ml of
cyclohexene. The precipitate was filtrated out using a G4 glass filter. The precipitate
was first washed with cyclohexene in chloroform (1/1 by volume) mixture then with
chloroform (300 ml) until the product was free from hydrogenated DBA and unreduced
Pt 2 (DBA) 3 as determined by IR spectroscopy. The sample is readily redispersible in a
solution of phenol/ethanol (1:1) with 10% water.
In order to prepare quiphos stabilized Pd nanoparticles 1084 mg Pd(DBA) 2 (1.9 mmol
Pd) and 659 mg quiphos (1.9 mmol) were dissolved in nitrogen flushed THF (80 ml).
Then the content was transferred into the metal reactor and 5 bar of hydrogen pressure
was applied for 4 hours under constant stirring. Then the hydrogen pressure was
released, reactor unplugged and the mixture was transferred into 300 ml cyclohexene.
After 24 hours of settling, black precipitate was filtrated by G4 filter and washed with
THF : cyclohexene (1:5) mixture until washing solvents mixture no longer became
yellow. Finally sample was washed with cyclohexene (200 ml) and black powder was
dried in air at room temperature for 12 hours.
Characterization
The samples were characterized by XRD, TEM, elemental analysis and DRIFTS in the
same ways as it is described e.g. in the chapter 4. However due to the limitations in
accessibility to TEM and elemental analysis apparatuses the corresponding
experiments: TEM characterization of quiphos modified Pd nanoclusters and elemental
analysis of this sample are missing.
Hydrogenation of ethyl pyruvate was carried out in a metal reactor (Parr Instrument
GmbH) under a pressure of 10 bar of hydrogen pressure at room temperature (22±1
°C). The stirring speed was 1000 rpm.
76

Typical conditions were such that 4 ml (36 mmol) ethyl pyruvate was dissolved in 70
ml THF solvent, when mentioned 24 mg (68 mmol) of free quiphos was added. 20 mg
of catalyst was added and mixture was sonicated for 10 min then transferred into the
reactor, flushed with Ar for 10 min and finally, after the pressure of hydrogen had been
stabilized for 10-15 sec, the initial time (t=0) was set and the reaction kinetics were
monitored. About one ml aliquots of the reaction mixture were taken at certain time
intervals, filtrated with a G4 glass filter and 1 μl was injected into the gas
chromatograph (GC). GC measurements were taken using a Varian 3900 instrument
with FID detector and Lipodex-E column.
Activation of conventional Pt/Al 2 O 3 has bee performed in the same was as is described
in the chapter 5.
7.3 Results and discussion
From the elemental analysis of quiphos modified Pt sample it has been found that the
sample has 4.5 % N and 28.3 % C or 37 % of quiphos as an average by carbon and
nitrogen and 56 % of metal (based on TGA), whereas metal content for quiphos
modified Pd sample was found to be 35 % (TGA).
XRD spectra of quiphos modified Pt and Pd nanoclusters clearly demonstrate Pt and Pd
crystalline nature of the samples. In fact, both spectra have a strong peak at 39°
corresponding (111) crystal face. In case of Pt small peaks at 46° and 67° correspond to
(200) and (220) faces, according to work of Swanson et al. [221]. The asymmetrical
peak at ~35° is difficult to assign unambiguously, since it appears not only in the
spectra of this sample but also in case of diphos on Pd nanoclusters (see chapter 9) we
believe that it is not due to the instrument fault. Here we would like to assign it to the
(222) Pt crystal face base on work of Hull [222].
350
300
250
1
200
150
100
50
2
0
-50
20 30 40 50 60 70 80
2Θ
Fig. 7-1. XRD of quiphos modified Pt (1) and Pd (2) nanoclusters.
XRD spectrum of quiphos modified Pd nanoclusters besides of mentioned already peak
at 39° ((111) face) shows wide peaks at 22° (200) [222] and 71° (620) [176].
77

Fig. 7-2. TEM of quiphos modified Pt nanoclusters.
TEM image of quiphos modified Pt nanoparticles (Fig. 7-2) shows that the average
particle’s size is 15 ± 0.5 nm nm, that differs (Table 7-1) from average size estimated
from the Scherrer equation (4.3 nm) and from the average size of quiphos modified Pd
nanoparticles (1.2 nm), probably due to the partial agglomeration of Pt nanoparticles.
Table 7-1. Comparison of the sizes of quiphos modified Pt and Pd particles found via
TEM and XRD.
Sample , nm (TEM) , nm (XRD)
Quiphos on Pt 15 4.3
Quiphos on Pd - 1.2
FTIR investigation of quiphos adsorbed on Pt and Pd nanoparticles
Two conformations of the quiphos molecule were found through molecular modelling
(Fig. 7-3). The energy difference between the more stable conformation (Fig 7-3-1) and
the second (Fig. 7-3-2) was found to be 11.3 kcal/mol in air and 5.2 kcal/mol in THF.
The geometry of the first conformation is in agreement with single-crystal XRD data
[11] and was chosen for analysis.
Fig. 7-3. Skeleton geometry of quiphos molecule in two conformations, hydrogen
atoms are omitted. Conformation “1” is energetically more preferred, ΔE 1,2 = -11.3
kcal/mol (in air).
78

The DRIFT spectra of the quiphos adsorbed on conventional Pt/Al 2 O 3 (Fig. 7-4)
demonstrates shifted peaks and shoulders whose analogs can be found in the spectra of
free quiphos. However due to the low quality (not all important peaks can be found) of
the spectra of this sample, probably because of low degree of surface coverage, this
spectra was not taken in to account for further interpretation in favor of spectra of
quiphos modified nanoparticles (please see below).
Relative units
1
2
1600 1400 1200 1000
Wavenumber, cm -1
Fig. 7-4. DRIFT spectra of quiphos adsorbed on conventional Pt/Al 2 O 3 (1) and free
quiphos (2).
Adsorbed quiphos on Pt and Pd produces IR spectra (Fig. 7-5) for which the peaks are
summarized in Table 7-2. Here, only peaks which can be unambiguously assigned to
free quiphos are presented.
In the same way as it is done in case of the cinchonidine, the DFT geometry
optimization and calculation of the IR spectrum of quiphos have been done to obtain
the orientation of the corresponding dipole moment. To facilitate spectra assignment,
the experimental and calculated IR spectra of quiphos (correlation coefficient is 1.03)
quinoline and aniline were compared with each other; see Fig. 7-6 and Table 7-2.
Aniline was chosen since quiphos also contains a C 6 H 5 N unit. The IR spectrum of
aniline was calculated with correlation coefficient 1.09.
All observed peaks corresponding to IR vibrations of quiphos adsorbed on Pt
demonstrate two orientations of the dipole moment: mostly perpendicular to the
quinoline ring and mostly perpendicularly to the phenyl ring. This leads us to propose
two orientations of adsorbed quiphos on Pt. In the first mode, quiphos is adsorbed on Pt
via the quinoline anchor (Fig. 7-7), in the similar way as cinchonidine does, but,
possibly more tilted; in the second mode - via the phenyl group (Fig. 7-8) and possibly
via 3,4 N and/or P atoms. Some degree of tilt is also possible in both cases. In contrast to
the cinchonidine, quiphos can not be adsorbed in a tilted orientation mode, for example,
through the nitrogen in quinoline part without significant changes in geometry.
The DRIFT spectra of adsorbed quiphos on Pt and Pd nanoparticles do not have
principal differences. In fact, positions of some peaks (see Table 7-2) are slightly
shifted that can be interpreted in slight differences in force constants of chemical bonds
79

of quiphos adsorbed on Pt and Pd and in the same time in differences in adsorption
strength. It is also can be seen (Fig. 7-5 and Table 7-2) that some new peaks appeared
or became stronger for quiphos being adsorbed on Pd with respect to Pt. This
phenomenon can be explained as increase of quality of the spectra (for example due to
the differences in size of Pt and Pd nanoparticles, or other secondary effects) or as
differences in relative concentration of spices adsorbed via the quinoline and phenyl
ring. It is difficult to conclude unambiguously which explanation is more valid here
without addition investigation in which we decided to do not focus on here.
1
Relative units
2
3
1600 1400 1200 1000 800
Wavenumber, cm -1
Fig. 7-5. IR spectrum of quiphos adsorbed on Pd (1) and Pt (2) nanoparticles and free
quiphos (intensity is reduced for clarity) (3).
80

3
A
Relative units
2
B
1
C
0
D
1600 1400 1200 1000 800 600
Wavenumber cm -1
Fig. 7-6. Comparison of IR peaks between free cinchonidine (A), quinoline (B),
quiphos (C) and aniline (D).
Fig. 7-7. Flat adsorbed quiphos demonstrates vibration mode at 1604 cm -1 with dipole
moment (big arrow) orientated mostly perpendicularly to the metal surface. The most
intensive displacement of atoms is shown by small arrows.
81

Fig. 7-8 Flat adsorbed quiphos via phenyl ring demonstrates vibration mode at 696 cm -1
with dipole moment (big arrow) orientated mostly perpendicularly to the metal surface.
The most intensive displacement of atoms is shown by small arrows.
Table 7-2. Assignment of frequencies (cm -1 ) of IR spectra of adsorbed and free
quiphos, free quinoline and free aniline.
Quiphos
on Pt
Quiphos
on Pd
Free
quiphos
Quiphos
theoretical
Quinoline Aniline Description Orientation
of
adsorbed
- 1610 1652 1620 - C-Qu.
strch., H-
Qu. b.
1604
(st)
1580
(sh)
1558
(med)
1503
(st)
1601
(st)
1574
(med)
1593 1650 - 1600
6,7,9,10 C-A.,
3,5 C strch.,
ip. def.
n.o. 1639 1596 - C-Qu.
strch., H-
Qu. b.
- 1568 1598 1572 - C-Qu.
strch., H-
Qu. b.
1534
(med)
1500
(st)
1538 - Unharmonic
vibr.
1498 1542 1501 1495 C-Qu.
strch., H-
Qu. b.
quiphos
-
Quinoline
flat
Phenyl flat
Phenyl flat
-
Phenyl flat
1468 1468 1464 1502 1471 - H-Qu. Phenyl flat
(med) (st)
asym. ip. b.
1422 1427 1422 1477 1432 1479 H-Qu. sym. Phenyl flat
(wk) (sh)
ip. b.
1390 1388 1383 1434 1394 -
37,42 H-Q. Phenyl flat
82

(wk) (med) sym. ip. b.,
C,
C 16 N-
Qu. strch.
1379
(sh)
1375
(med)
1369 1408 1371 - H,C-Qu.
comp.,
30,31,32 H b.
- 1324 1361 1314 1304 C-Qu.
strch., H-
Qu. ip. b.,
C-A. strch.
1309
(st)
1259
(sm)
1319
(st)
1258
(sm)
1311 1337 - 1273
1247 1292 - -
3 C 5 N strch.,
C,H-A.
comp.
H 11,13,14 2 C
b.
1 O 23 C
strch.,
40,41,42 H ip.
b., H 14 2 C b.
- 1225 1258 - - H 11,12,13,14 2 C
b.
- 1202 1240 - 1171
26,16 H-A,
28,29 H-A
sci.,
H 11,12,13,14 2 C
b.
- 1087 1113 1118 - H,C-Qu.
comp.,
11,12 C strch.,
- 1027 1027 - 992 H-Qu. oop.
wag.,
C 7 C,
8,10 C 9 C ip.
b.
- 988 1014 - - C-A. sym.
expan.,
H 11,12,13,14 2 C
b.
- 971 989 - -
- 954 951 - -
13,14 C,
12 C 4 N –
syn. strch.,
H 11,12,14 2 C
b.
13,14 C,
12 C 4 N –
asyn. strch.,
H 11,12,14 2 C
b.
- 901 926 - - H,C-A., -
Phenyl flat
-
Quinoline
flat
Phenyl flat
-
-
-
-
-
-
-
83

comp.,
H 11,12,13,14 2 C
b., 2 P 11 O 3 N,
N 14 C 13 C b.
- 887 912 - 871 C-Qu.
comp. def.,
H-A. oop.
wag.
- 864 884 - -
12,13 C 14 C
comp.,
11,12 C strch.,
3,4 NP asym.
strch., C-A.
comp.
833 823 829 849 805 - C,H-Qu.
(med) (med)
oop. wag.
-
-
Quinoline
flat
- 807 835 - - C-Qu. ip. b. -
- 803
(sm)
793 812 786 - C,H-Qu.
oop. wag.
Quinoline
flat
- 760 780 739 - C,H-Qu. -
oop. wag.
758 (st) 755 (st) 746 782 - 745 C,H-A. oop. Phenyl flat
wag.
696 (st) 697 (st) 690 712 - 688 C,H-A. oop. Phenyl flat
wag.
- 669 682 - -
7,9 C 8 C, -
6,10 C 5 C ip.
b. , 3 NP
strch.
Comments: The same as for the Table 4-3, plus the following n.o.- not observed, expan.
– expansion, asym. – asymmetrical, syn. – synchronously, asyn. – asynchronously, A-
aniline.
The results of ethyl pyruvate hydrogenation with quiphos modified Pt, Pd nanoparticles
and activated conventional Pt/Al 2 O 3 are summarized in the Table 7-3. As can be seen
from the Table 7-3 the addition of free quiphos stops the reaction completely. In other
words quiphos can be considered as a poison for this reaction with Pt or Pd based
catalysts. Since conversion was not observed at all, ee can not be defined.
Table 7-3. Results of ethyl pyruvate hydrogenation with quiphos modified Pt and Pd
nanoclusters.
Sample
name
Catalyst
loading, mg.
Reaction rate,
mmol·(min·g) -
Initial rate
(min) -1 Conversion A , [Free
% after time, quiphos],
1
min
mM
B
Pt/Al 2 O 3 20 0.6 16.5 12.5 (345) C 0
Pt/Al 2 O 3 20 ~0 ~0 ~0 (90) 1
Quiphos on
Pt
nanoparticles
20 ~0 ~0 ~0 (60) 0
84

Quiphos on 20 ~0 ~0 ~0 (90) 0
Pd
nanoparticles
Comments: A - initial amount of racemic ethyl lactate (as impurity in commercial ethyl
pyruvate) was taken into account. B- racemic hydrogenation without a modifier, C –
according to the curve development (conversion vs. time) reaction can go further.
7.4 Summary
It has been found that quiphos adsorbs on Pt nanoclusters in a similar mode as
cinchonidine through the quinoline moiety. An additional adsorption mode of the
quiphos is via the phenyl ring with possibly contribution from the 3,4 N and/or P atoms.
The presence of quiphos in the reaction mixture induces strong poison effect that stops
reaction completely, probably because of occupation of active surface sites by quiphos
molecule.
85

Chapter 8
“Quiphos-spider” modified Pt colloidal
nanoparticles
8.1 Introduction
The “quiphos-spider” (4-isopropyl-2 methyl cyclohexyl-1methyl-phosphonite -2-2'-1,5
binaphtyl) compound is the second of three ligands which is provided by our
collaborators (Marseille). The chirality of this ligand is conditioned by the fact that both
couple of aromatic rings do not lie in the one plane, making ~32° and the image mirror
can not be superimposed with the original molecule.
P O
Fig. 8-1. Quiphos-spider.
Here we investigate catalytic activity of Pt nanoparticles modified with quiphos-spider
ligand in the ethyl pyruvate hydrogenation and explain obtained results based on
proposed models of the molecule adsorption geometry.
8.2 Experimental
Materials
Quiphos-spider was provided by Prof. Gerard Buono and Dr. Didier Nuel of Université
Aix-Marseille III, France. Other chemicals were purchased and used as received.
Samples preparation
Preparation of quiphos-spider modified Pt nanoclusters had been done according to the
following procedure. 321 mg (0.73 mmol) quiphos-spider was dissolved in argon
purged 70 ml THF with 400 mg Pt 2 (DBA) 3 (0.37 mmol Pt) the mixture was transferred
in to the glass reactor and hydrogen pressure of 2 bar was applied under constant
stirring at room temperature for 19 hours. Then the mixture was transferred into 400 ml
cyclohexene, after 24 hours black-brown precipitate was collected and washed with
(THF/cyclohexene mixture 1/3) 200 ml and then with 200 ml cyclohexene, finally
sample was dried at 40°C over night.
Due to the limited amount of provided quiphos-spider ligand Pt modification of Pt
nanoclusters was done only.
Characterization
FTIR spectra were recorded on “Thermo Nicolet AVATAR – 370” instrument, with
DRIFTS accessories, measuring KBr background first.
TEM images were recorded according to the already described in the chapter 4
procedure.
86

Molecular modelling consisting of geometry optimization and IR spectra calculation
has been done with help of G03 program [175] for quiphos-spider and naphthalene.
Ethyl pyruvate hydrogenation (including activation of conventional Pt/Al 2 O 3 ) was
performed according to the procedure described in the chapter 6. Important parameters
are summarized in the Table 8-1.
8.3 Results and discussion
The average size of Pt nanoclusters was found to be 1.9±0.3 nm from TEM image (Fig.
8-2). Particles were found to be not uniformly distributed; however no significant
agglomeration was found.
Fig. 8-2. TEM image of quiphos-spider modified Pt nanoclusters.
As can be seen from comparison of DRIFT spectra of free and adsorbed on Pt
nanoclusters quiphos-spider ligand relative intensities and positions of some peaks have
changed. Since quiphos-spider molecule consists of two (not identical, due to their
orientation within skeleton geometry of the ligand) naphthalene rings, assignment of
the main adsorption peaks can be done through comparison IR spectra of naphthalene
and quiphos-spider. In fact in plane deformation of naphthalene ring(s) at 1595 cm -1 ,
1389 cm -1 , 1267 cm -1 having dipole moment orientated along short axis of naphthalene
and at 1509 cm -1 , 1209 cm -1 and 1013 cm -1 with dipole moment orientated along long
its axis, as well as out of plane vibrations at 956 cm -1 , 790 cm -1 and 617 cm -1 have
dipole moment along the normal vector of molecule’s plane. Since quiphos-spider
molecule have two naphthalene planes and reduced symmetry, the one peak in the
spectrum of naphthalene corresponds to two or more peaks (shoulders) in the spectrum
of quiphos-spider. Due to the fact that vibrations of free ligands having dipole moment
orientated along short naphthalene axis and along long naphthalene axis and along
normal vector of naphthalene planes are present as strong corresponding peaks in the
spectrum of adsorbed quiphos-spider, we would like to propose the following
adsorption models of the molecule (Fig. 8-5). In fact, both these models allow having
non zero projection of every mentioned orientation of dipole moment along normal
vector of the metal surface. It has to be mentioned that we do not exclude adsorption on
corners or edges of a nanocluster however we think that molecules adsorb mostly on a
87

flat crystal surface due to the higher number of atoms (making adsorption sites) on
planes of nanocluster than on corners.
Relative units
1
2
1600 1400 1200 1000 800 600
Wavenumber, cm -1
Fig. 8-3. Drift spectra of quiphos-spider on Pt nanoclusters (1) and (2) free quiphosspider.
1
2
3
1600 1400 1200 1000 800 600 400
Wavenumber, cm -1
Fig. 8-4. Comparison of experimentally obtained IR spectra of free quiphos-spider (1),
naphthalene (2) and calculated IR spectra (correlation coefficient 0.96) of free
naphthalene (3). The huge peak at ~800 cm -1 is not shown fully. The similar vibration
modes are marked.
88

Fig. 8-5. Proposed adsorption models of quiphos-spider on Pt nanoclusters.
The results of ethyl pyruvate hydrogenation with quiphos-spider modified Pt
nanoparticles and activated conventional (unmodified) Pt/Al 2 O 3 are summarized in the
Table 8-1. Quiphos-spider modified Pt nanoclusters sample demonstrates slow racemic
hydrogenation of ethyl pyruvate, but higher normalized on gram of a catalyst reaction
rate than unmodified activated with hydrogen (see experimental section) conventional
Pt/Al 2 O 3 , however it shows lower reaction rate normalized on number of surface metal
atoms, probably because some of them are “poisoned” by quiphos-spider ligand.
Zero enantiomeric excess was found probably due to the inability of lone pair of
phosphorus to “catch” the adsorbed ethyl pyruvate (as it is required for cinchonidineethyl
pyruvate system) because lone pair of phosphorus might be involved into the
adsorption binding.
Despite of the fact that the reaction rate normalized on gram of a catalyst was higher for
colloidal sample, reaction rate normalized on estimated mole of Pt surface atoms was
higher for conventional Pt/Al 2 O 3 , probably because the latter one was thermally
activated under hydrogen.
Initial rate A Conversion B , %
Table 8-1. Results of ethyl pyruvate hydrogenation with quiphos-spider modified Pt
nanoclusters.
Sample name Catalyst Reaction rate,
loading, mg. mmol·(min·g) -1 (min) -1 after time, min
C
Pt/Al 2 O 3 20 0.6 16.5 12.5 (345) D
Quiphos-spider 17 2 3.3 17 (160) D
on Pt
nanoparticles
Comments: A – values were estimated based on chemisorption data (Pt/Al 2 O 3 ), size of
Pt nanoclusters and metal content (30%), B - initial amount of racemic ethyl lactate (as
impurity in commercial ethyl pyruvate) was taken into account. C- racemic
hydrogenation without a modifier, D – according to the curve profile (conversion vs.
time) reaction can go further.
8.4 Summary
The adsorption models of quiphos spider on Pt nanocluster based on DRIFTS and
molecular modeling investigations has been proposed. According to these models
quiphos-spider adsorbs via its P atom and of one naphthalene ring. This system was not
found to be able to induce enantioselectivity in the hydrogenation of ethyl pyruvate
(racemic reaction was observed) probably due to the inability of adsorbed quiphosspider
to make semi-hydrogenated transition state by analogy with adsorbed
cinchonidine-hydrogen-ethyl pyruvate system.
89

Chapter 9
Diphos and diop modified Pt and Pd
colloidal nanoparticles
9.1 Introduction
In this chapter we used diphos and diop organophosphorus ligands as modifiers for Pt
and Pd nanoclusters. The chiral ligand diop attracts attention since it has been
successfully used in enantioselective homogeneous catalysis [24], whereas non chiral
diphos molecule has rather similar and simple structure that facilitates DRIFTS
investigation of ligands adsorbed on nanoclusters and relative comparisons e.g.
catalytic properties of diphos and diop modified Pt and Pd nanoclusters.
P
P
P
P
O
O
9.2 Experimental
Diphos
Diop
Fig. 9-1. Diphos and Diop molecules.
Materials
Ethylene bis(diphenylphosphine) (diphos) 96 % (Aldrich), (4S,5S)-4,5-
Bis(diphenylphosphinomethyl)-2,2-dimethyl-1,3-dioxolane ( or (S,S)-Diop) and
(4R,5R)-4,5-Bis(diphenylphosphinomethyl)-2,2-dimethyl-1,3-dioxolane ( or (R,R)-
Diop) 98 % (Aldrich) and Pd(DBA) 2 (Aldrich) were used as received, Pt 2 (DBA) 3 was
synthesize as described in the chapter 4.
Samples preparation
In order to prepare diphos on Pt nanoclusters, 560 mg Pt 2 (DBA) 3 (1 mmol Pt) in 40 ml
THF was added to 400 mg diphos (1 mmol) in 30 ml THF. The mixture was purged
with Ar for 10 min in metal reactor and hydrogen pressure of 4.5 bar was applied for
2.5 hours under constant stirring. After that time, hydrogen pressure was realized and
the mixture was transferred into 400 ml cyclohexene, after 30 hour supernatant was
decanted and precipitated black substance was transferred into 400 ml
THF:cyclohexene mixture (1:7), stirred for 10 min and after 4 hour precipitate was
collected and this procedure has been repeated until no trace of yellow DBA solution
left. Finally black powder was washed with 400 ml cyclohexene and dried in air at 65
°C.
The diphos on Pd nanoclusters sample was prepared according to the following
procedure. 1084 mg Pd(DBA) 2 (1.9 mmol Pd) in 40 ml THF was added to 752 mg (1.9
mmol) diphos in 20 ml THF. The mixture was purged with Ar for 10 min in the metal
reactor and then hydrogen (5 bar) was applied for 20 hours under constant stirring.
90

After that time the mixture was transferred to 450 ml cyclohexene and after 10 hour
black-brown precipitated was collected by G4 glass filter. 400 ml THF: cyclohexene
(1:1) mixture and then 200 ml cyclohexene were used for washing. Finally sample was
dried under vacuum at 40°C.
In order to prepare diop on Pd nanoclusters 400 mg (0.8 mmol) (S, S) diop in 40 ml
THF was mixed with 460 mg Pd(DBA) 2 (0.8 mmol Pd) in 40 ml THF in metal reactor.
The mixture was purged with Ar for 10 min and then 5 bar of hydrogen pressure was
applied for 6.5 hours. After that time, reaction mixture was transferred into 400 ml
cyclohexene, after 16 hours black precipitated was collected and washed with 200 ml
THF: cyclohexene mixture (1:4) and with 200 ml cyclohexene. Finally it was dried
under vacuum for during 10 hour at 40 °C.
(R, R) and (S, S) enantiomers of diop were used as modifiers for Pt nanocluster. 250
mg (0.5 mmol) e.g. (S, S) diop in 35 ml THF was transferred into the metal reactor with
274 mg Pt 2 (DBA) 3 in 35 ml THF. After 10 min for Ar purging hydrogen pressure of 4
bar has been set for 5.5 hours. After that time the mixture was transferred into 400 ml
cyclohexene, collected after 20 hours precipitate was washed with 400 ml THF:
cyclohexene mixture (1:4) and with 200 ml cyclohexene. Finally sample was dried at
40 °C under vacuum.
Characterization and experiments
Molecular modelling including geometry optimisation and IR spectra calculation has
been performed with Gaussian 03 software [175] for diop and diphos molecules by
using the DFT method (b3lyp) and (6-31G) basis set. Visualisation of the geometry and
IR vibrations was performed using Gaussview software. In order to facilitate assigment
of the peaks of the experimental IR spectrum for the free molecule, its IR spectrum was
compared (in terms of intensities of the signals and their vibrational frequencies) with
calculated IR spectra using MS Excel and Origin (intensities of both spectra were
normalized, frequencies of the calculated spectra were multiplied by a correlation
coefficient). The correlation cofficient dωtheory/dωexp of 1.02 was found to provide
the best fit.
XRD spectra were measured in the way described in the chapter 4.
Hydrogenation of ethyl pyruvate and DRIFTS investigations have been performed
according to the procedure described in details in the chapter 7 and 4 correspondingly.
Due to the absence of TEM instrument in the IUB and limited access to the instrument
outside the university, TEM investigation of diop modified Pt nanoclusters has been
done only.
9.3 Results and discussion
XRD analysis of diop and diphos modified Pt and Pd nanoclusters clearly demonstrate
Pt and Pd crystalline nature of the samples. In fact, all spectra have a strong peak at 39°
corresponding (111) crystal face (Fig. 9-2). The small peak at 46° (diop on Pt, Fig. 9-2,
curve 1) corresponds to (200) crystal face [221], where as in case of diphos on Pt
sample (same Fig. curve 4 this peak is not seen. The asymmetrical peak at ~35° (diphos
on Pd, curve 3) is difficult to assign unambiguously it probably might be assigned to
(311) Pd crystal face base on work of Hull [222].
Based on Scherrer equation [177] the average size of metal crystals was estimated
considering width on half height of the strongest peak at 39°. The corresponding sizes
of Pt and Pd nanoclusters are summarized in the Table 9-1 together with data obtained
from the TEM image (Fig. 9-3 ).
91

Table 9-1. Average crystal size of diphos and diop modified Pt and Pd nanoclusters
estimated according to the Scherrer model.
Sample
, nm (Scherrer , nm (TEM)
model)
Diop on Pt 1.7 2
Diop on Pd 1.4 -
Diphos on Pt 1.2 -
Diphos on Pd 1.8 -
Estimated values of the size of crystals lie within 1-2 nm range. We will use these
values for estimation of initial rate in further. Metal content in all samples was found to
be 43±5 % from TGA.
1
2
3
4
30 40 50 2Θ 60 70
Fig. 9-2. XRD spectra of diop modified Pd (1), Pt (2), diphos modified Pd (3) and Pt (4)
nanoclusters.
92

Fig. 9-3. TEM image of diop modified Pt nanoclusters.
Conformation analysis
As was found from molecular modelling the diphos molecule can exist in several
geometrical conformations (Fig. 9-4) at room temperature. The presence of different
conformations is conditioned by free rotation around Ph-P and P-C bonds.
93

Fig. 9-4. Conformation of diphos molecules.
At least three conformations of diphos were found with the energy gaps ΔE 2,1 =0.014
eV, ΔE 2,3 =0.003 eV that are less that energy corresponding to the heat motion of
molecules (K b T=0.023 eV, T = 300 °K), thus these three (and probably more)
conformation coexist at room temperature.
The very similar situation was found with diop molecule, in fact, the energy gap
ΔE 2,1 =7.5 meV results in coexisting of both conformations (Fig. 9-5) at room
temperature.
Fig. 9-5. Conformation of diop molecule. For classification of dipole moment
orientation please the text below.
DRIFTS investigation
94

Since the orientation of phenyl rings of diphos and diop molecules is not rigidly fixed
due to the free rotation around mentioned bonds, we have to do not take into account
the orientations of dipole moment induced by any vibration of any phenyl ring of
diphos and diop in determination of adsorption mode from analysis of DRIFT spectra
of corresponding samples. In fact, for example, in the case of conformation 1 of diop
the out of plane vibration of right upper phenyl ring (Fig. 9-5 left) induces dipole
moment orientated almost along the axis connecting two phosphorous atoms in the
molecule, whereas in the case of conformation 2 (Fig. 9-5 right) the dipole moment
induced by the same phenyl ring is orientated in almost perpendicular direction (out of
paper). However the vibrations produced by phenyl ring(s) are very strong and can be
use for comparison between DRIFT spectra of diop and diphos based samples.
Since diop and diphos have similar structure, their DRIFT spectra have very similar
groups of peaks (Fig. 9-6). In fact, for example, two strong peaks at 1479 cm -1 , 1432
cm -1 can be found in spectra of diop and diphos. It is important to note, that each of
them is a observed superposition of low resolved (due to the very close positions) group
of single peaks (red sticks) corresponding to in plane vibration of every four phenyl
ring, as is demonstrated from theoretical IR modelling of diphos molecule (Fig. 9-7).
Taking this fact into account, we would like to say, that it is not possible to answer on
the following question exactly. Which one of single peaks (red sticks, Fig. 9-7) does
correspond to the observed peak (1484 cm -1 ) in the spectra of adsorbed (Fig. 9-8,
curves 2 and 3) molecule diphos (as well as diop)?
2
1400 1200 1000 800
Wavenumber, cm -1 1
Fig. 9-6. DRIFT spectra of free diphos (1) and diop (2).
It is also possible that the observed peak (1484 cm -1 ) in the spectrum of adsorbed
molecule consists of several (or all) singe peaks (some of them present as shoulders)
and it is not clear which one of them dominates. At the same time the geometrical
orientations of phenyl rings are not fixed in the space due to the conversion between
conformations. The same objections are valid for other in plane vibrations (in the range
1160-1000 cm -1 ) and out of plane vibrations (in the range 846-675 cm -1 ) of phenyl
95

ings. Due to these reasons we can not successfully apply the metal adsorption selection
rule for determining the geometry of adsorption. This is why only some vibrations of
molecule which do not affect vibrations of phenyl rings were taken into the account.
Among these vibrations (at 1276 cm -1 , 1099 cm -1 and 907 cm -1 ) the vibration at 1099
cm -1 inducing dipole moment orientated (for all three conformations) along the line
connecting two phosphor atoms (Fig. 9-4 or 9-9) is the most strongest one. Since this
vibration presences in the spectrum of diphos adsorbed on Pt and Pd nanoclusters as a
strong peak, (both at 1105 cm -1 ) (Fig. 9-8) we propose that the axis connecting two
phosphor atoms is not parallel to the metal surface, i.e. it either parallel to the surface
normal vector, either has a significant angle of tilt, as schematically shown in Fig. 9-9.
We think that the adsorption on corners of Pt and Pd nanocluster can not be excluded,
however, in order to facilitate analysis and larger number of atoms (making adsorptions
sites) are on the plane of crystals we consider adsorption occurring on a flat surface
only.
Since no principal differences were found between spectra of adsorbed diphos on Pt
and Pd nanoclusters except that some shoulders in the spectrum of diphos on Pt are
represented as peaks in the spectrum of diphos on Pd and vice versa we think that there
are no principal differences in adsorption geometry. However some degree of tilt of the
adsorbed diphos is possible that might be a reason for slight difference in wavenumbers
of the corresponding peaks e.g. (out of plane vibration of phenyl ring at 754 cm -1 in
case of Pt and 747 cm -1 in case of Pd nanoparticles), that can also be explained as a
different adsorption strength.
Fig. 9-7. Fragment of calculated (low frequency axis) IR spectrum of diphos.
Corresponding frequencies of experimental observed peaks are written above, see also
text for details.
96

1106
1106
3
2
1099
1
1400 1200 1000 800
Wavenumber, cm -1
Fig. 9-8. Drift spectra of free (1) and adsorbed on Pt (2) and Pd (3) nanoparticles
diphos. Intensity of spectrum of free diphos is decreased for better performance.
We think that the lone pair of phosphor atom does not bind to the metal surface because
of steric hinder induced by both neighbour phenyl rings. The possible distance between
metal surface and phosphor atom can be roughly (without considering changes in the
geometry of adsorbed molecule) estimated as a half size of phenyl ring (~1.4 Å,
hydrogens were ignored).
The same logic was applied in the analysis of DRIFT spectra of adsorbed diop on Pt
and Pd nanoclusters (Fig. 9-10), thus only those vibrations which do not affect
vibrations of phenyl rings were considered. Among them the strongest vibrations
(marked in the Fig. 9-10) were chosen and analyzed as shown in the Table 9-2. As can
be seen from orientation of corresponding dipole moment the main axis (line
connecting both P-atoms) of the diop molecule can not have only exactly perpendicular
or parallel to the normal of surface orientations, otherwise all these vibrations
(absorption peaks) will not be observable simultaneously, in one spectrum. This means,
that the molecule has more then one adsorption mode, or what we think is the most
probable, the axis of the molecule has a significant tilt with respect to the all three
coordinate axis’s (one normal and two tangent vectors of the plane), schematically
adsorption mode is shown in the Fig. 9-11. Here again we would like to note that the
adsorption on the edges and corners of nanoparticle is also not excluded, however due
to the mentioned above reasons we consider adsorption on a plane.
97

Fig. 9-9. Proposed orientation of adsorbed diphos on Pt and Pd nanoclusters.
1087
3
1385
1371
1251
1216
891
2
1400 1200 1000
Wavenumber, cm -1 800
1
Fig. 9-10. DRIFT spectra of free (1) and adsorbed diop on Pt (2) and Pd (3)
nanoclusters.
Table 9-2. Assignment of some frequencies (cm -1 ) in DRIFT spectra of free, adsorbed
on Pt and Pd nanoclusters diop.
Free diop Diop on Pt Diop on Pd Diop theoretical A
1385 (st) 1376 B (st) 1381 (st)
C
1461 ↕ 1,5
1371 (st) 1376 B (st) 1372 (sh) 1447 ● 1,2
1251 (st) C 1236 (st) 1240 (st)
B
1282 ↕ 1,6
B
1275 ↕ 1,5
98

1216 (st) 1217 (sh) -
B
1210 ↔ 1,4
1087 (st) 1101 (st) 1104 (med)
B
1016 ↕ 1,5
891 (st) 885 (st) 889 (st)
B
874 ↔ 1,4
Comments: A – Type of the vibration summarized in this table and corresponding
orientation of the dipole moment are the same for all conformations of diop, B -
unambiguous assignment is difficult, C – orientation of the dipole moment was
characterized as “↔” - horizontal (in plane of the figure),“↕” – vertical and “●” -
horizontal (perpendicular to the figure’s plane) with numerical subscript according to
scheme on Fig. 9-5.
No principal differences between DRIFT spectra of diop adsorbed on Pt and Pd
nanoparticles were found. However the relative intensity of some peaks is slightly
different, some peaks have their analogs as shoulders, small shifts in position of some
peaks probably indicate on some space deviation of main axis of diop being adsorbed
on Pd with respect to Pt nanoclusters or/and different adsorption strength. The peaks at
721 cm -1 (out of plane vibration of phenyl ring) in the spectrum of diop on Pd has no
corresponding peak in case of Pt, probably due to the mentioned above reasons: some
degree of tilt of molecular axis and different bonding strength.
Here we again propose that the lone pair of closest to the metal surface phosphorous
atom does not take part in binding to metal (at least in case of adsorption on flat
surface), because of “big” distance (~1.4 Å, estimated in the same way as have been
done for diphos) to surface that is caused by space shielding of phenyl rings.
Fig. 9-11. Proposed adsorption model of diop on Pt and Pd nanoclusters, hydrogens are
not shown.
99

The model of adsorption of diop on Pt and Pd nanoclusters is very similar to model of
diphos adsorption. This is not surprising because both molecules have the same (Ph) 2 P-
X-P(Ph) 2 structure.
Results of ethyl pyruvate hydrogenation with diop and diphos modified Pt and Pd
catalysts are summarized in the Table 9-3.
Table 9-3. Results of ethyl pyruvate hydrogenation with diop and diphos modifiers of
Pt nanoclusters.
Sample
name
Catalyst
loading,
mg.
Reaction rate,
mmol·(min·g) -1
Conversion B ,
Initial
(min) -1 min
rate A % after time,
Ee, %
(dominated
enantiomer)
Diphos on 20 2 1.2 11 C (90) 0
Pt
Diphos on 20 ≈0 ≈0 0 D (75) -
Pd
(S,S)Diop 20 13 10 80 C (300) 4.4 (S)
on Pt
(R,R)Diop 20 10 7 50 C (90) 4.4 (R)
on Pt
Diop on Pd 20 ≈0 ≈0 0 D (90) -
E
Pt/Al 2 O 3 20 ≈0 ≈0 0 D (240) -
F
Pt/Al 2 O 3 20 0.6 16 2.3 C (90) 0
Comments: A – Values were estimated based on average crystal size (XRD) and metal
content (TGA) of the samples, B - initial amount of racemic ethyl lactate (as impurity
in commercial ethyl pyruvate) was taken into account, C – according to the curve
profile (conversion vs. time) reaction can go further, D - according to the curve profile
(conversion vs. time) reaction can not go further E- reaction was performed with
presence of 34 mg (S,S)diop (1 mM), F- racemic hydrogenation without a modifier.
As can be seen from the Table 9-3, surprisingly, diop modified Pt nanoclusters were
found to be able to induce low (4.4 %) but reproducible enantioselectivity in ethyl
pyruvate hydrogenation reaction. Moreover this type of catalyst demonstrates
enhancement of specific reaction rate and thus enhancement of the initial rate values in
factor of 7-8 with respect to Pt nanoclusters with similar average size modified by non
chiral diphos.
No conversion was found in ethyl pyruvate hydrogenation with activated (400°C,
H 2 /N 2 , 2 hours) conventional Pt/Al 2 O 3 with presence of free diop (1 mM), probably due
to the differences in surface chemistry (surface contaminations and external impurities)
between diop modified nanoclusters and conventional Pt/Al 2 O 3 .
Interestingly, that (S,S)diop leads to excess of S lactate and (R,R)diop to R lactate and
corresponding catalysts demonstrate comparable reaction rate. These facts clearly
indicate the importance of chiral nature of modifier on yielded direction of
enantioselectivity.
Taking into the account the model of enantioselectivity developed for cinchonidine on
Pt system we would like to propose the following model explaining the sense of
enantioselectivity. Author does not pretend to absolute rightfulness of the model,
however finds it to be sufficient for qualitative explanation of enantioselectivity, its
inverse with using opposite enantiomers of diop and enhancement of the reaction rate.
100

According to the determined adsorption geometry of diop, the lone pair of phosphorous
is not involved into binding to the metal surface (at least for adsorption on a flat
surface), thus it is possible that the phosphorous atom (blue marked) (Fig. 9-10 right)
can “catch” adsorbed on Pt ethyl pyruvate molecule via hydrogen bond (rate
enhancement), like unproronated cinchonidine does [50] in non-ionic solvents (toluene,
THF, etc.). In both cases α-carbonyl group of ethyl pyruvate molecule forms halfhydrogenated
state with modifier, is in asymmetric environment produced by red
marked OH and H groups of a modifier. It is known for cinchona alkaloids on Pt that
reverse of positions of OH and H groups leads to inverse of enantioselectivity
(cinchonidine “became” cinchonine) [87]. The same (or very similar) fact was found
for diop obtaining S lactate by using (S,S)diop and R lactate with (R,R) diop.
O
H
N
OH
H
H
P
H
O
O
O
O
*
= X
C
X
C
X
P
O
X
N
HHO
O H
N
O
P
H
H
O
X
Pt surface
Fig. 9-12. Models of half hydrogenated state in case of literature known [84-86] [ethyl
pyruvate-cinchonidine] ad complex (left bottom) on Pt and proposed [ethyl pyruvatediop]
ad complex (right bottom) on Pt. Schematic positions of OH and H groups of
modifiers with respect to α-carbonyl C=O bond of ethyl pyruvate for cinchonidine (left
up) and diop (right up) modifier.
It is interesting to note, that due to the geometry of modifiers OH and H groups are
located significantly closer to “caught” C=O group of ethyl pyruvate in case of
cinchonidine than for diop. Probably because of remoteness of OH and H groups their
asymmetrical influence on C=O bond is not as strong as it is in the case of cinchona
system, that results in low ee. It is also possible that the different nature of H-N and H-
P hydrogen bonds also influences on ee. The electric (electrostatic) influence of closed
to the metal surface phenyl rings, in principal, can also cause asymmetrical influence of
C=O bond, since one of them in tilted. However we do not think their influence and
resulted asymmetric environment are significant, because they have very similar nature
(with respect to OH, H pair), in fact, by keeping (Ph) 2 P group in the same state and
reversing positions of OH and H groups (using another enantiomer of diop) only, we
change direction of enantioselectivity of the system.
101

The calculation of energy of forming pro-R and pro-S half-hydrogenated diop-ethyl
pyruvate complexes can confirm or disprove this model; however we leave this out of
the subject of the current thesis.
9.4 Summary
The models of adsorption geometry of diop and diphos molecules on Pt and Pd
nanoclusters have been proposed. According to these models both molecules are
adsorbed via their phenyl rings pointing molecular axis (a line connecting two
phosphorous atoms) out from the surface, some degree of tilt of the axis is also
possible. In this model lone pair of phosphorous does not take part in adsorption,
probably due to this fact and asymmetrical (with respect to C=O bond) location of OH
and H groups of diop result in observed enantioselectivity in hydrogenation of ethyl
pyruvate. Diop modified Pt nanoclusters demonstrate enhancement of the reaction rate
with respect to Pt nanoclusters modified with non chiral ligand in a factor of 7-8. The
absence enhanced reaction rate and enantioselectivity in ethyl pyruvate hydrogenation
with conventional Pt/Al 2 O 3 with presence of diop shows on differences between
catalysts, probably in surface modification.
102

Chapter 10
Binap and synphos modified Pt
colloidal nanoparticles
10.1 Introduction
In the current chapter we use chiral binap and synphos molecules (Fig. 10-1) as
modifiers for Pt (synphos, binap) and Pd (binap) nanoclusters.
O
PPh 2 O PPh 2
PPh 2 O PPh 2
Binap
Synphos
Fig. 10-1. Binap and synphos molecules.
As it was already mentioned binap attracts attention due to the numerous successful
applications of it in enantioselective homogeneous catalysis [223]. The Synphos
molecule is similar to binap structure and synphos based organometalic complex as a
homogeneous catalyst has been found to be even more efficient in some of reactions
[33]. Immobilization of this type of chiral ligand on a surface of metal nanoparticles is
one strategy for heterogenation of enantioselective homogeneous catalysts. Recently,
binap was immobilized on Au and Pd nanoclusters with size 1.7 and 2 nm,
correspondingly, interestingly, that the latter system was found to be able to induce 95
% enantiomeric excess in hydrosilylation of styrene [224]. Another example is the
immobilization of Ru complex with phosphonic acid substituted binap on Fe 3 O 4
nanoparticles with 10 nm size [225]. The obtained quasi-homogeneous catalyst
demonstrated even higher ee’s in hydrogenation of aromatic ketones. It is very
important to mention that magnetic properties of the nanoparticles allow their
separation from reaction products by permanent magnet.
In the current chapter we test binap and synphos modified Pt nanocluster in
hydrogenation of ethyl pyruvate only, whereas the further test of these (and all other)
samples is supposed to be conducted by our collaborators from Paris and Bucharest,
according to the mentioned in the chapter 3 EU-COST project.
10.2 Experimental
Materials
Synphos ligand was provided by Virginie Ratovelomanana-Vidal from the Laboratoire
de Synthese Selective Organiquie et Produits Naturels, UMR (Ecole Nationale
Superieure de Chemie de Paris), binap was purchased from Aldrich.
Samples preparation
O
103

For the preparation of synphos modified Pt nanoclusters 273 mg Pt 2 (DBA) 3 (0.51 mmol
of Pt) was dissolved in nitrogen flushed THF (40 ml), 300 mg (0.47 mmol) synphos
dissolved in 20 ml nitrogen flushed THF. Both mixtures were transferred into the metal
reactor, purged with argon for 10 min and kept under 3 bar of hydrogen pressure. After
five hours the pressure was released and the mixture was transferred into 300 ml
cyclohexene, then additional 250 ml cyclohexene was added. The brown-back
precipitate was collected by G4 filter. The precipitate in the filter was washed three
times with 100 ml THF/cyclohexene (2/3 by volume) mixture, then with 300 ml
cyclohexene. The brown-black powder was collected mechanically from the filter and
finally fried on air at 35°C for 24 hours. The sample was found to be redispersible in
e.g. chloroform or THF.
Due to the limited amount of synphos, immobilization of it on Pd nanoclusters has not
been done.
In order to prepare binap modified Pt nanoclusters 560 mg Pt 2 (DBA) 3 (0.51 mmol Pt)
in nitrogen purged for 10 min 40 ml THF was added to nitrogen purged 30 ml THF
with 635 mg (1.02 mmol) binap. The mixture was transferred into the metal reactor and
after purging with Ar for 10 min hydrogen pressure of 4 bar has been applied. After 5
hours the mixture was transferred into 500 ml cyclohexene. After 24 hours black
precipitate was collected and washed with 400 ml THF: cyclohexene mixture (1:4) and
with 200 ml cyclohexene. Finally sample was dried at 40 °C under vacuum.
Characterization
The free (not adsorbed) ligand and quasi-homogeneous synphos modified Pt colloidal
sample (3 mg) were dissolved (separately) in 1.3 ml of CDCl 3 and 31 P-NMR
measurements were performed on JEOL ECX 400 MHz instrument making 20 and
20000 scans, respectively.
Molecular modelling, XRD spectra measuring, DRIFTS investigations and
hydrogenation of ethyl pyruvate have been performed as it is described e.g. in the
chapter 9.
Due to the absence of TEM instrument in the IUB and limited access to the instrument
outside the university, TEM investigation of prepared samples is missing.
10.3 Results and discussion
31 P-NMR investigation
NMR investigation of free ligand and ligand modified colloidal Pt nanoclusters is
illustrated on example of synphos.
Synphos molecule has two phosphorous atoms; however 31 P-NMR spectrum (Fig. 10-2
up) has only one peak at -14.8 ppm due to the specific structure of the synphos
molecule (identical surrounding of P atoms). The absolute absence of this peak in the
spectrum of colloidal sample unambiguously says that the obtained after written above
purification procedure sample contains no trace of free (not adsorbed) synphos.
Whereas, in case of adsorbed synphos, significantly shifted peak to 8.2 ppm says about
changes in chemical environment of phosphorous atoms.
The obtain spectrum is very similar to the spectra obtained by Hyeon and co workers
[226] for binap (and other phosphor organic compounds) adsorbed on colloidal Pd
nanoclusters.
104

-14.8
Free synphos
8.2
Synphos on Pt
40 30 20 10 0 -10 -20 -30 -40
PPM
Fig. 10-2. 31 P-NMR spectra of free synphos (up) and synphos modified colloidal Pt
nanoclusters (bottom). Both samples were dissolved in CDCl 3 .
It is important to note, that due to the small size of colloidal Pt nanoclusters estimated
as 1 nm based on XRD (see below) and 2 nm based on dynamic light scattering (data
are not included) it was estimated that random rotation of single nanocluster (in
solution) is fast enough to do not consider anisotropic NMR effects and the system can
be considered within liquid phase NMR theory, i.e. free random rotation of nanocluster
corresponds to free random rotation of a molecule. In fact, rotation correlation time of
an object with radius R (2 nm) in a solvent with viscosity η (0.58 cp or 5.8·10 -4
kg/(m·sec)) at room temperature T (300 K) can be determined from the following
equation, based on Debye-Stokes-Einstein formula.
4 3 η
−
τ = ≈ 5 ⋅10
9
c
πR
sec
3 KT
From another hand the chemical shift anisotropy (CSA) effect can be roughly estimated
as a 10-15 ppm, that is ≈6 kHz in terms of frequencies or τ SCA = 0.2·10 -3 sec. Thus, the
rotation correlation time is much less than typical CSA time that verifies our mentioned
above assumption.
The absent of the peak corresponding to the free state of binap in the spectrum of the
binap (as well as e.g. diop) modified Pt colloidal sample again clearly demonstrate that
used preparation and purification procedures allow having ligand in adsorbed state only
and no trace of free ligand. However due to the technical and organisation problems
with NMR instrument it was not possible to perform required continuous measurement
(≈ 20000 scans, ~3 days) for obtaining sufficient signal/noise ratio for resolving peak
corresponding to the adsorbed ligand for all samples having P atoms.
XRD spectra of binap and synphos modified Pt nanoclusters demonstrate crystalline
nature of the samples.
105

1
2
20 40 60 80
2Θ
Fig. 10-3. XRD spectra of binap (1) and synphos (2) modified Pt nanoclusters. Intensity
of (1) was decreased for better performance.
In fact, all spectra have a strong peak at 39° corresponding (111) crystal face (Fig. 10-
3). In case of binap on Pt sample peaks at 36°, 46° and 67° correspond to (222), (200)
and (220) faces, according to work of Swanson et al. [221]. The last tree peaks are not
resolved in the spectrum of synphos on Pt sample, probably, due to the low sample
loading. Based on the width of the peak at 39° the average crystal size was estimated
with help of Scherrer model, data are summarized in the Table 10-1.
Table 10-1. Average crystal size of binap and synphos modified Pt estimated according
to the Scherrer model.
Sample
, nm
Binap on Pt 1.6
Synphos on Pt 1.0
Estimated values of the size of crystals lie within 1-2 nm range. We will use these
values for estimation of initial rate in further. Metal content in all samples was found to
be 43±5 % from TGA.
Conformation analysis
It was found from molecular modelling that at least two conformations of synphos
molecule are possible due two rotation around P-Ph and C-P(Ph) 2 bonds, like it was
found for diphos and diop molecules.
106

Fig. 10-4. Conformations of synphos.
The energy difference between conformations was found to be 0.027 eV, that is almost
the same value as energy of thermal motion (KT=0.026 eV) is, thus both conformations
can coexist, especially in a solution at room temperature. Because of this fact, we do
not take into the consideration molecular vibrations induced by phenyl rings.
Regardless of the fact that such type of conformations were not found for binap
molecule, we ignore any vibration of a phenyl ring in interpretation of DRIFT spectrum
of adsorbed binap as well.
DRIFTS investigation
From analysis and comparison of DRIFT spectra of free synphos, binap and diphos, all
(in and out of plane) vibrations of phenyl rings were excluded from analysis due to the
mentioned above reasons.
1
2
1600 1400 1200 1000 800
Wavenumber, cm -1
Fig. 10-5. DRIFT spectra of free binap (1) and synphos (2).
107

The DRIFT spectra of synphos (Fig. 10-6) and modified Pt nanoclusters clearly show
shifts in wavenumbers of 2-10 cm -1 that indicates to slight changes in skeleton
geometry or/and changes of force constants of bonds. The relative intensity of some
corresponding peaks is also changed. In order to facilitate spectra analysis, frequencies
of some peaks allowing unambiguous assignment in terms of orientation of induced
dipole moments are summarized in the Table 10-2. Left peaks were not considered due
to the fact that some peaks are superposition of two or more vibrations, as was
illustrated in Fig. 9-5.
2
1600 1400 1200 1000 800 600
Wavenumber, cm -1 1
Fig. 10-6. DRIFT spectra of free (1) and adsorbed on Pt nanoclusters (2) synphos.
Intensity of free synphos was decreased for better performance.
Table 10-2. Assignment of some frequencies (cm -1 ) of DRIFT spectra of free and
adsorbed on Pt nanoclusters synphos.
Free synphos
(experimental)
Synphos on Pt Calculated synphos A Dipole moment
orientation B
1452 (str) 1463 (wk. sh.) 1501 (str) ↔ 1,4
1433 (str) 1441 (str) 1485 (str) ↕ 1,5
1401 (med) 1404 (med) 1449 (wk) ↔ 1,3
1289 (str) 1301 (str) 1316 (str) ↔ 1,4 or ↕ 1,5
1255 (med) 1257 (med) 1281 (med) ↔ 1,4
1150 (str) 1159 (med) 1172 (med) ↔ 2,4
938 (med) 939 (med) 975 (med) ↕ 1,5
873 (med) 877 (med) 906 (med) C
902 (wk) C ↔ 1,4
↕ 1,5
Comments: A - conformation from Fig. 10-4 (left) was used, calculated frequencies for
other conformation were just slightly (± max. 10cm -1 ) different, B - orientation of the
dipole moment was the same for both conformations and was characterized as “↔” -
horizontal (in plane of the figure),“↕” – vertical and with numerical subscript according
to scheme in Fig. 10-4, C – unambiguous assignment is difficult.
108

Below in the text we assume that only one adsorption mode of synphos exists on Pt
surface and adsorption takes place on a flat surface.
Based on the data from the Table 10-2 it follows, that the synphos molecule can not
adsorb (see denied examples N1, 2 and 3 in Fig. 10-7) on the flat surface without
significant tilt (with respect to depicted in Fig. 10-4 orientation, assuming that the metal
surface has a normal parallel to ↕ 1,5 direction). At the same time it allows a coexistence
of non zero projection to the surface’s normal of all orientations of dipole moment from
this table. However “denied” orientations become possible if they coexist on the Pt
surface, e.g. 1 coexist with 2 or 3.
Due to the isotropic shape of the molecule it is not possible to determine the orientation
of the adsorbed molecule unambiguously; in fact, at least two adsorption species (see
allowed examples (4 and 5) in Fig. 10-7-bottom) can be proposed, they can also coexist
on the surface together.
Fig. 10-7. Schematic illustration of denied mono modes (up) and allowed mono and
multi modes (bottom) of adsorption of synphos molecule on Pt nanoclusters. Hydrogen
atoms are not shown.
It is interesting to note that proposed adsorption mode N5 (Fig. 10-7) is similar to
adsorption modes found for diphos and diop molecules. In fact, all three molecules
have P(Ph) 2 functional group which can be considered as a common anchor for them, in
the similar way as quinoline is an anchor for cinchonidine and quiphos molecules.
In spite of similarity in the structure of synphos and binap, IR spectrum of the latter is
difficult for full analysis because it has a lot of peaks which have no analogs in the IR
spectrum of synphos (probably due to the enhanced inharmonic vibrations for binap).
Unfortunately, almost every strong peak in the spectrum of binap which can be
unambiguously assigned has to be excluded from consideration due the mentioned
above reasons (different conformations and/or contribution from vibration of the
similar parts of the molecule, e.g. another phenyl rings). From another hand, those
109

peaks which can be used for analysis of orientation of adsorbed binap can not be
assigned unambiguously.
However ambiguity in assignment and interpretation of spectra very often results in
difficulties (impossibilities) in discrimination between ↕ 1,5 and ↔ 1,4 orientations of
induced dipole moment. For example the strong peaks at 816 cm -1 in the spectrum of
binap on Pt (Fig. 10-8) corresponds to the peak at 818 cm -1 in the spectrum of free
binap which consists of two components (due to the symmetrical and asymmetrical out
of plane bending of hydrogens in naphthalene rings, both are not shown). The
symmetrical vibration has dipole moment along the ↕ 1,5 direction, whereas
asymmetrical one has it along the ↔ 1,4 direction. The similar situation is with all other
strong peaks.
816
2
1600 1400 1200 1000 800
Wavenumber, cm -1 1
Fig. 10-8. DRIFT spectra of free (1) and adsorbed on Pt nanoclusters (2) binap.
Intensity of free binap was decreased for better performance.
Since spectra of adsorbed binap on Pt nanoclusters allows interpretation in terms of
orientation of dipole moment in ↕ 1,5 or/and ↔ 1,4 directions, we would like to use the
same logic for determination of orientations of adsorbed binap. This results to the same
adsorption modes like for synphos molecule (Fig. 10-7).
In hydrogenation of ethyl pyruvate both catalysts: binap and synphos did not show
detectable enantioselectivity. The specific reaction rate and initial rate (normalized on
the number of metal surface atoms) were similar for both samples. Reaction rate
normalized on number of surface Pt atoms was higher for conventional Pt/Al 2 O 3 ,
probably due to the thermal activation of the latter under hydrogen. The reason of
absence of enantioselectivity with synphos and binap modified Pt nanoclusters
probably lies in principal difference in their structure, i.e. both molecules do not meet
the specific requirements developed for cinchona on Pt system, whereas, we think, diop
molecule (see chapter 9) fits partly, that results in low ee.
Table 10-3. Results of ethyl pyruvate hydrogenation with binap and synphos modifiers
of Pt nanoclusters
110

Initial rate A Conversion B , %
Sample name Catalyst Reaction rate,
loading, mg. mmol·(min·g) -1 (min) -1 after time, min
Binap on Pt 20 5 3.7 30 (120) D
Synphos on Pt 10 4.5 2.8 16 (120) D
C
Pt/Al 2 O 3 20 0.6 16.5 12.5 (345) D
Comments: A – values were estimated based on chemisorption (Pt/Al 2 O 3 ), size of Pt
nanoclusters and metal content (43%) data, B - initial amount of racemic ethyl lactate
(as impurity in commercial ethyl pyruvate) was taken into account. C- hydrogenation
without a modifier, D – according to the curve profile (conversion vs. time) reaction
can go further.
10.4 Summary
The models of adsorption geometry of synphos and binap molecules on 1 and 1.6 nm Pt
nanoclusters have been proposed. According to these models both molecules are
adsorbed via their phenyl rings, however existence of other adsorption modes is also
possible. In any model of adsorption on a flat surface, the lone pair of phosphorous
does not take part in adsorption, due to steric hinder of induced by e.g. phenyl rings. No
enantioselectivity was found with binap and synphos modified Pt nanoclusters in the
hydrogenation of ethyl pyruvate, probably due the fact that both molecules do not meet
specific requirements which were developed based on cinchona on Pt system.
111

References
[1] G.E. Tranter, J. Chem. Soc., Chem. Commun. (1986) 60-61.
[2] A. Streitwieser, W.D. Schaeffer, J. Am. Chem. Soc. 78 (1956) 5597-5599.
[3] J. March, Advanced Organic Chemistry, John Wiley & Sons, New Your, 1992.
[4] Corriu, Guerin, Moreau, The Chemistry of Organic Silicon Compounds, John
Wiley & Sons, New York, 1998.
[5] M. Gielen, Top. Curr. Chem. 104 (1982) 57-105.
[6] Stackhou.J, R.D. Baechler, K. Mislow, Tetrahedron Lett. (1971) 3437.
[7] Stackhou.J, R.D. Baechler, K. Mislow, Tetrahedron Lett. (1971) 3441.
[8] J.D. Andose, J.M. Lehn, K. Mislow, J. Wagner, J. Am. Chem. Soc. 92 (1970)
4050.
[9] G. Delapierre, J.M. Brunel, T. Constantieux, G. Buono, Tetrahedron:
Asymmetry 12 (2001) 1345-1352.
[10] G. Delapierre, T. Constantieux, J.M. Brunel, G.R. Buono, Eur. J. Org. Chem.
(2000) 2507.
[11] J.M. Brunel, T. Constantieux, G. Buono, J. Org. Chem. 64 (1999) 8940-8942.
[12] J.M. Brunel, T. Constantieux, A. Labande, F. Lubatti, G. Buono, Tetrahedron
Lett. 38 (1997) 5971-5974.
[13] S.J. Brois, J. Am. Chem. Soc. 90 (1968) 506.
[14] S.J. Brois, J. Am. Chem. Soc. 90 (1968) 508.
[15] H. Hamill, M.A. McKervey, J. Chem. Soc., Chem. Commun. (1969) 864.
[16] A. Vargos, Phd Thesis, Zurich, 2004.
[17] V. Prelog, M. Kovacevic, M. Egli, Angew. Chem., Int. Ed. Engl. 28 (1989)
1147-1152.
[18] V.S. Martin, S.S. Woodard, T. Katsuki, Y. Yamada, M. Ikeda, K.B. Sharpless,
J. Am. Chem. Soc. 103 (1981) 6237-6240.
[19] J.D. Morrison, H.S. Mosher, Asymmetric Organic Reactions, Englewood Cliffs,
New Jersey, 1971.
[20] I. Ojima, Catalytic Asymmetric Synthesis, Willey-VCH, 1999.
[21] D.E. De Vos, I.F.J. Vankelecom, P.A. Jacobs, Chiral catalyst immobilization
and recycling, Weinheim, 2000.
[22] H.-U. Blaser, M. Eissen, P. Fauquex, K. Hungerbühler, E. Schmidt, G.
Sedelmeir, M. Studer, in: H.U. Blaser, E. Schmidt (Eds.), Asymmetric Catalysis
on Industrial Scale: Challenges, Approaches and Solutions, WILEY-VCH,
Weinheim, 2004, pp. 91-103.
[23] R. Noyori, M. Koizumi, D. Ishii, T. Ohkuma, Pure Appl. Chem. 73 (2001) 227-
232.
[24] T.P. Dang, H.B. Kagan, J. Chem. Soc., Chem. Commun. (1971) 481-481.
[25] T. Hayashi, M. Kumada, Acc. Chem. Res. 15 (1982) 395-401.
[26] S. Jeulin, N. Champion, P. Dellis, V. Ratovelomanana-Vidal, J.P. Genet,
Synthesis-Stuttgart (2005) 3666-3671.
[27] R. Touati, T. Gmiza, S. Jeulin, C. Deport, V. Ratovelamanana-Vidal, C. Ben
Hassine, J.P. Genet, Synlett (2005) 2478-2482.
[28] R. Le Roux, N. Desroy, P. Phansavath, J.P. Genet, Synlett (2005) 429-432.
[29] C. Mordant, P. Dunkelmann, V. Ratovelomanana-Vidal, J.P. Genet, Eur J Org
Chem (2004) 3017-3026.
[30] C. Mordant, P. Dunkelmann, V. Ratovelomanana-Vidal, J.P. Genet, Chem
Commun (2004) 1296-1297.
112

[31] S. Jeulin, S.D. de Paule, V. Ratovelomanana-Vidal, J.P. Genet, N. Champion, P.
Dellis, P Natl Acad Sci USA 101 (2004) 5799-5804.
[32] J.P. Genet, Accounts Chem Res 36 (2003) 908-918.
[33] S.D. de Paule, S. Jeulin, V. Ratovelomanana-Vidal, J.P. Genet, N. Champion, P.
Dellis, Eur. J. Org. Chem. (2003) 1931-1941.
[34] S.D. de Paule, S. Jeulin, V. Ratovelomanana-Vidal, J.P. Genet, N. Champion,
G. Deschaux, P. Dellis, Org Process Res Dev 7 (2003) 399-406.
[35] S.D. de Paule, S. Jeulin, V. Ratovelomanana-Vidal, J.P. Genet, N. Champion, P.
Dellis, Tetrahedron Lett 44 (2003) 823-826.
[36] K. Inoguchi, K. Achiwa, Synlett (1991) 49-51.
[37] K. Achiwa, J. Am. Chem. Soc. 98 (1976) 8265-8266.
[38] G.M. Schwab, F. Rost, L. Rudolph, Kolloid Z 68 (1934) 157-159.
[39] G.M. Schwab, L. Rudolph, Naturwisserschaft 20 (1932) 363-364.
[40] G.J. Hutchings, Annu. Rev. Mater. Res. 35 (2005) 143-166.
[41] A. Tai, T. Kikukawa, T. Sugimura, Y. Inoue, S. Abe, T. Osawa, T. Harada, Bull.
Chem. Soc. Japan 67 (1994) 2473-2477.
[42] T. Harada, T. Osawa, in: G. Jannes, V. Dubois (Eds.), Chiral Reactions in
Heterogeneous Catalysis, Plenum Press, New York, 1995, p. 83.
[43] T. Osawa, A. Tai, Y. Imachi, S. Takasaki, in: G. Jannes, V. Dubois (Eds.),
Chiral Reactions in Heterogeneous Catalysis, Plenum Press, New York, 1995, p.
75.
[44] M. Studer, H.U. Blaser, C. Exner, Adv. Synth. Catal. 345 (2003) 45-65.
[45] Y. Orito, S. Imai, S. Niwa, J. Chem. Soc. Jpn. (1979) 1118.
[46] Y. Orito, S. Imai, S. Niwa, J. Chem. Soc. Jpn. (1980) 670.
[47] Y. Orito, S. Imai, S. Niwa, Nippon Kagaku Kaishi (1979) 1118-1120.
[48] N. Kunzle, R. Hess, T. Mallat, A. Baiker, J. Catal. 186 (1999) 239-241.
[49] P.B. Wells, A.G. Wilkinson, Top. Catal. 5 (1998) 39-50.
[50] A. Baiker, J. Mol. Catal. A: Chem. 163 (2000) 205-220.
[51] B. Törok, K. Felfoldi, G. Szakonyi, K. Balazsik, M. Bartok, Catal. Lett. 52
(1998) 81-84.
[52] M. Schürch, N. Künzle, T. Mallat, A. Baiker, J. Catal. 176 (1998) 569-571.
[53] O.J. Sonderegger, T. Burgi, A. Baiker, J. Catal. 215 (2003) 116-121.
[54] E. Toukoniitty, P. Maki-Arvela, J. Kuusisto, V. Nieminen, J. Paivarinta, M.
Hotokka, T. Salmi, D.Y. Murzin, J. Mol. Catal. A: Chem. 192 (2003) 135-151.
[55] E. Toukoniitty, P. Maki-Arvela, M. Kuzma, A. Villela, A.K. Neyestanaki, T.
Salmi, R. Sjoholm, R. Leino, E. Laine, D.Y. Murzin, J. Catal. 204 (2001) 281-
291.
[56] J.A. Slipszenko, S.P. Griffiths, P. Johnston, K.E. Simons, W.A.H. Vermeer,
P.B. Wells, J. Catal. 179 (1998) 267-276.
[57] W.A.H. Vermeer, A. Fulford, P. Johnston, P.B. Wells, J. Chem. Soc., Chem.
Commun. (1993) 1053-1054.
[58] B. Torok, K. Felfoldi, K. Balazsik, M. Bartok, Chem. Commun. (1999) 1725-
1726.
[59] M. Studer, S. Burkhardt, H.-U. Blaser, Chem. Commun. (1999) 1727-1728.
[60] M. von Arx, T. Mallat, A. Baiker, Catal. Lett. 78 (2002) 267-271.
[61] M. von Arx, T. Mallat, A. Baiker, Tetrahedron: Asymmetry 12 (2001) 3089-
3094.
[62] M. von Arx, T. Mallat, A. Baiker, J. Catal. 193 (2000) 161-164.
[63] K. Balazsik, K. Szori, K. Felfoldi, B. Torok, M. Bartok, Chem. Commun.
(2000) 555-556.
113

[64] A. Szabo, N. Kunzle, T. Mallat, A. Baiker, Tetrahedron: Asymmetry 10 (1999)
61-76.
[65] N. Kunzle, A. Szabo, M. Schurch, G.Z. Wang, T. Mallat, A. Baiker, Chem.
Commun. (1998) 1377-1378.
[66] G.Z. Wang, T. Mallat, A. Baiker, Tetrahedron: Asymmetry 8 (1997) 2133-2140.
[67] W.R. Huck, T. Mallet, A. Baiker, J. Catal. 193 (2000) 1-4.
[68] M. von Arx, T. Mallat, A. Baiker, Top. Catal. 19 (2002) 75-87.
[69] D.Y. Murzin, P. Maki-Arvela, E. Toukoniitty, T. Salmi, Catal. Rev. - Sci. Eng.
47 (2005) 175-256.
[70] T. Bürgi, A. Baiker, J. Am. Chem. Soc. 120 (1998) 12920-12926.
[71] M. Schürch, O. Schwalm, T. Mallat, J. Weber, A. Baiker, J. Catal. 169 (1997)
275-286.
[72] A.F. Carley, M.K. Rajumon, M.W. Roberts, P.B. Wells, Journal Of The
Chemical Society-Faraday Transactions 91 (1995) 2167-2172.
[73] K.E. Simons, P.A. Meheux, S.P. Griffiths, I.M. Sutherland, P. Johnston, P.B.
Wells, A.F. Carley, M.K. Rajumon, M.W. Roberts, A. Ibbotson, Recueil Des
Travaux Chimiques Des Pays-Bas-Journal Of The Royal Netherlands Chemical
Society 113 (1994) 465-474.
[74] T. Evans, A.P. Woodhead, A. Gutierrez-Sosa, G. Thornton, T.J. Hall, A.A.
Davis, N.A. Young, P.B. Wells, R.J. Oldman, O. Plashkevych, O. Vahtras, H.
Agren, V. Carravetta, Surf. Sci. 436 (1999) L691.
[75] G. Bond, P.B. Wells, J. Catal. 150 (1994) 329.
[76] D. Ferri, T. Burgi, J. Am. Chem. Soc. 123 (2001) 12074-12084.
[77] W. Chu, R.J. LeBlanc, C.T. Williams, J. Kubota, F. Zaera, J. Phys. Chem. B 107
(2003) 14365-14373.
[78] C. LeBlond, J. Wang, J. Liu, A.T. Andrews, Y.K. Sun, J. Am. Chem. Soc. 121
(1999) 4920.
[79] M. Bartok, B. Torok, K. Balazsik, T. Bartok, Catal. Lett. 73 (2001) 127-131.
[80] X.B. Li, N. Dummer, R. Jenkins, R.P.K. Wells, P.B. Wells, D.J. Willock, S.H.
Taylor, P. Johnston, G.J. Hutchings, Catal. Lett. 96 (2004) 147-151.
[81] M. Garland, H.-U. Blaser, J. Am. Chem. Soc. 112 (1990) 7048-7050.
[82] D. Ferri, T. Burgi, A. Baiker, J. Catal. 210 (2002) 160-170.
[83] H.-U. Blaser, H.P. Jalett, D.M. Monti, A. Baiker, J.T. Wehrli, Stud. Surf. Sci.
Catal. 67 (1991) 147.
[84] O. Schwalm, B. Minder, J. Weber, A. Baiker, Catal. Lett. 23 (1994) 271-279.
[85] O. Schwalm, J. Weber, B. Minder, A. Baiker, Theochem-Journal Of Molecular
Structure 330 (1995) 353-357.
[86] O. Schwalm, J. Weber, B. Minder, A. Baiker, Int. J. Quantum Chem. 52 (1994)
191-197.
[87] P.B. Wells, R.P.K. Wells, in: D.E. De Vos, I.F.J. Vankelecom, P.A. Jacobs
(Eds.), Chiral Catalyst Immobilization and Recycling, Wiley-VCH, Weinheim,
2000, pp. 123-138.
[88] G. Bond, P.A. Meheux, A. Ibbotson, P.B. Wells, Catal. Today 10 (1991) 371.
[89] H.-U. Blaser, H.P. Jalett, M. Garland, M. Studer, H. Thies, A. Wirth-Tijani, J.
Catal. 173 (1998) 282.
[90] J.L. Margitfalvi, M. Hegedus, J. Mol. Catal. A: Chem. 107 (1996) 281-289.
[91] J.L. Margitfalvi, M. Hegedus, E. Tfirst, 11th International Congress On
Catalysis - 40th Anniversary, Pts A And B, 1996, pp. 241-250.
[92] J.L. Margitfalvi, M. Hegedus, E. Tfirst, Tetrahedron: Asymmetry 7 (1996) 571-
580.
114

[93] J.L. Margitfalvi, E. Talas, E. Tfirst, C.V. Kumar, A. Gergely, Appl. Catal., A
191 (2000) 177.
[94] H.-U. Blaser, H.P. Jalett, D.M. Monti, J.F. Reber, J.T. Wehrli, Stud. Surf. Sci.
Catal. 41 (1988) 153.
[95] K.E. Simons, A. Ibbotson, P. Johnston, H. Plum, P.B. Wells, J. Catal. 150
(1994) 321-328.
[96] E. Toukoniitty, P. Maki-Arvela, A.N. Villela, A.K. Neyestanaki, T. Salmi, R.
Leino, R. Sjoholm, E. Laine, J. Vayrynen, T. Ollonqvist, P.J. Kooyman, Catal.
Today 60 (2000) 175-184.
[97] H.-U. Blaser, H.P. Jalett, M. Müller, M. Studer, Catal. Today 37 (1997) 441-
463.
[98] Y. Nitta, K. Kobiro, Chem. Lett. (1996) 897-898.
[99] Y. Nitta, Y. Okamoto, Chem. Lett. (1998) 1115-1116.
[100] P.A. Meheux, A. Ibbotson, P.B. Wells, J. Catal. 128 (1991) 387-396.
[101] Y. Nitta, T. Imanaka, S. Teranishi, J. Catal. 80 (1983) 31-39.
[102] A. Baiker, J. Mol. Catal. A: Chem. 115 (1997) 473-493.
[103] M. Schürch, T. Heinz, R. Aeschimann, T. Mallat, A. Pfaltz, A. Baiker, J. Catal.
173 (1998) 187.
[104] T. Heinz, Phd Thesis, Universität Basel, Basel, 1997.
[105] H.A. Wierenga, L. Soethout, J.W. Gerritsen, B.E.C. Vandeleemput, H.
Vankempen, G. Schmid, Advanced Materials 2 (1990) 482-484.
[106] X.B. Zuo, H.F. Liu, D.W. Guo, X.Z. Yang, Tetrahedron 55 (1999) 7787-7804.
[107] A. Kockritz, S. Bischoff, V. Morawsky, U. Prusse, K.D. Vorlop, J. Mol. Catal.
A: Chem. 180 (2002) 231-243.
[108] A. Gamez, J. Kohler, J. Bradley, Catal. Lett. 55 (1998) 73-77.
[109] J.U. Kohler, J.S. Bradley, Langmuir 14 (1998) 2730-2735.
[110] J.U. Kohler, J.S. Bradley, Catal. Lett. 45 (1997) 203-208.
[111] J.S. Bradley, J. Kohler, D. deCaro, Abstracts Of Papers Of The American
Chemical Society 212 (1996) 14-CATL.
[112] Y.L. Huang, J.R. Chen, H. Chen, R.X. Li, Y.Z. Li, L.E. Min, X.J. Li, J. Mol.
Catal. A: Chem. 170 (2001) 143-146.
[113] J. Zhang, X.P. Yan, H.F. Liu, J. Mol. Catal. A: Chem. 175 (2001) 125-130.
[114] H. Bönnemann, G.A. Braun, Angew. Chem., Int. Ed. 35 (1996) 1992-1995.
[115] H. Bönnemann, G.A. Braun, Chem. Eur. J. 3 (1997) 1200-1202.
[116] X.B. Zuo, H.F. Liu, M.H. Liu, Tetrahedron Lett. 39 (1998) 1941-1944.
[117] X.B. Zuo, H.F. Liu, J.F. Tian, J. Mol. Catal. A: Chem. 157 (2000) 217-224.
[118] S. Jansat, M. Gomez, K. Philippot, G. Muller, E. Guiu, C. Claver, S. Castillon,
B. Chaudret, J. Am. Chem. Soc. 126 (2004) 1592-1593.
[119] H.H. Helcher, J. Herbord Klossen, 1718.
[120] M. Faraday, Philos. Trans. R. Soc. London 147 (1857) 145.
[121] U. Kolb, S.A. Quaiser, M. Winter, M.T. Reetz, Chem. Mater. 8 (1996) 1889-
1894.
[122] N. Toshima, Y. Wang, Langmuir 10 (1994) 4574-4580.
[123] G. Schmid, Endeavour 14 (1990) 172-178.
[124] DEFINITION AND CLASSIFICATION OF COLLOIDS, IUPAC,
http://www.iupac.org/reports/2001/colloid_2001/manual_of_s_and_t/node33.ht
ml
[125] Classification of colloids, Wikipedia, http://en.wikipedia.org/wiki/Colloid
[126] M.A. Watzky, R.G. Finke, Chem. Mater. 9 (1997) 3083-3095.
[127] M.A. Watzky, R.G. Finke, J. Am. Chem. Soc. 119 (1997) 10382-10400.
115

[128] J.D. Aiken III, R.G. Finke, J. Mol. Catal. A: Chem. 145 (1999) 1-44.
[129] B.K. Teo, N.J.A. Sloane, Inorg. Chem. 24 (1985) 4545-4558.
[130] N.J.A. Sloane, B.K. Teo, J. Chem. Phys. 83 (1985) 6520-6534.
[131] A.J. Bard, L.R. Faulkner, Electrochemical Methods, Wiley, New York, 1980.
[132] A. Roucoux, J. Schulz, H. Patin, Chemical Reviews 102 (2002) 3757-3778.
[133] A. Amulyavichus, A. Daugvila, R. Davidonis, C. Sipavichus, Fiz. Met.
Metalloved. 85 (1998) 111-117.
[134] E. Gaffet, M. Tachikart, O. ElKedim, R. Rahouadj, Mater. Charact. 36 (1996)
185-190.
[135] H. Bönnemann, W. Brijoux, R. Brinkmann, R. Fretzen, T. Joussen, R. Koppler,
B. Korall, P. Neiteler, J. Richter, J. Mol. Catal. 86 (1994) 129-177.
[136] R. Houbertz, T. Feigenspan, F. Mielke, U. Memmert, U. Hartmann, U. Simon,
G. Schon, G. Schmid, Europhys. Lett. 28 (1994) 641-646.
[137] G. Schmid, S. Peschel, New J. Chem. 22 (1998) 669-675.
[138] U. Simon, R. Flesch, H. Wiggers, G. Schon, G. Schmid, J. Mater. Chem. 8
(1998) 517-518.
[139] M. Bardaji, O. Vidoni, A. Rodriguez, C. Amiens, B. Chaudret, M.J. Casanove,
P. Lecante, New J. Chem. 21 (1997) 1243-1249.
[140] G. Schmid, H. West, J.O. Malm, J.O. Bovin, C. Grenthe, Chem. Eur. J. 2 (1996)
1099-1103.
[141] M.N. Vargaftik, V.P. Zagorodnikov, I.P. Stolarov, Moiseev, II, D.I. Kochubey,
V.A. Likholobov, A.L. Chuvilin, K.I. Zamaraev, J. Mol. Catal. 53 (1989) 315-
348.
[142] G. Schmid, B. Morun, J.O. Malm, Angew. Chem., Int. Ed. Engl. 28 (1989) 778-
780.
[143] G. Schmid, A. Lehnert, Angew. Chem., Int. Ed. Engl. 28 (1989) 780-781.
[144] G. Schmid, Polyhedron 7 (1988) 2321-2329.
[145] G. Schmid, Applied Homogeneous Catalysis with Organometallic Compounds,
Wiley-VCH, 1996, p. 677.
[146] R. Franke, J. Rothe, J. Pollmann, J. Hormes, H. Bönnemann, W. Brijoux, T.
Hindenburg, J. Am. Chem. Soc. 118 (1996) 12090-12097.
[147] O. Vidoni, K. Philippot, C. Amiens, B. Chaudret, O. Balmes, J.O. Malm, J.O.
Bovin, F. Senocq, M.J. Casanove, Angew. Chem., Int. Ed. 38 (1999) 3736-
3738.
[148] K. Patel, S. Kapoor, D.P. Dave, T. Mukherjee, J. Chem. Sci. 117 (2005) 311-
316.
[149] S.H. Sun, C.B. Murray, D. Weller, L. Folks, A. Moser, Science 287 (2000)
1989-1992.
[150] P.H. Hess, P.H. Parker, J. Appl. Polym. Sci. 10 (1966) 1915.
[151] Y. Kato, S. Sugimoto, K. Shinohara, N. Tezuka, T. Kagotani, K. Inomata,
Mater. Trans. 43 (2002) 406-409.
[152] M.T. Reetz, W. Helbig, J. Am. Chem. Soc. 116 (1994) 7401-7402.
[153] S.I. Stoeva, B.L.V. Prasad, S. Uma, P.K. Stoimenov, V. Zaikovski, C.M.
Sorensen, K.J. Klabunde, J. Phys. Chem. B 107 (2003) 7441-7448.
[154] W.T. Nichols, T. Sasaki, N. Koshizaki, Euromat, Prague, 2005.
[155] K.M. Beck, T. Sasaki, N. Koshizaki, Chem. Phys. Lett. 301 (1999) 336-342.
[156] X.Y. Zeng, N. Koshizaki, T. Sasaki, A. Yabe, F. Rossignol, H. Nagai, Y.
Nakata, T. Okutani, M. Suzuki, Appl. Surf. Sci. 140 (1999) 90-98.
[157] Turkevic.J, G. Kim, Science 169 (1970) 873.
116

[158] H. Bönnemann, G. Braun, W. Brijoux, R. Brinkmann, A.S. Tilling, K. Seevogel,
K. Siepen, J. Organomet. Chem. 520 (1996) 143.
[159] H. Bönnemann, W. Brijoux, Active Metals, VCH, Weinheim, 1996.
[160] H. Bönnemann, R.M. Richards, Eur. J. Inorg. Chem. (2001) 2455-2480.
[161] H. Bönnemann, W. Brijoux, R. Brinkmann, E. Dinjus, T. Joussen, B. Korall,
Angewandte Chemie-International Edition In English 30 (1991) 1312-1314.
[162] L. D'Souza, Phd Thesis, International University Bremen, Bremen, 2005.
[163] COST Chemistry Action D24, Sustainable Chemical Processes: Stereoselective
Transition Metal-Catalysed Reactions,
http://costchemistry.epfl.ch/load/d24_mid_term_report.pdf
[164] T. Mallat, A. Baiker, J. Catal. 176 (1998) 271-274.
[165] D.G. Blackmond, J. Catal. 176 (1998) 267-270.
[166] T. Mallat, Z. Bodnar, B. Minder, K. Borszeky, A. Baiker, J. Catal. 168 (1997)
183-193.
[167] U.K. Singh, R.N. Landau, Y.K. Sun, C. Leblond, D.G. Blackmond, S.K.
Tanielyan, R.L. Augustine, J. Catal. 154 (1995) 91-97.
[168] Y.K. Sun, J. Wang, C. Leblond, R.N. Landau, D.G. Blackmond, J. Catal. 161
(1996) 759-765.
[169] K. Balazsik, M. Bartok, J. Catal. 224 (2004) 463-472.
[170] X.B. Li, R.P.K. Wells, P.B. Wells, G.J. Hutchings, Catal. Lett. 89 (2003) 163.
[171] V. Morawsky, Phd Thesis, Braunschweig, 2002.
[172] T.O. Ely, C. Pan, C. Amiens, B. Chaudret, F. Dassenoy, P. Lecante, M.J.
Casanove, A. Mosset, M. Respaud, J.M. Broto, J. Phys. Chem. B 104 (2000)
695-702.
[173] K. Moseley, P.M. Maitlis, J. Chem. Soc. D (1971) 982-983.
[174] P. Yuan, D.Q. Wu, H.P. He, Z.Y. Lin, Appl. Surf. Sci. 227 (2004) 30-39.
[175] M. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R.
Cheeseman, V.G. Zakrzewski, J.A. Montgomery, Jr., R.E. Sratmann, J.C.
Burant, S. Dapprich, J.M. Millam, A.D. Daniels, R. Cammi, B. Mennucci, C.
Pomelli, C. Adamo, S. Clifford, J. Ochterski, G.A. Petersson, P.Y. Ayala, Q.
Ciu, K. Morokuma, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B.
Foresman, J.P. Cioslowski, I. Komaromi, R. Gompets, R.L. Martin, D.J. Fox, T.
Keith, M. Al-Laham, P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, J.L.
Andres, C. Gonzalez, M. Head-Gordon, E.S. Replogle, J.A. Pople, Gaussian
Inc., Pittsburgh, 2003.
[176] W.P. Davey, Physical Review 25 (1925) 763.
[177] P. Scherrer, Göttingen Nachrichten 2 (1918) 98.
[178] J. Kubota, Z. Ma, F. Zaera, Langmuir 19 (2003) 3371-3376.
[179] R.G. Greenler, J. Chem. Phys. 44 (1966) 310.
[180] H.-U. Blaser, M. Garland, H.P. Jallet, J. Catal. 144 (1993) 569-578.
[181] H.-U. Blaser, H.P. Jalett, J. Wiehl, J. Mol. Catal. 68 (1991) 215-222.
[182] E. Orglmeister, T. Mallat, A. Baiker, J. Catal. 233 (2005) 333-341.
[183] B. Minder, T. Mallat, A. Baiker, G. Wang, T. Heinz, A. Pfaltz, J. Catal. 154
(1995) 371.
[184] D. Ferri, T. Burgi, K. Borszeky, T. Mallat, A. Baiker, J. Catal. 193 (2000) 139-
144.
[185] V. Morawsky, U. Pruße, L. Witte, K.D. Vorlop, Catal. Commun. 1 (2000) 15.
[186] K. Balazsik, M. Bartok, J. Catal. 224 (2004) 463.
[187] J. Kubota, F. Zaera, J. Am. Chem. Soc. 123 (2001) 11115.
[188] B.E. Spiewak, R.D. Cortright, J.A. Dumesic, J. Catal. 176 (1998) 405-414.
117

[189] K. Umezawa, T. Ito, M. Asada, S. Nakanishi, P. Ding, W.A. Lanford, B.
Hjorvarsson, Surf. Sci. 387 (1997) 320-327.
[190] M.R. Voss, H. Busse, B.E. Koel, Surf. Sci. 414 (1998) 330-340.
[191] J. Fearon, G.W. Watson, J. Mater. Chem. 16 (2006) 1989-1996.
[192] Asymmetric Hydrogenation of Activated Ketones with catASium® F214, R.
Hartung, J.G.E. Krauter, S. Seebald, P. Kukula, U. Nettekoven, M. Studer,
http://www.degussa-catalysts.com/asp/documents/79_Document.pdf
[193] CatASium® F214 Data Sheet heterogeneous catalyst for the enantioselective
hydrogenation of -functionalised ketones, http://www.degussacatalysts.com/asp/documents/125_Document.pdf
[194] K. Balazsik, M. Bartok, J. Mol. Catal. A: Chem. 219 (2004) 383-389.
[195] Z. Ma, J. Kubota, F. Zaera, J. Catal. 219 (2003) 404-416.
[196] C.P. Hwang, C.T. Yeh, J. Mol. Catal. A: Chem. 112 (1996) 295-302.
[197] C.E. Smith, J.P. Biberian, G.A. Somorjai, J. Catal. 57 (1979) 426-443.
[198] T. Matsushima, D.B. Almy, J.M. White, Surf. Sci. 67 (1977) 89-108.
[199] H.H. Rotermund, J. Lauterbach, G. Haas, Appl. Phys. A: Mater. Sci. Process. 57
(1993) 507-511.
[200] H. Niehus, G. Comsa, Surf. Sci. 93 (1980) L147-L150.
[201] P.R. Norton, R.L. Tapping, J.W. Goodale, J. Vac. Sci. Technol. 14 (1977) 446-
451.
[202] D. Ferri, T. Burgi, A. Baiker, J. Phys. Chem. B 105 (2001) 3187-3195.
[203] B. Minder, M. Schurch, T. Mallat, A. Baiker, T.P.A. Heinz, J. Catal. 160
(1996) 261.
[204] C. Keresszegi, T. Burgi, T. Mallat, A. Baiker, J. Catal. 211 (2002) 244-251.
[205] R.L. Augustine, S.K. Tanielyan, J. Mol. Catal. A: Chem. 118 (1997) 79-87.
[206] U. Bohmer, F. Franke, K. Morgenschweis, T. Bieber, W. Reschetilowski, Catal.
Today 60 (2000) 167-173.
[207] X.H. Li, X. You, P.L. Ying, J.L. Xiao, C. Li, Top. Catal. 25 (2003) 63-70.
[208] J.R.G. Perez, J. Malthete, J. Jacques, Comptes Rendus De L Academie Des
Sciences Serie Ii 300 (1985) 169-172.
[209] Y. Nitta, A. Shibata, Chem. Lett. (1998) 161-162.
[210] Y. Nitta, K. Kobiro, Y. Okamoto, Stud. Surf. Sci. Catal. 108 (1997) 191-198.
[211] Y. Nitta, K. Kobiro, Chem. Lett. (1995) 165-165.
[212] Y. Nitta, Y. Ueda, T. Imanaka, Chem. Lett. (1994) 1095-1098.
[213] Y. Nitta, K. Kobiro, Y. Okamoto, Proc. 70th Ann. Meeting Chem. Soc. Japan,
1996, p. 573.
[214] H.-U. Blaser, M. Müller, Stud. Surf. Sci. Catal. 59 (1991) 73.
[215] Y. Nakamura, Bull. Chem. Soc. Japan 16 (1941) 367.
[216] T.J. Hall, P. Johnston, W.A.H. Vermeer, S.R. Watson, P.B. Wells, 11th
International Congress On Catalysis - 40th Anniversary, Pts A And B, 1996, pp.
221-230.
[217] P.J. Collier, T.J. Hall, J.A. Iggo, P. Johnston, J.A. Slipszenko, P.B. Wells, R.
Whyman, Chem. Commun. (1998) 1451-1452.
[218] J. Haglund, A.F. Guillermet, G. Grimvall, M. Korling, Physical Review B 48
(1993) 11685-11691.
[219] T.J. Hall, P. Johnston, W.A.H. Vermeer, S.R. Watson, P.B. Wells, Stud. Surf.
Sci. Catal. 101 (1996) 221.
[220] F. Gao, L. Chen, M. Garland, J. Catal. 238 (2006) 402-411.
[221] Swanson, Tatge, Natl. Bur. Stand. (U.S.), Circ. 539 (1953) 31.
[222] Hull, Physical Review 17 (1921) 571.
118

[223] R. Noyori, H. Takaya, Acc. Chem. Res. 23 (1990) 345-350.
[224] M. Tamura, H. Fujihara, J. Am. Chem. Soc. 125 (2003) 15742-15743.
[225] A.G. Hu, G.T. Yee, W.B. Lin, J. Am. Chem. Soc. 127 (2005) 12486-12487.
[226] S.U. Son, Y. Jang, K.Y. Yoon, E. Kang, T. Hyeon, Nano Lett. 4 (2004) 1147-
1151.
119

I declare that the current Ph.D. thesis has been done by myself.
Alexander Kraynov,
Date: 24.07.2006
120