School of Engineering and Science - Jacobs University
School of Engineering and Science - Jacobs University
School of Engineering and Science - Jacobs University
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
Preparation <strong>and</strong> Characterization <strong>of</strong> Pt <strong>and</strong> Pd Nanoclusters Modified with<br />
Chiral Lig<strong>and</strong>s: Examination <strong>of</strong> Catalytic Activity in Hydrogenation <strong>of</strong><br />
Ethyl Pyruvate<br />
by<br />
Alex<strong>and</strong>er Kraynov<br />
A thesis submitted in partial fulfilment<br />
<strong>of</strong> the requirements for the degree <strong>of</strong><br />
Doctor <strong>of</strong> Philosophy<br />
in Chemistry<br />
Approved, Thesis Committee<br />
Pr<strong>of</strong>. Dr. Ryan M. Richards<br />
Name <strong>and</strong> title <strong>of</strong> chair<br />
Pr<strong>of</strong>. Dr. Thomas Nugent<br />
Name <strong>and</strong> title <strong>of</strong> committee member<br />
Pr<strong>of</strong>. Dr. Stefan Tautz<br />
Name <strong>and</strong> title <strong>of</strong> committee member<br />
Dr. N. Waldoefner<br />
Name <strong>and</strong> title <strong>of</strong> committee member<br />
27 September 2006<br />
<strong>School</strong> <strong>of</strong> <strong>Engineering</strong> <strong>and</strong> <strong>Science</strong>
Table <strong>of</strong> Contents<br />
Acknowledgment<br />
Abstract<br />
IV<br />
V<br />
Chapter 1<br />
Introduction to the enantioselective catalysis 1<br />
1.1 General aspects: optical activity <strong>and</strong> chirality 2<br />
1.2 Chirality in our life 4<br />
1.3 Obtaining pure enantiomers 4<br />
1.3.1 Separation <strong>of</strong> enantiomers 4<br />
1.3.2 Asymmetric synthesis 7<br />
1.3.3 Asymmetric catalysis 7<br />
Chapter 2<br />
Introduction to nanoscale materials 18<br />
2.1 General information <strong>and</strong> definitions 18<br />
2.2 Methods <strong>of</strong> colloid <strong>and</strong> supported nanoparticles preparation 21<br />
2.3 Preparation <strong>of</strong> nanoclusters on a heterogeneous support 24<br />
Chapter 3<br />
Current thesis in the EU-COST project: aims <strong>of</strong> the work 25<br />
3.1 Introduction <strong>and</strong> goals setting 25<br />
Chapter 4<br />
Cinchonidine modified Pt colloidal nanoparticles: characterization <strong>and</strong><br />
catalytic properties 28<br />
4.1 Introduction 28<br />
4.2 Experimental 29<br />
4.3 Results <strong>and</strong> discussion 31<br />
4.3.1 General characterization 31<br />
4.3.2 FTIR investigation <strong>of</strong> cinchonidine adsorbed on Pt 35<br />
4.3.3 Investigation <strong>of</strong> catalytic behavior <strong>of</strong> cinchonidine<br />
modified Pt nanoclusters 43<br />
4.4 Summary 53<br />
Chapter 5<br />
Cinchonidine modified Pt colloidal nanoparticles immobilized on a<br />
heterogeneous support 55<br />
5.1 Introduction 55<br />
5.2 Experimental 56<br />
5.3 Results <strong>and</strong> discussion 57<br />
5.3.1 Role <strong>of</strong> catalyst activation in obtaining enantiomeric<br />
excess 57<br />
5.3.2 Comparison <strong>of</strong> catalysts <strong>and</strong> proposed model <strong>of</strong><br />
catalyst activation 58<br />
5.3.3 Catalyst stability <strong>and</strong> reuse tests 61<br />
5.3.4 IR investigation <strong>and</strong> proposed model <strong>of</strong> activation 63<br />
ii
5.4 Summary 66<br />
Chapter 6<br />
Cinchonidine modified Pd colloidal nanoparticles 67<br />
6.1 Introduction 67<br />
6.2 Experimental 68<br />
6.3 Results <strong>and</strong> discussion 69<br />
6.4 Summary 74<br />
Chapter 7<br />
Quiphos modified Pt <strong>and</strong> Pd colloidal nanoparticles 75<br />
7.1 Introduction 75<br />
7.2 Experimental 76<br />
7.3 Results <strong>and</strong> discussion 77<br />
7.4 Summary 85<br />
Chapter 8<br />
“Quiphos-spider” modified Pt colloidal nanoparticles 86<br />
8.1 Introduction 86<br />
8.2 Experimental 86<br />
8.3 Results <strong>and</strong> discussion 87<br />
8.4 Summary 89<br />
Chapter 9<br />
Diphos <strong>and</strong> diop modified Pt <strong>and</strong> Pd colloidal nanoparticles 90<br />
9.1 Introduction 90<br />
9.2 Experimental 90<br />
9.3 Results <strong>and</strong> discussion 91<br />
9.4 Summary 102<br />
Chapter 10<br />
Binap <strong>and</strong> synphos modified Pt colloidal nanoparticles 103<br />
10.1 Introduction 103<br />
10.2 Experimental 103<br />
10.3 Results <strong>and</strong> discussion 104<br />
10.4 Summary 111<br />
References 112<br />
iii
Acknowledgment<br />
I extend my sincere gratitude <strong>and</strong> appreciation to many people who made this doctoral<br />
thesis possible. Special thanks are due to my supervisor Pr<strong>of</strong>. Ryan Richards for his<br />
timely advices <strong>and</strong> cooperation.<br />
Thanks are also due to my former <strong>and</strong> current lab mates Mr. A. Suchopar for help with<br />
experiment organisation, Dr. K. Zhu <strong>and</strong> Dr. L. D’Souza <strong>and</strong> Dr. F. Pozgan, for many<br />
useful advices <strong>and</strong> directing, Dr. J. C. Hu <strong>and</strong> Mrs. L. F. Chen for general help <strong>and</strong><br />
cooperation in lab work, Mr. L. Zhi <strong>and</strong> Mrs. Helga van Eys-Schaefer with Mrs. Kirsten<br />
Töbe (Fraunh<strong>of</strong>er-Institut für Fertigungstechnik und Angew<strong>and</strong>te Materialforschung,<br />
Bremen) for very important <strong>and</strong> high quality TEM photographs.<br />
Author thanks Pr<strong>of</strong>. Gerard Buono <strong>and</strong> Dr. Didier Nuel <strong>of</strong> Université Aix-Marseille III,<br />
France for providing the quiphos <strong>and</strong> quiphos-spider lig<strong>and</strong>s <strong>and</strong> for useful discussion<br />
concerning the quiphos modifier, Dr. Virginie Ratovelomanana-Vidal (Laboratoire de<br />
Chimie Organique, E.N.S.C.P., UMR) for providing the synphos lig<strong>and</strong> <strong>and</strong> her<br />
colleague Dr. Véronique Michelet for useful discussion about synphos lig<strong>and</strong>.<br />
I would like to thank Pr<strong>of</strong>. U. Kortz (International <strong>University</strong> Bremen, Bremen) <strong>and</strong> Dr.<br />
E. Kirilina (Physikalisch-Technische Bundesanstalt, Berlin) for useful discussions<br />
about NMR spectra, Pr<strong>of</strong>. Thomas Nugent (International <strong>University</strong> Bremen, Bremen)<br />
<strong>and</strong> Pr<strong>of</strong>. S. Tautz (International <strong>University</strong> Bremen, Bremen) for useful discussions<br />
about organic chemistry <strong>and</strong> surface science, correspondingly, as well as for being my<br />
work supervisors, Dr. T. Mallat <strong>and</strong> Pr<strong>of</strong>. A. Baiker (Swiss Federal Institute <strong>of</strong><br />
Technology, Zurich) for useful discussions about cinchonidine on Pd <strong>and</strong> Pt systems,<br />
Dr. Achim Gelessus (International <strong>University</strong> Bremen, Bremen) for help with work<br />
with Gaussian 03 s<strong>of</strong>tware, Dr. L. G. Gordeeva (Boreskov Institute <strong>of</strong> Catalysis,<br />
Novosibirsk) for discussions about surface chemistry <strong>of</strong> alumina <strong>and</strong> silica materials,<br />
Dr. A. G. Okunev (Boreskov Institute <strong>of</strong> Catalysis, Novosibirsk) for discussion about<br />
kinetics <strong>of</strong> heterogeneous reactions, Dr. Serguei Soubatch (International <strong>University</strong><br />
Bremen, Bremen) for discussions about chemical physics <strong>of</strong> surfaces.<br />
I am highly indebted to Dr. Yu.I.Aristov (Boreskov Institute <strong>of</strong> Catalysis, Novosibirsk),<br />
Pr<strong>of</strong>. S. A. Dzuba <strong>and</strong> Mrs. R. I. Ratushkova (Chair <strong>of</strong> Chemical <strong>and</strong> Biological<br />
Physics, Novosibirsk State <strong>University</strong>, Novosibirsk) for introducing me to the world <strong>of</strong><br />
science <strong>and</strong> right directing <strong>and</strong> general support.<br />
My gratitude goes to the International <strong>University</strong> Bremen for providing me a financial<br />
support during my Ph.D. studies <strong>and</strong> making this work possible.<br />
Finally, my special gratitude goes to my wife Mrs. Liliya Kulishova <strong>and</strong> my family for<br />
their cooperation, support <strong>and</strong> encouragement.<br />
iv
Abstract<br />
Studies <strong>of</strong> the interactions <strong>of</strong> organic substrates with nanoscale metals <strong>and</strong> metal oxides<br />
are <strong>of</strong> fundamental importance for both fundamental scientific underst<strong>and</strong>ing <strong>and</strong> the<br />
development <strong>of</strong> commercial applications. Additionally, the development <strong>of</strong><br />
heterogeneous (solid/liquid or solid/gas) enantioselective catalysts is <strong>of</strong> great interest<br />
for the pharmaceutical industry. Here, the adsorption <strong>of</strong> chiral lig<strong>and</strong>s on metal surfaces<br />
to develop heterogeneous asymmetric catalysts is presented. Functionalizing metal<br />
nanoparticles with chiral lig<strong>and</strong>s facilitates increased surface coverage <strong>and</strong> the observed<br />
activity <strong>of</strong> these catalysts is significantly higher than their conventional phase<br />
counterparts.<br />
The first two chapters are an introduction to the enantioselective catalysis <strong>and</strong><br />
nanosized materials. As described in the chapter 3, the primary objectives <strong>of</strong> this work<br />
are<br />
1. to synthesize Pt <strong>and</strong> Pd nanoclusters modified by specific chiral lig<strong>and</strong>s by<br />
`bottom up' chemical synthesis,<br />
2. to determine the geometric orientation <strong>of</strong> adsorbed modifier with respect to<br />
the metal surface via DRIFTS (Diffuse Reflectance Infra-red Fourier<br />
Transform Spectroscopy) <strong>and</strong> molecular modeling,<br />
3. test enantioselective activity <strong>of</strong> the obtained samples in the hydrogenation <strong>of</strong><br />
ethyl pyruvate. The special effort has been given for investigation <strong>of</strong><br />
cinchonidine as modifier (chapter 4, 5, 6).<br />
Among the following chiral lig<strong>and</strong>s: quiphos (chapter 7), “quiphos-spider” (chapter 8),<br />
binap <strong>and</strong> synphos (chapter 10), only diop (chapter 9) was found to be able to induce<br />
low enantioselectivity, that was qualitatively explained through comparison with<br />
cinchonidine/Pt system.<br />
v
Chapter 1<br />
Introduction to the enantioselective<br />
catalysis<br />
1.1 General aspects: optical activity <strong>and</strong> chirality<br />
Any material that rotates the plane <strong>of</strong> polarized light is said to be optically active <strong>and</strong><br />
molecule <strong>of</strong> this materials is nonsuperimposable on its mirror image (Fig. 1-1). If a<br />
molecule is superimposable on its mirror image, the compound does not rotate the<br />
plane <strong>of</strong> polarized light; it is optically inactive. The property <strong>of</strong> nonsuperimposability <strong>of</strong><br />
an object on its mirror image is called chirality. If a molecule is not superimposable on<br />
its mirror image, it is chiral. Inverse logic is also true, i.e. if a molecule is<br />
superimposable on its mirror image, it is achiral. This is a necessary <strong>and</strong> sufficient<br />
requirement.<br />
Fig. 1-1. Superimposable (up) <strong>and</strong> nonsuperimposable (bottom) models <strong>of</strong> molecules.<br />
If a molecule is nonsuperimposable on its mirror image, the mirror image must be a<br />
different molecule, since superimposability is the same as identity. Both real molecule<br />
<strong>and</strong> its mirror image are called enantiomers. Enantiomers have identical (without<br />
considering interaction between elemental particles i.e. weak interaction [1]) physical<br />
<strong>and</strong> chemical properties except in two important respects:<br />
1. They rotate the plane <strong>of</strong> polarized light in opposite direction, though in equal<br />
amounts.<br />
2. They react at different rate with other chiral compounds. These rates may be so<br />
close together that the distinction is practically useless, or reaction rate <strong>of</strong> one<br />
enantiomer might be significantly far from another one. This is the reason that
many compounds are biologically active while their enantiomers are not active<br />
or lead to different reaction results.<br />
The mixture <strong>of</strong> enantiomers in equal amounts is called racemic mixture or racemate.<br />
The separation <strong>of</strong> a racemic mixture into its two enantiomers is called resolution. It is<br />
important to mention that the presence <strong>of</strong> optical activity always proves that a given<br />
compound is chiral, but its absence does not prove that the compound is achiral. A<br />
compound that is optically inactive may be achiral, or it may be a racemic mixture.<br />
Optically active compound may be classified into several categories.<br />
1. Compound with a chiral carbon atom. If there is only one carbon atom with<br />
four different substitutes (so called chiral carbon atom), the molecule must be<br />
optically active. The optical activity has been detected even for compounds,<br />
where one group is hydrogen <strong>and</strong> another deuterium i.e. 1-butanol-1-d ( Fig 1-2)<br />
[2].<br />
C 2 H 5<br />
H<br />
C<br />
D<br />
Fig. 1-2. 1-butanol-1-d [3].<br />
2. Compound with other quadrivalent chiral atoms. A molecule containing an<br />
atom that has four bonds pointing to the corner <strong>of</strong> a tetrahedron will be optically<br />
active if the four groups are different. Among atoms in this category are Si [4],<br />
Ge [5], Sn [5] <strong>and</strong> N (in quaternary salts or N-oxides).<br />
3. Compounds with tervalent chiral atoms. Atoms with pyramidal bonding might<br />
be expected to be optically active if the atom is connected to three different<br />
groups, since the unsaturated pair <strong>of</strong> electrons is analogous to a fourth group.<br />
For example, a secondary or tertiary amine (Fig. 1-3) where X, Y <strong>and</strong> Z are<br />
different. However due to the rapid oscillation <strong>of</strong> the lone pair from one side <strong>of</strong><br />
the XYZ plane to the other <strong>and</strong> following conversion <strong>of</strong> the molecule into its<br />
enantiomers it is very difficult to isolate <strong>and</strong> even detect an enantiomer. The<br />
inversion is less rapid in substituted ammonias [6-8] <strong>and</strong> phosphates [9-12] <strong>and</strong><br />
in suitable cases the inversion energetic barrier can be increased <strong>and</strong> such<br />
enantiomers could be detected <strong>and</strong> separated at room temperature [13, 14].<br />
OH<br />
Z<br />
N<br />
X Y<br />
Fig. 1-3. Chiral amine [3].<br />
4. Suitably substituted adamantine. Adamantines bearing four different substitutes<br />
at the bridgehead positions are chiral <strong>and</strong> optically active <strong>and</strong> could be resolved<br />
[15]. This type <strong>of</strong> molecule (Fig. 1-4) is a kind <strong>of</strong> exp<strong>and</strong>ed tetrahedron <strong>and</strong> has<br />
the same symmetry properties as any other tetrahedron.<br />
CH 3<br />
H<br />
COOH<br />
Br<br />
Fig. 1-4. Example <strong>of</strong> adamantane [3].<br />
2
5. Restricted rotation giving rise to perpendicular dissymmetric planes. Some<br />
compound what do not have asymmetric atoms are nevertheless chiral because<br />
their structure could be represented as two perpendicular planes neither <strong>of</strong><br />
which can be bisected by a plane <strong>of</strong> symmetry (Fig. 1-5). There are few<br />
examples for illustration below. Biphenyls containing four large groups can not<br />
freely rotate about the central bond because <strong>of</strong> steric hindrance (Fig. 1-6). In<br />
such compound the two rings are in perpendicular planes, that fits to the<br />
structure in Fig. 1-5.<br />
Fig. 1-5. Two perpendicular planes [3].<br />
There are few examples for illustration below.<br />
Cl<br />
O 2 N<br />
COOH<br />
HOOC<br />
NO 2<br />
Cl<br />
O 2 N<br />
COOH<br />
HOOC<br />
NO 2<br />
A<br />
P(Ph) 2<br />
P(Ph) 2<br />
(Ph) 2 P<br />
(Ph) 2 P<br />
Mirror<br />
Fig. 1-6. Chiral 3'-chloro-2',6'-dinitrobiphenyl-2,6-dicarboxylic acid (A) <strong>and</strong> binap (B)<br />
[3].<br />
6. Chirality due to a helical shape. Some molecules existing in helical shape with<br />
left- or right-h<strong>and</strong>ed orientation (Fig. 1-7). The entire molecule is usually less<br />
that one full turn <strong>of</strong> the helix, but this does not alter the possibility <strong>of</strong> left- <strong>and</strong><br />
right-h<strong>and</strong>edness.<br />
B<br />
3
Fig. 1-7. Hexahelicene [3].<br />
7. Chirality caused by restriction rotation <strong>of</strong> other types. In this case (e.g. Fig. 1-8.)<br />
chirality results because the benzene ring can not rotate in such a way that the<br />
carboxyl group goes through the alicyclic ring.<br />
HOOC<br />
(CH 2 ) 10<br />
Fig. 1-8. Example <strong>of</strong> chiral molecule with restricted rotation [3].<br />
1.2 Chirality in our life<br />
Chemical compound like drugs, fragrances, agrochemicals, etc. are needed in<br />
enantiopure form, because the all living organisms are built from chiral molecules, e.g.<br />
DNA. The enantiopure drugs, for instance, thus can work more efficiently than in<br />
racemate form because they influence on specific biochemical reactions by a lock-key<br />
mechanism. When one enantiomer is needed as an active compound, the other one is<br />
ballast, in the best case, or can induce a poison effect, like the (S)-thalidomide (where it<br />
was in racemic mixture with (R)- thalidomide), which caused birth defects in the<br />
beginning <strong>of</strong> 1960s. In present days drugs manufacturers need enantiopure chemical<br />
compound for producing enantiopure drugs. Thus the market related to the singleenantiomer<br />
drug business reached 40 % <strong>of</strong> total drug market in 2000. Whereas it<br />
covered 4 % only in 1985.<br />
1.3 Obtaining pure enantiomers<br />
Since huge number <strong>of</strong> natural compound are (e.g. cinchonidine etc.) chiral thus one <strong>of</strong><br />
the simplest way for obtaining chiral compounds is through extraction from nature i.e.<br />
from plants, living organisms etc. However it is clear that not all needed chiral<br />
compounds are available from the natural “chiral pool”. At the same time difficulties in<br />
obtaining chiral compound in sufficient amounts might lead to increases <strong>of</strong> products<br />
cost. Therefore methods for obtaining required enantiopure compounds have to be<br />
developed.<br />
1.3.1 Separation <strong>of</strong> enantiomers<br />
The resolution <strong>of</strong> the racemic mixture could be performed through removal or<br />
decomposition partially or completely <strong>of</strong> one enantiomer.<br />
Mechanical separation<br />
In 1848 L. Pasteur has shown an opportunity <strong>of</strong> mechanical separation <strong>of</strong> crystals <strong>of</strong> Na<br />
salt <strong>of</strong> racemic tartaric acid. In this example crystals <strong>of</strong> enantiomers form separately<br />
<strong>and</strong> visually look different, like mirror reflections <strong>of</strong> each other (Fig. 1-9), obtaining<br />
4
crystals are called enantiomorphic crystals. This method is known to as spontaneous<br />
resolution by crystallization.<br />
Fig. 1-9. Enantiomorphic crystals. Abstract model [16].<br />
Preferential crystallization by inoculation<br />
The selective crystallization <strong>of</strong> a single enantiomer is possible to perform in the<br />
presence <strong>of</strong> traces <strong>of</strong> the same enantiomer that is to be separated. In this case an<br />
oversaturated racemic solution is "seeded" in small amount <strong>of</strong> pure enantiomer <strong>and</strong> the<br />
resulting precipitate will have domination <strong>of</strong> the enantiomer in the seed. However, with<br />
the exception <strong>of</strong> the separation <strong>of</strong> glutamic acid <strong>and</strong> <strong>of</strong> a copper complex <strong>of</strong> DLaspartic<br />
acid, this method has been found to be impractical or resulted in portion<br />
resolution only.<br />
Biochemical processes<br />
The chiral compounds that react at different rates with the two enantiomers may be<br />
present in a living organism. For example, in 1858 Pasteur discovered that aqueous<br />
solution <strong>of</strong> racemic tartaric acid after contact with the mould fungus (Penicillium<br />
glaucum) slowly became enantioenriched. The mould was preferentially metabolizing<br />
the right rotatory enantiomer. This method is limited, since it is necessary to find the<br />
proper organism <strong>and</strong> since one <strong>of</strong> the enantiomer is destroyed in the process. However,<br />
when the proper organism is found, the method leads to a high extent <strong>of</strong> resolution<br />
since biological processes are usually very stereoselective. Another disadvantage <strong>of</strong> this<br />
method is that only dilute solutions can be used.<br />
Conversion in diastereomers<br />
If the racemic mixture to be resolved contains a carboxyl group (<strong>and</strong> no strongly basic<br />
group), it is possible to form a salt with an optically active base. Fig. 1-10 illustrates<br />
reaction <strong>of</strong> S-brucine with racemic hydroxypropanoic acid produced a mixture <strong>of</strong> two<br />
salts having the configuration SS <strong>and</strong> RS. Although the acids are enantiomers, the salts<br />
are diastereomers <strong>and</strong> have different properties e.g. solubility. The mixture <strong>of</strong><br />
diastereometric salts is allowed to crystallize from a suitable solvent.<br />
5
R<br />
H<br />
COOH<br />
C OH<br />
CH 3<br />
H<br />
COO - Bricine-H +<br />
C OH<br />
CH 3<br />
S-brucine<br />
R<br />
S<br />
S<br />
COO - Bricine-H +<br />
COOH<br />
HO C H<br />
HO C H<br />
CH 3<br />
CH 3<br />
S S<br />
Fig. 1-10 formation <strong>of</strong> diastereometric salts [3].<br />
Since the solubilities are different, the initial crystals formed will be richer in one<br />
diastereomer. Once the two diastereomers have been separated, it is easy to convert the<br />
salt back to the free acids <strong>and</strong> the recovered base can be used again. Advantages <strong>of</strong> this<br />
approach are that naturally occurring optically active bases (mostly alkaloids: brucine,<br />
ephedrine, morphine, etc.) are readily available. However, unfortunately, the difference<br />
in solubilities is rarely if ever great enough to effect total separation with one<br />
crystallization.<br />
Differential absorption<br />
When a racemic mixture is placed on a chromatographic column consisting <strong>of</strong> chiral<br />
phase, then, in principal the enantiomers should move along the column at different<br />
rates <strong>and</strong> should be separated. This technique is widely in use in different<br />
chromatography applications: GC, HPLS.<br />
Chiral recognition<br />
In order to separate enantiomers from racemic mixture the use <strong>of</strong> chiral host to form<br />
diastereometric inclusion is possible. Some hosts can form an inclusion with one<br />
enantiomer only or this formation is faster then with another enantiomer. For example,<br />
chiral host (Fig. 1-11-A) can partially resolve the racemic amine salt [17] (Fig. 1-11-B).<br />
O<br />
O<br />
O<br />
O<br />
O<br />
O<br />
Ph<br />
CH 3<br />
C<br />
+<br />
NH 3<br />
-<br />
PF 6<br />
H<br />
A<br />
B<br />
Fig. 1-11. Chiral host-A <strong>and</strong> guest molecule –B [3].<br />
Kinetic resolution<br />
Since enantiomers react with chiral compounds at different rates, it is sometimes<br />
possible to effect a partial separation by stopping the reaction before completion. An<br />
example is the resolution <strong>of</strong> allylic alcohol [18] (Fig. 1-12-A) with one enantiomer <strong>of</strong> a<br />
chiral epoxidizing agent, in this case the discrimination was extreme. One enantiomer<br />
was converted to the epoxide <strong>and</strong> the other was not, the ratio being more than 100.<br />
Reactions catalyzed by enzymes can be utilized for this kind <strong>of</strong> resolution.<br />
6
Bu<br />
O<br />
Bu<br />
Bu<br />
OH<br />
OH<br />
OH<br />
racemic (R)-enantiomer (S)-enantiomer<br />
A B<br />
Fig. 1-12. Kinetic resolution <strong>of</strong> racemic allylic alcohol –A, yielding its (S) enantiomer<br />
(B) [3].<br />
1.3.2 Asymmetric synthesis<br />
There are two basic ways to synthesize a chiral compound. The first way is to begin<br />
with single stereoisomer <strong>and</strong> to use a synthesis that does not affect the chiral centre.<br />
The other basic method is called asymmetric synthesis. The definition <strong>of</strong> asymmetric<br />
synthesis is as follows: "Asymmetric synthesis is a reaction in which an achiral unit in<br />
an ensemble <strong>of</strong> substrate molecules is converted by a reactant into a chiral unit in such<br />
a manner that the stereoisomeric products are formed in unequal amounts" [19].<br />
1.3.3 Asymmetric catalysis<br />
As was mentioned above in order to synthesize a chiral compound another chiral<br />
substance is needed. Before such a chiral substance was considered as a reactant,<br />
however it can also be a chiral catalyst in asymmetrical synthesis. According to the<br />
definition catalyst influences on reaction but keeps its structure unchanged in the end <strong>of</strong><br />
reaction. Thus, a small amount <strong>of</strong> specific chiral catalyst can induce synthesis <strong>of</strong> chiral<br />
product in large scales. This strategy attracts much attention from industry <strong>and</strong> science,<br />
in fact, the number <strong>of</strong> publication on asymmetric syntheses has been increasing<br />
exponentially every year since the 80s [20].<br />
Willian S. Knowles, Ryoji Noyori <strong>and</strong> Barry Sharpless were awarded the Nobel Prize<br />
for chemistry in 2001 for their contribution in asymmetrical catalysis, sealing the<br />
importance <strong>of</strong> this field in contemporary chemistry <strong>and</strong> chemical technology.<br />
Asymmetrical reactions catalyzed with natural <strong>and</strong> man-made chiral catalysts becomes<br />
main-stream in different fields <strong>of</strong> industry (chemical, pharmaceutical, etc.).<br />
Asymmetrical catalysis can be divided into two main areas: homogeneous <strong>and</strong><br />
heterogeneous. It is clear from terms that in the case <strong>of</strong> homogeneous catalysis the<br />
important reactions are occurring in one phase, e.g. molecules in a solution. Whereas,<br />
in the case <strong>of</strong> heterogeneous catalysis reactions are occurring on the interface <strong>of</strong> two<br />
phases e.g. solid-liquid, solid-gas or in other words, on the surface <strong>of</strong> solid bodies.<br />
It is important to mention here briefly the important requirements [21] for<br />
enantioselective catalysts, since they are applied either for heterogeneous <strong>and</strong><br />
homogeneous catalysts.<br />
Enantioselectivity, expressed as enantiopurity <strong>of</strong> a system in terms <strong>of</strong> enantiomeric<br />
excess (ee, %). A catalyst has to be able to induce enantioselectivity in the range <strong>of</strong> 99<br />
% for further pharmaceutical applications. However ee’s > 80 % are acceptable also if<br />
further enantiopurification could be easily performed, e.g. via recrystallization.<br />
Catalyst productivity, characterized in terms <strong>of</strong> turnover number (TON). For example,<br />
in case <strong>of</strong> homogeneous enantioselective hydrogenation reactions the following values<br />
<strong>of</strong> TON can be used for the catalyst productivity classification. TON’s > 1000 for<br />
small-scale, high-value products <strong>and</strong> > 50000 for large-scale or less expensive products.<br />
7
Catalyst activity, expressed in the turnover frequency (TOF). For large-scale synthesis<br />
TOF > 10000 1/h are acceptable (under high pressure).<br />
Separation, should be able to perform by a simple low cost operation e.g. settling,<br />
filtration, distillation <strong>and</strong> at least 95 % <strong>of</strong> catalyst should be recovered.<br />
Stability, catalyst should be chemically <strong>and</strong> mechanically stable without loss <strong>of</strong> activity<br />
<strong>and</strong> enantioselectivity.<br />
Price <strong>of</strong> catalyst, consisting <strong>of</strong> price <strong>of</strong> expensive chiral lig<strong>and</strong> (homogeneous<br />
catalysis) <strong>and</strong> depending on type <strong>of</strong> catalyst in heterogeneous catalysis.<br />
Availability <strong>of</strong> the catalyst at the right time in an appropriate quantity.<br />
These criteria determine the choice <strong>of</strong> chiral catalyst for the specific reaction, as was<br />
illustrated in the synthesis <strong>of</strong> HPB (Ethyl (R)-2-Hydroxy-4-Phenylbutyrate) ester [22].<br />
In the following paragraphs the description <strong>of</strong> the most important aspects <strong>and</strong> ideas is<br />
summarized, however more detailed reviews can be found in the references that are<br />
given below.<br />
Homogeneous asymmetric catalysis<br />
The homogeneous enantioselective catalysis is widely used in industry covers the<br />
following asymmetric reaction [20]: hydrogenation, hydrosilylation (<strong>and</strong> related<br />
reaction), isomerisation <strong>of</strong> allylamines, carbometallation, oxidation (<strong>and</strong> related),<br />
carbonylation, polymerization <strong>and</strong> other that are not mentioned here.<br />
The origin <strong>of</strong> homogeneous asymmetric catalysis is in chiral organometallic complex<br />
catalyzing enantioselectivly specific types <strong>of</strong> reactions performed in one phase, e.g. in<br />
solutions. The electronic properties <strong>of</strong> the transition metals promote the formation <strong>of</strong><br />
organometallic complex (e.g. binap-Ru Fig. 1-13-A) with lower activation barrier for<br />
the reaction, whereas steric properties <strong>of</strong> used lig<strong>and</strong> create a chiral environment around<br />
the metal, leading to the specific reaction pathway <strong>and</strong> ending in enantioselectivity. The<br />
model <strong>of</strong> the reaction pathway for the C=O bond hydrogenation in the original work <strong>of</strong><br />
Noyori [23], here we only mention that formation <strong>of</strong> the transition state Fig 1-13-B<br />
which is responsible for enantioselectivity results in excellent, up to 98-99 % <strong>of</strong> ee [23].<br />
H<br />
Ar 2 R<br />
H 1 O<br />
P Cl 2<br />
CH H H<br />
N R 2<br />
H<br />
Ru<br />
N<br />
P Cl<br />
N R 3<br />
X(R 3 P) 2 Ru<br />
H 2<br />
Ar 2<br />
R 4<br />
H 2 N<br />
Ar=C 6 H 5 ; R 1 =R 4 =H; R 2 =R 3 =C 6 H 5<br />
X=H, OR, etc<br />
A<br />
B<br />
Fig. 1-13 Chiral binap-Ru complex (A) <strong>and</strong> based on it transition state (B) in<br />
mechanism <strong>of</strong> asymmetrical hydrogenation <strong>of</strong> C=O bond [23].<br />
Binap-Ru complex is not only one example <strong>of</strong> chiral homogeneous catalysis, in general,<br />
the suitable combination <strong>of</strong> a metal species <strong>and</strong> chiral lig<strong>and</strong> might be able to induce<br />
enantioselectivity. Here (Fig. 1-14) we give some examples <strong>of</strong> lig<strong>and</strong>s: quiphos [9-12],<br />
diop [24], BPPFA [25], synphos [26-35], PPCP [36], BPPM [37] which were<br />
successfully applied in corresponding asymmetrical reactions. Some <strong>of</strong> these lig<strong>and</strong>s<br />
were used as chiral modifiers <strong>of</strong> surface <strong>of</strong> Pt <strong>and</strong> Pd nanoparticles in the current thesis,<br />
please see corresponding chapters below.<br />
8
N<br />
N<br />
P<br />
Fe P(C 6H 5 ) 2<br />
O<br />
P<br />
P<br />
N<br />
O O<br />
P(C 6 H 5 ) 2<br />
Quiphos<br />
BPPFA: X= (CH 3 ) 2 N<br />
Diop<br />
O<br />
O PPh 2<br />
(C 6 H 5 ) 2 P<br />
P(C 6 H 5 ) 2<br />
O PPh 2<br />
P(C 6 H 5 ) 2<br />
P(C 6 H 5 ) N<br />
2<br />
O<br />
X<br />
BPPM: X=(CH 3 ) 3 COCO<br />
PPCP<br />
Synphos<br />
Fig. 1-14. Some chiral lig<strong>and</strong>s used in homogeneous asymmetrical catalysis.<br />
In spite <strong>of</strong> the fact that chiral homogeneous catalysts demonstrate excellent<br />
enantioselectivity, due to the separation criterion mentioned above, the disadvantage<br />
this type <strong>of</strong> catalysts appears. In fact, these soluble catalyst are more difficult to<br />
separate <strong>and</strong> h<strong>and</strong>le that the heterogeneous ones. One promising strategy to combine<br />
the best properties <strong>of</strong> the two catalyst types in the heterogenization or immobilization <strong>of</strong><br />
active metal complexes on supports or carriers, which may be separated by filtration or<br />
precipitation. However we leave this approach out <strong>of</strong> the current thesis, but keeping it<br />
in EU-COST project, please see below. Consequently, emphasis has to be given to the<br />
design <strong>of</strong> stable heterogeneous catalysts that are capable <strong>of</strong> high enantioselectivity.<br />
Heterogeneous asymmetric catalysis<br />
To date there have been numerous approaches to the design <strong>of</strong> heterogeneous<br />
asymmetric catalysts, since Schwab <strong>and</strong> coworkers first demonstrated that Cu <strong>and</strong> Ni<br />
could be supported on chiral silica surfaces [38, 39] <strong>and</strong> that the resulting catalysts<br />
could give low enantioselection in the dehydration <strong>of</strong> butan-2-ol.<br />
There are several strategies for the design <strong>of</strong> chiral heterogeneous catalysts, the three<br />
mentioned below, lead to stable catalysts that allow high enantioselectivity.<br />
• Modification <strong>of</strong> metal interfaces with chiral molecules;<br />
• Tethering <strong>of</strong> active homogeneous catalysts to a three-dimensional structure;<br />
• Electrostatic interaction as a means <strong>of</strong> immobilization.<br />
An overview <strong>of</strong> the about three strategies can be found in the work <strong>of</strong> Hutchings [40],<br />
whereas the current thesis focuses more on modification <strong>of</strong> the surface <strong>of</strong> the metal<br />
nanoclusters by chiral molecules.<br />
In contrast to homogeneous enantioselective catalysis the heterogeneous one is rather<br />
young <strong>and</strong> has a limited success in applications. Several factors contributed to this<br />
situation: a) the more complex structure <strong>of</strong> the heterogeneous catalyst surface on which<br />
coexist centers with different catalytic activity <strong>and</strong> selectivity, which can lead to<br />
undesired secondary reactions, b) an increased difficulty to create an effective<br />
asymmetric environment <strong>and</strong> to accommodate it with the multitude <strong>of</strong> reactions that are<br />
interesting to be carried out under enantioselective restrictions.<br />
X<br />
9
In fact, there are only two heterogeneous catalysts that reliably give high<br />
enantioselectivities (90 % e.e. or above). These are Raney nickel (or Ni/SiO 2 ) system<br />
modified with tartaric acid or alanine for hydrogenation <strong>of</strong> β-ketoesters [41] <strong>and</strong> β-<br />
diketones [41], methyl ketones [42, 43] <strong>and</strong> others summarized in the reviews <strong>of</strong> Blaser<br />
et al. [44] <strong>and</strong> platinum or platinum-on-support modified with cinchona alkaloids first<br />
reported by Orito [45-47] for the hydrogenation <strong>of</strong> α-ketoesters (Orito reaction), α-<br />
ketolactones [48-52], α-diketones [53-57], α-keto acetals [58, 59], α- α- α-<br />
trifluoroketones [60-63] <strong>and</strong> linear <strong>and</strong> cyclic α-ketoamides [64-66]. Cinchonidine<br />
modification <strong>of</strong> Pd surface was found to be successful in enantioselective<br />
hydrogenation <strong>of</strong> C=C bonds [67].<br />
Below we give basic aspects <strong>and</strong> models concerning behavior <strong>of</strong> cinchonidine based<br />
heterogeneous catalysts <strong>and</strong> origin <strong>of</strong> its enantioselectivity, since significant part <strong>of</strong> the<br />
current thesis is devoted to cinchonidine modified Pt <strong>and</strong> Pd systems, more detailed<br />
reviews can be found in work <strong>of</strong> Baiker [50, 68], Blazer [44], Murzin [69], Hutchings<br />
[40] <strong>and</strong> others [49].<br />
Conformations <strong>of</strong> cinchonidine molecule<br />
Several geometrical conformations were found for the cinchonidine molecule. The most<br />
important ones are in the relative orientation <strong>of</strong> quinoline <strong>and</strong> quinuclidine moieties<br />
(Fig. 1-15) via changing torsion angles τ 3’,4’,9,8 <strong>and</strong> τ 4’,9,8,N . The two types <strong>of</strong> major<br />
conformers are called closed (Fig. 1-16-A) <strong>and</strong> open (Fig. 1-16-B). The terms “open”<br />
<strong>and</strong> “closed” indicate whether the lone pair <strong>of</strong> quinuclidine nitrogen point away or<br />
towards the quinoline ring, respectively.<br />
Chiral center<br />
8<br />
HO<br />
9<br />
4'<br />
3'<br />
N<br />
N<br />
Quinuclidine<br />
Quinoline<br />
Fig. 1-15. Scheme <strong>of</strong> cinchonidine molecule<br />
It was shown by NMR spectroscopy [70], that cinchonidine molecule can exist in open<br />
<strong>and</strong> closed conformations at room temperature in different solvents. It has to be noted<br />
that there are other conformations (e.g. closed 1, 2, 4 they are not shown here) <strong>of</strong><br />
cinchonidine which belong to the open or closed types.<br />
10
Fig. 1-16. Closed –A <strong>and</strong> open –B types <strong>of</strong> cinchonidine conformers. Hydrogens are<br />
not shown for clarity.<br />
The relative population <strong>of</strong> these conformers depends on polarity <strong>of</strong> the solvent <strong>and</strong><br />
possible chemical interaction (cinchonidine-solvent), in fact as was show by<br />
calculations <strong>and</strong> NMR spectroscopy the open 3 conformer is most abundant in the<br />
solvent with low dielectric constant.<br />
Fig. 1-17. Dependence <strong>of</strong> relative abundance <strong>of</strong> conformer (Open 3) on solvent polarity<br />
determined by NMR spectroscopy <strong>and</strong> calculated using reaction field model [41].<br />
Solvents: (1) benzene, (2) toluene, (e) ethyl ether, (4) tetrahydroguran, (5) aceton, (6)<br />
dimethylformamide, (7) dimethyl sulfoxide, (8) water, (9) ethanol. The calculated<br />
model, does not take into account specific interaction such as the hydrogen boning<br />
between cinchonidine <strong>and</strong> ethanol. The figure is adapted from review <strong>of</strong> Baiker [50].<br />
It is important to note, that in the hydrogenation <strong>of</strong> ketopantolactone over cinchonidinemodified<br />
Pt ee decreases with increasing the dielectric constant <strong>of</strong> a solvent in the very<br />
11
similar way as relative abundance <strong>of</strong> the Open 3 conformer. Such relation is in<br />
agreement with suggestion that the Open 3 plays a crucial role in the mechanism <strong>of</strong> the<br />
enantioselectivity, see below.<br />
Fig. 1-18. Combined plot <strong>of</strong> the enantiomeric excess achieved in the hydrogenation <strong>of</strong><br />
ketopantolactone over cinchonidine-modified Pt [52, 71] (left axis) <strong>and</strong> the population<br />
<strong>of</strong> conformer Open(3) as calculated by density functional theory in combination with a<br />
reaction field model (P Open(3) , right axis) versus the dielectric constant <strong>of</strong> the solvent.<br />
The axis scale is arbitrarily chosen. The solvents are (1) cyclohexane, (2) hexane, (3)<br />
toluene, (4) diethyl ether, (5) tetrahydr<strong>of</strong>urane, (6) acetic acid, (7) ethanol, (8) water,<br />
(9) formamide. The picture is adapted from [70].<br />
Adsorption modes on Pt <strong>and</strong> Pd macro support<br />
Adsorption <strong>of</strong> quinoline (part <strong>of</strong> cinchonidine) <strong>and</strong> dihydrocinchonidine on the surface<br />
<strong>of</strong> Pt (single or polycrystal) was studied by XPS [72, 73], LEED [72], NEXAFS [74],<br />
H/D exchange experiments [75] in vacuum. It was found that adsorption <strong>of</strong> these<br />
compounds takes place preferentially in the flat mode through the π-aromatic system at<br />
room <strong>and</strong> lower temperatures, whereas even at 50 °C on average the quinoline ring was<br />
considerably tilted with respect to the surface [74]. Since experimental conditions there<br />
were far from used in the reactions, Ferri <strong>and</strong> Bürgi [76] performed ATR studies <strong>of</strong> the<br />
cinchonidine adsorption on solid-liquid interface, where 100 nm layer <strong>of</strong> Pt was<br />
deposited on alumina <strong>and</strong> cinchonidine in different concentration in dichloromethane<br />
was used. It was found that at low concentration (10 -6 M) a predominantly flat<br />
adsorption mode prevails (Fig. 1-19-1), at increasing coverage two different tilted<br />
species, α-H abstracted (Fig. 1-19-2) <strong>and</strong> N lone pair bonded (Fig. 1-19-3) cinchonidine<br />
were observed, at higher (10 -3 – 10 -4 M) concentrations all three species coexist on the<br />
Pt surface, this was also confirmed by the RAIRS experiments <strong>of</strong> Zaera <strong>and</strong> et al. [77].<br />
Adsorption <strong>of</strong> cinchonidine on Pt (111) was also studied by STM in vacuum <strong>and</strong> with<br />
presence <strong>of</strong> hydrogen, where mobility <strong>of</strong> cinchonidine was observed at high hydrogen<br />
pressure.<br />
12
Fig. 1-19. Three schematically favored adsorbed orientations <strong>of</strong> cinchonidine on wide<br />
Pt support [76]. Species 1: “π-bonded”, 2: “α-H abstracted”, 3: “N lone pair bonded”.<br />
Hydrogen atoms are omitted.<br />
It is interesting to note, that maximum observed ee values were found at low<br />
cinchonidine concentration, for example, in the works <strong>of</strong> LeBlond [78] <strong>and</strong> Bartok [51,<br />
79].<br />
The Fig. 1-20 illustrates dependence <strong>of</strong> ee on concentration <strong>of</strong> cinchonidine above the<br />
Pt/Al 2 O 3 catalyst.<br />
Fig. 1-20. Effect <strong>of</strong> cinchonidine concentration <strong>of</strong> the ee <strong>and</strong> the initial rate <strong>of</strong><br />
hydrogenation. Filled squares - ee , filled cicles - initial rate in ethyl pyruvate<br />
hydrogenation, open squares - ee <strong>and</strong> open cycles initial rate in bute-2,3-dione<br />
hydrogenation. Reaction conditions: dichlormethane (12.5 ml), reactant (66 mmol),<br />
catalyst 0.25 g, 50 bar H 2 , 20 C, 1200 rpm stirring. The picture is adapted from [80].<br />
The observed dependence <strong>of</strong> ee on cinchonidine concentration can be explained in the<br />
following way, in fact, at low concentration <strong>of</strong> modifier <strong>and</strong> small part <strong>of</strong> Pt surface is<br />
modified, whereas on unmodified sites reaction goes racemically (ee =0). When the<br />
13
surface concentration <strong>of</strong> cinchonidine becomes optimum ([cinchonidine]/[Pt sur ] ≈ 1/20<br />
[78, 81]) the maximum <strong>of</strong> ee is reached. Adsorption <strong>of</strong> cinchonidine in tilted modes<br />
becomes dominated at further increase <strong>of</strong> cinchonidine’s concentration in solution, that<br />
leads to decrease <strong>of</strong> relative concentration <strong>of</strong> π-bonded cinchonidine (active chiral site,<br />
see below) <strong>and</strong> thus decrease <strong>of</strong> ee.<br />
With respect to the adsorption on Pt, cinchonidine demonstrates differences <strong>and</strong><br />
similarities in adsorption on the Pd/Al 2 O 3 (as a thin film) surface [82]. In fact,<br />
cinchonidine was found in adsorption mode with the quinoline moiety nearly parallel to<br />
the Pd surface, likely through the π-system (flat adsorption mode) <strong>and</strong> in the mode with<br />
interaction with Pd through the lone pair <strong>of</strong> the quinoline N. Compared to Pt, the<br />
relative abundance <strong>of</strong> the two species under similar conditions was found to be different<br />
on the two surfaces, such that the π-bonded species are less stable on Pd than on Pt.<br />
Based on the comparison with adsorbed pyridine on Pd <strong>and</strong> Pt, the most obvious<br />
difference in the adsorption <strong>of</strong> cinchonidine on the two metals is that the : “α-H<br />
abstracted” species observed on Pt are absent on Pd. The difference in the diffuseness<br />
<strong>of</strong> the d orbital <strong>of</strong> the metal was proposed to be the origin <strong>of</strong> the different behavior with<br />
respect to the adsorption <strong>of</strong> cinchonidine.<br />
Origin <strong>of</strong> enantioselectivity <strong>and</strong> rate enhancement <strong>and</strong> two-cycle model<br />
Through the systematic variation <strong>of</strong> the cinchona alkaloid structure it was found that<br />
changing the absolute configuration at C-8 (S) <strong>and</strong> C-9 (R) <strong>of</strong> cinchonidine, i.e.,<br />
substituting cinchonidine by the diastereomer cinchonine, alters the chirality <strong>of</strong> the<br />
product from (R)- to (S)-lactate. The most important is the finding that the<br />
enantioselectivity is completely lost upon alkylation <strong>of</strong> the quinuclidine nitrogen atom<br />
which indicates that this center plays a crucial role in the mechanism <strong>of</strong><br />
enantioselection [83].<br />
According to the modern model, based on theoretical studies using quantum chemistry<br />
techniques, at both ab initio <strong>and</strong> semiemperical levels, <strong>and</strong> molecular mechanics [84-<br />
86] enantioselectivity is assumed to be thermodynamically controlled, i.e. by the<br />
difference in the energy <strong>of</strong> adsorption <strong>and</strong> complex formation ΔG 0 ad, <strong>of</strong> the pro-(R),<br />
[EP-CD] R ad <strong>and</strong> pro-(S), [EP-CD] S ad (Fig. 1-21). When adsorption equilibrium is<br />
established the relative surface coverage <strong>of</strong> [EP-CD] R ad <strong>and</strong> [EP-CD] S ad can be<br />
0<br />
ΔΔGad<br />
θ * −(<br />
)<br />
R<br />
RT<br />
expressed as = e . If no kinetic factor controls, i.e. if activation energies <strong>and</strong><br />
θ *<br />
S<br />
pre-exponential factors <strong>of</strong> the subsequent hydrogen addition are the same for pro-(R)<br />
[ R]<br />
[ θ * ]<br />
R<br />
<strong>and</strong> pro-(S) complexes, the ratio <strong>of</strong> surface coverage = . Thus, under these<br />
[ S]<br />
[ θ * ]<br />
S<br />
conditions the origin <strong>of</strong> enantioselectivity should be determined by stability <strong>of</strong> the<br />
intermediate pro-(R) <strong>and</strong> pro-(S) transition states. The question regarding whether the<br />
mechanism is kinetically or thermodynamically controlled, or whether the<br />
thermodynamic <strong>and</strong> kinetic factors are working in concert with each other is not yet<br />
defined.<br />
Enantioselectivity <strong>and</strong> rate enhancement (in factor <strong>of</strong> 10-100 compared with racemic<br />
EP hydrogenation) occur always together in this system. The H-bonding (in Pro-R <strong>and</strong><br />
Pro-S state, Fig. 1-21) would stabilize the half-hydrogenated state, increasing its life<br />
time <strong>and</strong>, hence, its concentration, thereby increasing the observed reaction rate [87].<br />
14
Fig. 1-21. The optimized structures <strong>of</strong> the proposed transition state formed on<br />
interaction <strong>of</strong> cinchonidine’s protonated nitrogen <strong>and</strong> the pyruvate ester [88].<br />
It has to be noted, that in the transition states the α-ketoesters are stabilized in their<br />
half-hydrogenated form [88] <strong>and</strong> cinchonidine (in the flat adsorption mode) is<br />
represented in open 3 conformation in these transition states as was suggested from the<br />
investigation <strong>of</strong> obtained ee with conformation analysis.<br />
Today it is practically accepted that enantioselective hydrogenation proceeds through<br />
the “two-cycle” mechanism as presented in Fig. 1-22. It is assumed, that the active<br />
chiral sites associated with adsorbed cinchonidine [CD] ad on Pt surface. Ethyl pyruvate<br />
from the liquid phase adsorb reversibly on these sites in its two enanti<strong>of</strong>acial<br />
configurations forming the diastereomeric intermediate complexes [EP-CD] R ad <strong>and</strong><br />
[EP-CD] S ad, which upon hydrogenation afford the (R)- <strong>and</strong> (S)- ethyl lactate,<br />
respectively. It has been suggested that the adsorbed cinchonidine interacts with the<br />
adsorbed ethyl pyruvate via hydrogen bonding between quinuclidine N <strong>and</strong> O atom <strong>of</strong><br />
the α-carbonyl moiety [73, 84]. Blaser et al. [89] suggested that for hydrogenation on<br />
unmodified sites, leading to the racemic products, addition <strong>of</strong> the first hydrogen is ratedetermining<br />
whereas for hydrogenation on modified sites, leading to the major<br />
enantiomer, the rate determining step is the addition <strong>of</strong> the second hydrogen.<br />
Fig. 1-22. Two-cycle mechanism suggested for enantiomeric hydrogenation <strong>of</strong> ethyl<br />
pyruvate [EP] over Pt modified by cinchonidine. The [CD] ad present an active chiral<br />
15
site on Pt surface formed by a adsorbed cinchonidine, [EP-CD] R ad <strong>and</strong> [EP-CD] S ad,<br />
represent half-hydrogenated complexes formed by interaction <strong>of</strong> chiral sites with EP,<br />
affording (R)- <strong>and</strong> (S)-ethyl lactate (EL). The picture is adapted from [50].<br />
It has to be noted, that the alternative models (e.g. the Chemical Shielding Model [90-<br />
93]) explaining the origin <strong>of</strong> enantioselectivity were developed, but it was found to be<br />
less successful with respect to the model mentioned above. The detailed comparison <strong>of</strong><br />
existing models with underlining their strong <strong>and</strong> weak point can be found in the work<br />
<strong>of</strong> Vargas [16].<br />
Several features with cinchonidine based systems<br />
Here we briefly mention some features <strong>and</strong> parameters <strong>of</strong> cinchona alkaloid based<br />
heterogeneous catalyst <strong>and</strong> give corresponding references for detailed examination,<br />
whereas some <strong>of</strong> the below mentioned features <strong>and</strong> reaction parameters (metal particle<br />
size, type <strong>of</strong> metal, pressure, conversion dependence on ee <strong>and</strong> catalyst reuse) <strong>and</strong> their<br />
role in enantioselectivity will be under investigation <strong>and</strong> discussion in the following<br />
paragraphs <strong>and</strong> chapters.<br />
Instead <strong>of</strong> Pt, Ir [94, 95] is also a suitable metal for enantioselective hydrogenation <strong>of</strong><br />
α-ketoesters. The Rh/Al 2 O 3 [94] gave 30 % lower ee than Pt/Al 2 O 3 in the hydrogenation<br />
<strong>of</strong> α-ketoesters. The Ru/Al 2 O 3 [96] afforded no enantioselectivity with cinchonidine in<br />
the hydrogenation <strong>of</strong> methyl pyruvate. The use <strong>of</strong> Pd will be discussed in a later<br />
chapter.<br />
The most convenient support material, like Al 2 O 3 , SiO 2 <strong>and</strong> carbon, have been reported<br />
to be suitable for ethyl pyruvate hydrogenation <strong>and</strong> have no particular effect <strong>of</strong><br />
enantioselection [97]. However an ee <strong>of</strong> 72 % was reported in the hydrogenation <strong>of</strong> E-<br />
α-phenylcinnamic support texture <strong>and</strong> a large effect on enantioselectivity [98, 99].<br />
The hydrogen pressure in the range <strong>of</strong> 1-100 bar was found to effect on reaction rate as<br />
the first order dependence, whereas ee increases with plateau [44].<br />
Hydrogenation <strong>of</strong> ethyl pyruvate started to decrease at temperature <strong>of</strong> 70 °C [89],<br />
whereas the drop in ee has been observed above 50 °C [100]. Moreover the adsorption<br />
mode <strong>of</strong> the quinoline ring is changed above 50-60 °C towards Pt (111) surface [89].<br />
Effect <strong>of</strong> concentration <strong>of</strong> ethyl pyruvate induces changes in the reaction order 1 → 0<br />
(saturation) or sometimes maximum [44, 89, 101].<br />
The increase <strong>of</strong> ee in the beginning <strong>of</strong> the reaction is the so called initial transient<br />
period. The nature <strong>of</strong> the effect is still not well understood [40], whereas its<br />
underst<strong>and</strong>ing might be useful in developing a catalyst which induces constant high<br />
enantiomeric excess during reaction, at low hydrogen pressure. This effect will be<br />
considered in details in the chapter 4.<br />
Additivies can effect on enantioselective hydrogenation in different ways, e.g., act as<br />
poison, interact with the products, change the chemical nature <strong>of</strong> modifier or interact<br />
with the modifier. All this effects are summarized in the review <strong>of</strong> Murzin [69].<br />
The very interesting work has been done by Baiker <strong>and</strong> co-workers through variation <strong>of</strong><br />
the modifier structure, briefly the new alternative to quinuclidine chiral moisture<br />
attached to the anchor part was designed [50, 102-104], the new obtained modifier<br />
results in high ee in the hydrogenation <strong>of</strong> α-ketoesters. At the same time it was show<br />
that the anthracene anchor <strong>of</strong> the new chiral lig<strong>and</strong> results in ~ 10 % higher ee with<br />
respect to naphthalene or quinoline anchors.<br />
Quasi-homogeneous catalysis<br />
16
The term quasi-homogeneous (it is also known as colloidal or soluble heterogeneous or<br />
soluble hybrid heterogeneous-homogeneous) was introduced first by Schmid [105]. It<br />
means that the catalyst forms the same (homogeneous) phase with the reaction mixture<br />
i.e. it is soluble in the used solvent however catalyst is not consisting <strong>of</strong> molecules or<br />
complexes like e.g. mentioned above binap-Ru catalyst is. The quasi-homogeneous<br />
catalysis consists <strong>of</strong> small particles (generally few nm in size) which are soluble since<br />
their surface is modified by a specific lig<strong>and</strong>.<br />
Colloidal catalysts are <strong>of</strong> interest for two reasons. Firstly the support effect is<br />
eliminated <strong>and</strong> secondly, it is possible to better control the morphology (size <strong>and</strong> shape)<br />
<strong>of</strong> the metal particles compared to classical supported catalysts. Among several studies,<br />
where Pt <strong>and</strong> other colloids were used for the cinchonidine-modified hydrogenation <strong>of</strong><br />
pyruvates, diketones [106-112] <strong>and</strong> trifluoroacetophenone [113] we would like to mark<br />
the works <strong>of</strong> Bönneman [114, 115] since in his work dihydrocinchonidine was used as<br />
chiral modifier <strong>and</strong> stabilizer, simultaneously, that has allowed to investigate this<br />
system <strong>and</strong> compare with conventional supported Pt catalysts. Others interesting works<br />
[106, 116, 117] where Pt colloidal catalyst stabilized by PVP (poly-N-vinyl-2-<br />
pyrrolidone) (about stabilization see the next chapter) demonstrated the maximum<br />
obtained 98 % ee in methyl pyruvate hydrogenation. Very recently Pd colloids<br />
modified with specific chiral lig<strong>and</strong>s demonstrated up to 98 % <strong>of</strong> ee in allylic alkylation<br />
reaction [118].<br />
Similar rate acceleration [109] <strong>and</strong> solvent effect [106] which are known for<br />
conventional Pt/Al 2 O 3 were observed for colloidal systems.<br />
The very interesting questions arise, wherever the size <strong>of</strong> Pt nanoparticles influence on<br />
ee or not <strong>and</strong> how adsorption <strong>of</strong> cinchonidine on small ~1-2 nm nanoparticles is<br />
different from adsorption <strong>of</strong> big macroscopic Pt support (plane). Thus, in the current<br />
thesis these questions will be under focus <strong>of</strong> our attention.<br />
17
Chapter 2<br />
Introduction to the nanoscale materials<br />
2.1 General information <strong>and</strong> definitions<br />
Introduction<br />
Colloidal metal (originally called "sols") particles are known at lease since work [119]<br />
<strong>of</strong> Hans Heinrich Helcher (1718) about gold colloids. In the Middle Ages glass-blower<br />
workers used gold <strong>and</strong> other metals in colloidal form in order to make coloured glass<br />
(Fig. 2-1) which were used e.g. as stained-glass window in churches or as service<br />
dishes. In 1857, Faraday reported the formation <strong>of</strong> deep red solutions <strong>of</strong> colloidal Au by<br />
reduction <strong>of</strong> an aqueous solution <strong>of</strong> AuCl 4 using phosphorus in CS 2 (a two-phase<br />
system) in a well-known work [120].<br />
Fig. 2-1. Red coloured glass based on gold colloids. Both pictures were found via<br />
google.com<br />
Now the field <strong>of</strong> nanomaterials is booming with scientific investigations <strong>and</strong><br />
application <strong>of</strong> the different areas <strong>of</strong> technology. In fact, for example the catalytic<br />
properties <strong>and</strong> the electronic structure <strong>of</strong> the nanomaterials can be considered from<br />
changing the cluster size, composition <strong>and</strong> structure [121, 122]. Herein, we will provide<br />
an overview <strong>of</strong> colloidal methods <strong>and</strong> properties as well as the approaches being used<br />
to bridge the fields <strong>of</strong> nanotechnology <strong>and</strong> catalysis. In our opinion, it is better to start<br />
acquaintance with nanomaterials from definitions developed by Schmid (Table 2-1)<br />
[123] as well st<strong>and</strong>ardized by IUPAC [124].<br />
Classification<br />
Nanoparticles are constituted <strong>of</strong> several tens or hundreds <strong>of</strong> atoms or molecules <strong>and</strong> can<br />
have a variety <strong>of</strong> sizes <strong>and</strong> morphologies (amorphous, crystalline, spherical, needles,<br />
etc.). According to the IUPAC definition the term colloidal refers to a state <strong>of</strong><br />
subdivision, implying that the molecules or polymolecular particles dispersed in a<br />
medium have at least in one direction a dimension, roughly between 1 nm <strong>and</strong> 1µm, or<br />
that in a system discontinuities are found at distances <strong>of</strong> that order [124].<br />
Table 2-1. Classification <strong>of</strong> nanoscale materials.<br />
18
Discrete, single Metal clusters, Traditional metal Bulk metal<br />
metal a complexes including<br />
nanoclusters<br />
colloids<br />
1-10 nm > 10 nm<br />
Comments: A - the classification can be used not only for metal materials.<br />
Another useful classification <strong>of</strong> variety form <strong>of</strong> colloids is illustrated in the Table 2-2,<br />
adapted form online encyclopaedia (Wikipedia) [125].<br />
Table 2-2. Examples <strong>of</strong> variety <strong>of</strong> colloids.<br />
Dispersed medium<br />
Gas Liquid Solid<br />
Continu Gas None (all gases are Liquid Aerosol Solid Aerosol<br />
ous<br />
soluble)<br />
(fog, mist) (smoke, dust)<br />
media Liquid Foam (cream) Emulsion (milk, Sol (paint, ink)<br />
blood)<br />
Solid Solid foam (aerogel, Gel (gelatine, Solid sol (ruby glass)<br />
pumice)<br />
cheese)<br />
Full shell clusters model<br />
According to the Schmid [123], metal clusters which have a complete, regular outer<br />
geometry are designated as full-shell, or “magic number”, clusters (Table 2-3).<br />
It is believed, that full-shell clusters have added stability, as their densely-packed<br />
structures provide the maximum number <strong>of</strong> metal-metal bonds [126, 127].<br />
Full-shell clusters are constructed by successively packing layers – or shells- <strong>of</strong> metal<br />
atoms around a single metal atom. The total number <strong>of</strong> metal atoms, y, per n th shell is<br />
given by the equation y 2 n<br />
= A ⋅ n + 2 (n>0). In this formula parameter A depends on<br />
geometry <strong>of</strong> the current crystal, the values <strong>of</strong> parameter A are given in the Table 2-4 for<br />
several types <strong>of</strong> crystals.<br />
Table 2-3. Full shell model on the example <strong>of</strong> cuboctahedrone crystal. The table is<br />
adapted from [128].<br />
Table 2-4. Full shell model applied for different polyhedrons [129, 130].<br />
Polyhedron<br />
Illustrated A values<br />
index a<br />
Truncated tetrahedron 1 14<br />
Cuboctahedron or twinned cuboctahedron 2 10<br />
Truncated octahedron 3 30<br />
Truncated cube 4 46<br />
19
Triangular prism 5 4<br />
Comments: A – The geometry <strong>of</strong> the corresponding polyhedron can be found in the Fig.<br />
2-3.<br />
Fig. 2-3. Geometrical models <strong>of</strong> some polyhedrons. Numeration corresponds to<br />
illustration index from Table 2-4.<br />
The full shell model is very useful in calculation <strong>of</strong> concentration <strong>of</strong> surface atoms for a<br />
nanocluster with known size, that can be considered as number <strong>of</strong> active catalytic sites<br />
e.g. for estimation <strong>of</strong> initial rate values normalized on number <strong>of</strong> active sites.<br />
Stabilization <strong>of</strong> nanoparticles<br />
Colloidal particles generally are not thermodynamically stable, because <strong>of</strong> high free<br />
surface energy. As any physical system, nanoparticles always have a tendency to<br />
minimize their full energy that leads to their agglomeration <strong>and</strong> losing nanoscale size, it<br />
should be noted that Van der Waals attraction (see below) can also take part in inducing<br />
an agglomeration. In order to prevent agglomeration particles have to be stabilized.<br />
However before considering stabilization principals it makes sense to mention different<br />
type <strong>of</strong> interactions between colloidal particles [125], they are summarized below.<br />
Excluded Volume Repulsion: This refers to the impossibility <strong>of</strong> any overlap between<br />
hard particles.<br />
Electrostatic interaction: Colloidal particles <strong>of</strong>ten carry an electrical charge <strong>and</strong><br />
therefore attract or repel each other. The charge <strong>of</strong> both the continuous <strong>and</strong> the<br />
dispersed phase, as well as the mobility <strong>of</strong> the phases are factors affecting this<br />
interaction.<br />
Van der Waals forces: This interaction is due to induced dipole-dipole interaction. Even<br />
if the particles don't have a permanent dipole, fluctuations <strong>of</strong> the electron gas give rise<br />
to a temporary dipole, meaning that Van der Waals forces are always present, although<br />
possibly at a much lower magnitude than others.<br />
Entropic forces: According to the second law <strong>of</strong> thermodynamics, a system progresses<br />
to a state in which entropy is maximized. This can result in effective forces even<br />
between hard spheres.<br />
Steric forces between polymer-covered surfaces or in solutions containing nonadsorbing<br />
polymer can modulate interparticle forces, producing an additional repulsive<br />
steric stabilization force or attractive depletion force between them.<br />
Steric stabilization <strong>and</strong> electrostatic stabilization are the two main mechanisms for<br />
colloid stabilization. Electrostatic stabilization is based on the mutual repulsion <strong>of</strong> like<br />
20
electrical charges; in fact the total electric potential is the sum <strong>of</strong> electrostatic <strong>and</strong> Van<br />
der Waals potentials <strong>and</strong> results in energy barrier. Electrostatic repulsion (Fig. 2-4 left)<br />
is generally caused by an electrical double (really multi) layer [131] formed by ions<br />
adsorbed on the particle surface (e.g. halides) <strong>and</strong> the corresponding counter ions (e.g<br />
tetraalkylammonium). Small particle sizes lead to enormous surface areas, <strong>and</strong> this<br />
effect is greatly amplified in colloids. In a stable colloid, the mass <strong>of</strong> a dispersed phase<br />
is so low that its buoyancy or kinetic energy is too little to overcome the electrostatic<br />
repulsion between charged layers <strong>of</strong> the dispersing phase. The charge on the dispersed<br />
particles can be observed by applying an electric field: all particles migrate to the same<br />
electrode <strong>and</strong> therefore must all have the opposite sign charge with respect to the<br />
electrode’s charge.<br />
Fig. 2-4. Schematic image <strong>of</strong> two electrostatically-stabilized nanoparticles (top-left) <strong>and</strong><br />
plot <strong>of</strong> electrical potential space distribution (bottom-left). Two sterically-stabilized<br />
nanoparticles (top-right) <strong>and</strong> caused a large energy barrier (bottom-right) against<br />
particle interaction. Pictures are adapted from [128, 132].<br />
Steric stabilization (Fig. 2-4-right) is achieved by the coordination <strong>of</strong> sterically<br />
dem<strong>and</strong>ed molecules that act as protective shields on the metallic surface. In this way,<br />
nanometallic cores are separated from each other <strong>and</strong> agglomeration is prevented.<br />
2.2 Methods <strong>of</strong> colloid <strong>and</strong> supported nanoparticles<br />
preparation<br />
Nanoparticles can be obtained by two general approaches: “top down” <strong>and</strong> “bottom up”.<br />
In “top down” method bulk materials are mechanically ground to the nanosized scale<br />
<strong>and</strong> stabilized by a suitable molecule [133, 134]. The problem with this method is<br />
difficulty in achieving the narrow size distribution <strong>and</strong> control both shape <strong>and</strong> average<br />
size <strong>of</strong> the particles.<br />
Moreover, bimetallic nanoparticles with core shell structures cannot be obtained by this<br />
method. In “bottom up” methods, nanoparticles are obtained by starting with molecular<br />
21
precursors <strong>and</strong> building-up. This method allows control <strong>of</strong> size <strong>and</strong> shape <strong>of</strong><br />
nanoparticles <strong>and</strong> their size distribution. Taking into account these reasons we will<br />
focus on “bottom up” methods only.<br />
The “bottom up” method can be classified conditionally as methods occurring in<br />
condensed phase <strong>and</strong> gas (or vacuum) phase. In condensed phase synthesis,<br />
nanoparticles are prepared by means <strong>of</strong> chemical synthesis; this method is also called as<br />
wet chemical preparation. For example, in order to prepare metal nanoparticles, metal<br />
salts are usually reduced in the presence <strong>of</strong> suitable capping agents (stabilizer) that<br />
stabilize the nanoparticles from agglomeration. In the gas phase synthesis, metal is<br />
vaporized <strong>and</strong> then vaporized metal atoms are condensed. The same idea can be<br />
realized in vacuum where the metal <strong>of</strong> interest is vaporized with high energy ions or by<br />
heating or by laser beam (the latter method can also be used in liquid media) <strong>and</strong> thus<br />
generated metal vapour is deposited on a support. Below we describe methods for<br />
preparation metal (due to their particular interest in the current thesis) nanoparticles in<br />
condensed <strong>and</strong> briefly in gas phase.<br />
Wet chemical preparation method<br />
Wet chemical reduction procedures have been widely used to stabilize the different<br />
types <strong>of</strong> metal nanoparticles with different kinds <strong>of</strong> organic as well as inorganic<br />
stabilizers.<br />
As was mentioned above, metal precursors are usually reduced in the presence <strong>of</strong> a<br />
stabilizer in order to prevent the nanoparticles from agglomeration. The choice <strong>of</strong> metal<br />
precursor, stabilizer, reducing agent <strong>and</strong> reaction conditions (relative<br />
precursor/stabilizer concentration, solvent, temperature <strong>and</strong> even stirring rate) strongly<br />
influence in properties <strong>of</strong> obtained nanoparticles, i.e. average particle size.<br />
The main classes <strong>of</strong> stabilizer predominant in the literature are macromolecules<br />
containing P, N, <strong>and</strong> S (e.g. phosphanes, amines, thioethers) atoms [105, 135-145].<br />
Such solvents as THF [135, 146] or THF/MeOH [147] also can act as stabilizers.<br />
Molecules like ionic surfactants stabilize the nanoparticles by means <strong>of</strong> both<br />
electrostatic repulsion <strong>and</strong> steric hindrance. Their polar group forms an electrostatic<br />
layer around them <strong>and</strong> the long organic moiety gives steric hindrance. In general,<br />
lipophilic protecting agents give metal colloids that are soluble in organic media<br />
(“organosols”), while hydrophilic agents yield water-soluble colloids (“hydrosols”).<br />
Reducing agents play an important role in directing the size <strong>and</strong> shape <strong>of</strong> particles.<br />
Among different kinds <strong>of</strong> reducers (NaBH 4 , R-CHO, etc.) we would like to underline<br />
hydrogen. In fact, it has been applied in the synthesis <strong>of</strong> various metal clusters. The<br />
advantage <strong>of</strong> using hydrogen lies in the avoidance <strong>of</strong> by-products, the disadvantage is a<br />
weak reducing agent that is unable to reduce many metal complexes <strong>and</strong> salts.<br />
However, the use <strong>of</strong> hydrogen makes possible to perform reduction under mild<br />
condition that might be important for sensitive stabilizers.<br />
It is interesting to note, that reduction <strong>of</strong> metal salts e.g. Na 2 PtCl 4 by microwave<br />
radiation, thus in presence <strong>of</strong> stabilizer gives 20-30 Pt nanoparticles [148].<br />
The decomposition <strong>of</strong> organometallic complexes with null valence metal atom(s) in<br />
presence <strong>of</strong> stabilizer can also be applied as a method <strong>of</strong> preparation <strong>of</strong> nanoparticles.<br />
This decomposition is typically initiated by an external force e.g. heat, ultrasonic or by<br />
a chemical agent capable to induce decomposition e.g. H 2 or CO. Classical example is<br />
in thermal decomposition <strong>of</strong> iron Fe(CO) 6 [149] or cobalt complex Co 2 (CO) 8 [150] at<br />
130-170 °C in presence <strong>of</strong> polymer yielding 45 nm Co nanoparticles [151].<br />
22
The advantage <strong>of</strong> use <strong>of</strong> organometallic route in preparation <strong>of</strong> nanoclusters is obvious<br />
when it is needed to stabilize (modify) nanoparticles by hydrophobic surfactant or in<br />
other words, when required stabilizer is water insoluble.<br />
An interesting method <strong>of</strong> preparation <strong>of</strong> metal nanoparticles is through electrochemical<br />
approach established by Reetz [152]. In this method, an anode is made up <strong>of</strong> the<br />
specific metal (required for nanoparticles) <strong>and</strong> a cathode is <strong>of</strong> any other metal. Under<br />
the suitable applied current density the anode sacrificially dissolves in the electrolyte,<br />
the metal ions migrate towards the cathode <strong>and</strong> reduction occurs. The nucleation <strong>and</strong><br />
growth <strong>of</strong> reduced metal atoms occurs at the electrode surface. The stabilizing agent in<br />
the reaction vessel arrests the growth <strong>of</strong> the nanoparticles <strong>and</strong> facilitates the formation<br />
<strong>of</strong> stable nanostructures. The final step is dissolution <strong>of</strong> nanoparticles from the<br />
electrode surface into the bulk <strong>of</strong> the solution. The advantages <strong>of</strong> the electrochemical<br />
pathway are that the contamination with by-products resulting from chemical reducing<br />
agents is avoided <strong>and</strong> that the products are easily isolated from the precipitate. Further,<br />
the electrochemical preparation allows for size-selective particle formation. The<br />
particle size obtained by the electrochemical route depends on many factors, the<br />
distance between the electrodes, reaction time, temperature, <strong>and</strong> polarity <strong>of</strong> the solvent<br />
contribute to the particle size. Experiments have also shown that the applied current<br />
density also has a major influence on the particle size.<br />
Vapor phase synthesis<br />
As was mentioned above metal atoms can be obtained not only through reduction <strong>of</strong><br />
metal salts, but also through solid to gas phase transformation <strong>of</strong> metal. The Solvated<br />
metal atom dispersion (SMAD) techniques is based on this method. Typically, the metal<br />
<strong>of</strong> interest is heated in a crucible at elevated temperature evaporated under vacuum <strong>and</strong><br />
then co-condensed with a specified amount <strong>of</strong> liquid lig<strong>and</strong> substrate on the liquid<br />
nitrogen cooled walls <strong>of</strong> the reactor vessel. The reactor is removed from the liquid<br />
nitrogen <strong>and</strong> the reactor is allowed to warm slowly. After the warming stage, particles<br />
can be stabilized either sterically (by solvation) or electrostatically (by incorporation <strong>of</strong><br />
negative charge). The SMAD techniques has been successfully applied e.g. in<br />
preparation <strong>of</strong> thiol stabilized gold colloid [153]. The main advantage <strong>of</strong> this “clean”<br />
method <strong>of</strong> preparing nanoclusters in gram scale is in fact, that it does not involve<br />
starting materials containing counter ions like nitrate, chlorides, sulfates which are<br />
common in reduction synthetic methods. However, the use <strong>of</strong> SMAD method is limited<br />
because <strong>of</strong> difficulties in the operation <strong>of</strong> the apparatus <strong>and</strong> it is difficult to obtain<br />
narrow particle size distributions.<br />
Another interesting method is through the laser ablation <strong>of</strong> bulk metal target crystal<br />
being in liquid phase. The recent example is in preparation <strong>of</strong> Pt nanoparticles by laser<br />
ablation <strong>of</strong> Pt crystal in water [154-156]. Due to the laser induced evaporation <strong>of</strong> Pt<br />
atoms into the liquid phase with formation <strong>of</strong> 1-15 nm nanoparticles, it was found that<br />
the yield <strong>of</strong> nanoparticles depends on the wavelength <strong>of</strong> the used laser beam whereas it<br />
has no significant influence on the average size <strong>of</strong> produced nanoparticles. The laser<br />
ablation method has the same important advantages as SMAD method demonstrates,<br />
moreover there is not need to work with very high temperatures (for metal evaporation)<br />
<strong>and</strong> keep system under high vacuum. However with respect to the chemical methods<br />
the laser ablation has the following disadvantages. There is more insight about the<br />
reaction at plasma <strong>and</strong> liquid interface <strong>and</strong> it is relatively more expensive to obtain the<br />
large amount <strong>of</strong> products, in fact the rate <strong>of</strong> nanoparticles production was about 4.4<br />
mg/h.<br />
23
2.3 Preparation <strong>of</strong> nanoclusters on a heterogeneous<br />
support<br />
Metal clusters especially in nanoscale size are widely used in heterogeneous catalysis.<br />
Heterogeneous catalysts are traditionally obtained by impregnation methods consisting<br />
<strong>of</strong> immersing a solid support like Al 2 O 3 , SiO 2 or ZrO 2 in a solution <strong>of</strong> metal salts <strong>of</strong><br />
interest <strong>and</strong> inducing deposition. After some specified time, the solid substance is<br />
filtered <strong>and</strong> dried. The dried product is calcinated at high temperature in the presence <strong>of</strong><br />
gaseous reducing agents such as H 2 or CO to yield the final catalytic material. The<br />
problem with impregnation methods is the difficulty in controlling the particle size <strong>and</strong><br />
that the metal particle formed after the reduction may be located where it is not<br />
accessible for further application.<br />
An alternative method for the preparation <strong>of</strong> heterogeneous catalysis is the “precursor<br />
concept“ which was originally developed by Turkevich in 1970 [157] <strong>and</strong> was further<br />
developed in the 1990's by many groups [135, 158-161].<br />
This method consists <strong>of</strong> depositing the pre-prepared metal colloid on a solid support; by<br />
applying this method it is possible to control the size <strong>and</strong> shape <strong>of</strong> the nanoparticles.<br />
Application <strong>of</strong> the precursor concept also allows full pre-characterization <strong>of</strong> the catalyst<br />
components (support <strong>and</strong> colloid) prior to deposition, thus facilitating the possibilities<br />
for comprehensive study <strong>of</strong> heterogeneous catalyst support/metal effects. Moreover,<br />
almost all <strong>of</strong> the particles on the inert support are accessible because the preformed<br />
colloids generally do not “fill” cracks <strong>and</strong> other defects <strong>of</strong> the support’s surface as<br />
molecular precursors <strong>of</strong>ten do.<br />
Metal nanoparticles have a high active metal surface area (50-100 m 2 /g for 6-3 nm<br />
particles) that facilitates proper surface modification (roughly, due to the high activity<br />
<strong>of</strong> “freshly” prepared nanoclusters) <strong>and</strong> sometimes unique size dependence properties.<br />
The catalytic properties <strong>of</strong> supported <strong>and</strong> not supported metal nanoclusters were found<br />
in many important reactions such as oxidation (Co, Pt, Pd, Ag <strong>and</strong> Au), hydrogenation<br />
(Pt, Pd, Pt/Pd, Au/Pt, Au/Pd <strong>and</strong> Rh) <strong>and</strong> others. The detailed overview can be found in<br />
the work <strong>of</strong> D’Souza [162].<br />
24
Chapter 3<br />
Current thesis in the EU-COST project:<br />
aims <strong>of</strong> the work<br />
3.1 Introduction <strong>and</strong> goals setting<br />
The idea to use metal nanoclusters modified with chiral lig<strong>and</strong>s is a much promising<br />
strategy in research <strong>and</strong> development <strong>of</strong> enantioselective catalysts. In fact, making<br />
nanoparticles stabilized with chiral lig<strong>and</strong>s it is possible to reach maximum chiral<br />
coverage <strong>of</strong> the metal surface, that is very important for obtaining high<br />
enantioselectivity. Since Pt <strong>and</strong> Pd (generally alumina or silica supported) were found<br />
to be the best metals for heterogeneous enantioselective hydrogenation, in this thesis,<br />
we focused on the chiral modification <strong>of</strong> the surface <strong>of</strong> Pt <strong>and</strong> Pd nanoclusters.<br />
The current thesis is a part <strong>of</strong> the “EU COST D24 WORKING GROUP 7<br />
D24/0007/02” working project, which besides International <strong>University</strong> Bremen (Pr<strong>of</strong>.<br />
R. Richards) also involves research groups from the Marseille <strong>University</strong> (Pr<strong>of</strong>. G.<br />
Buono), Ecole Nationale Supérieure de Chimie de Paris Institute (Pr<strong>of</strong>. J.-P. Genet),<br />
<strong>University</strong> <strong>of</strong> Ljubljana (Pr<strong>of</strong>. M. Kocevar) <strong>and</strong> <strong>University</strong> <strong>of</strong> Bucharest (Pr<strong>of</strong>. V. I.<br />
Parvulescu).<br />
The working scheme (Fig. 3-1) shows that the groups from Paris <strong>and</strong> Marseille provide<br />
us with specific chiral lig<strong>and</strong>s (synphos <strong>and</strong> quiphos with quiphos-spider, respectively).<br />
At the International <strong>University</strong> Bremen we have prepared Pt <strong>and</strong> Pd nanoclusters<br />
modified with obtained from our colleagues (<strong>and</strong> purchased from Fluka <strong>and</strong> Aldrich)<br />
chiral lig<strong>and</strong>s for chiral modification <strong>of</strong> them on the surface <strong>of</strong> metal nanoclusters. The<br />
groups in Paris, Ljubljana <strong>and</strong> Bucharest study catalytic activity (as well as<br />
enantioselectivity) over obtained samples in corresponding reactions. The details about<br />
their work can be found in the description <strong>of</strong> the EU-COST project [163] <strong>and</strong> from the<br />
corresponding research groups.<br />
25
Paris<br />
Marseille<br />
Ljubljana<br />
Synphos<br />
Quiphos<br />
Quiphos - spider<br />
IUB catalysts<br />
preparation <strong>and</strong><br />
characterization<br />
Bucharest<br />
Aldrich<br />
Fluka<br />
Cinchonidine,<br />
binap, diop, diphos<br />
Fig. 3-1. The working scheme <strong>of</strong> EU-COST project.<br />
The all used in the current work lig<strong>and</strong> for the surface modification are summarized in<br />
the Fig. 3-2.<br />
O<br />
HO<br />
N<br />
N<br />
N<br />
P<br />
N<br />
O<br />
PPh 2<br />
O<br />
PPh 2<br />
PPh 2<br />
O PPh 2<br />
N<br />
Cinchondine<br />
Quiphos<br />
(2R, 5S)-3-Phenyl-2-(8-quinolinoxy)<br />
-1,3-diaza-2-phosphabicyclo-octane<br />
Binap<br />
2,2'-Bis(diphenylphosphino)-<br />
1,1'-binaphthyl<br />
O<br />
Synphos<br />
(2,3,2',3'-tetrahydro-5,5'-bi(1,4-<br />
benzodioxin)<br />
-6,6'-diyl)-bis(diphenylphosphane)<br />
P<br />
P<br />
P<br />
P O<br />
P<br />
O<br />
O<br />
Diphos<br />
Bis(diphenylphosphino)ethane<br />
Diop<br />
4,5-Bis(diphenylphosphinomethyl)-<br />
2,2-dimethyl -1,3-dioxolane<br />
Quiphos-Spider<br />
4-isopropyl-2 methyl cyclohexyl-1methylphosphonite<br />
-2-2'-1,5 binaphtyl<br />
Fig. 3-2. Lig<strong>and</strong>s used for modification <strong>of</strong> surface <strong>of</strong> Pt <strong>and</strong> Pd nanoclusters.<br />
26
Besides preparation <strong>of</strong> chiral lig<strong>and</strong> modified Pt <strong>and</strong> Pd nanoclusters another aim <strong>of</strong><br />
the current thesis is in determination the geometrical models <strong>of</strong> adsorption <strong>of</strong> chiral<br />
lig<strong>and</strong>s (Fig. 3-2) on the surface <strong>of</strong> Pt <strong>and</strong> Pd nanoclusters via DRIFTS (Diffuse<br />
Reflectance Infra-red Fourier Transform Spectroscopy) <strong>and</strong> molecular modeling. Since<br />
the ability <strong>of</strong> the adsorbed cinchonidine to induce enantioselectivity e.g. in<br />
hydrogenation <strong>of</strong> α-ketoesters is well known in heterogeneous enantioselective<br />
catalysis we placed high emphasis on the examination <strong>of</strong> cinchonidine modified Pt <strong>and</strong><br />
Pd nanoclusters in ethyl pyruvate hydrogenation. The work is also focused on<br />
obtaining active, highly enantioselective stable catalyst for the performing<br />
hydrogenation under mild conditions that is very important from application point <strong>of</strong><br />
view. Furthermore we investigated catalytic <strong>and</strong> enantioselective activity <strong>of</strong> all<br />
synthesized samples in typical Orito reaction.<br />
27
Chapter 4<br />
Cinchonidine modified Pt colloidal<br />
nanoparticles: characterization <strong>and</strong><br />
catalytic properties<br />
4.1 Introduction<br />
Cinchonidine modified transition metal surfaces (primarily Pt) represent one approach<br />
towards asymmetric heterogeneous catalysis. Since the first report by Orito [45, 46],<br />
numerous experimental <strong>and</strong> theoretical investigations into the cinchona on Pt <strong>and</strong> Pd<br />
systems have revealed a great deal <strong>of</strong> insight into the system <strong>and</strong> its applicability. As<br />
was mentioned above, adsorption geometry <strong>of</strong> the cinchonidine molecule on metal<br />
surface has a significant role in the enantioselectivity. It is known from the work <strong>of</strong><br />
Ferri [76], Zaera [77] <strong>and</strong> Wells [74], that cinchonidine can adsorb on the surface <strong>of</strong><br />
macroscopic Pt (plane, single crystal with sizes are much bigger than the size <strong>of</strong> the<br />
cinchonidine molecule) in three different orientations (Fig. 1-19): “π-bonded”, “α-H<br />
abstracted” <strong>and</strong> “N lone pair bonded”. It is also considers, that only cinchonidine in the<br />
flat “π-bonded” orientation is able to induce enantioselective hydrogenation. However<br />
adsorption mode <strong>of</strong> cinchonidine on the surface <strong>of</strong> nanosized Pt crystals has never been<br />
investigated <strong>and</strong> it attracts special attention, for instance, because the maximum<br />
reported enantiomeric excess (98 %, in the hydrogenation <strong>of</strong> methyl pyruvate) was<br />
obtained on small (1.2 nm) Pt colloids. Thus the beginning <strong>of</strong> this chapter is dedicated<br />
to the determination <strong>of</strong> adsorption modes <strong>of</strong> cinchonidine on Pt nanoclusters.<br />
The initial transient period (ITP) is an interesting phenomenon which consists <strong>of</strong> an<br />
increase <strong>of</strong> enantiomeric excess in the beginning <strong>of</strong> e.g. ethyl pyruvate hydrogenation.<br />
The ITP effect was observed in the examination <strong>of</strong> enantioselectivity <strong>of</strong> cinchonidine<br />
modified Pt nanoclusters <strong>and</strong> discussed in the second part <strong>of</strong> the current chapter. The<br />
initial transient period (ITP) for the Pt/Al 2 O 3 system has been reported <strong>and</strong> investigated<br />
by numerous groups <strong>and</strong> has been the subject <strong>of</strong> intense discussions [164, 165]. ITP<br />
was observed by Baiker <strong>and</strong> co-workers [166] <strong>and</strong> was attributed to an impurity effect<br />
(from destructive adsorption <strong>of</strong> EP <strong>and</strong> alcoholic solvents on Pt) which could be<br />
eliminated by pre-activation with hydrogen. Blackmond <strong>and</strong> co-workers proposed the<br />
“reaction-driven equilibration" explanation [167] <strong>and</strong> demonstrated the significant<br />
influence <strong>of</strong> mass transport (gas-liquid diffusion <strong>of</strong> hydrogen) limitation on the<br />
enantioselectivity [168]. Recently, it was shown that ITP is not only the effect <strong>of</strong><br />
impurities (lactates in ethyl pyruvate) but could also be a result <strong>of</strong> competitive<br />
adsorption <strong>of</strong> reactant, modifier <strong>and</strong> solvent [169]. From the work <strong>of</strong> Hutchings et al.<br />
[170], we know, that the ITP effect was found with cinchonidine premodified<br />
conventional Pt/Al 2 O 3 in ethanol <strong>and</strong> dichloromethane solution <strong>and</strong> explanation through<br />
“removal <strong>of</strong> adsorbed oxygen species from the surface” was proposed.<br />
Here, we investigate the behavior <strong>of</strong> enantiomeric excess (ee) during hydrogenation <strong>of</strong><br />
ethyl pyruvate with conventional Pt/Al 2 O 3 <strong>and</strong> with cinchonidine Pt nanoclusters<br />
prepared by two different methods, an aqueous (first reported by Bönnemann [114])<br />
<strong>and</strong> a novel organometallic method through decomposition <strong>of</strong> Pt 2 (DBA) 3 (DBA is bisdibenzylidene<br />
acetone) complex. The nature <strong>of</strong> the increase in ee at initial reaction<br />
time, further decreases <strong>and</strong> differences in the behavior <strong>of</strong> conventional Pt/Al 2 O 3 <strong>and</strong><br />
28
“quasi-homogeneous” cinchonidine stabilized Pt nanoclusters are discussed below.<br />
Further, the experimental conditions selected for these experiments were chosen to be<br />
mild: room temperature <strong>and</strong> low hydrogen pressure (2-10 bar) since these are likely to<br />
be the most relevant conditions for hydrogenation <strong>of</strong> α-ketoesters on an industrial scale.<br />
4.2 Experimental<br />
Materials<br />
Unless otherwise mentioned the following materials: Pt on alumina (Aldrich 5% Pt),<br />
cinchonidine (> 98%, Fluka), acetic acid (99.8% Fluka), ethyl pyruvate (98%, Aldrich),<br />
THF (Applichem, 99.8%), formic acid (Aldrich, 99%), bis-dibenzylidene acetone<br />
(Fluka, ≥96.0%), K 2 PtCl 4 (Acros Organics) <strong>and</strong> ethanol (Applichem, 99.9%) were used<br />
as received in all chapters.<br />
The 5 % Pt on alumina (E4759) catalyst was kindly provided by Engelhard company by<br />
request.<br />
Sample preparation<br />
In order to activate the conventional Pt/Al 2 O 3, it was heated at 300 °C under vacuum<br />
(10 -3 mbar) for 1 hour, then kept for 3 hours under flowing hydrogen (100 ml/min) <strong>and</strong><br />
cooled to room temperature. Then the required amount was weighed <strong>and</strong> transferred<br />
into a glass reactor with acetic acid, ethyl pyruvate <strong>and</strong> cinchonidine.<br />
Chiral modification <strong>of</strong> conventional Pt/Al 2 O 3 . 40 mg <strong>of</strong> conventional Pt/Al 2 O 3 was<br />
heated to 400 ºC under vacuum (10 -3 mbar) for 1 hour, then the mixture <strong>of</strong> N 2 (93%)<br />
<strong>and</strong> H 2 (7%) gases were passed (100 ml min -1 ) over the sample for 3 hours. Then, under<br />
flowing gases (H 2 <strong>and</strong> N 2 ), the support was cooled to room temperature <strong>and</strong> the solution<br />
<strong>of</strong> cinchonidine in CHCl 3 (6.0 ml, 3 mM) was added by syringe through a stopper. The<br />
system was kept at 10-15 ºC for 24 hours, then washed 3 times with CHCl 3 <strong>and</strong> finally<br />
dried under vacuum at room temperature.<br />
Preparation <strong>of</strong> 10.11-dihydrocinchonidine (DHCIN) has been done according to the<br />
protocol from work <strong>of</strong> Morawsky [171]. 10 g cinchonidine (33 mmol) was dissolved in<br />
100 ml 0.5 M H 2 SO 4 <strong>and</strong> placed in a metal reactor. 0.5 g Pd/C (5 % wt, Degussa<br />
E101/0/W) was added to the solution. The H 2 pressure was set to 6 bars, reaction<br />
started by stirring. After 1.5 h reaction was finished. The catalyst was removed via a G4<br />
filter, 50 ml 2 M NaOH was added to precipitate the hydrated alkaloid. The white<br />
precipitate was washed with 500 ml H 2 O <strong>and</strong> recrystallized in ethanol. It has to be<br />
noted that resulted product had some amount <strong>of</strong> cinchonidine, as was detected as<br />
presence <strong>of</strong> attenuated peak at 1636 cm -1 in IR spectrum <strong>of</strong> the product, which<br />
corresponds to the stretching <strong>of</strong> C=C bond.<br />
The preparation <strong>of</strong> Pt 2 (DBA) 3 organometallic complex has been done according to the<br />
literary protocol [172, 173]. Typically, to a solution <strong>of</strong> 2.8 g sodium acetate<br />
<strong>and</strong> 2.36 g <strong>of</strong> bis-dibenzylidene acetone (dba) in 60 mL <strong>of</strong> ethanol at 50 °C was added a<br />
solution <strong>of</strong> 1.40 g K 2 PtCl 4 in 12 mL <strong>of</strong> freshly distilled water. The initial pale yellow<br />
suspension redissolved while the mixture was heated to the refluxing temperature (90<br />
°C). After refluxing for 1 h a dark violet precipitate formed. The mixture was allowed<br />
to settle overnight. The solution was eliminated by filtration <strong>and</strong> the solid washed three<br />
times with 200 mL <strong>of</strong> water, dried overnight under vacuum, further washed three times<br />
with 200 mL <strong>of</strong> pentane <strong>and</strong> finally dried under vacuum overnight.<br />
A water soluble Pt nanoclusters modified with DHCIN was prepared based on the<br />
protocol published by Bönnemann [114]. Typically, 0.208 g (0.6 mmol) PtCl 4 in 160 ml<br />
distilled water was reduced with 0.1 M aqueous formic acid (15 ml) with dissolved<br />
29
dihydrocinchonidine (355 mg or 1.2 mmol) under reflux, at ~95 ºC. After 10-15<br />
minutes the product was precipitated into 200 ml <strong>of</strong> aqueous saturated NaHCO 3<br />
solution, then the precipitate was filtered with a G4 fritted glass filter <strong>and</strong> washed first<br />
with a half-saturated NaHCO 3 (400 ml) solution followed by water. Finally, the black<br />
powder was taken up from the filter using an acetic acid solution. The acetic acid – Ptcinchonidine<br />
system was dried at room temperature under vacuum. This catalyst was<br />
found to be redispersible in water, acetic acid <strong>and</strong> toluene. This catalyst will be referred<br />
to as B – henceforth in this chapter. Physical state: black powder.<br />
The cinchonidine modified Pt nanosized catalyst was prepared by decomposition <strong>of</strong> the<br />
organometallic complex Pt 2 (DBA) 3 under hydrogen in the presence <strong>of</strong> cinchonidine.<br />
Pt 2 (DBA) 3 (402 mg or 0.37 mmol) <strong>and</strong> an excess <strong>of</strong> cinchonidine (4.5 g or 15 mmol)<br />
were dissolved in 160 ml THF <strong>and</strong> flushed with Ar for 20 min. The mixture was kept in<br />
the glass reactor under 2 bar <strong>of</strong> H 2 with constant stirring at 300 rpm. After 18 hours the<br />
black precipitate was collected <strong>and</strong> washed first with 50 ml pentane, then with a<br />
pentane/THF mixture (5:1) <strong>and</strong> chlor<strong>of</strong>orm until all hydrogenated DBA <strong>and</strong> excess<br />
cinchonidine was removed as monitored by IR <strong>and</strong> XRD spectroscopy. Thus prepared<br />
catalysts were found to be redispersible in acetic acid <strong>and</strong> THF <strong>and</strong> will be referred as<br />
D – henceforth in this chapter. Physical state: black powder.<br />
In order to purify ethyl pyruvate, 20 ml <strong>of</strong> it was added to 80 ml <strong>of</strong> water/n-pentane<br />
(1/1 by volume) mixture <strong>and</strong> kept for two days in a separation funnel. The n-<br />
pentane/ethyl pyruvate fraction was separated <strong>and</strong> n-pentane was evaporated in a rotary<br />
evaporator. The procedure was repeated two times <strong>and</strong> the collected purified ethyl<br />
pyruvate was found to contain 0.05 % R, S ethyl lactate (from the initial 0.5%) by GC.<br />
Characterization <strong>and</strong> experiments<br />
Hydrogen chemisorption on conventional Pt/Al 2 O 3 (Aldrich) catalyst was performed on<br />
the Autosorb instruments supplied by Quantachrome, Germany. Typically the sample<br />
was preheated at 400 °C for 1 h under vacuum (10 -2 mm. Hg.) <strong>and</strong> for 3 hour under<br />
flow <strong>of</strong> hydrogen then chemisorption has started at 40 °C in the 0-586 mm. Hg.<br />
hydrogen pressure interval.<br />
Transmission electron microscopy (TEM) micrographs were obtained by using a<br />
TECNAI F20 instrument in the IFAM Bremen. Specimens were prepared by placing a<br />
drop <strong>of</strong> the colloidal onto a copper grid with a perforated carbon film <strong>and</strong> then allowing<br />
the solvent to evaporate. Around 100-150 particles were considered for size distribution<br />
calculations.<br />
X-ray diffraction (XRD): Powder XRD analyses were performed using a Siemens<br />
D5000 instrument with Cu Kα radiation (wavelength 1.54·10 -10 m) operated at 40 kV<br />
<strong>and</strong> 40 mA.<br />
Gas chromatography: (GC) Measurements were performed at 88 °C (isothermal) on a<br />
Varian GC 3900, FID with a Lipodex-E chiral column.<br />
Catalytic Reactor: The catalytic reactor was supplied by Parr Instrument (Germany)<br />
GmbH. The apparatus consists <strong>of</strong> five major components; reactor, H 2 reservoir, oil<br />
reservoir, a temperature <strong>and</strong> pressure controller <strong>and</strong> computer. The batch type reactor<br />
(300 ml) is made <strong>of</strong> double walled borosilicate glass with a maximum pressure rating <strong>of</strong><br />
10 bars. A regulator situated between the reactor <strong>and</strong> reservoir maintains a constant<br />
hydrogen pressure inside the reactor. The reactor chamber also consists <strong>of</strong> a water<br />
cooler, an electric stirrer <strong>and</strong> thermocouple.<br />
Hydrogenation <strong>of</strong> ethyl pyruvate was carried out in the described above 300 ml glass<br />
reactor under a pressure <strong>of</strong> either 2 or 10 bars <strong>of</strong> hydrogen at room temperature (22±1<br />
30
°C). The stirring speed was 300 rpm. The ethyl pyruvate was added to 70 ml <strong>of</strong> acetic<br />
acid, then the mixture was sonicated for 10 min, Ar was then flushed through the<br />
solution for 10 min to remove dissolved air. The required amount <strong>of</strong> catalyst was<br />
added, the mixture was sonicated for 5 min, transferred into the reactor, flushed with Ar<br />
for 10 min <strong>and</strong> finally after the pressure <strong>of</strong> hydrogen had been stabilized for 10-15 sec<br />
the initial time (t=0) was set <strong>and</strong> the reaction kinetics were monitored. Around one ml<br />
aliquots <strong>of</strong> the reaction mixture were taken at certain time intervals, filtrated with a G4<br />
filter <strong>and</strong> 1 μl was injected into the gas chromatograph (GC).<br />
FTIR spectra <strong>of</strong> D type <strong>of</strong> catalyst were measured with a “Thermo Nicolet Avatar”<br />
spectrometer in transmission mode, resolution 4 cm -1 , number <strong>of</strong> scans 500. IR pellets<br />
were prepared by mixing 100 mg KBr <strong>and</strong> 1.5 mg sample then pressing into a pellet<br />
under a pressure <strong>of</strong> 10 tons. For IR investigation <strong>of</strong> the B type <strong>of</strong> catalyst a “Thermo<br />
Nicolet 4700 FT-IR” spectrometer with liquid nitrogen cooled detector <strong>and</strong> DRIFT<br />
accessory was used with the following parameters: 3000 scans, 600-2000 cm -1 scan<br />
range, 4 cm -1 resolution. The DRIFT spectra were chosen for analysis because it is the<br />
more applicable technique for observing adsorbed species on finely divided metal<br />
powder [174].<br />
The carbon <strong>and</strong> nitrogen elemental analysis measurements were done on a Carlo Erba<br />
NA2500 C/N analyzer in the <strong>School</strong> <strong>of</strong> Geo<strong>Science</strong>s at Grant Institute in the <strong>University</strong><br />
<strong>of</strong> Edinburgh. Measurements were repeated several times with an error ~1%. The metal<br />
content was found from thermogravimetric Analysis (TGA) in 30-1200 °C temperature<br />
interval under nitrogen 100 ml/min with an SDTQ 600 instrument from TA<br />
Instruments.<br />
The cumulative enantiomeric excess (ee) was calculated according to the formula<br />
[ R]<br />
− [ S]<br />
ee = ⋅100%<br />
.<br />
[ S]<br />
+ [ R]<br />
The actual enantiomeric excess (ee*) was calculated by the previously established<br />
[169] formula<br />
ee y − ee y<br />
ee<br />
*<br />
i+<br />
1 i+<br />
1 i i<br />
i+<br />
1 / 2<br />
=<br />
,<br />
yi+<br />
1<br />
− yi<br />
where y i - mol <strong>of</strong> both ethyl lactates, ee –cumulative (observed) enantiomeric excess.<br />
Computational methods: Geometry optimization <strong>and</strong> calculation <strong>of</strong> IR spectra <strong>of</strong><br />
cinchonidine <strong>and</strong> quinoline were carried out using the Gaussian 03 program [175].<br />
Density functional theory (DFT) calculations were performed by using the B3LYP/6-<br />
31G method <strong>and</strong> basis set. The vibrations <strong>and</strong> orientations <strong>of</strong> dipole moment were<br />
visualized with GaussView <strong>and</strong> MOLDEN programs.<br />
4.3 Results <strong>and</strong> discussion<br />
4.3.1 General characterization<br />
Choice <strong>of</strong> conventional catalyst<br />
The Engelhard 4759 (5% Pt/Al 2 O 3 ) is known from 90 th to be one <strong>of</strong> the most studied<br />
catalyst for Orito reaction, however the Engelhard company does not produce one any<br />
more <strong>and</strong> it is not widely available. Due to this reason <strong>and</strong> the fact that under the same<br />
reaction conditions (Table 4-1) catalyst from Aldrich demonstrates higher<br />
enantioselectivity <strong>and</strong> reaction rate, we chose it to be the basic conventional catalyst for<br />
further comparisons.<br />
31
Table 4-1. Comparison between 5 % Pt/Al 2 O 3 catalysts from Aldrich <strong>and</strong> Engelhard<br />
under the same reaction conditions A .<br />
Supplier Initial rate, Conversion, % after Ee, % (dominated<br />
mmol·(min·g) -1 (time, min) enantiomer)<br />
Aldrich 28 80 (60) B 91 (R)<br />
Engelhard 2 34 (250) B 70 (R)<br />
Comments: A - To 4 ml ethyl pyruvate (36 mmol) <strong>and</strong> 20 mg cinchonidine (1 mM) 20<br />
mg thermally activated catalyst (please find the activation procedure in the chapter 5)<br />
was added, the mixture purged with Ar for 10 min <strong>and</strong> hydrogen pressure <strong>of</strong> 10 bar was<br />
set under the constant stirring (1000 rpm), B - according to the curve pr<strong>of</strong>ile<br />
(conversion vs. time) reaction can go further.<br />
From the hydrogen chemisorption experiment (Fig. 4-1) with conventional 5% Pt/Al 2 O 3<br />
(here <strong>and</strong> further Aldrich catalyst is mentioned as a conventional) it was determined<br />
that the metal dispersion is 29 %, active metal surface area is 3.6 m 2 g -1 <strong>and</strong><br />
concentration <strong>of</strong> active sites is 37 μmol g -1 , average crystal size is 3.8 nm.<br />
Fig. 4-1. Combined, weak <strong>and</strong> their difference (strong) curves <strong>of</strong> hydrogen<br />
chemisorption data on conventional 5% Pt/Al 2 O 3 .<br />
X-ray diffraction phase analysis <strong>of</strong> the cinchonidine modified Pt nanoclusters was<br />
carried out using powder XRD to ascertain the polycrystalline nature <strong>of</strong> the particles.<br />
32
800<br />
111<br />
Intensity, a.u.<br />
400<br />
200<br />
220 311<br />
0<br />
2Θ<br />
40 60 80<br />
Fig. 4-2. XRD spectrum <strong>of</strong> B sample.<br />
600<br />
111<br />
Intensity, a.u.<br />
400<br />
200<br />
200<br />
220<br />
0<br />
40 50 60 70<br />
2Θ<br />
Fig. 4-3. XRD spectrum <strong>of</strong> D sample.<br />
As can be seen from XRD spectra <strong>of</strong> sample B (Fig. 4-2) <strong>and</strong> D (Fig. 4-3) they both<br />
have dominant peak at ≈ 39 °, which corresponds to the (111) crystal face, as well as<br />
peak at 66 °, which corresponds to (220) crystal phase. The peak at 46 ° (corresponding<br />
to the (220) crystal phase) is clearly seen in the spectra <strong>of</strong> D sample, however it is<br />
present as poorly observable shoulder in the spectra <strong>of</strong> B sample, probable due to the<br />
widening <strong>of</strong> the dominant peak at 39 ° <strong>and</strong> their overlapping. The assignment <strong>of</strong> the<br />
peaks was based on the work <strong>of</strong> Davey [176]. The quantitative value <strong>of</strong> a peak<br />
33
widening can be used in estimation <strong>of</strong> the average crystal size, according to the<br />
Scherrer equation [177]. The obtained values can be found in the Table 4-1 below.<br />
K ⋅λ<br />
d hkl<br />
= , where K is a constant whose value depends on the shape <strong>of</strong> the<br />
Δ ⋅cos(Θ)<br />
particle <strong>and</strong> the method used to measure the peak width, in this work K was taken to be<br />
0.94, according to the work <strong>of</strong> Mukherjee [148], Δ – Peak width, taken on a half height.<br />
TEM images <strong>of</strong> the B <strong>and</strong> D samples (Fig. 4-4) <strong>and</strong> their corresponding size distribution<br />
histograms (Fig. 4-5) clearly show that D type <strong>of</strong> sample contains Pt nanoparticles with<br />
average size <strong>of</strong> 2.3 nm, whereas average size <strong>of</strong> nanoparticles <strong>of</strong> B sample is 1.4 nm.<br />
Fig. 4-4. TEM <strong>of</strong> B <strong>and</strong> D-sample.<br />
The very similar values <strong>of</strong> the average crystal size were estimated from the Scherrer<br />
equation based on XRD data are summarized in the Table 4-2 together with data<br />
obtained from TEM.<br />
Table 4-2. The average size <strong>of</strong> cinchonidine modified Pt nanoclusters obtained by<br />
aqueous (B) <strong>and</strong> organometallic (D) methods determined via TEM <strong>and</strong> XRD<br />
measurements.<br />
Type <strong>of</strong> sample (TEM), nm. (XRD), nm.<br />
B 1.4 1.2<br />
D 2.3 2.3<br />
34
Number <strong>of</strong> particles<br />
30 Types <strong>of</strong> sample<br />
D<br />
B<br />
25<br />
20<br />
15<br />
10<br />
5<br />
0<br />
1.0 1.5 2.0 2.5 3.0 3.5<br />
Particle size, nm<br />
Fig. 4-5. Particles size distribution for B <strong>and</strong> D types <strong>of</strong> sample.<br />
From the elemental analysis it was found, that the B-series sample contains 43.5 % C,<br />
4.3 % N or 45 % cinchonidine, based on nitrogen content. From the TGA, the Pt<br />
content was found to be 25 % thus differing from the 40 % metal content in the original<br />
work [11] for 1.5 nm sized Pt cluster, probably because <strong>of</strong> the presence <strong>of</strong><br />
moisture/CO/CO2 <strong>and</strong> possibly acetic acid (30 %).<br />
Elemental analysis <strong>of</strong> the D-catalysts before reaction (please see below) showed 16.6 %<br />
C <strong>and</strong> 1.8 % N or 20 % cinchonidine, by making average between C <strong>and</strong> N. 63 % Pt<br />
was found from TGA. The remaining content was attributed to moisture/CO/CO 2 . The<br />
content <strong>of</strong> C <strong>and</strong> N on the D catalyst after 50 min <strong>of</strong> reaction was found to be 5.3 % <strong>and</strong><br />
0.8 %, correspondingly or 7 % <strong>of</strong> cinchonidine.<br />
4.3.2 FTIR investigation <strong>of</strong> cinchonidine adsorbed on Pt<br />
FTIR investigation <strong>of</strong> cinchonidine adsorbed on Pt was performed with cinchonidine<br />
modified conventional Pt/Al 2 O 3 (Fig. 4-6) <strong>and</strong> cinchonidine modified Pt nanoclusters B<br />
(Fig. 4-7) <strong>and</strong> D (Fig. 4-8) samples. The use <strong>of</strong> different Pt sample allows comparison<br />
<strong>of</strong> adsorbed modes <strong>of</strong> the cinchonidine molecule <strong>of</strong> Pt crystals with different sizes,<br />
starting from 1.2 (B-sample), passing 2.3 nm (D-sample) with 3.8 nm (conventional<br />
alumina supported Pt) <strong>and</strong> ending with comparison <strong>of</strong> adsorption mode <strong>of</strong> cinchonidine<br />
on microscopic Pt plane from the work <strong>of</strong> Ferri [76] <strong>and</strong> Zaera [77, 178]. In the Fig. 4-6<br />
the region below 1300 cm -1 is not accessible <strong>of</strong> IR investigation, due to the contribution<br />
<strong>of</strong> alumina background, whose FTIR spectrum can be found as an inset in the Fig. 5-5<br />
from the chapter 5.<br />
35
1<br />
Absorbance<br />
0.1<br />
2<br />
1600 1400 1200 1000 800<br />
Wavenumber, cm -1<br />
Fig. 4-6. DRIFT spectra <strong>of</strong> the cinchonidine modified conventional Pt/Al 2 O 3 (1) <strong>and</strong><br />
free cinchonidine (2), intensity <strong>of</strong> that spectrum is decreased for better presentation.<br />
Absorbance<br />
0.13<br />
1<br />
2<br />
1600 1400 1200 1000 800<br />
Wavenumber cm -1<br />
Fig. 4-7. DRIFT spectra <strong>of</strong> the cinchonidine modified Pt nanoclusters (1, sample B) <strong>and</strong><br />
free cinchonidine (2), intensity <strong>of</strong> that spectrum is decreased for better presentation.<br />
It is important to note, that the absence <strong>of</strong> Al 2 O 3 (normally used as a support for Pt<br />
[76]) allows investigation down to the range <strong>of</strong> 700-900 cm -1 thus facilitating<br />
assignment <strong>of</strong> the important out <strong>of</strong> plane vibrations <strong>of</strong> aromatic rings.<br />
36
Absorbance<br />
0.05<br />
1<br />
2<br />
1600 1400 1200 1000<br />
Wavenumber, cm -1<br />
Fig. 4-8. FTIR spectra <strong>of</strong> the cinchonidine modified Pt nanoclusters (1, sample D) <strong>and</strong><br />
free cinchonidine (2), intensity <strong>of</strong> that spectrum is decreased for better presentation.<br />
The spectra measured in transmission mode with KBr pressed pellet.<br />
The 3-9 cm -1 blue <strong>and</strong> red shifts in the wavenumber <strong>of</strong> the corresponding peaks in the<br />
spectra <strong>of</strong> free <strong>and</strong> adsorbed modifier clearly indicate on the fact that cinchonidine is<br />
adsorbed on the Pt cluster. The quality <strong>of</strong> shown spectra increases significantly with<br />
absence <strong>of</strong> a support due <strong>of</strong> the mentioned background contribution <strong>and</strong> because <strong>of</strong><br />
higher cinchonidine concentration <strong>of</strong> surface <strong>of</strong> Pt for B <strong>and</strong> D samples with respect to<br />
modified conventional Pt/Al 2 O 3 . However, despite <strong>of</strong> the fact that not all peaks in the<br />
spectrum <strong>of</strong> cinchonidine modified conventional Pt/Al 2 O 3 (Fig. 4-6) are resolved well<br />
(due to the alumina background effect) <strong>and</strong> some corresponding peaks from<br />
cinchonidine modified Pt nanoclusters (Fig. 4-7, 4-8) presence in the Fig. 4-6 as<br />
shoulders, it is possible to conclude, that the spectra <strong>of</strong> adsorbed cinchonidine on all<br />
three samples have no principal difference. And only difference is in quality <strong>of</strong> the<br />
spectra. Through comparison <strong>of</strong> all three spectra is also possible to conclude that the<br />
cinchonidine molecule is still on the Pt surface <strong>and</strong> did not decompose after samples<br />
preparation, that is especially important to note for B <strong>and</strong> D sample. Since the spectrum<br />
quality <strong>of</strong> B type <strong>of</strong> nanosized sample is better that the quality <strong>of</strong> modified conventional<br />
Pt/Al 2 O 3 <strong>and</strong> D sample, this spectrum had been chosen for further detailed analysis,<br />
where as brief analysis <strong>of</strong> two left spectra can be found below in this chapter (sample<br />
B) <strong>and</strong> in the chapter 5 (cinchonidine modified conventional Pt/Al 2 O 3 ).<br />
The DRIFT spectrum <strong>of</strong> sample B (Fig. 4-7) has many peaks <strong>and</strong> shoulders, thus it<br />
gives difficulties to find corresponding peaks in the spectrum <strong>of</strong> free cinchonidine<br />
unambiguously for all <strong>of</strong> them. This is due to the adsorption selection rule [179] by<br />
which intensities <strong>of</strong> some vibration bonds <strong>of</strong> adsorbed molecules are changed. That is<br />
why only some <strong>of</strong> the peaks <strong>and</strong> shoulders (which are well separated <strong>and</strong>/or with high<br />
intensity) were chosen for further analysis, they are summarized in Table 4-3. Some <strong>of</strong><br />
the peaks from Table 4-3 could be easily assigned to the ‘flat’ or/<strong>and</strong> ‘tilted’ orientation<br />
mode <strong>of</strong> cinchonidine by using data from the works <strong>of</strong> Ferri [76] <strong>and</strong> Zaera [77].<br />
However, not all vibrations from Table 4-3 could be correlated with prior work due to<br />
37
the focus <strong>of</strong> these prior experiments on selective regions <strong>of</strong> the spectra. In the work <strong>of</strong><br />
Ferri <strong>and</strong> co-workers the interval 1321-1635 cm -1 is particularly covered <strong>and</strong> in the<br />
work <strong>of</strong> Zaera <strong>and</strong> co-workers, the description <strong>of</strong> many peaks is given, but some <strong>of</strong><br />
them, which were observed in our experiments, are not mentioned. This why the<br />
assignment <strong>and</strong> interpretation <strong>of</strong> summarized in the Table 4-3 peaks <strong>of</strong> adsorbed<br />
cinchonidine has been performed with help <strong>of</strong> modelling <strong>of</strong> the IR vibrations <strong>of</strong> the<br />
molecule <strong>and</strong> metal adsorption selection rule [179]. The metal adsorption selection rule<br />
[179] states, if a vibration <strong>of</strong> a molecule adsorbed on a metal surface produces a dipole<br />
moment orientated perpendicularly to the surface, the intensity <strong>of</strong> this vibration is<br />
enhanced (or can be observed) <strong>and</strong> vice versa; i.e. if the dipole moment is orientated<br />
parallel to the metal surface, then intensity <strong>of</strong> the corresponding vibration is reduced (or<br />
not present). It means that dipole moments corresponding to all vibrations from Table<br />
4-3 are orientated mostly perpendicularly to the metal surface.<br />
Fig. 4-9. Illustration <strong>of</strong> the metal adsorption selection rule. Image charges are drawn<br />
dotted.<br />
It is important to note, that intensity <strong>of</strong> the IR radiation adsorbed by a molecule is<br />
proportional to the first time derivative <strong>of</strong> dipole moment in power <strong>of</strong> two, or, roughly,<br />
it is proportional to the squared dipole moment. At the same time the absolute value <strong>of</strong><br />
electric field induced by any dipole moment is inversely proportional to the distance in<br />
power <strong>of</strong> three. Thus contribution <strong>of</strong> the electric field from dipole moment <strong>of</strong> the<br />
molecule adsorbed on the neighbor nanoparticle is significantly less (approximately in<br />
the factor <strong>of</strong> 10 3 ) that electric field from dipole moment <strong>of</strong> the molecule adsorbed on<br />
the current nanoparticle. Here we took into account, that size <strong>of</strong> the molecule is ~ 10-20<br />
% <strong>of</strong> the size <strong>of</strong> nanocluster <strong>and</strong> typical distance between nanoparticles (in the solid<br />
state) can be estimated as a size <strong>of</strong> nanoparticle. Since, an intensity <strong>of</strong> any<br />
electromagnetic field is proportional to the squared amplitude <strong>of</strong> electric field, we can<br />
conclude, that observed IR peaks belong to the molecule adsorbed on the current metal<br />
nanoparticle only.<br />
In the case <strong>of</strong> cinchonidine molecule, its optimized geometry was chosen in the open 3<br />
conformation [77] <strong>and</strong> vibration modes with corresponding dipole moments orientation<br />
were found from DFT calculation (with correlation coefficient dω theory /dω exp = 1.02).<br />
Then every observed vibrational frequency from Table 4-3 was analyzed with respect<br />
to the metal adsorption selection rule. The analysis results are also summarized in the<br />
38
Table 4-3 <strong>and</strong> the ‘flat’ <strong>and</strong> ‘tilted’ modes <strong>of</strong> adsorbed cinchonidine are indicated by a<br />
vibration b<strong>and</strong> e.g. at 768 cm -1 <strong>and</strong> at 1514 cm -1 are illustrated in the Fig. 4-10 <strong>and</strong> Fig.<br />
4-11, respectively.<br />
Fig. 4-10. Flat adsorbed cinchonidine demonstrates vibration mode at 768 cm -1 with<br />
dipole moment (big arrow) orientated mostly perpendicularly to the metal surface. The<br />
most intensive displacement <strong>of</strong> atoms is shown by small arrows.<br />
39
Fig. 4-11. Tilted adsorbed cinchonidine demonstrates vibration mode at 1514 cm -1 with<br />
dipole moment (big arrow) orientated mostly perpendicularly to the metal surface. The<br />
most intensive displacement <strong>of</strong> atoms is shown by small arrows.<br />
It should be noted, that observed “blue <strong>and</strong> red shifts” in frequencies (see Table 4-3)<br />
indicate the changes in the bond force constants <strong>and</strong> thus changes in the skeletal<br />
geometry <strong>of</strong> an adsorbed molecule with respect to the free species. However, for the<br />
purpose <strong>of</strong> this work, we will leave these considerations aside, since they do not hinder<br />
determination <strong>of</strong> the orientation <strong>of</strong> adsorbed cinchonidine <strong>and</strong> other used molecules<br />
(see below). Calculated (correlation coefficient 1.06) <strong>and</strong> measured IR spectra <strong>of</strong> free<br />
quinoline (mostly medium <strong>and</strong> strong peaks were chosen) were used as a particular<br />
reference for facilitating assignment <strong>of</strong> the spectra <strong>of</strong> the cinchonidine (as well as<br />
quiphos with aniline from the next chapter), Fig. 4-12.<br />
3<br />
A<br />
Relative units<br />
2<br />
B<br />
1<br />
C<br />
0<br />
D<br />
1600 1400 1200 1000 800 600<br />
Wavenumber cm -1<br />
Fig. 4-12 Comparison <strong>of</strong> IR peaks between free cinchonidine (A), quinoline (B),<br />
quiphos (C) <strong>and</strong> aniline (D).<br />
Table 4-3. Assignment <strong>of</strong> some frequencies (cm -1 ) from IR spectrum <strong>of</strong> adsorbed <strong>and</strong><br />
free cinchonidine <strong>and</strong> free quinoline.<br />
Cinchonidine<br />
on Pt<br />
Free<br />
cinchonidine<br />
Quinoline Cinchonidine<br />
theoretical<br />
Vibration<br />
description<br />
Orientation<br />
<strong>of</strong> adsorbed<br />
cinchonidine<br />
1634 (sh) 1636 (med) - 1709<br />
19,20 C strch. Flat<br />
1613 (sh) 1619 (wk3) 1620<br />
(med)<br />
1594 (med) 1590 (str4) 1596<br />
(med)<br />
1670 C-Qu.<br />
strch., H-<br />
Qu. b.<br />
1643 C-Qu.<br />
strch., H-<br />
Qu. b.<br />
Tilted<br />
Tilted<br />
40
1576 (sh) 1567 (med) 1571<br />
(med)<br />
1622 C-Qu.<br />
strch., H-<br />
Qu. b.<br />
1514 (str) 1507 (str) 1502 (str) 1556 C-Qu.<br />
strch., H-<br />
Qu. b.<br />
1463 (str) 1452 (str) - 1494<br />
1427 (sh) 1420 (med) 1431 (wk) 1445<br />
15,17,22 CH 2 -<br />
QL. sci.<br />
2,3 C strch.,<br />
C-H-Qu.,<br />
ip. def.,<br />
15 CH 2 ,<br />
HO 2 C sci.,<br />
H 2 C b.<br />
1208 (med) 1207 (med) - 1240 C-H-Qu.<br />
ip. b., 2,3 C<br />
strch., HO,<br />
2 CH b.,<br />
QL. comp.<br />
1179 (wk) 1180 (wk) - 1223<br />
1168 (wk) 1163 (wk) - 1189<br />
30 H-Qu.,<br />
23,24 H b.,<br />
H- 17 C-Ql.<br />
comp.<br />
25,26,30 H-<br />
Qu. ip. b.,<br />
23,24 H b.,<br />
QL. comp.<br />
1145 (wk) 1133 (wk) - 1143 H-Qu. ip.<br />
b.,<br />
23,24 H<br />
b., Ql.<br />
comp.<br />
1118 (med) 1099 (wk) - 1128 H-Qu. ip.<br />
b., O 2 C<br />
strch., QL.<br />
comp.<br />
856 (med) 862 (wk) - 873<br />
29,10 H-Qu.<br />
oop. wag.,<br />
QL. comp.<br />
831 (med) 823 (med) 807 (str) 843 H-Qu. oop.<br />
wag, H 20 2 C<br />
sci., QL.<br />
Comp.<br />
809 (med) 803 (med) 787 (str) 824 H-Qu. oop.<br />
Wag.,<br />
40,41 H 20 C<br />
sci., QL.<br />
comp.<br />
782 (str) 778 (wk) 761 (wk) 791 H-Qu. oop.<br />
wag.,<br />
40,41 H 20 C<br />
Tilted , flat<br />
Tilted<br />
Tilted<br />
Flat<br />
Tilted<br />
Flat<br />
Flat, tilted<br />
Flat<br />
Flat<br />
Flat<br />
Flat<br />
Flat<br />
Flat<br />
41
sci., QL.<br />
comp.<br />
768 (str) 756 (str) 740 (str) 775<br />
25,26,27,28 H- Flat<br />
Qu. oop.<br />
wag.<br />
Comments: sh. – shoulder, med. – medium, wk – weak, st. – strong., 15C –carbon atom<br />
number 15. Qu – quinoline, QL – quinuclidine. C-Qu. – only carbons in quinoline, H-<br />
Qu. – only hydrogens in quinoline, skel. – skeleton, comp. – complex: more then one<br />
type <strong>of</strong> vibrations (stretching, wagging, rocking, <strong>and</strong> scissoring) are present, def. –<br />
deformation, C-H-QL. – carbons <strong>and</strong> hydrogens in quinuclidine, b. – bending, strch. –<br />
stretching, sci. – scissoring, wag. – wagging, ip.- in plane, oop. – out <strong>of</strong> plane vibration.<br />
As it is seen from the Table 4-3, cinchonidine on Pt nanoclusters was found in both the<br />
‘flat’ <strong>and</strong> ‘tilted’ adsorption modes. Here, we did not differentiate between the two<br />
types <strong>of</strong> tilted adsorption modes, the so called “N-lone pair bonded” <strong>and</strong> “α-H<br />
abstracted" because we found assignment <strong>of</strong> the corresponding species is somewhat<br />
ambiguous <strong>and</strong> since the flat adsorption mode is more determinant in enantioselective<br />
catalysis.<br />
It is very important to compare the adsorbed modes <strong>of</strong> cinchonidine on macroscopic<br />
<strong>and</strong> nanocluster surfaces <strong>of</strong> Pt. In fact, through comparison <strong>of</strong> IR investigations <strong>of</strong><br />
cinchonidine adsorbed on bulk Pt crystals (ATR, RAIRS) <strong>and</strong> on Pt nanoclusters<br />
(DRIFTS) we found no principal difference in the spectra as well in the spectra <strong>of</strong><br />
sample D <strong>and</strong> cinchonidine modified conventional Pt/Al 2 O 3 (the only difference we<br />
found was in the quality <strong>of</strong> the spectra). However, in the spectra <strong>of</strong> cinchonidine on Pt<br />
nanoclusters small shifts <strong>of</strong> the spectra from free cinchonidine were found <strong>and</strong> between<br />
the spectra <strong>of</strong> cinchonidine adsorbed on Pt. For example: in case <strong>of</strong> free cinchonidine,<br />
the peak at 1593 cm -1 in work <strong>of</strong> Ferri [76] was observed at 1590 cm -1 in the current<br />
thesis <strong>and</strong> 1591 cm -1 in the work <strong>of</strong> Zaera [77]; in case <strong>of</strong> cinchonidine on Pt, the peak<br />
at 1458 cm -1 in work <strong>of</strong> Ferri was found at 1463 cm -1 in the current work <strong>and</strong> at 1463<br />
cm -1 in the work <strong>of</strong> Zaera. It seems most likely that these small shifts are due to the role<br />
<strong>of</strong> solvents/purity <strong>of</strong> cinchonidine used <strong>and</strong> other secondary effects. Another difference<br />
is that the spectra <strong>of</strong> the lig<strong>and</strong>s on nanoclusters contains less noise <strong>and</strong>, mostly, have<br />
better quality with respect to the spectra <strong>of</strong> lig<strong>and</strong>s on bulk metal. The reason for this<br />
may be due to the high surface area <strong>of</strong> nanoclusters <strong>and</strong> thus higher concentration <strong>of</strong><br />
adsorbed lig<strong>and</strong> in the IR beam. If size <strong>of</strong> the cinchonidine supported Pt nanoclusters<br />
(2.3 nm sample D, or 4 nm conventional Pt/Al 2 O 3 ) is bigger than typical size <strong>of</strong><br />
molecule (about 0.5 nm, the length <strong>of</strong> the quinoline anchor), the support can be<br />
considered as a plane <strong>and</strong> molecule behaves similarly as to being adsorbed on a<br />
continuous surface <strong>of</strong> a (bulk) metal crystal.<br />
However, if the size <strong>of</strong> the clusters becomes comparable with the size <strong>of</strong> the molecule,<br />
as we have in case <strong>of</strong> cinchonidine on Pt (1.2 nm in sample B) the lig<strong>and</strong> might behave<br />
differently (especially in terms <strong>of</strong> enantioselectivity) <strong>and</strong> would be expected to<br />
experience less steric hindrance due to the curvature <strong>of</strong> the surface.<br />
In fact, Yang <strong>and</strong> co-authors [106] reported enantiomeric excess in hydrogenation <strong>of</strong><br />
methyl pyruvate to be 90 % over 1.2 nm Pt cluster <strong>and</strong> 85 % over 3.4 nm under 40 bar<br />
<strong>of</strong> H 2 pressure. They explained this difference as a size effect resulting from<br />
cinchonidine <strong>and</strong> methyl pyruvate forming a half-hydrogenated complex which is<br />
adsorbed on different crystal faces. However, these authors did not focus on stability <strong>of</strong><br />
Pt nanoclusters during hydrogenation <strong>and</strong> changes in size, especially after removal <strong>of</strong><br />
42
PVP stabilizer. In the work <strong>of</strong> Bönnemann [114, 115] , no such dependence was found<br />
(ee = 76% over 1.5 nm Pt cluster <strong>and</strong> 78 % over 3.9 nm, under 1 bar <strong>of</strong> H 2 pressure).<br />
In our opinion, one <strong>of</strong> the primary effects causing enhanced ee is a higher concentration<br />
<strong>of</strong> active chiral sites (“π-bonded” cinchonidine) on the catalyst surface. In the work <strong>of</strong><br />
Yang <strong>and</strong> co-authors cinchonidine adsorbs on the Pt surface from solution (optimum<br />
concentration <strong>of</strong> cinchonidine [80, 102, 180]) whereas high hydrogen pressure (40 bar)<br />
promotes a surface cleaning (from PVP, AcOH <strong>and</strong> etc.) for obtaining more space for<br />
flat adsorbed cinchonidine ( cinchonidine in any tilted mode needs, definitely, less free<br />
space for the adsorption). The catalyst prepared according to Bönnemann’s procedure<br />
(sample B) possesses adsorbed cinchonidine in flat <strong>and</strong> in two tilted modes as<br />
demonstrated in this manuscript <strong>and</strong> a high pressure <strong>of</strong> hydrogen was not applied.<br />
Moreover ee=94-95 % was obtained under high (100 bar) H 2 pressure with<br />
conventional 5R94 [180] <strong>and</strong> Engelhard 4759 [181]- Pt/Al 2 O 3 catalysts (average<br />
particles size 3.6 nm [182]) where the above mentioned size effect seems to be less<br />
probable.<br />
Based on these facts, we would like to conclude here, that it was found that the size <strong>of</strong><br />
Pt clusters (in range 1.4- ∞ nm) has no observed influence on the adsorption mode <strong>of</strong><br />
the cinchonidine (at least in case <strong>of</strong> sample B (1.4 nm), sample D (2.3 nm),<br />
conventional Pt/Al 2 O 3 (4 nm) <strong>and</strong> macroscopic samples from work <strong>of</strong> Zaera [77] <strong>and</strong><br />
Ferri [76]). This is important because it has been reported in the literature that cluster<br />
size does influence both catalytic activity <strong>and</strong> the observed ee [106].<br />
Thus, it was found that both ‘flat’ <strong>and</strong> ‘tilted’ modes <strong>of</strong> adsorbed cinchonidine are<br />
present on Pt nanoclusters <strong>and</strong> could be an explanation <strong>of</strong> the fact that this type <strong>of</strong><br />
catalyst does not provide the maximum possible enantiomeric excess (99 %).<br />
4.3.3 Investigation <strong>of</strong> catalytic behaviour <strong>of</strong> cinchonidine<br />
modified Pt nanoclusters<br />
Investigation <strong>of</strong> catalytic behavior <strong>of</strong> cinchonidine modified Pt nanoclusters was<br />
studied with B, D <strong>and</strong> conventional modified Pt/Al 2 O 3 in the hydrogenation <strong>of</strong> ethyl<br />
pyruvate.<br />
Initial transient period<br />
Under all reaction conditions, summarized in Table 4-4, enantiomeric excess was found<br />
to depend on the reaction time in a manner similar to that shown in the Fig. 4-13.<br />
During the first moments <strong>of</strong> hydrogenation (the period <strong>of</strong> time termed the initial<br />
transient period) enantiomeric excess increases to a maximum value (ee max ) <strong>and</strong> then<br />
decreases until the entire amount <strong>of</strong> ethyl pyruvate is converted in the reaction mixture.<br />
43
Conversion, %<br />
100<br />
80<br />
60<br />
40<br />
ITP time<br />
ee max<br />
80<br />
70<br />
Actual enantiomeric excess, %<br />
20<br />
0<br />
Y =83.7-0.19 X<br />
60<br />
0 25 50 Time, min75 100 125<br />
Fig. 4-13. Typical catalytic behaviour <strong>of</strong> cinchonidine modified Pt nanoclusters during<br />
ethyl pyruvate hydrogenation.<br />
Effect <strong>of</strong> hydrogen pressure<br />
The effect (Exp. 1 <strong>and</strong> 2, see Table 4-4) <strong>of</strong> changing the hydrogen pressure in the<br />
reaction system is shown in the Fig. 4-14 (other graphs are similar, thus they are not<br />
shown in further, important features are mentioned in the text, in the corresponding<br />
paragraphs <strong>and</strong> in the Table 4-4). With an increase <strong>of</strong> hydrogen pressure from 2 to 10<br />
bar the initial rate increases by a factor <strong>of</strong> ten from 14 min -1 to 140 min -1 An increase <strong>of</strong><br />
initial rate could be qualitatively explained as an increase <strong>of</strong> hydrogen migration from<br />
the gas to the liquid phase, but the fact that the maximum enantiomeric excess reached<br />
significantly increases from 36 % to 78 % seems to indicate a specific hydrogen –<br />
catalyst interaction, as a consequence <strong>of</strong> this, the catalyst demonstrates increasing<br />
enantioselectivity. A decrease by 4 fold in the time required to reach ee max is observed<br />
by changing the pressure from 2 (110 min) to 10 (30 min) bar.<br />
44
80<br />
Conversion, %<br />
100<br />
80<br />
60<br />
40<br />
Y =83 - 0.19 X<br />
Hydrogen pressure<br />
10 bar<br />
2 bar<br />
70<br />
60<br />
50<br />
40<br />
Actual enantiomeric excess, %<br />
20<br />
30<br />
0<br />
20<br />
0 50 100 150 200 250 300<br />
Time, min<br />
Y =44 - 0.07 X<br />
Figure 4-14. Effect <strong>of</strong> hydrogen pressure.<br />
Experiment 1: 5 mg (13 μmol Pt) D- type <strong>of</strong> catalyst, [EP 0 ] =2.571 mol l -1 , no free<br />
cinchonidine, P H2 = 2 bar. Full squares (left axis) – ethyl pyruvate conversion, crossed<br />
squares (right axis) – actual EE, line – linear approximation <strong>of</strong> actual ee decrease.<br />
Experiment 2: 4 mg (13 μmol Pt) D- type <strong>of</strong> catalyst, [EP 0 ] = 2.571 mol l -1 , no free<br />
cinchonidine, P H2 = 10 bar. Full triangles (left axis) – ethyl pyruvate conversion,<br />
crossed triangles (right axis) – actual ee, line – linear approximation <strong>of</strong> actual EE<br />
decrease.<br />
Type <strong>of</strong><br />
[EP ],<br />
Table 4-4. Details <strong>of</strong> catalysts <strong>and</strong> reaction parameters.<br />
Exp. P H2 , Pt surf.<br />
ee * max, Cinchonidine, Time slope Initial<br />
0<br />
N. bar Atoms, catalyst % concentration, ee * max, min -1 mol l - rate b ,<br />
μmol a <strong>and</strong><br />
mmol l -1<br />
1<br />
min<br />
min -1<br />
loading,<br />
mg<br />
1 2 5.8 D, 4 36 0 110 0.07 2.571 14<br />
2 10 5.8 D, 4 78 0 30 0.07 2.571 140<br />
3 c 10 5.8 D, 4 76 0 20 0.19 0.514 99<br />
4 c 10 29 D, 20 65 0 9 0.22 0.514 37<br />
5A c 10 1.11 Pt/Al 2 O 3 , 76 1 - 0.24 0.257 128<br />
20<br />
5B c 10 0.56 Pt/Al 2 O 3 ,<br />
10<br />
76 1 - 0.21 0.257 132<br />
45
6A c 10 19.5 B, 20 61 0 3 0.18 0.514 24<br />
6B c 10 3.9 B, 4 76 0 - - 0.514 -<br />
7A d 10 2.2 D, 1.5 79 1 30 - 0.514 e 130<br />
7B 10 2.2 D, 1.5 78 1 30 - 0.514 120<br />
8 c 10 5.8 D, 4 76 1 55 0.04 0.514 121<br />
9 c 10 3.9 B, 4 74 1 34 0.07 0.514 42<br />
Comments: a – The number <strong>of</strong> Pt surface atoms was estimated from the medium<br />
particles size according to “full shell” model <strong>of</strong> Pt nanoclusters <strong>and</strong> from chemisorption<br />
on conventional Pt/Al 2 O 3 , b – Initial rate values [min -1 ] were calculated as<br />
normalization a reaction rate [mmol min -1 ] on number <strong>of</strong> Pt surface atoms [mmol] in<br />
the reaction mixture, c – The graph is not shown; important parameters <strong>of</strong> reaction are<br />
presented, d – The purified EP was used.<br />
Effect <strong>of</strong> catalyst loading<br />
The decrease <strong>of</strong> initial rate from 99 min -1 to 37 min -1 with an increase <strong>of</strong> catalyst<br />
loading from 4 mg (13 μmol Pt) to 20 mg (65 μmol Pt) was found in Exp. 3 <strong>and</strong> 4.<br />
Also, the maximum <strong>of</strong> actual enantiomeric excess (ee* max ) has also decreased from 76<br />
% to 65 %. A possible explanation for this is that the time necessary to reach the<br />
maximum enantioselective activation level may not be the same for various catalyst<br />
loadings <strong>and</strong> the full conversion <strong>of</strong> ethyl pyruvate has already been accomplished. In<br />
other words, the number <strong>of</strong> chiral sites (π-bonded cinchonidine) did not reach the<br />
maximum possible value under the 10 bars <strong>of</strong> hydrogen pressure. It should be noted,<br />
that this effect can not be explained as mass transport phenomena in liquid phase, since<br />
according to Minder, Baiker <strong>and</strong> co-workers [183] the kinetically controlled region lays<br />
below 15 g l -1 <strong>of</strong> catalyst concentration while in the Exp. 3 <strong>and</strong> 4 the maximum catalyst<br />
concentration is 0.28 g l -1 . In fact, we performed the reactions with 20 mg (Exp. 5A,<br />
ee max = 76 %, initial rate = 128 min -1 ) <strong>and</strong> 10 mg (Exp. 5B, ee max = 76 %, initial rate =<br />
132 min -1 ) with conventional Pt/Al 2 O 3 <strong>and</strong> concluded that in our experiments mass<br />
transport effect in liquid phase does not limit the reaction. The rate <strong>of</strong> migration <strong>of</strong><br />
hydrogen from gas to the liquid phases was the same, since both experiments were<br />
performed under the same hydrogen pressure (10 bar).<br />
A similar decrease <strong>of</strong> ee max with an increase <strong>of</strong> catalyst loading was found for the B<br />
type catalyst (see Exp. 6A <strong>and</strong> 6B in Table 4-4).<br />
By comparing these results with those reviewed by Studer, Blaser <strong>and</strong> Exner in 2003<br />
[44] we can see that there are intrinsic differences between the colloids stabilized by<br />
chiral lig<strong>and</strong>s <strong>and</strong> the conventional Pt/Al 2 O 3 heterogeneous catalysts which are<br />
unmodified prior to introduction to the reaction vessel. In fact, decrease <strong>of</strong> initial rate<br />
<strong>and</strong> maximum enantiomeric excess with an increase <strong>of</strong> catalyst loading are in<br />
contradiction to the results observed for the conventional Pt/Al 2 O 3 system which may<br />
be attributed to the fact, that our catalysts have adsorbed cinchonidine already on them,<br />
whereas the catalyst systems reviewed in the literature are exposed to cinchonidine<br />
during the reaction (in situ).<br />
Role <strong>of</strong> the ethyl pyruvate in nature <strong>of</strong> ITP<br />
The ITP time was found to be almost independent <strong>of</strong> the initial concentration <strong>of</strong> ethyl<br />
pyruvate (Exp. 2 <strong>and</strong> 3). The ITP was found to be 20-30 min in Exp. 2 (20 ml EP) <strong>and</strong><br />
20-25 min in the case <strong>of</strong> Exp. 3 (4 ml EP). Ee* max is almost the same value (76-78 %)<br />
in both experiments, but initial rate increased by ≈ 30 % with an increase <strong>of</strong> the initial<br />
concentration <strong>of</strong> ethyl pyruvate.<br />
46
In order to clarify the role in ee <strong>of</strong> the initial presence <strong>of</strong> ethyl lactate (0.5 % each S <strong>and</strong><br />
R isomers) as an impurity in EP, the catalytic hydrogenation reaction with the D type <strong>of</strong><br />
catalyst was performed with both the unpurified <strong>and</strong> purified EP. Fig. 4-15 (Exp. 7A,<br />
7B) shows that in this experiment the ee observed is similar for both the purified <strong>and</strong><br />
unpurified experiments with respect to the reaction time <strong>and</strong> both demonstrate the same<br />
ITP = 30 min <strong>and</strong> initial rate = 120-130 min -1 .<br />
80<br />
Conv. 47 %<br />
Enantiomeric excess, %<br />
76<br />
72<br />
Purified EP<br />
Normal EP<br />
Conv. 4 %<br />
68<br />
0 20 40 60 80<br />
Time, min<br />
Fig. 4-15. Role <strong>of</strong> the impurities in the ethyl pyruvate on EE.<br />
Experiment 7A: 1.5 mg (4.8 μmol Pt) D type <strong>of</strong> catalyst, [EP 0 ] = 0.514 mol l -1<br />
(purified), [Cinchonidine 0 ] =1 mmol l -1 , P H2 = 10 bar.<br />
Experiments 7B: 1.5 mg (4.8 μmol Pt) D type <strong>of</strong> catalyst, [EP 0 ] = 0.514 mol l -1 ,<br />
[Cinchonidine 0 ] =1 mmol l -1 , P H2 = 10 bar.<br />
However, the initial presence <strong>of</strong> R, S ethyl lactates gives a systematic shift in ee. This<br />
shift has a maximum value (≈ 3 %) in initial times (at low conversion) <strong>and</strong> decreases<br />
with time (with increase <strong>of</strong> conversion). Taking into account the initial concentration <strong>of</strong><br />
both ethyl lactates, the values <strong>of</strong> the shift lie within the estimated limit.<br />
In order to show that the presence <strong>of</strong> EP does not play a significant role in increasing<br />
ee, for example as a reagent in side reactions (aldol condensation) reaction [184] we<br />
performed a blank experiment in the absence <strong>of</strong> EP with free cinchonidine under<br />
reaction conditions (acetic acid, 10 bar <strong>of</strong> H 2 ). However, both B <strong>and</strong> D samples are<br />
unstable with prolonged hydrogen exposure especially at high pressure. Thus, to<br />
improve stability, we prepared cinchonidine modified Pt nanoclusters deposited on<br />
alumina. This catalyst was found to be stable under the reaction conditions for a long<br />
time <strong>and</strong> demonstrates (P H2 = 10 bar, [EP 0 ] = 0.514 mol l -1 , [cinchonidine] = 1 mmol l -<br />
1 ) ee = 79 %, however, after pretreating the catalyst for 16 hours in acetic acid it<br />
demonstrates ee = 84 %. This experiment obviously shows that the ethyl pyruvate does<br />
not influence significantly to the increase <strong>of</strong> ee after exposing catalyst under reaction<br />
condition. Another interesting fact is that under optimized conditions (hydrogen<br />
47
pressure, exposing time, cinchonidine concentration, etc.) the maximum obtained ee<br />
was found to be 88 %.<br />
In order to keep clarity in description <strong>of</strong> the samples <strong>and</strong> their characterization, it is<br />
recommended to the reader to see Chapter 5 for further information about cinchonidine<br />
modified Pt nanoclusters deposited on alumina, where we give detailed models <strong>of</strong> effect<br />
<strong>of</strong> ee increasing, whereas here we would like to continue description <strong>of</strong> observed ITP<br />
effect.<br />
Role <strong>of</strong> free cinchonidine in ITP<br />
The addition <strong>of</strong> 20 mg (0.068 mmol) <strong>of</strong> free cinchonidine into the reaction mixture<br />
(Exp. 3 <strong>and</strong> 8) increases ITP by a factor <strong>of</strong> almost three (from 20 min to 55 min) while<br />
ee* max ≈ 76 % remained nearly constant <strong>and</strong> the initial rate increases only slightly. The<br />
stability in ee with addition <strong>of</strong> free cinchonidine was also found for B type <strong>of</strong> catalyst in<br />
out experiments <strong>and</strong> in original work <strong>of</strong> Bönnemann [114]. This fact might be<br />
explained by competitive formation <strong>and</strong> disappearance <strong>of</strong> new chiral sites, as is<br />
discussed below.<br />
Post ITP decrease <strong>of</strong> enantiomeric excess<br />
The decrease <strong>of</strong> ee during the hydrogenation <strong>of</strong> ethyl pyruvate has been previously<br />
observed both at high (40-100 bar) [50, 185] <strong>and</strong> low (~1 bar) [186] pressures <strong>of</strong><br />
hydrogen <strong>and</strong> has been shown to result from the partial hydrogenation <strong>of</strong> the quinoline<br />
(anchor) portion <strong>of</strong> the cinchonidine. The partially hydrogenated quinoline ring is<br />
unable to function as an anchor any longer due to the loss <strong>of</strong> the π-aromaticity <strong>and</strong> as a<br />
result, the cinchonidine desorbs from the surface resulting in an unmodified catalyst<br />
which is not capable <strong>of</strong> inducing enantioselectivity. Our results support the possibility<br />
that this effect also takes place with ‘quasi-homogeneous’ catalysts in the 2-10 bar<br />
range.<br />
To the best <strong>of</strong> our underst<strong>and</strong>ing <strong>and</strong> characterization <strong>of</strong> the effect, the decay <strong>of</strong> the<br />
actual ee after ITP was linearly approximated <strong>and</strong> the absolute value <strong>of</strong> the line’s slope<br />
was used to determine the stability <strong>of</strong> catalyst enantioselectivity with respect to the<br />
catalyst type <strong>and</strong> reaction conditions. Thus, in experiments with the same initial<br />
concentration <strong>of</strong> ethyl pyruvate we can easily see where no additional cinchonidine was<br />
added (Exp. 3, 4, slope ~ 0.18-0.22 min -1 ) the rate <strong>of</strong> decrease <strong>of</strong> actual enantiomeric<br />
excess was observed to be 3-4.5 times greater than with (Exp. 8, 9) free (20 mg or 1<br />
mmol l -1 ) cinchonidine (slope ~ 0.04-0.07 min -1 ).<br />
Qualitatively, this effect can be explained by maintaining a more or less constant π-<br />
bonded cinchonidine concentration on Pt surface due to the replacement <strong>of</strong> “old”<br />
molecules (with partially hydrogenated quinoline anchor) by “new” molecules <strong>of</strong><br />
cinchonidine from solution.<br />
Comparison <strong>of</strong> D, B ‘quasi-homogeneous’ <strong>and</strong> conventional<br />
heterogeneous catalysts<br />
Comparison <strong>of</strong> catalytic behavior<br />
The important principal difference in catalytic behavior <strong>of</strong> D, B <strong>and</strong> conventional<br />
Pt/Al 2 O 3 catalyst systems is the negligibly small increase <strong>of</strong> actual (non-cumulative)<br />
enantiomeric excess during the reaction in the case <strong>of</strong> the previously non-modified<br />
48
Pt/Al 2 O 3 (Exp. 5A). This increase <strong>of</strong> enantiomeric excess was found to be significantly<br />
smaller than in the series <strong>of</strong> experiments with pre-modified catalysts. This indicates the<br />
ITP effect is nearly non-existent for the systems which are modified in situ under our<br />
reaction conditions. That the ITP effect is hardly observed in acetic acid for catalysts<br />
without cinchonidine pre-modification is in agreement with the findings <strong>of</strong> Balazsik<br />
<strong>and</strong> Bartok [169]. In our system, the conventional Pt/Al 2 O 3 shows ee* max = 76 % which<br />
is the same value as was found in Exp. 3 <strong>and</strong> 8 with the D type <strong>of</strong> catalysts <strong>and</strong> also<br />
demonstrated similar values <strong>of</strong> initial rate. However, the rate <strong>of</strong> decrease <strong>of</strong> the actual<br />
enantiomeric excess is 3-6 times higher for the conventional catalyst than for the B <strong>and</strong><br />
D types <strong>of</strong> catalyst with the presence <strong>of</strong> the same amount <strong>of</strong> free cinchonidine in the<br />
reaction mixture (Exp. 8 <strong>and</strong> 9). Also, the D type <strong>of</strong> catalyst gave slightly higher<br />
maximum enantiomeric excess <strong>and</strong> initial rate than the B type under the same reaction<br />
conditions (compare Exp. 3 with 6B, Exp. 4 with 6A <strong>and</strong> Exp. 8 with 9) possibly due to<br />
the different preparation methods (used solvents, washing procedure, preparation<br />
temperature, etc.).<br />
The two-cycle mechanism proposed by Blaser, Garl<strong>and</strong> <strong>and</strong> Jallet [81, 180], based on<br />
the assumption that the reaction goes with [R]/[S] = 1 on unmodified sites <strong>and</strong> with<br />
[R]/([R]+[S]) = s on modified sites (near flat adsorbed cinchonidine), explains the<br />
observed correlation between ee max <strong>and</strong> initial rate (R) as ee max = A-B*R -1 , where<br />
A=(2s-1)k m /(k m -k u ), B=A*k u , k u <strong>and</strong> k m are pseudo-first-order rate constants <strong>of</strong><br />
unmodified <strong>and</strong> modified cycles. In fact, the graph ee max vs. R -1 (Fig. 4-16) shows a<br />
good linear correlation for all our experiments. This means that on modified sites the<br />
ratio [R]/[S]=(1/s)-1 is the same for all catalysts <strong>and</strong> does not depend (within the<br />
margin <strong>of</strong> error) on reaction conditions, this seems reasonable because the [R]/[S] ratio<br />
is determined by the nature <strong>of</strong> the interactions between molecules <strong>of</strong> cinchonidine, ethyl<br />
pyruvate <strong>and</strong> hydrogen on Pt surface. However the ratio <strong>of</strong> modified/unmodified sites<br />
might strongly depend on experimental conditions, hydrogen pressure (cleanliness <strong>of</strong><br />
the surface), type <strong>of</strong> catalyst, concentration <strong>of</strong> modifier, etc.<br />
80<br />
ee max<br />
, %<br />
60<br />
ee max<br />
= A - B*R -1<br />
40<br />
0.00 0.02 0.04 0.06 0.08<br />
Rate -1 , min.<br />
49
Fig. 4-16. Linear relationship between maximum reached enantiomeric excess <strong>and</strong><br />
inversed initial rate for Exp.1-9.<br />
Comparison <strong>of</strong> the stability <strong>of</strong> samples<br />
It is obvious that for the high performance, long activity <strong>and</strong> reuse <strong>of</strong> the catalyst the<br />
stability <strong>of</strong> its structure is very important aspect to focus on. Here we qualitatively<br />
compare stability <strong>of</strong> B type <strong>of</strong> colloidal sample before (Fig. 4-16-1) <strong>and</strong> after (Fig. 4-<br />
16-2) ethyl pyruvate hydrogenation. Precipitation <strong>of</strong> Pt particles due to the<br />
agglomeration says about low stability <strong>of</strong> the sample under exposure to hydrogen for a<br />
long time (approximately 2-3 hours, 10 bars <strong>of</strong> H 2 , no free cinchonidine). It should be<br />
notes, that sample can be kept in colloidal form in acetic acid solution for long (~1<br />
year) time if no hydrogen is applied.<br />
Fig. 4-17. Soluble (stable) cinchonidine modifier colloidal Pt nanoparticles in acetic<br />
acid before hydrogenation reaction (1) <strong>and</strong> precipitated (2) after exposure under 10 bars<br />
<strong>of</strong> hydrogen (in AcOH) for 2 hours.<br />
IR investigation<br />
The orientation <strong>of</strong> the cinchonidine species on Pt during the reaction was investigated<br />
by measuring the IR spectra <strong>of</strong> D type catalyst prior to <strong>and</strong> after 50 min <strong>and</strong> 24 hours<br />
under reaction conditions, at 10 bar. After 50 min, the reaction in Exp. 4 (P H2 = 10 bar,<br />
[cinchonidine] = 0 ml l -1 , [EP] = 0.514 mmol l -1 ) was terminated, the hydrogen pressure<br />
was released <strong>and</strong> the reactor opened to ambient conditions. After 1 hour the colloidal<br />
catalyst precipitated <strong>and</strong> a black powder was collected, washed three times with acetic<br />
acid, dried at 30 ºC under vacuum <strong>and</strong> finally was used for IR investigation. To insure<br />
that the washing procedure does not influence results, a control experiment was<br />
conducted both before <strong>and</strong> after washing B <strong>and</strong> D type catalysts (without exposure to<br />
the catalytic reaction conditions). Because the spectra revealed the presence <strong>of</strong> both the<br />
tilted <strong>and</strong> flat orientations <strong>of</strong> cinchonidine for all samples it was determined that the<br />
washing procedure does not influence the results.<br />
The reaction in Exp. 4 was repeated <strong>and</strong> when finished, the reaction mixture was kept<br />
in the reactor under 10 bar <strong>of</strong> hydrogen pressure for 24 hours after which the colloid<br />
solution formed a black precipitate that was collected, washed three times with acetic<br />
acid <strong>and</strong> dried at 30 ºC under vacuum <strong>and</strong> finally was used for IR investigation.<br />
For better visualization, intensity <strong>of</strong> the spectra (Fig. 4-18) <strong>of</strong> free cinchonidine was<br />
decreased. Intensities <strong>of</strong> the spectra <strong>of</strong> the D type catalyst are presented as measured.<br />
To aid in spectral analysis, the observed vibration frequencies are summarized in Table<br />
4-4.<br />
50
0.16<br />
0.12<br />
1<br />
1 Free cinchondine<br />
2 Before reaction<br />
3 After 1 h<br />
4 After 24 h<br />
Absorbance<br />
0.08<br />
2<br />
0.04<br />
3<br />
4<br />
0.00<br />
1700 1600 1500 1400 1300 1200 1100 1000<br />
Wavenumber, cm -1<br />
Fig. 4-18. FTIR spectra <strong>of</strong> the D type <strong>of</strong> catalyst at different reaction time <strong>and</strong> spectra<br />
<strong>of</strong> free cinchonidine.<br />
Table 4-4. Vibration b<strong>and</strong> assignment (cm -1 ) <strong>of</strong> the cinchonidine spectra <strong>of</strong> the free<br />
cinchonidine, <strong>and</strong> D type <strong>of</strong> catalyst before reaction, at 50 min <strong>and</strong> after 24 h under<br />
reaction condition.<br />
Free<br />
Before reaction At 50 min At 24 h Assignment a<br />
cinchonidine<br />
1162 1152 1152 1152 F: 1<br />
1207 1209 1209 1209 T: 2, 3<br />
1383 1383 - 1382 F: 1, T: 2, 3,<br />
acetic acid<br />
1420 1402 1402 - F: 1, acetic acid<br />
1451 1456 - - T: 2, 3<br />
1507 1507 - - T: 3<br />
1569 1570 1567 - F: 1, T: 2, 3<br />
1590 1590 - - T: 3<br />
Comments: a – F: 1 – flat adsorbed cinchonidine, T: 2 <strong>and</strong> T: 3 – cinchonidine adsorbed<br />
in tilted mode, according to the Fig. 1-19. Assignment <strong>of</strong> peaks had been done<br />
according to the work <strong>of</strong> Ferri <strong>and</strong> Burgi [76], Zaera [77, 178, 187] <strong>and</strong> Table 4-3 .<br />
Analysis <strong>of</strong> the IR spectra <strong>of</strong> the D system shows that prior to reaction, cinchonidine is<br />
adsorbed in the flat (1) <strong>and</strong> in two tilted modes: α-H abstracted (2) <strong>and</strong> in N-lone pair<br />
bonded (3). It was found that after 50 min under 10 bar <strong>of</strong> hydrogen the peaks at 1456<br />
cm -1 <strong>and</strong> 1383 cm -1 corresponding to the tilted (2 <strong>and</strong>/or 3) <strong>and</strong> peaks at 1407 cm -1 <strong>and</strong><br />
1590 cm -1 tilted (3) form <strong>of</strong> cinchonidine are absent <strong>and</strong> only those IR peaks at 1152<br />
cm -1 , 1209 cm -1 , 1402 cm -1 <strong>and</strong> 1567 cm -1 <strong>and</strong> indicating the flat-mode (1) <strong>and</strong> tilted<br />
mode (2) are observed. After 24 hours under 10 bar <strong>of</strong> hydrogen only peaks at 1152 cm -<br />
51
1 , 1209 cm -1 <strong>and</strong> 1382 cm -1 remain. The peaks at 1152 cm -1 <strong>and</strong> 1209 cm -1 could be<br />
assigned to flat (1) <strong>and</strong> tilted (2) modes, whereas the peak at 1382 cm -1 corresponds<br />
here to the acetic acid.<br />
The absence <strong>of</strong> the tilted (3) mode <strong>of</strong> the adsorbed cinchonidine (absence <strong>of</strong> the peaks<br />
at 1590 cm -1 , 1507 cm -1 <strong>and</strong> possibly 1456 cm -1 ) after approximately one hour <strong>of</strong><br />
hydrogen exposure indicates that N-lone pair bonded cinchonidine on Pt nanoclusters<br />
has a weak adsorption strength with respect to the flat (π-bonded) <strong>and</strong> α-H abstracted<br />
species. This fact has been demonstrated previously in the work <strong>of</strong> Ferri <strong>and</strong> Burgi [76]<br />
for a thin film <strong>of</strong> Pt. Further, we have recently demonstrated above that the adsorption<br />
modes for cinchonidine for both the thin film <strong>and</strong> nanocluster systems are similar<br />
(please the beginning <strong>of</strong> this chapter).<br />
Collectively, these experimental data <strong>of</strong> the system have led us to propose the model<br />
illustrated in the Fig. 4-19 for the observed phenomena. During ITP for our Pt<br />
nanocluster systems, desoprtion <strong>of</strong> weakly bound cinchonidine molecules (tilted<br />
orientation) is occurring as a result <strong>of</strong> interaction with hydrogen <strong>and</strong> possibly solvent,<br />
making free space for the formation <strong>of</strong> new chiral sites via adsorption <strong>of</strong> cinchonidine<br />
in the ‘flat’ conformation thus leading to an increase <strong>of</strong> ee. When cinchonidine<br />
coverage <strong>of</strong> the Pt surface reaches some minimum (due to hydrogenation <strong>and</strong>/or<br />
desorption) the nanoclusters become unstable <strong>and</strong> agglomeration takes place as was<br />
observed with B <strong>and</strong> D catalysts.<br />
cinchonidine<br />
in solution<br />
H 2<br />
1507, 1590 cm -1<br />
N<br />
N<br />
HO<br />
H<br />
N<br />
OH<br />
H<br />
H<br />
H<br />
N<br />
N<br />
H H CO/CO 2 /H 2 O<br />
OH<br />
H<br />
H<br />
N<br />
1 2 3<br />
Pt surface<br />
Fig. 4-19. Proposed model <strong>of</strong> behaviour <strong>of</strong> cinchonidine on Pt under hydrogen<br />
exposure. 1-“π-bonded”, 2- “α-H abstracted”, 3- “N lone pair bonded”.<br />
The addition <strong>of</strong> free cinchonidine leads to an increase <strong>of</strong> enantioselective stability <strong>of</strong><br />
these catalysts <strong>and</strong> increase <strong>of</strong> ITP likely due to the replacement “old” to “new”<br />
cinchonidine molecules on Pt.<br />
It is interesting to compare adsorption energies <strong>of</strong> hydrogen <strong>and</strong> cinchonidine on Pt<br />
surface in order to check a principal ability <strong>of</strong> hydrogen to replace cinchonidine from<br />
the surface. It is well known that hydrogen adsorb on Pt surface with further<br />
dissociation to two hydrogen atoms (here we consider strong hydrogen adsorption<br />
only), however adsorption energy <strong>of</strong> hydrogen atom on Pt depends on many factors<br />
such as type <strong>of</strong> Pt crystal plane, adsorption site (atop, bridge, etc.), size <strong>of</strong> the Pt cluster,<br />
nature <strong>of</strong> a support (Al 2 O 3 etc.), presence <strong>of</strong> electric charge, etc. The experimental<br />
values for hydrogen adsorption on Pt (111) range from 29 to 90 kJ mol -1 [188-190] .<br />
52
We do not consider an extreme values, <strong>and</strong> would like to take the average 45 kJ mol -1<br />
values for rough comparison, since it is considered for the uncharged, unsupported Pt<br />
(111) plane [191]. The adsorption energy <strong>of</strong> cinchonidine on Pt continuous plane<br />
deposited on Al 2 O 3 support in CH 2 Cl 2 was estimated to be 31 kJ mol -1 [76], note that<br />
this is average value assigned for all three species <strong>of</strong> cinchonidine adsorption. Thus, we<br />
can conclude, that hydrogen atoms, in principal, can remove all adsorbed species <strong>of</strong><br />
cinchonidine from Pt surface, however there is a high chance, that weakly adsorbed<br />
mode <strong>of</strong> cinchonidine (N-lone pair bonding) will be removed (desorbed) faster that<br />
others.<br />
4.4 Summary<br />
The orientation <strong>of</strong> the chiral modifier cinchonidine on the surfaces <strong>of</strong> Pt nanoclusters<br />
has been investigated for the first time using a combination <strong>of</strong> DRIFTS <strong>and</strong> molecular<br />
modeling methods. It was found that the size <strong>of</strong> Pt clusters (in range 1.2- ∞ nm) has no<br />
observed influence on the adsorption mode <strong>of</strong> the cinchonidine. Thus, it was found that<br />
both ‘flat’ <strong>and</strong> ‘tilted’ modes <strong>of</strong> adsorbed cinchonidine are present on Pt nanoclusters<br />
<strong>and</strong> could be an explanation <strong>of</strong> the fact that this type <strong>of</strong> catalyst does not provide the<br />
maximum possible enantiomeric excess (98 %).<br />
For the first time, the ITP was observed <strong>and</strong> studied for cinchonidine modified Pt<br />
‘quasi-homogeneous’ systems. The ITP was observed at room temperature (22 ± 1 ºC),<br />
in acetic acid media with catalysts prepared by two different methods (D <strong>and</strong> B type),<br />
with different catalyst loadings (1.5 - 20 mg), with different ethyl pyruvate<br />
concentration (0.257, 0.514 <strong>and</strong> 2.57 mol l -1 ) under 2 <strong>and</strong> 10 bar <strong>of</strong> hydrogen pressure<br />
both with (1 mmol l -1 ) <strong>and</strong> without addition <strong>of</strong> free cinchonidine into the reaction<br />
mixture.<br />
The ITP was found to be dependent on hydrogen pressure <strong>and</strong> catalyst loading,<br />
increasing from 30 min to 100 min with a decrease <strong>of</strong> hydrogen pressure from 10 bar to<br />
2 bar. Further, ITP decreases from 30 min to 9 min with an increase <strong>of</strong> catalyst (D-type)<br />
loading from 4 to 20 mg. ITP was also found to be independent <strong>of</strong> the initial<br />
concentration <strong>of</strong> ethyl pyruvate in the range <strong>of</strong> 0.257-2.57 mol l -1 . It was also found that<br />
cinchonidine modified Pt nanoclusters on alumina catalyst demonstrates (5-10%)<br />
higher ee after treatment under reaction condition without EP.<br />
Addition <strong>of</strong> free cinchonidine (1 mmol l -1 ) shows almost no detectable influence on the<br />
maximum value <strong>of</strong> ee <strong>and</strong> slight increase <strong>of</strong> initial rate, but increases ITP by a factor <strong>of</strong><br />
3 <strong>and</strong> enantioselective stability <strong>of</strong> catalyst by nearly 5 fold. The highest ee was also<br />
found to be independent on the initial concentration <strong>of</strong> the ethyl pyruvate <strong>and</strong> decreases<br />
with an increase <strong>of</strong> the catalyst loading from 78 % to 65 % in the range <strong>of</strong> 4 mg to 20<br />
mg, probably due to the incomplete chiral activation <strong>of</strong> the catalyst.<br />
The racemic ethyl lactates as an impurity in EP does not play a significant role in<br />
behavior <strong>of</strong> enantiomeric excess, however they make systematical down shift in ee,<br />
which decreases with conversion.<br />
From IR investigation, it was found that after approximately one hour under 10 bar <strong>of</strong><br />
hydrogen pressure in acetic acid, the weakly bound species which correspond to<br />
cinchonidine adsorption in the tilted (“N lone pair bonded”) mode are not observed <strong>and</strong><br />
only the strongly bound species corresponding to cinchonidine adsorption in the flat<br />
<strong>and</strong> in tilted (“α-H abstracted”) mode remain. In other words, “N lone pair bonded”<br />
species <strong>of</strong> the adsorbed cinchonidine are desorbed due to the influence <strong>of</strong> hydrogen <strong>and</strong><br />
possibly solvent.<br />
53
Our results are in line with previously published work <strong>of</strong> Balazsik <strong>and</strong> Bartok [169] in<br />
which the authors assigned the nature <strong>of</strong> the ITP to competitive adsorption <strong>of</strong> reactants,<br />
modifier <strong>and</strong> solvent <strong>and</strong> supplement it in desorption <strong>of</strong> weakly bonded cinchonidine<br />
adsorbed in tilted “N lone pair bonded” mode. It also has to be noted, that the hydrogen<br />
effect <strong>of</strong> cleaning Pt surface from previously adsorbed impurities: H 2 O/CO/CO 2 <strong>and</strong><br />
oxygen plays a major role in ITP [166].<br />
54
Chapter 5<br />
Cinchonidine modified Pt colloidal<br />
nanoparticles immobilized on a<br />
heterogeneous support<br />
5.1 Introduction<br />
Orito [45, 46] first reported the system <strong>of</strong> cinchonidine on Pt/Al 2 O 3 as a catalyst for<br />
enantioselective hydrogenation <strong>of</strong> α-ketoesters. The ability <strong>of</strong> Pt modified with<br />
cinchonidine to induce enantioselectivity has great fundamental <strong>and</strong> practical interest<br />
[21], for example in the synthesis <strong>of</strong> HPB ester (Ethyl (R)-2-hydroxy-4-phenylbutyrate)<br />
[22]. In order to obtain high enantiomeric excess, the reaction conditions were<br />
optimized by Orito <strong>and</strong> Blaser <strong>and</strong> summarized in a review <strong>of</strong> Studer, Blaser <strong>and</strong> Exner<br />
in 2003 [44]. As can be seen from this review, that high (≥ 40 bar) hydrogen pressure<br />
<strong>and</strong> thermal pretreatment (generally at 400 °C under flow <strong>of</strong> H 2 ) <strong>of</strong> the conventional<br />
Pt/Al 2 O 3 are critical requirements to obtain high ee. However, the use <strong>of</strong> high hydrogen<br />
pressure <strong>and</strong> thermal pretreatment <strong>of</strong> the Pt/Al 2 O 3 catalyst can induce technical<br />
difficulties, safety issues <strong>and</strong> increase production cost especially on an industrial scale.<br />
The available literature dealing with mild conditions (no thermal treatment, low (≤ 10<br />
bar) hydrogen pressure) are limited [51, 78, 192, 193].<br />
In fact, ee <strong>of</strong> 94 % was obtained in ethyl pyruvate hydrogenation by Sun <strong>and</strong> coworkers<br />
[78], under 5.8 bar H 2 , 17 °C, however, this system required maintaining low<br />
dihydrocinchonidine concentration through dosing <strong>of</strong> modifier (otherwise ee<br />
significantly decreases) throughout the reaction <strong>and</strong> a very specific reaction procedure<br />
which included “add catalyst <strong>and</strong> AcOH to reactor, hydrogenate for 1 h under the<br />
hydrogenation condition (e.g. 17 °C, 5.8 bar). Evacuate H 2 <strong>and</strong> add the modifier to the<br />
reactor under a N 2 purge. Apply H 2 to desired pressure <strong>and</strong> quickly add pyruvate to<br />
reactor under H 2 pressure using a syringe pump”. Despite the fact that the authors<br />
obtained 94 % ee without thermal pretreatment <strong>of</strong> the catalyst, such a reaction<br />
procedure seems to present its own set <strong>of</strong> difficulties.<br />
In another interesting work <strong>of</strong> Bartok <strong>and</strong> coworkers [51], 97 % ee was obtained in<br />
ethyl pyruvate hydrogenation under 10 bar <strong>of</strong> H 2 with ultrasonic activation (20 kHz, 30<br />
W) in a closed hydrogen atmosphere, the activated catalyst thus requires h<strong>and</strong>ling<br />
without exposure to air. Further, using this ultrasonic activation it was reported that<br />
changes in the size <strong>and</strong> morphology <strong>of</strong> Pt nanoclusters on alumina support occurred.<br />
It is also should be noted, that the “CatASium F214” [192] catalyst developed by<br />
Degussa <strong>and</strong> Solvias does not require thermal activation <strong>and</strong> can induce e.g. 94 % ee in<br />
ethyl pyruvate hydrogenation under 60 bar <strong>of</strong> H 2 . “Alternatively, the reaction can also<br />
be carried out at 1.1 bar hydrogen pressure in a glass shaker <strong>and</strong> the hydrogen can be<br />
supplied from a balloon. However, activity <strong>and</strong> selectivity don't reach an optimum”<br />
[193].<br />
Herein, we present the detailed preparation <strong>of</strong> a novel heterogeneous catalyst:<br />
cinchonidine modified Pt nanoclusters deposited on non-porous alumina which<br />
demonstrates 80-88 % ee in hydrogenation <strong>of</strong> ethyl pyruvate under relatively low<br />
hydrogen pressure (2.5-10 bar). The reaction procedure does not require thermal or<br />
ultrasonic activation <strong>of</strong> the catalyst <strong>and</strong> allows storage <strong>of</strong> the catalyst in air before<br />
55
eaction. The stability <strong>of</strong> the catalyst structure under reaction <strong>and</strong> preparation<br />
conditions is considered as well as catalyst recycling <strong>and</strong> reuse. The influence <strong>of</strong> the<br />
catalyst nature towards obtaining high enantiomeric excess in this pressure range is also<br />
considered.<br />
5.2 Experimental<br />
Materials<br />
5% Pt on alumina (Aldrich, dispersion 29 %, Pt particles size 3.9 nm), cinchonidine<br />
(98%, Fluka), acetic acid (99.8% Fluka), ethyl pyruvate (98%, Aldrich), formic acid<br />
(Aldrich, 99%), acetone <strong>and</strong> aluminum oxide (4 m 2 g -1 ) 90 active neutral (0.063-0.2<br />
mm) by Applichem were used as received.<br />
Samples preparation<br />
Adsorption <strong>of</strong> PtCl 4 on alumina was performed according to the following procedure. 4<br />
g Al 2 O 3 was heated at 180 °C under vacuum (10 -3 mbar) for 16 hours <strong>and</strong> then was<br />
cooled to room temperature under vacuum. To this alumina, 16 mg PtCl 4 dissolved in<br />
40 ml <strong>of</strong> acetone solution was added by syringe. After 5-10 minutes stirring, the<br />
initially dark supernatant acetone solution became colorless <strong>and</strong> an additional 10 ml<br />
aliquot <strong>of</strong> acetone (with 30 mg PtCl 4 dissolved) was added by syringe. This procedure<br />
was repeated several times, until 480 mg PtCl 4 was added, at which point the<br />
supernatant no longer became colorless. The mixture was then refluxed for 12 hours<br />
<strong>and</strong> the supernatant was evaporated under vacuum. Thus 480 mg PtCl 4 on 4000 mg<br />
alumina (~11 % PtCl 4 ) was obtained.<br />
In order to reduce Pt The 600 mg <strong>of</strong> alumina with adsorbed PtCl 4 was transferred into a<br />
round bottom flask with distilled water (100 ml). The mixture was refluxed with<br />
constant stirring. 160 mg dihydrocinchonidine dissolved in 25 ml aqueous solution <strong>of</strong><br />
formic acid 0.1 mol·L -1 was added rapidly by syringe <strong>and</strong> Pt was reduced. The heat was<br />
removed after 10 min. The mixture was allowed to stir for 24 hours, the black powder<br />
was isolated by filtration <strong>and</strong> washed with water, then with acetic acid <strong>and</strong> finally dried<br />
in air at room temperature. Thus prepared catalysts will be referred to as A1 henceforth<br />
in this manuscript.<br />
For catalyst activation140 mg <strong>of</strong> the A1 catalyst was transferred into a metal reactor<br />
with 70 ml acetic acid <strong>and</strong> 20 mg cinchonidine. The mixture was maintained under<br />
hydrogen pressure (10 bar) for 1 hour with constant stirring (700 rpm). The pressure<br />
was then released <strong>and</strong> the black powder was collected, washed 3 times with acetic acid<br />
(100 ml), <strong>and</strong> then dried in air at room temperature. Thus prepared catalysts will be<br />
referred as A2 henceforth in this manuscript.<br />
Unless otherwise mentioned, the conventional Pt/Al 2 O 3 was heated to 400 °C under<br />
vacuum (10 -3 mbar) for 1 hour, then kept for 3 hours under a flowing mixture <strong>of</strong> N 2<br />
(93%) <strong>and</strong> H 2 (7%) gases (100 ml min -1 ) <strong>and</strong> subsequently cooled to the room<br />
temperature. The required amount was then weighed (in open atmosphere) <strong>and</strong><br />
immediately transferred into a glass reactor with acetic acid, cinchonidine <strong>and</strong> ethyl<br />
pyruvate.<br />
Modification <strong>of</strong> conventional Pt/Al 2 O 3 has been done, according to the following<br />
procedure. 40 mg <strong>of</strong> conventional Pt/Al 2 O 3 was heated to 400 ºC under vacuum (10 -3<br />
mbar) for 1 hour, then the mixture <strong>of</strong> N 2 (93%) <strong>and</strong> H 2 (7%) gases were passed (100 ml<br />
min -1 ) over the sample for 3 hours. Then, under flowing gases, the support was cooled<br />
to room temperature <strong>and</strong> the solution <strong>of</strong> cinchonidine in CHCl 3 (6.0 ml, 3 mM) was<br />
56
added by syringe through a stopper. The system was kept at 10-15 ºC for 24 hours, then<br />
washed 3 times with CHCl 3 <strong>and</strong> finally dried under vacuum at room temperature.<br />
Characterization <strong>and</strong> experiments<br />
Transmission electron microscopy (TEM) micrographs were taken using a TECNAI<br />
F20 instrument operating at 200 kV. Specimens were prepared by placing a drop <strong>of</strong> an<br />
ethanolic sample suspension onto a copper grid with a perforated carbon film <strong>and</strong> then<br />
allowing the ethanol to evaporate.<br />
For IR investigation a “Nicolet 4700 FT-IR” spectrometer with liquid nitrogen cooled<br />
detector <strong>and</strong> nitrogen purged DRIFT accessory was used with the following parameters:<br />
200 scans, 600-4000 cm -1 scan range, 4 cm -1 resolution. KBr background was acquired<br />
before spectra measurements. The DRIFT spectra were chosen for analysis because it is<br />
the more applicable technique for observing adsorbed species on finely divided metal<br />
powders [174].<br />
Hydrogenation <strong>of</strong> ethyl pyruvate<br />
Hydrogenation <strong>of</strong> ethyl pyruvate was carried out in a metal reactor (Parr Instrument<br />
GmbH) under a pressure <strong>of</strong> 2.5-10 bar <strong>of</strong> hydrogen at room temperature (22±1 °C). The<br />
stirring speed was 1000 rpm.<br />
Typical conditions were such that 4 ml (36 mmol) ethyl pyruvate was added to 20 mg<br />
(68 µmol) cinchonidine in acetic acid (70 ml) solution. 20 mg <strong>of</strong> catalyst was added<br />
into the mixture <strong>and</strong> then transferred into the reactor, flushed with Ar for 10 min <strong>and</strong><br />
finally, after the pressure <strong>of</strong> hydrogen had been stabilized for 10-15 sec, the initial time<br />
(t=0) was set <strong>and</strong> the reaction kinetics were monitored. About one ml aliquots <strong>of</strong> the<br />
reaction mixture were taken at certain time intervals, filtrated with a G3/G4 glass filter<br />
<strong>and</strong> 1 μl was injected into the gas chromatograph (GC). GC measurements were taken<br />
using a Varian 3900 instrument with FID detector <strong>and</strong> Lipodex-E column.<br />
Catalyst stability <strong>and</strong> reuse tests<br />
In order to test the stability <strong>of</strong> the catalysts under the typical reaction conditions, 30 mg<br />
A1 was transferred into the reactor with cinchonidine (250 mg) in 70 ml acetic acid<br />
solution <strong>and</strong> left with continuous stirring under 10 bar <strong>of</strong> hydrogen pressure for 16<br />
hours. The pressure was then released <strong>and</strong> samples were collected for TEM analysis.<br />
In the case <strong>of</strong> catalyst reuse experiments, after the reaction is finished (conversion<br />
reached 100 %), hydrogen pressure was released <strong>and</strong> the reactor was unplugged. The<br />
black powder was collected through centrifugation <strong>and</strong> washed three times with acetic<br />
acid, then was dried in air at room temperature; the catalyst was then used again in the<br />
identical manner to that described above.<br />
5.3 Results <strong>and</strong> discussion<br />
5.3.1 Role <strong>of</strong> catalyst activation in obtaining enantiomeric<br />
excess<br />
The A1 catalyst (Table 5-1) demonstrates 79-80 % <strong>of</strong> ee which is similar to the data<br />
reported by Bönnemann [114, 115] for alumina supported cinchonidine modified Pt<br />
nanoclusters. The same catalyst after activation (A2) demonstrates 85 % <strong>of</strong> ee <strong>and</strong> an<br />
increase <strong>of</strong> reaction rate by a factor <strong>of</strong> ~9. The A2 catalyst demonstrates stable ee <strong>and</strong><br />
100 % <strong>of</strong> conversion after 20-25 min (Fig. 5-1).<br />
57
Table 5-1. Ee <strong>and</strong> reaction rate induced by cinchonidine modified catalysts before <strong>and</strong><br />
after activation.<br />
Catalyst A Ee, % Reaction rate, mmol<br />
(min·g) -1<br />
A1- Before activation 79-80 3<br />
A2 -After activation 85 28<br />
100<br />
100<br />
Conversion, %<br />
80<br />
60<br />
40<br />
20<br />
80<br />
60<br />
40<br />
20<br />
Enantiomeric excess, %<br />
0<br />
0 5 10 15 20<br />
Time, min<br />
0<br />
Fig. 5-1. Evolution <strong>of</strong> conversion (squares) <strong>and</strong> enantiomeric excess (triangles) for A2<br />
catalyst under 10 bar <strong>of</strong> hydrogen pressure.<br />
A similar effect <strong>of</strong> increased ee (10-20%) after catalyst exposure to reaction condition<br />
was found by Bartok <strong>and</strong> Balàzsik [194] for conventional Pt/Al 2 O 3 (E4759) in toluene<br />
but was not found in acetic acid, probably, due to the differences in the catalysts. The<br />
proposed origin <strong>of</strong> ee <strong>and</strong> rate enhance for our catalyst is discussed below.<br />
5.3.2 Comparison <strong>of</strong> catalysts <strong>and</strong> proposed model <strong>of</strong><br />
catalyst activation<br />
In order to compare cinchonidine modified catalysts A1 <strong>and</strong> A2 with modified <strong>and</strong><br />
unmodified conventional catalyst (Aldrich) (the latter is modified with cinchonidine in<br />
situ) we tested them in ethyl pyruvate hydrogenation in the range <strong>of</strong> 2.5-10 bar <strong>of</strong><br />
hydrogen pressure.<br />
The major difference in selectivity between cinchonidine modified conventional<br />
Pt/Al 2 O 3 (see modification procedure in experimental part) <strong>and</strong> A1 catalyst was<br />
observed in the ethyl pyruvate hydrogenation without addition <strong>of</strong> free cinchonidine. In<br />
fact, the A1 catalyst induces 71 % <strong>of</strong> ee, whereas cinchonidine modified conventional<br />
Pt/Al 2 O 3 demonstrates only 40 % ee.<br />
It should be noted that without free cinchonidine, the A1 catalyst shows 71 % <strong>of</strong> ee <strong>and</strong><br />
ee decreases with reaction time (Fig. 5-2), whereas in the experiment with the addition<br />
58
<strong>of</strong> free cinchonidine, the catalyst demonstrates 80 % <strong>of</strong> ee <strong>and</strong> no decrease <strong>of</strong> ee was<br />
found during reaction. Another important fact is that just 10 % <strong>of</strong> conversion after 80<br />
min (Fig. 5-2) was reached in the ethyl pyruvate hydrogenation without free<br />
cinchonidine, whereas conversion <strong>of</strong> 70 % after 80 min (not shown) was reached with<br />
free cinchonidine with the same catalyst <strong>and</strong> under the same conditions. This<br />
phenomena can be explained as sorption/desorption <strong>of</strong> cinchonidine between Pt <strong>and</strong><br />
solution. Desorption <strong>of</strong> cinchonidine from Pt surface has been already observed even<br />
under 1 bar <strong>of</strong> hydrogen <strong>and</strong> explained as partial hydrogenation <strong>of</strong> quinoline anchor<br />
[169].<br />
10<br />
72<br />
Conversion, %<br />
8<br />
6<br />
4<br />
70<br />
Enantiomeric excess, %<br />
2<br />
68<br />
0<br />
0 20 40 60 80<br />
Time, min<br />
Fig. 5-2. Evolution <strong>of</strong> conversion (squares) <strong>and</strong> enantiomeric excess (triangles) for A1<br />
catalyst without free cinchonidine, hydrogen pressure 10 bar.<br />
Based on IR investigations, we would like to suggest here that A1 (<strong>and</strong> A2) possess a<br />
higher concentration <strong>of</strong> adsorbed cinchonidine with respect to modified conventional<br />
Pt/Al 2 O 3 . Despite the fact that both catalysts have Pt particles <strong>of</strong> similar size (3 (TEM)<br />
<strong>and</strong> 3.9 nm (chemisorption)) the A1 catalyst (made by reduction <strong>of</strong> the Pt salt) is more<br />
active than conventional Pt/Al 2 O 3 , most probably due to the fact the latter has a lower<br />
level <strong>of</strong> cleanliness <strong>of</strong> the Pt surface (even after treatment with H 2 at 400 °C) with<br />
respect to freshly reduced <strong>and</strong> modified Pt.<br />
For ethyl pyruvate hydrogenation with the addition <strong>of</strong> free cinchonidine (Fig. 5-3) it<br />
was found that in the range <strong>of</strong> 2.5- 10 bar <strong>of</strong> hydrogen pressure, A2 demonstrates<br />
approximately 7-8 % higher enantiomeric excess <strong>and</strong> a higher reaction rate by a factor<br />
<strong>of</strong> 1.7-2.3 than the conventional (non modified in prior) Pt/Al 2 O 3 at the same<br />
conditions.<br />
59
Enantiomeric excess, %<br />
85<br />
80<br />
75<br />
A2<br />
Conventional Pt/Al 2<br />
O 3<br />
40<br />
30<br />
20<br />
Reaction rate, mmol/min*g cat<br />
10<br />
70<br />
2 4 6 8 10<br />
Hydrogen Pressure, bar<br />
Fig. 5-3. Influence <strong>of</strong> hydrogen pressure on enantiomeric excess (filled symbols) <strong>and</strong><br />
reaction rate (open symbols) for conventional Pt/Al 2 O 3 (triangles) <strong>and</strong> A2 catalysts<br />
(squares).<br />
Our proposal that this is probably due to higher surface concentration <strong>of</strong> active chiral<br />
sites for the A2 catalyst is based on the fact that the A2 catalyst has already adsorbed<br />
cinchonidine (in flat <strong>and</strong> in tilted adsorption mode, see section 5.3.4 for IR<br />
investigation) <strong>and</strong> adsorbs additional cinchonidine (in both modes) from solution <strong>and</strong> at<br />
the same time the conventional Pt/Al 2 O 3 (which is previously unmodified) adsorbs<br />
cinchonidine only from solution (also in both modes) during reaction (in situ).<br />
An increase <strong>of</strong> ee with increased hydrogen pressure was found for both A2 <strong>and</strong> the<br />
conventional catalyst. This increase could be explained through the hydrogen cleaning<br />
effect. In fact, it was shown that hydrogen cleans the Pt surface <strong>of</strong> different impurities<br />
from solvent, ethyl pyruvate, <strong>and</strong> air (O 2 , CO/CO 2 ) [166, 195]. It is known, that the<br />
likelihood <strong>of</strong> the occurrence for flat adsorption mode <strong>of</strong> the cinchonidine (active chiral<br />
site) depends on the concentration <strong>of</strong> modifier in the solution [76, 187], however it is<br />
clear, that cinchonidine adsorbed in flat mode requires more free space on the Pt<br />
surface with respect to any tilted adsorption mode. Therefore, we suppose that with an<br />
increase <strong>of</strong> concentration <strong>of</strong> free sites (due the hydrogen cleaning effect) the chance <strong>of</strong><br />
flat adsorption increases.<br />
A1 after activation (see experimental part) is referred to as A2 was found to have a<br />
lower concentration <strong>of</strong> CO contamination on the Pt surface (see IR investigation section<br />
5.3.4) with respect to its state before activation, <strong>and</strong> demonstrates a plateau in the plot<br />
<strong>of</strong> ee vs P H2 (Fig. 5-3). The presence <strong>of</strong> the plateau for the A2 catalyst <strong>and</strong> its absence<br />
for conventional catalysts indicates differences in the processes associated with each<br />
system.<br />
It should be noted that, since all catalysts were kept in air <strong>and</strong> thermal treatment at 400<br />
°C is not possible for those catalysts modified during preparation, these catalysts as<br />
well as the conventional Pt/Al 2 O 3 catalyst were used without treatment at high<br />
60
temperature in all <strong>of</strong> the above mentioned experiments. However, thermally treated<br />
(with hydrogen according to procedure described above) conventional Pt/Al 2 O 3<br />
demonstrates 91 % ee (10 bar H 2 ) which is 3 % higher than maximum obtained with A2<br />
catalyst under the same reaction conditions. This significant (~14 %) increase in ee<br />
after catalyst treatment clearly demonstrates importance <strong>of</strong> surface cleaning.<br />
It is known, that platinum can be oxidized by O 2 <strong>and</strong> form PtO or PtO x [196] under red<br />
heat <strong>and</strong> Pt is able to dissociatively chemisorb oxygen at room temperature, other<br />
proposed models <strong>of</strong> oxygen adsorption on Pt can be found in the works <strong>of</strong> Smith [197]<br />
<strong>and</strong> Matsushima [198] <strong>and</strong> others [199-201]. In our experiments with <strong>and</strong> without<br />
thermal pretreatment (catalyst was transferred to reactor in air at room temperature) <strong>of</strong><br />
the Pt catalyst, chemisorption <strong>of</strong> oxygen on Pt likely occurs. However, it does not<br />
prevent cinchonidine <strong>and</strong> ethyl pyruvate adsorption with the presence <strong>of</strong> hydrogen in a<br />
solvent (in situ) <strong>and</strong> min. 77 % <strong>of</strong> ee (conventional Pt/Al 2 O 3 ) was observed due to the<br />
surface cleaning effect, induced by hydrogen [76, 82, 195, 202]. The significant<br />
increase <strong>of</strong> ee (up to 91 %, in our work) with thermally treated Pt/Al 2 O 3 has been<br />
known since report <strong>of</strong> Orito et al. [45] <strong>and</strong> was explained as further surface cleaning<br />
[203] e.g. from organic species, H 2 O, CO, CO 2 , etc.<br />
The global role <strong>of</strong> oxygen in the reaction is rather ambiguous [69], for example, it was<br />
shown that oxygen induces a cleaning effect through CO removal [195, 204] <strong>and</strong><br />
oxygen was to reduce cinchonidine coverage on Pt in such a way that a larger fraction<br />
<strong>of</strong> the Pt surface would be available for ethyl pyruvate hydrogenation [205].<br />
5.3.3 Catalyst stability <strong>and</strong> reuse tests<br />
The TEM investigation <strong>of</strong> sample A1 before the stability test (Fig. 5-4) shows that<br />
alumina globules are fully covered with Pt nanoclusters with an average size about 2.7<br />
nm. The small increase <strong>of</strong> average particles size (maximum <strong>of</strong> 0.5 nm) was observed<br />
from the sample following the stability test; in fact the average size <strong>of</strong> Pt nanoclusters<br />
after the stability test was found to be about 3.2 nm. It should be noted that the surface<br />
<strong>of</strong> an alumina globule after the stability test is still fully covered with Pt nanoclusters<br />
<strong>and</strong> ultrasonic treatment (at least for half an hour) does not influence the size <strong>and</strong><br />
stability <strong>of</strong> alumina supported Pt nanoclusters.<br />
61
35<br />
30<br />
Before<br />
After<br />
Percent <strong>of</strong> particles, %<br />
25<br />
20<br />
15<br />
10<br />
5<br />
0<br />
1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5<br />
Size, nm<br />
Fig. 5-4. TEM images <strong>of</strong> fully covered alumina globule before (above) <strong>and</strong> after<br />
(below) stability test <strong>and</strong> a histogram representing the Pt particle size distribution<br />
before <strong>and</strong> after catalyst stability test.<br />
Since, there is no significant change in the size <strong>of</strong> Pt nanoclusters after the reaction time<br />
was exceeded; the possibility <strong>of</strong> recycling was explored. The A2 catalyst (see Table 5-<br />
2) was found to be fully reusable (at least four times) without loss <strong>of</strong> activity <strong>and</strong><br />
enantioselectivity. Moreover, a slight increase <strong>of</strong> ee was observed, which could be<br />
explained as the result <strong>of</strong> additional activation after first reaction.<br />
Table 5-2: A2 catalyst reuse data under 10 bar <strong>of</strong> hydrogen pressure.<br />
Number <strong>of</strong> use Ee, % Reaction rate, mmol<br />
(min·g) -1<br />
1 85 28<br />
2 87 25<br />
3 86 29<br />
4 88 30<br />
It is known from the work <strong>of</strong> Reschetilowski <strong>and</strong> co-workers [206] (home-made<br />
Pt/zeolite) <strong>and</strong> work <strong>of</strong> Bartok <strong>and</strong> co-workers [194] (conventional Pt/Al 2 O 3 ), that ee<br />
could be constant (Reschetilowski) or decrease <strong>of</strong> 5 % in catalyst reuse experiments,<br />
where “fresh” (dihydro)cinchonidine was added in every reuse. However, a significant<br />
decrease <strong>of</strong> reaction rate was detected in both works, which might indicate a decrease<br />
<strong>of</strong> catalyst activity, whereas the A2 catalyst does not demonstrate a decrease <strong>of</strong> ee or<br />
62
eaction rate, moreover, the ee <strong>and</strong> reaction rate were slightly increased, probably due<br />
to the further activation.<br />
5.3.4 IR investigation <strong>and</strong> proposed model <strong>of</strong> activation<br />
As we have shown already (chapter 4) cinchonidine demonstrates similarity in<br />
adsorption on Pt nanoclusters (1.2 nm size) with respect to the macroscopic Pt support.<br />
In fact, cinchonidine was found in both the tilted <strong>and</strong> flat adsorption modes in both<br />
cases. By using previously obtained data from modeling <strong>of</strong> IR vibrations <strong>of</strong> the<br />
cinchonidine molecule in open 3 conformation (chapter 4) <strong>and</strong> data from the work <strong>of</strong><br />
Ferri [76] <strong>and</strong> Zaera [77] the assignment <strong>of</strong> the observed peaks has been accomplished.<br />
Fig. 5-5 shows the presence <strong>of</strong> peaks <strong>and</strong> shoulders that are in correlation with the<br />
spectrum <strong>of</strong> free cinchonidine. Here, modification <strong>of</strong> both: conventional catalysts <strong>and</strong><br />
A1 prepared as described above is clearly established. It should be noted that, in the<br />
case <strong>of</strong> A1, peaks corresponding to the adsorbed cinchonidine are better resolved <strong>and</strong><br />
have higher intensity with respect to cinchonidine modified conventional Pt/Al 2 O 3 due<br />
to the higher concentration <strong>of</strong> adsorbed cinchonidine on the Pt surface.<br />
1640<br />
0.5<br />
D<br />
F<br />
A - 71 % ee<br />
1500 1000<br />
Adsorbance<br />
0.25<br />
B - 40 % ee<br />
C<br />
1700 1600 1500 1400<br />
Wavenumber,<br />
1300<br />
cm<br />
1200 1100 1000<br />
Fig. 5-5. DRIFTS <strong>of</strong> cinchonidine modified A1 (A), conventional Pt/Al 2 O 3 (B) <strong>and</strong> free<br />
cinchonidine (C). The unmodified Pt/Al 2 O 3 (Aldrich) <strong>and</strong> alumina presented in inset as<br />
F <strong>and</strong> D, correspondingly. Numbers near symbols A <strong>and</strong> B show maximum obtained<br />
enantiomeric excess with respective catalyst without addition <strong>of</strong> free cinchonidine<br />
under 10 bar <strong>of</strong> H 2 . The intensity <strong>of</strong> spectra <strong>of</strong> free cinchonidine was decreased for<br />
better comparison.<br />
A DRIFTS investigation <strong>of</strong> cinchonidine modified Pt nanoclusters on alumina support<br />
was performed before (A1) <strong>and</strong> after catalyst activation (A2). The practical absence <strong>of</strong><br />
the peaks at ~2100 cm -1 (Fig. 5-6) which corresponds to adsorbed CO [202, 207] after<br />
63
catalyst activation seems to indicate a Pt surface cleaning effect. This cleaning effect<br />
was also observed by Baiker <strong>and</strong> co-authors for similar systems [166, 202].<br />
Absorbance<br />
0.004<br />
A1<br />
A2<br />
2300 2200 2100 2000 1900<br />
Wavenumber, cm -1<br />
Fig.5-6. DRIFT spectra <strong>of</strong> CO adsorbed on A1 <strong>and</strong> A2 catalysts.<br />
It should be noted, that the spectral range below 1200 cm -1 is not accessible for<br />
investigation, due to the alumina background (Fig. 5-5, 5-7) which covers the frequency<br />
range below 1200 cm -1 where the most useful (for unambiguous assignment <strong>of</strong> the flat<br />
adsorbed species <strong>of</strong> the cinchonidine) out <strong>of</strong> plane vibrations <strong>of</strong> quinoline ring are<br />
located. Further, the presence <strong>of</strong> alumina “dilutes” the cinchonidine in the IR beam <strong>and</strong><br />
lowers peak resolution (especially around 1580 cm -1 ) with respect to cinchonidine<br />
stabilized Pt nanoclusters (support free) (please see chapter 4). Both <strong>of</strong> these effects<br />
hinder unambiguous interpretation <strong>of</strong> the change in spectra. However, some significant<br />
changes are clearly observed, they are summarized in the Table 5-4 <strong>and</strong> we provide the<br />
following interpretation.<br />
64
A2 -88 % ee<br />
Adsorbance<br />
1709<br />
1613<br />
1583<br />
1572<br />
1470<br />
1463<br />
1422<br />
1422<br />
1371<br />
1340 1336<br />
1280<br />
A1 -80 % ee<br />
C<br />
0.5<br />
1700 1600 1500 1400 1300 1200 1100 1000<br />
Wavenumber, cm -1<br />
Fig. 5-7. DRIFT spectra <strong>of</strong> A1 <strong>and</strong> A2 catalysts <strong>and</strong> free cinchonidine (C). Numbers<br />
near symbols A <strong>and</strong> B show enantiomeric excess obtained with respective catalyst with<br />
addition <strong>of</strong> free cinchonidine (1 mol L -1 ) under 10 bar <strong>of</strong> H 2 . The intensity <strong>of</strong> spectra <strong>of</strong><br />
free cinchonidine was decreased for better performance.<br />
The appearance <strong>of</strong> the shoulder at 1613 cm -1 after activation could be explained as an<br />
increase <strong>of</strong> concentration <strong>of</strong> cinchonidine in the “N lone pair bonded” adsorption mode.<br />
The peak at 1572 cm -1 before <strong>and</strong> 1583 cm -1 after activation are super positions <strong>of</strong><br />
peaks/shoulders (which could be seen separately in DRIFT spectra <strong>of</strong> unsupported<br />
cinchonidine on Pt nanocluster samples (chapter 4)) corresponding to 1619, 1591 <strong>and</strong><br />
1568 cm -1 peaks in the spectra <strong>of</strong> free cinchonidine. Such a shift as that from 1572<br />
before to 1583 cm -1 after activation can be explained as partial interconversion <strong>of</strong> one<br />
or more adsorption mode (flat 1, tilted 2, tilted 3) adsorption mode to tilted 3. Also, an<br />
increase <strong>of</strong> the relative intensity <strong>and</strong> blue shift in wavenumber from 1463 (wk) to 1470<br />
(med) cm -1 allows interpretation as an interconversion <strong>of</strong> adsorption modes (flat 1,<br />
tilted 2, tilted 3) between each other <strong>and</strong> taking into the account the assumption that a<br />
larger shift in wavelength results from stronger adsorption, thus it is possible to propose<br />
that the concentration <strong>of</strong> tilted 2 mode decreased with respect to flat 1 or/<strong>and</strong> tilted 3<br />
modes. No significant change in intensity or in wavenumber was found for the peak at<br />
1422 cm -1 , corresponding to the flat 1 adsorption mode, whereas an increase <strong>of</strong><br />
intensity at 1371 cm -1 led to the appearance <strong>of</strong> the shoulder which corresponds to the<br />
flat 1 adsorption mode. The decrease <strong>of</strong> intensity at 1340 cm -1 before activation <strong>and</strong> red<br />
shift to 1336 cm -1 after activation is difficult to unambiguously explain since there is<br />
not a clear assignment <strong>of</strong> this vibration b<strong>and</strong> with an adsorption mode. The situation is<br />
similar with the peak at 1264 cm -1 before activation <strong>and</strong> the weak shoulder at 1264 cm -1<br />
after activation. The latter is almost covered by the broad peak at 1275 cm -1 which can<br />
be assigned to acetic acid.<br />
65
Table 5-3. Assignment <strong>of</strong> some vibrations (cm-1) <strong>of</strong> cinchonidine adsorbed on Pt<br />
before (A1) <strong>and</strong> after activation (A2).<br />
A1 A2 Free Assignment <strong>and</strong> orientation <strong>of</strong><br />
cinchonidine. adsorbed cinchonidine<br />
- 1709 (st. wd.) - Adsorbed acetic acid<br />
- 1613 (sh) 1619 (wk) Tilted 3<br />
1572 (st) 1583 (st) 1591 (st), Tilted 3<br />
1568 (med) Flat 1, tilted 2 <strong>and</strong> tilted 3<br />
1463 (wk) 1470 (med) 1452 (str) Flat 1, tilted 2 <strong>and</strong> tilted 3<br />
1422 (med) 1422 (med) 1422 (med) Flat 1<br />
- 1371 (sh) 1355 (med) Flat 1<br />
1340 (med) 1336 (sh) 1327 (med), -<br />
1337 (med) Flat 1<br />
- 1275 (wd) - Adsorbed acetic acid<br />
1264 (med) 1264 (wk, sh.) - -<br />
Comments: st. – strong, wd. – wide, sh. – shoulder, wk. – weak, med. – medium, flat: -1<br />
“π-bonded” adsorption mode, tilted: 2 - “α-H abstracted” adsorption mode, tilted: 3 -<br />
“N lone pair bonded” adsorption.<br />
In order to assess the ability <strong>of</strong> the cinchonidine to adsorb on used alumina (pure<br />
alumina passed all preparation <strong>and</strong> activation steps) the control experiments were<br />
performed <strong>and</strong> it was found from DRIFT spectra that no cinchonidine peaks were found<br />
on alumina after washing with acetic acid. Further, no conversion <strong>of</strong> ethyl pyruvate was<br />
observed by using cinchonidine <strong>and</strong> alumina without Pt.<br />
Based on the data described above we would like to propose a model for catalyst<br />
activation. Thus, activation results from a decrease <strong>of</strong> CO coverage <strong>of</strong> Pt surface <strong>and</strong> an<br />
increase <strong>of</strong> the concentration <strong>of</strong> flatly bound cinchonidine (active chiral site), probably<br />
through reorganization between adsorbed cinchonidine species (partial desorption <strong>of</strong><br />
weakly bound cinchonidine in tilted mode to solution) <strong>and</strong> additional adsorption <strong>of</strong><br />
cinchonidine from solution onto the produced free sites as well <strong>of</strong> adsorption <strong>of</strong> acetic<br />
acid, which could be competitive.<br />
5.4 Summary<br />
High enantiomeric excess in ethyl pyruvate hydrogenation was obtained at low – 2.5-10<br />
bar hydrogen pressure with conventional Pt/Al 2 O 3 - 91 % ee (after thermal pretreatment<br />
with hydrogen) <strong>and</strong> 80-88 % ee (without thermal pretreatment) with cinchonidine<br />
modified Pt nanoclusters deposited on nonporous alumina (after activation with<br />
hydrogen <strong>and</strong> cinchonidine in acetic acid at room temperature) without requirement <strong>of</strong><br />
avoiding catalyst contact with air. The stability <strong>of</strong> the catalyst under reaction conditions<br />
<strong>and</strong> recycling <strong>and</strong> reuse <strong>of</strong> the catalyst was successfully demonstrated.<br />
The positive effect <strong>of</strong> activation under reaction conditions was demonstrated <strong>and</strong><br />
explained as removal <strong>of</strong> contaminants from the Pt surface <strong>and</strong> reorganization <strong>of</strong> the<br />
adsorbed cinchonidine species <strong>and</strong> adsorption <strong>of</strong> a new portion <strong>of</strong> lig<strong>and</strong> from solution.<br />
These effects led to an increase <strong>of</strong> the concentration <strong>of</strong> active chiral sites on Pt surface<br />
<strong>and</strong> thus obtained ee <strong>and</strong> rate enhancement.<br />
66
Chapter 6<br />
Cinchonidine modified Pd colloidal<br />
nanoparticles<br />
6.1 Introduction<br />
It is known, that Pd is a suitable catalyst for hydrogenation <strong>of</strong> double bonds.<br />
Cinchonidine modified Pd system was found to be able to induce enantioselectivity in<br />
hydrogenation <strong>of</strong> substituted alkenes. In 1985 Perez <strong>and</strong> co-workers reported the<br />
hydrogenation <strong>of</strong> (E)-α-phenyl-cinnamic acid to 2,3-diphenylpropionic acid over<br />
cinchonidine-modified Pd/C; the enantiomeric excess was 30 % in favour <strong>of</strong> the S-<br />
product [208]. The role <strong>of</strong> a support for Pd, metal loading, dielectric constant <strong>of</strong> the<br />
solvent, concentration <strong>of</strong> cinchonidine <strong>and</strong> substrate, hydrogen pressure <strong>and</strong><br />
temperature has been studied by Nitta <strong>and</strong> co-workers [98, 209-213]. Enantioselective<br />
hydrogenation <strong>of</strong> the carbon-nitrogen double bond on an example <strong>of</strong> Ph(C=NOH)Me<br />
(ee 18 % ) is known from reports <strong>of</strong> Blaser [214] <strong>and</strong> Nakamura [215]. More examples<br />
for enantioselective hydrogenations <strong>of</strong> double carbon <strong>and</strong> carbon-nitrogen bond can be<br />
found from the review <strong>of</strong> Wells [87].<br />
In this chapter we focus on hydrogenation activated ketones over cinchonidine<br />
modified Pd catalysts. Hydrogenation <strong>of</strong> ethyl (or methyl) pyruvate on cinchonidine<br />
modified Pd catalyst leads to the interesting phenomena, so called inversed<br />
enantioselectivity. In fact, in 1988 Blaser <strong>and</strong> co-workers reported that cinchonidinemodified<br />
Pd/carbon was mildly enantioselective for the hydrogenation <strong>of</strong> ethyl<br />
pyruvate, the reaction giving an enantiomeric excess <strong>of</strong> 4% in favour <strong>of</strong> the S-<br />
enantiomer [94], where as hydrogenation over cinchonidine modified Pt catalyst results<br />
in enantiomeric excess in favour <strong>of</strong> R-ethyl lactate (see chapters 4 <strong>and</strong> 5). Closer<br />
examination showed that reaction over conventionally supported Pd differed very<br />
significantly from the corresponding Pt system [49, 216]. In fact, rate enhancement was<br />
absent over cinchonidine modified Pd, enantioselective reaction was half order in<br />
hydrogen whereas it was first-order for Pt, reactions over Pd were only achieved in<br />
certain solvents <strong>and</strong> Pd was enantioselective only if reduced at low temperature.<br />
It is known, that in the case <strong>of</strong> Pt catalyst the use <strong>of</strong> cinchonidine modifier results in<br />
excess <strong>of</strong> R enantiomer <strong>and</strong> cinchonine gives S lactate. However very interesting<br />
phenomenon takes place with use <strong>of</strong> Pd based catalyst, in fact, the use <strong>of</strong> cinchonidine<br />
over Pd catalysts leads to positive (excess <strong>of</strong> R enantiomers) or negative (excess <strong>of</strong> S<br />
enantiomers) ee, at the same time cinchonine over Pd catalysts is able to induce<br />
negative <strong>and</strong> positive ee [49]. Wells <strong>and</strong> Whyman [217] first reported this phenomenon,<br />
they found, that the direction <strong>of</strong> the enantioselectivity being solvent <strong>and</strong>/or substrate<br />
(methyl <strong>and</strong> ethyl pyruvate) dependent.<br />
In this chapter we define the following aims: to investigate orientation <strong>of</strong> the<br />
cinchonidine molecule being adsorbed on Pd nanoclusters (~1.4 nm, see below) via<br />
DRIFTS similarly to the Pt nanoclusters (please see the chapter 4) <strong>and</strong> to show<br />
differences in the direction <strong>of</strong> enantioselectivity between modified with cinchonidine in<br />
situ conventional Pd/Al 2 O 3 catalyst <strong>and</strong> cinchonidine modified Pd nanoclusters in ethyl<br />
pyruvate <strong>and</strong> ethyl benzoyformate hydrogenation reactions. The choice <strong>of</strong> ethyl<br />
benzoyformate was caused because it can not, in principal, form an enolate tautomer<br />
state (please find details below).<br />
67
6.2 Experimental<br />
Materials<br />
Cinchonidine (>98% Fluka), 5 % Pd/Al 2 O 3 (Fluka), Pd(DBA) 2 (Aldrich), THF (> 99.5<br />
% Fluka), cyclohexene (> 99% Fluka) <strong>and</strong> ethyl benzoylformate (95 % Aldrich) were<br />
used as received.<br />
Samples preparation<br />
Cinchonidine modified Pd nanoclusters were prepared according to the following<br />
procedure. 1084 mg Pd(DBA) 2 (1.88 mmol Pd) <strong>and</strong> 1658 mg cinchonidine (5.64 mm)<br />
in 70 ml THF were transferred into the metal reactor. After purging with Ar for 10 min,<br />
3 bar <strong>of</strong> hydrogen pressure was applied under constant stirring. After 3 hours pressure<br />
was released <strong>and</strong> the mixture was transferred into 500 ml cyclohexene <strong>and</strong> stirred for<br />
24 h, then stirring was stopped <strong>and</strong> after 10 hours a precipitate was collected <strong>and</strong><br />
washed with cyclohexene (200 ml), then dried in air at room temperature. The washing<br />
procedure was repeated one more time using 50 ml THF only, for complete removal <strong>of</strong><br />
side products. Obtained black precipitate was dried in air <strong>and</strong> transferred into 40 ml<br />
AcOH (sample was completely dissolved). After stirring (30 min), the mixture was<br />
transferred (Bönnemann’s procedure) into the 500 ml saturated NaHCO 3 water<br />
solution, after 2 h filtrated with G4 glass filter, then washed with 200 ml half saturated<br />
water solution <strong>of</strong> NaHCO 3 , then with water. Then it was picked up by AcOH (10-15<br />
ml) (with few drops <strong>of</strong> water) <strong>and</strong> left for drying on the glass plate in air at room<br />
temperature. Finally a black powder was scratched from the glass.<br />
The thermal activation <strong>of</strong> conventional Pd/Al 2 O 3 was performed at 400 °C under<br />
vacuum (10 -3 mbar) for 1 hour, then kept for 3 hours under a flowing mixture <strong>of</strong> N 2<br />
(93%) <strong>and</strong> H 2 (7%) gases (100 ml min -1 ) <strong>and</strong> subsequently cooled to the room<br />
temperature. The required amount was then weighed (in open atmosphere) <strong>and</strong><br />
immediately transferred into a glass reactor with acetic acid, cinchonidine <strong>and</strong> ethyl<br />
pyruvate.<br />
Characterization<br />
Molecular modelling, GC, XRD spectra measuring <strong>and</strong> DRIFTS investigations have<br />
been done as it is described e.g. in the chapter 4 <strong>and</strong> 5.<br />
Hydrogenation <strong>of</strong> ethyl pyruvate (EP) <strong>and</strong> ethyl benzoylformate (EB) was conducted in<br />
the metal reactor under 10 bars <strong>of</strong> hydrogen. Typically, to 70 ml AcOH with 36 mmol<br />
substrate <strong>and</strong> (unless otherwise mentioned) 20 mg cinchonidine (1 mM) <strong>and</strong> 20 mg<br />
catalyst was added. In the case <strong>of</strong> colloidal catalyst the mixture was sonicated for 10<br />
min. Finally it was transferred into the metal reactor, purged with Ar for 10 min <strong>and</strong><br />
hydrogen pressure (10 bar) has been set, the stirring frequency was 1000 rpm. In the<br />
case <strong>of</strong> the conventional Pt/Al 2 O 3 (was used for comparison) <strong>and</strong> Pd/Al 2 O 3 the catalyst<br />
was thermally activated under hydrogen (as described in details in the e.g. chapter 5)<br />
before use.<br />
68
O<br />
OH<br />
OH<br />
O<br />
O<br />
O<br />
Catalyst, AcOH<br />
O<br />
P H2 = 10 bar<br />
O<br />
O<br />
O<br />
OH<br />
OH<br />
O<br />
O<br />
O<br />
Ph<br />
Ph<br />
Ph<br />
O<br />
O<br />
O<br />
Fig. 6-1. Scheme <strong>of</strong> ethyl pyruvate <strong>and</strong> ethyl benzoylformate hydrogenation.<br />
6.3 Results <strong>and</strong> discussion<br />
XRD spectrum <strong>of</strong> cinchonidine modified Pd nanoclusters (Fig. 6-2) shows strong peaks<br />
at ~40° which corresponds to the (111) crystal face <strong>and</strong> small peak at 67° (220) face,<br />
according to work <strong>of</strong> Haglund [218].<br />
400<br />
300<br />
200<br />
100<br />
0<br />
-100<br />
20 30 40 50 60 70 80<br />
2Θ<br />
Fig. 6-2. XRD spectrum <strong>of</strong> cinchonidine modified Pd nanoclusters.<br />
The average crystal size has been estimated to be 1.6 nm based on the Scherrer’s model<br />
[177] from the width on a half height <strong>of</strong> the strongest peaks (2Θ=40°). The TEM image<br />
(Fig. 6-3) shows particles with 1.4 <strong>and</strong> 2.9 nm size. It is important to note, that at close<br />
examination it is seen that “big” particles consist <strong>of</strong> at least two “small” particles. The<br />
average size <strong>of</strong> “small” particles is very similar to the size obtained from XRD analysis<br />
<strong>and</strong> in further we consider 1.4 nm as the average size <strong>of</strong> basic Pd nanoclusters.<br />
69
Fig. 6-3. TEM image <strong>of</strong> cinchonidine modified Pd nanoclusters.<br />
Determination <strong>of</strong> adsorption geometry <strong>of</strong> cinchonidine on Pd nanoclusters via DRIFTS<br />
It is convenient to perform FTIR investigations <strong>of</strong> cinchonidine modified Pd<br />
nanoclusters through direct comparison with the spectrum <strong>of</strong> cinchonidine modified Pt<br />
nanocluster (considered in the chapter 4). Fig. 6-4 shows that DRIFT spectra <strong>of</strong><br />
cinchonidine modified Pt <strong>and</strong> Pd nanoclusters are very similar, however some<br />
distinctions are present <strong>and</strong> they are discussed below. The most important peaks <strong>and</strong><br />
shoulders (from both DRIFT spectra) for determination the adsorption modes <strong>of</strong><br />
cinchonidine are summarised in the Table 6-1.<br />
Table 6-1. Assignment <strong>of</strong> some frequencies (cm -1 ) from DRIFT spectra <strong>of</strong> cinchonidine<br />
modified Pt <strong>and</strong> Pd nanoclusters.<br />
Cinchonidine on Pd Cinchonidine on Pt Free cinchonidine Direction <strong>of</strong> dipole<br />
moment A .<br />
n.o. 1637 (sh) 1636 (med) (0, 0, 1)<br />
1613 (sh) 1613 (sh) 1619 (wk) (-1, 1, 0)<br />
1586 (med) 1594 (med) 1590 (st) (1, 1, 0)<br />
1570 (sh) 1576 (sh) 1567 (med) (1, -1, 0)<br />
1515 (st) 1514 (st) 1507 (st) (-1, 1, 0)<br />
1462 (med) 1463 (st) 1452 (st) (-1, 1, 0)<br />
811 (wk) 812 (med) 806 (med) (1, -1, -1)<br />
768 (med) 766 (str) 758 (st) (0, 0, 1)<br />
687 (st) 697 (med) 665 (wk) (-1, 1, 0)<br />
652 (sh) 655 (st) 648 (wk) or<br />
638 (med)<br />
(1, 1, -1) or<br />
(1, 0, -1)<br />
Comments: A – Orientations <strong>of</strong> dipole moments were characterized as directions <strong>of</strong> the<br />
unit vector (X, Y, Z) in the coordinate system having X <strong>and</strong> Y axes along the long <strong>and</strong><br />
short axis <strong>of</strong> the quinoline <strong>and</strong> Z axis along the normal <strong>of</strong> the quinoline plane towards<br />
to quinuclidine.<br />
70
The presence <strong>of</strong> in plane vibrations <strong>of</strong> cinchonidine on Pd nanoclusters e.g. at 1515 cm -<br />
1 clearly indicates the tilted adsorption mode. The out <strong>of</strong> plane vibrations <strong>of</strong> quinoline<br />
moiety e.g. 811 cm -1 were also observed on Pd that confirms the flat adsorption mode.<br />
The proposed adsorption modes (schematically shown in Fig. 4-10 <strong>and</strong> 4-11) <strong>of</strong> the<br />
cinchonidine on Pd nanoclusters (1.4 nm) were found to be very similar to the<br />
adsorption modes determined from the work <strong>of</strong> Ferri, Burgi <strong>and</strong> Baiker [82] over<br />
macroscopic film <strong>of</strong> Pd on Al 2 O 3 .<br />
1609<br />
1640<br />
3<br />
2<br />
1<br />
1600 1400 1200 1000 800<br />
Wavenumber, cm -1<br />
Fig. 6-4. DRIFT spectra <strong>of</strong> free (1) <strong>and</strong> adsorbed on Pd (2) <strong>and</strong> Pt (3) nanoclusters<br />
cinchonidine. Intensity <strong>of</strong> spectrum <strong>of</strong> free cinchonidine was decreased for better<br />
performance.<br />
Comparison <strong>of</strong> adsorption modes <strong>of</strong> cinchonidine on Pt <strong>and</strong> Pd nanoclusters<br />
The relative intensity <strong>of</strong> the peaks (811 cm -1 <strong>and</strong> 768 cm -1 ) corresponding to the out <strong>of</strong><br />
plane vibrations <strong>of</strong> quinoline ring is significantly lower with respect to the cinchonidine<br />
on Pt system, probably due to the lower concentration <strong>of</strong> the π-bonded cinchonidine<br />
with respect to the tilted mode. In fact, in the preparation <strong>of</strong> the samples the used ratio<br />
cinchonidine/Pd was 3, whereas in the case <strong>of</strong> Pt based sample the ratio cinchonidine/Pt<br />
was 2, at the same time it is known that with increase <strong>of</strong> concentration <strong>of</strong> cinchonidine<br />
the chance <strong>of</strong> flat adsorption decreases, as was shown for Pt [76] <strong>and</strong> Pd [82]. It is also<br />
can be expressed as a favour <strong>of</strong> tilted adsorption on Pd with respect to Pt, due to the<br />
nature <strong>of</strong> the adsorption <strong>of</strong> cinchonidine on metal.<br />
It is interesting to note the absence <strong>of</strong> the peaks (or shoulder) at 1640 cm -1 which<br />
corresponds to the stretching <strong>of</strong> C=C bond, that might indicate to its hydrogenation.<br />
However the parallel to the Pd surface orientation <strong>of</strong> this bond due to the specific<br />
adsorption mode <strong>of</strong> cinchonidine can not be fully excluded, especially at high<br />
concentration <strong>of</strong> modifier, whose adsorption modes are less studied in high<br />
concentration range.<br />
Hydrogenation <strong>of</strong> α-ketones<br />
71
From the results <strong>of</strong> EP <strong>and</strong> EB hydrogenations which are summarized in the Table 6-2,<br />
it is seen that presence <strong>of</strong> the cinchonidine (entry 2) results in 14 % <strong>of</strong> conversion (in<br />
contrast with 0 % without cinchonidine, entry 1) <strong>and</strong> in excess <strong>of</strong> R lactate (37 % ee). It<br />
is clear that cinchonidine induces a reaction rate <strong>and</strong> positive ee (excess <strong>of</strong> R<br />
enantiomer) similarly to the Pt catalyst, but reaction is significantly less intensive over<br />
Pd/Al 2 O 3 , In fact only 14 % <strong>of</strong> conversion has been achieved after 19 hours (entries 1<br />
<strong>and</strong> 2), whereas under the same conditions Pt/Al 2 O 3 results 91 % ee (R lactate) <strong>and</strong><br />
80% conversion after ~60 min (chapter 5). The different situation was found with<br />
cinchonidine modified Pd nanoclusters (entry 3), this catalyst without presence <strong>of</strong> free<br />
cinchonidine demonstrates 11 % enantiomeric excess in favour <strong>of</strong> S lactate <strong>and</strong> high<br />
conversion after 60 min. Surprisingly the normalized on number <strong>of</strong> metal surface atoms<br />
reaction rates were found to be almost the same as for conventional Pd/Al 2 O 3 , however<br />
the direct comparison between colloidal <strong>and</strong> thermally activated catalyst is laboured<br />
due to the presence <strong>of</strong> impurities blocking active sites.<br />
Conv. B , %<br />
EP ≈0 ≈0 0 D (75) -<br />
EP ≈0.2 ≈0.1 1.3 D (90) -<br />
EP ≈0 ≈0 0 D (60) -<br />
Table 6-2. Results <strong>of</strong> ethyl pyruvate <strong>and</strong> ethyl benzoylformate hydrogenation over Pd<br />
based catalysts.<br />
Run Catalyst Substra Initial rate, Initial<br />
Ee, %<br />
(min) -1 (time, enantiomer)<br />
te mmol·(min·g) -1 rate A after (dominated<br />
min)<br />
1<br />
C<br />
Pd/Al 2 O 3 EP ≈0 ≈0 ~0 (16 h) -<br />
2 Pd/Al 2 O 3 EP 0.36 9.8 14 D (19 h) 37 (R)<br />
3 Cinchonidine EP 13 9.6 59 E (60) 11 (S)<br />
on Pd coll. C<br />
4 Cinchonidine EP 9 6.7 71 E (120) 19 (S)<br />
on Pd coll.<br />
5 Diphos on<br />
Pd coll. C,F<br />
6 Diop on Pd<br />
coll. C,F<br />
7 Quiphos on<br />
Pd coll. C,F<br />
8 Cinchonidine EB 3.9 2.9 61 E (280) 0<br />
on Pd coll.<br />
9 Pt/Al 2 O 3 EB ≈0 ≈0 0 D (19 h) -<br />
Comments: A – Values were estimated based on TEM, metal content <strong>of</strong> the samples<br />
<strong>and</strong> chemisorption data, B - initial amount <strong>of</strong> racemic ethyl lactate (as impurity in<br />
commercial ethyl pyruvate) was taken into account, C – no free cinchonidine was<br />
added, D – according to the curve pr<strong>of</strong>ile (conversion vs. time) reaction can not go<br />
further, E - according to the curve pr<strong>of</strong>ile (conversion vs. time) reaction can go further,<br />
F – the sample was show for comparison, detailed information about it can be found in<br />
the corresponding chapter.<br />
Addition <strong>of</strong> free cinchonidine (entry 4, 1 mM) increases the enantiomeric excess to the<br />
maximum reported value <strong>of</strong> 19 % with respect to the S enantiomer, but practically does<br />
not influence on the reaction rate. The change <strong>of</strong> direction <strong>of</strong> enantioselectivity towards<br />
to the excess <strong>of</strong> S enantiomer over cinchonidine modified Pd nanoclusters in particular<br />
(entries 3 <strong>and</strong> 4) <strong>and</strong> over different types <strong>of</strong> Pd catalysts in general [217] can be<br />
explained through formation <strong>of</strong> enol intermediate state in a rate-determining step (Fig.<br />
72
6-5) [87], the latter was verified after reaction with deuterium yielded the product<br />
exchanged at the methyl group near α-keto group, i.e. giving the product<br />
CX 3 CX(OX)COOCH 3 (where X = H or D) [219]. It is assumed that due to the<br />
intermolecular repulsion between the >C=C< <strong>and</strong> >C=O moiety <strong>of</strong> the adsorbed state 3<br />
(Fig. 6-5) can be released by rotation around the C-C bond to give species 3´ which<br />
yields S enantiomer upon hydrogenation. According to this mechanism, these is not any<br />
supposed enhancement <strong>of</strong> the reaction rate (it is assumed that intermediate state 2<br />
which is involved in the rate=determining step 2-3, can not participate in hydrogen<br />
bonding making semi hydrogenated complex with cinchonidine). However Pd colloids<br />
modified with other chiral <strong>and</strong> non chiral lig<strong>and</strong>s (entries 5, 6 <strong>and</strong> 7) demonstrate very<br />
low (practically zero) reaction rate. The similar effect <strong>of</strong> enhancement <strong>of</strong> reaction rate<br />
was observed by Wells <strong>and</strong> co-workers [217] over Pd colloids but was not explained. In<br />
the current thesis we also do not give exact explanation, because this problem requires<br />
addition investigation that can not be covered by current PhD thesis having other aims,<br />
we give only an idea <strong>of</strong> interpretation <strong>of</strong> the effect which can be found below.<br />
O OEt<br />
1<br />
*<br />
H 3 C O<br />
*<br />
+H -H<br />
2<br />
O<br />
*<br />
H 2 C<br />
OEt<br />
* O<br />
+H slow<br />
3<br />
HO OEt<br />
H 2 C * * O<br />
+2H<br />
fast<br />
HO<br />
H<br />
H 3 C<br />
OEt<br />
O<br />
*<br />
(R)<br />
fast<br />
H 2 C OEt<br />
+2H<br />
H OEt<br />
HO<br />
(S)<br />
3' *<br />
HO O<br />
*<br />
fast<br />
H 3 C O<br />
Fig. 6-5. Mechanism for the enantioselective hydrogenation <strong>of</strong> the EP over Pd, adapted<br />
from [87].<br />
It is very interesting to note, that in our experiments (entries 2 <strong>and</strong> 4) under the same<br />
reaction conditions enantiomeric excess in favour <strong>of</strong> R enantiomer (37 %) <strong>and</strong> in favour<br />
<strong>of</strong> S enantiomer (19 %) was found for unmodified in advance (modified in situ)<br />
Pd/Al 2 O 3 <strong>and</strong> already modified with cinchonidine Pd nanoclusters, respectively. The<br />
similar inverse <strong>of</strong> ee with cinchonidine Pd system has been observed for different type<br />
<strong>of</strong> Pd catalysts, solvents <strong>and</strong> substrates [217]. This clearly indicates on the fact that<br />
there are competing mechanisms <strong>of</strong> hydrogenations e.g. the presence <strong>of</strong> both enols <strong>and</strong><br />
keto forms on the Pd surface. The keto- mechanism results in excess <strong>of</strong> R product (in<br />
the similar way to the cinchonidine-Pt system, probably with rate enhancement) <strong>and</strong><br />
enol- mechanism gives excess <strong>of</strong> S enantiomer as it is shown in Fig. 6-5 without rate<br />
73
enhancement. However the exact factor (nature <strong>of</strong> solvent, Pd catalyst (e.g. presence <strong>of</strong><br />
Pd δ+ ), substrate, hydrogen pressure or present <strong>of</strong> chiral or non chiral impurities) playing<br />
a major role in domination <strong>of</strong> one mechanism with respect to another one is not known<br />
still. As was mentioned above the origin <strong>of</strong> the rate enhancement is also unknown <strong>and</strong> it<br />
can not be explained through domination <strong>of</strong> the keto – mechanism because it will give<br />
an excess <strong>of</strong> R lactate. Our idea is based on another adsorption mode <strong>of</strong> cinchonidine<br />
on Pd surface i.e. via its double carbon bond (Fig. 6-6-right).<br />
Fig. 6-6. Proposed adsorption modes: π-bonded (left) <strong>and</strong> via C=C bond (right) <strong>of</strong><br />
cinchonidine on Pd surface results in different asymmetrical environments caused by H<br />
<strong>and</strong> OH groups.<br />
In fact, in this adsorption mode (opposite to adsorption via quinoline moiety)<br />
cinchonidine might behave in the similar manner to cinchonine, i.e. having replaced<br />
OH <strong>and</strong> H groups, thus induces excess <strong>of</strong> S enantiomer <strong>and</strong> rate enhancement (third<br />
mechanism). The proposed adsorption mode probably is not stable on the metal surface<br />
due the C=C bond hydrogenation. Probably because <strong>of</strong> the addition <strong>of</strong> free cinchonidine<br />
(entry 4) ee has increased from 11 to 19 % in favour <strong>of</strong> S enantiomer.<br />
Interestingly that hydrogenation <strong>of</strong> ethyl benzoylformate (where formation <strong>of</strong> enole<br />
state is not possible in principal, thus enol- mechanism is forbidden) with the same<br />
cinchonidine modified Pd nanoclusters results in zero enantiomeric excess. This means<br />
that the reaction was performed racemically (on unmodified sites) or equal contribution<br />
from keto- mechanism (yielding excess <strong>of</strong> R lactate) <strong>and</strong> the third mechanism (excess<br />
<strong>of</strong> S lactate). At the same time, conventional activated Pt/Al 2 O 3 (entry 9) with presence<br />
<strong>of</strong> free cinchonidine (1 mM) showed no conversion <strong>of</strong> EB in contrast with cinchonidine<br />
modified Pd colloids under 10 bars <strong>of</strong> hydrogen.<br />
The observations show that the Pd catalyzed enantioselective hydrogenation <strong>of</strong><br />
pyruvate esters is much more complicated system than the corresponding Pt catalyzed<br />
reaction <strong>and</strong> co adsorption <strong>of</strong> modifier, solvent <strong>and</strong> substrate [217] <strong>and</strong> possible<br />
impurities on Pd surface might be critical in determining the direction <strong>of</strong><br />
enantioselectivity. The Pd cinchona system requires more investigation for concluding<br />
reaction mechanisms especially in the light <strong>of</strong> recent work <strong>of</strong> Garl<strong>and</strong> et al. [220] where<br />
74
inverse <strong>of</strong> ee in hydrogenation <strong>of</strong> EB <strong>and</strong> EP was observed even over Pt/Al 2 O 3 <strong>and</strong> Pt/C<br />
catalysts.<br />
6.4 Summary<br />
The orientation <strong>of</strong> the cinchonidine on the surfaces <strong>of</strong> 1.4 nm Pd nanoclusters has been<br />
investigated for the first time using a combination <strong>of</strong> DRIFTS <strong>and</strong> molecular modeling<br />
methods. It was found that both ‘flat’ <strong>and</strong> ‘tilted’ modes <strong>of</strong> adsorbed cinchonidine are<br />
present on Pd nanoclusters having reduced relative concentration <strong>of</strong> flat bonded<br />
cinchonidine with respect to the cinchonidine modified Pt nanoclusters having the<br />
similar size (1.4 nm).<br />
The highest known enantiomeric excess <strong>of</strong> 19 % with respect to the S ethyl lactate was<br />
reported over cinchonidine modified Pd nanoclusters whereas under the same reaction<br />
conditions enantiomeric excess induced by conventional Pd/Al 2 O 3 modified with<br />
cinchonidine in situ was found to be 37 % with excess <strong>of</strong> R ethyl lactate.<br />
The inverse <strong>of</strong> enantioselectivity with respect to the Pt catalyzed hydrogenation was<br />
explained with literature model through the enol- mechanism, whereas rate<br />
enhancement effect <strong>and</strong> dependence <strong>of</strong> direction <strong>of</strong> enantiomeric excess on the nature<br />
<strong>of</strong> Pd catalyst <strong>and</strong> reaction conditions require additional investigation.<br />
Chapter 7<br />
Quiphos modified Pt <strong>and</strong> Pd colloidal<br />
nanoparticles<br />
7.1 Introduction<br />
The Quinoline diazaphospholidine ‘quiphos’ (Fig. 7-1) family <strong>of</strong> lig<strong>and</strong>s represents a<br />
possible alternative to the cinchonidine class <strong>of</strong> lig<strong>and</strong>s for application in<br />
enantioselective catalysis. Quiphos is a chiral molecule because <strong>of</strong> the chiral<br />
phosphorus atom due to the inversion high barrier at room temperature. In this work we<br />
used enantiopure quiphos that was confirmed from synthesis <strong>and</strong><br />
31 P-NMR<br />
spectroscopy, for the further information about its synthesis we recommend to contact<br />
to Dr. Didier Nuel <strong>of</strong> Université Aix-Marseille III, France.<br />
These lig<strong>and</strong>s have attracted attention as a result <strong>of</strong> their ease <strong>of</strong> synthesis <strong>and</strong> stability<br />
to air <strong>and</strong> moisture as well as possessing the same quinoline anchor as the cinchonidine<br />
systems <strong>and</strong> thus may be able to assume a flat “π-bonded” geometrical orientation<br />
which is known to be necessary for inducing enantioselectivity in the cinchonidine<br />
system.<br />
N<br />
N<br />
N<br />
P<br />
O<br />
Fig. 7-1. Quiphos molecule.<br />
75
Further, this family <strong>of</strong> man-made lig<strong>and</strong>s <strong>of</strong>fers numerous possibilities for structural<br />
modification which may enable ‘tuning’ <strong>of</strong> the resulting system <strong>and</strong> exploration <strong>of</strong> a<br />
wide variety <strong>of</strong> derivitization. This family <strong>of</strong> lig<strong>and</strong>s has also been successfully applied<br />
to a variety <strong>of</strong> transition metal catalysts for homogeneous enantioselective synthesis [9,<br />
11, 12].<br />
7.2 Experimental<br />
Materials<br />
Quiphos lig<strong>and</strong> was provided by Pr<strong>of</strong>. Gerard Buono <strong>and</strong> Dr. Didier Nuel <strong>of</strong> Université<br />
Aix-Marseille III, France. Other chemicals were purchased <strong>and</strong> used as received.<br />
Samples preparation<br />
The modification <strong>of</strong> conventional Pt/Al 2 O 3 with quiphos was performed in the similar<br />
way, how it was described for cinchonidine in the chapter 4. Typically, 40 mg Pt/Al 2 O 3<br />
was heated at 200-220 ºC under vacuum (~10 -2 mm) for 2-3 hours, then the gases N 2<br />
(93%) <strong>and</strong> H 2 (7%) were passed (50 ml/min) over the support for 2-3 hours without<br />
contact with air at the same temperature. While still under flowing gases, the support<br />
was cooled to room temperature <strong>and</strong> lig<strong>and</strong> was added as a solution in CHCl 3 (6.0 ml, 3<br />
mM). Then the system was kept at 10-15 ºC for one day, washed 3 times with CHCl 3<br />
<strong>and</strong> finally dried under vacuum at room temperature.<br />
For the preparation <strong>of</strong> quiphos stabilized Pt nanoclusters 70 ml <strong>of</strong> chlor<strong>of</strong>orm was<br />
flushed with nitrogen for 10 minutes, then 405 mg <strong>of</strong> Pt 2 (DBA) 3 (0.74 mmol <strong>of</strong> Pt) was<br />
added together with 775 mg <strong>of</strong> quiphos (2.22 mmol). The solution was flushed with Ar<br />
for 15 min under stirring. Then the contents were transferred into the glass reactor <strong>and</strong><br />
3 bar <strong>of</strong> hydrogen pressure was applied under constant stirring. After 18 hours a black<br />
reaction mixture was precipitated by adding the reaction mixture to 230 ml <strong>of</strong><br />
cyclohexene. The precipitate was filtrated out using a G4 glass filter. The precipitate<br />
was first washed with cyclohexene in chlor<strong>of</strong>orm (1/1 by volume) mixture then with<br />
chlor<strong>of</strong>orm (300 ml) until the product was free from hydrogenated DBA <strong>and</strong> unreduced<br />
Pt 2 (DBA) 3 as determined by IR spectroscopy. The sample is readily redispersible in a<br />
solution <strong>of</strong> phenol/ethanol (1:1) with 10% water.<br />
In order to prepare quiphos stabilized Pd nanoparticles 1084 mg Pd(DBA) 2 (1.9 mmol<br />
Pd) <strong>and</strong> 659 mg quiphos (1.9 mmol) were dissolved in nitrogen flushed THF (80 ml).<br />
Then the content was transferred into the metal reactor <strong>and</strong> 5 bar <strong>of</strong> hydrogen pressure<br />
was applied for 4 hours under constant stirring. Then the hydrogen pressure was<br />
released, reactor unplugged <strong>and</strong> the mixture was transferred into 300 ml cyclohexene.<br />
After 24 hours <strong>of</strong> settling, black precipitate was filtrated by G4 filter <strong>and</strong> washed with<br />
THF : cyclohexene (1:5) mixture until washing solvents mixture no longer became<br />
yellow. Finally sample was washed with cyclohexene (200 ml) <strong>and</strong> black powder was<br />
dried in air at room temperature for 12 hours.<br />
Characterization<br />
The samples were characterized by XRD, TEM, elemental analysis <strong>and</strong> DRIFTS in the<br />
same ways as it is described e.g. in the chapter 4. However due to the limitations in<br />
accessibility to TEM <strong>and</strong> elemental analysis apparatuses the corresponding<br />
experiments: TEM characterization <strong>of</strong> quiphos modified Pd nanoclusters <strong>and</strong> elemental<br />
analysis <strong>of</strong> this sample are missing.<br />
Hydrogenation <strong>of</strong> ethyl pyruvate was carried out in a metal reactor (Parr Instrument<br />
GmbH) under a pressure <strong>of</strong> 10 bar <strong>of</strong> hydrogen pressure at room temperature (22±1<br />
°C). The stirring speed was 1000 rpm.<br />
76
Typical conditions were such that 4 ml (36 mmol) ethyl pyruvate was dissolved in 70<br />
ml THF solvent, when mentioned 24 mg (68 mmol) <strong>of</strong> free quiphos was added. 20 mg<br />
<strong>of</strong> catalyst was added <strong>and</strong> mixture was sonicated for 10 min then transferred into the<br />
reactor, flushed with Ar for 10 min <strong>and</strong> finally, after the pressure <strong>of</strong> hydrogen had been<br />
stabilized for 10-15 sec, the initial time (t=0) was set <strong>and</strong> the reaction kinetics were<br />
monitored. About one ml aliquots <strong>of</strong> the reaction mixture were taken at certain time<br />
intervals, filtrated with a G4 glass filter <strong>and</strong> 1 μl was injected into the gas<br />
chromatograph (GC). GC measurements were taken using a Varian 3900 instrument<br />
with FID detector <strong>and</strong> Lipodex-E column.<br />
Activation <strong>of</strong> conventional Pt/Al 2 O 3 has bee performed in the same was as is described<br />
in the chapter 5.<br />
7.3 Results <strong>and</strong> discussion<br />
From the elemental analysis <strong>of</strong> quiphos modified Pt sample it has been found that the<br />
sample has 4.5 % N <strong>and</strong> 28.3 % C or 37 % <strong>of</strong> quiphos as an average by carbon <strong>and</strong><br />
nitrogen <strong>and</strong> 56 % <strong>of</strong> metal (based on TGA), whereas metal content for quiphos<br />
modified Pd sample was found to be 35 % (TGA).<br />
XRD spectra <strong>of</strong> quiphos modified Pt <strong>and</strong> Pd nanoclusters clearly demonstrate Pt <strong>and</strong> Pd<br />
crystalline nature <strong>of</strong> the samples. In fact, both spectra have a strong peak at 39°<br />
corresponding (111) crystal face. In case <strong>of</strong> Pt small peaks at 46° <strong>and</strong> 67° correspond to<br />
(200) <strong>and</strong> (220) faces, according to work <strong>of</strong> Swanson et al. [221]. The asymmetrical<br />
peak at ~35° is difficult to assign unambiguously, since it appears not only in the<br />
spectra <strong>of</strong> this sample but also in case <strong>of</strong> diphos on Pd nanoclusters (see chapter 9) we<br />
believe that it is not due to the instrument fault. Here we would like to assign it to the<br />
(222) Pt crystal face base on work <strong>of</strong> Hull [222].<br />
350<br />
300<br />
250<br />
1<br />
200<br />
150<br />
100<br />
50<br />
2<br />
0<br />
-50<br />
20 30 40 50 60 70 80<br />
2Θ<br />
Fig. 7-1. XRD <strong>of</strong> quiphos modified Pt (1) <strong>and</strong> Pd (2) nanoclusters.<br />
XRD spectrum <strong>of</strong> quiphos modified Pd nanoclusters besides <strong>of</strong> mentioned already peak<br />
at 39° ((111) face) shows wide peaks at 22° (200) [222] <strong>and</strong> 71° (620) [176].<br />
77
Fig. 7-2. TEM <strong>of</strong> quiphos modified Pt nanoclusters.<br />
TEM image <strong>of</strong> quiphos modified Pt nanoparticles (Fig. 7-2) shows that the average<br />
particle’s size is 15 ± 0.5 nm nm, that differs (Table 7-1) from average size estimated<br />
from the Scherrer equation (4.3 nm) <strong>and</strong> from the average size <strong>of</strong> quiphos modified Pd<br />
nanoparticles (1.2 nm), probably due to the partial agglomeration <strong>of</strong> Pt nanoparticles.<br />
Table 7-1. Comparison <strong>of</strong> the sizes <strong>of</strong> quiphos modified Pt <strong>and</strong> Pd particles found via<br />
TEM <strong>and</strong> XRD.<br />
Sample , nm (TEM) , nm (XRD)<br />
Quiphos on Pt 15 4.3<br />
Quiphos on Pd - 1.2<br />
FTIR investigation <strong>of</strong> quiphos adsorbed on Pt <strong>and</strong> Pd nanoparticles<br />
Two conformations <strong>of</strong> the quiphos molecule were found through molecular modelling<br />
(Fig. 7-3). The energy difference between the more stable conformation (Fig 7-3-1) <strong>and</strong><br />
the second (Fig. 7-3-2) was found to be 11.3 kcal/mol in air <strong>and</strong> 5.2 kcal/mol in THF.<br />
The geometry <strong>of</strong> the first conformation is in agreement with single-crystal XRD data<br />
[11] <strong>and</strong> was chosen for analysis.<br />
Fig. 7-3. Skeleton geometry <strong>of</strong> quiphos molecule in two conformations, hydrogen<br />
atoms are omitted. Conformation “1” is energetically more preferred, ΔE 1,2 = -11.3<br />
kcal/mol (in air).<br />
78
The DRIFT spectra <strong>of</strong> the quiphos adsorbed on conventional Pt/Al 2 O 3 (Fig. 7-4)<br />
demonstrates shifted peaks <strong>and</strong> shoulders whose analogs can be found in the spectra <strong>of</strong><br />
free quiphos. However due to the low quality (not all important peaks can be found) <strong>of</strong><br />
the spectra <strong>of</strong> this sample, probably because <strong>of</strong> low degree <strong>of</strong> surface coverage, this<br />
spectra was not taken in to account for further interpretation in favor <strong>of</strong> spectra <strong>of</strong><br />
quiphos modified nanoparticles (please see below).<br />
Relative units<br />
1<br />
2<br />
1600 1400 1200 1000<br />
Wavenumber, cm -1<br />
Fig. 7-4. DRIFT spectra <strong>of</strong> quiphos adsorbed on conventional Pt/Al 2 O 3 (1) <strong>and</strong> free<br />
quiphos (2).<br />
Adsorbed quiphos on Pt <strong>and</strong> Pd produces IR spectra (Fig. 7-5) for which the peaks are<br />
summarized in Table 7-2. Here, only peaks which can be unambiguously assigned to<br />
free quiphos are presented.<br />
In the same way as it is done in case <strong>of</strong> the cinchonidine, the DFT geometry<br />
optimization <strong>and</strong> calculation <strong>of</strong> the IR spectrum <strong>of</strong> quiphos have been done to obtain<br />
the orientation <strong>of</strong> the corresponding dipole moment. To facilitate spectra assignment,<br />
the experimental <strong>and</strong> calculated IR spectra <strong>of</strong> quiphos (correlation coefficient is 1.03)<br />
quinoline <strong>and</strong> aniline were compared with each other; see Fig. 7-6 <strong>and</strong> Table 7-2.<br />
Aniline was chosen since quiphos also contains a C 6 H 5 N unit. The IR spectrum <strong>of</strong><br />
aniline was calculated with correlation coefficient 1.09.<br />
All observed peaks corresponding to IR vibrations <strong>of</strong> quiphos adsorbed on Pt<br />
demonstrate two orientations <strong>of</strong> the dipole moment: mostly perpendicular to the<br />
quinoline ring <strong>and</strong> mostly perpendicularly to the phenyl ring. This leads us to propose<br />
two orientations <strong>of</strong> adsorbed quiphos on Pt. In the first mode, quiphos is adsorbed on Pt<br />
via the quinoline anchor (Fig. 7-7), in the similar way as cinchonidine does, but,<br />
possibly more tilted; in the second mode - via the phenyl group (Fig. 7-8) <strong>and</strong> possibly<br />
via 3,4 N <strong>and</strong>/or P atoms. Some degree <strong>of</strong> tilt is also possible in both cases. In contrast to<br />
the cinchonidine, quiphos can not be adsorbed in a tilted orientation mode, for example,<br />
through the nitrogen in quinoline part without significant changes in geometry.<br />
The DRIFT spectra <strong>of</strong> adsorbed quiphos on Pt <strong>and</strong> Pd nanoparticles do not have<br />
principal differences. In fact, positions <strong>of</strong> some peaks (see Table 7-2) are slightly<br />
shifted that can be interpreted in slight differences in force constants <strong>of</strong> chemical bonds<br />
79
<strong>of</strong> quiphos adsorbed on Pt <strong>and</strong> Pd <strong>and</strong> in the same time in differences in adsorption<br />
strength. It is also can be seen (Fig. 7-5 <strong>and</strong> Table 7-2) that some new peaks appeared<br />
or became stronger for quiphos being adsorbed on Pd with respect to Pt. This<br />
phenomenon can be explained as increase <strong>of</strong> quality <strong>of</strong> the spectra (for example due to<br />
the differences in size <strong>of</strong> Pt <strong>and</strong> Pd nanoparticles, or other secondary effects) or as<br />
differences in relative concentration <strong>of</strong> spices adsorbed via the quinoline <strong>and</strong> phenyl<br />
ring. It is difficult to conclude unambiguously which explanation is more valid here<br />
without addition investigation in which we decided to do not focus on here.<br />
1<br />
Relative units<br />
2<br />
3<br />
1600 1400 1200 1000 800<br />
Wavenumber, cm -1<br />
Fig. 7-5. IR spectrum <strong>of</strong> quiphos adsorbed on Pd (1) <strong>and</strong> Pt (2) nanoparticles <strong>and</strong> free<br />
quiphos (intensity is reduced for clarity) (3).<br />
80
3<br />
A<br />
Relative units<br />
2<br />
B<br />
1<br />
C<br />
0<br />
D<br />
1600 1400 1200 1000 800 600<br />
Wavenumber cm -1<br />
Fig. 7-6. Comparison <strong>of</strong> IR peaks between free cinchonidine (A), quinoline (B),<br />
quiphos (C) <strong>and</strong> aniline (D).<br />
Fig. 7-7. Flat adsorbed quiphos demonstrates vibration mode at 1604 cm -1 with dipole<br />
moment (big arrow) orientated mostly perpendicularly to the metal surface. The most<br />
intensive displacement <strong>of</strong> atoms is shown by small arrows.<br />
81
Fig. 7-8 Flat adsorbed quiphos via phenyl ring demonstrates vibration mode at 696 cm -1<br />
with dipole moment (big arrow) orientated mostly perpendicularly to the metal surface.<br />
The most intensive displacement <strong>of</strong> atoms is shown by small arrows.<br />
Table 7-2. Assignment <strong>of</strong> frequencies (cm -1 ) <strong>of</strong> IR spectra <strong>of</strong> adsorbed <strong>and</strong> free<br />
quiphos, free quinoline <strong>and</strong> free aniline.<br />
Quiphos<br />
on Pt<br />
Quiphos<br />
on Pd<br />
Free<br />
quiphos<br />
Quiphos<br />
theoretical<br />
Quinoline Aniline Description Orientation<br />
<strong>of</strong><br />
adsorbed<br />
- 1610 1652 1620 - C-Qu.<br />
strch., H-<br />
Qu. b.<br />
1604<br />
(st)<br />
1580<br />
(sh)<br />
1558<br />
(med)<br />
1503<br />
(st)<br />
1601<br />
(st)<br />
1574<br />
(med)<br />
1593 1650 - 1600<br />
6,7,9,10 C-A.,<br />
3,5 C strch.,<br />
ip. def.<br />
n.o. 1639 1596 - C-Qu.<br />
strch., H-<br />
Qu. b.<br />
- 1568 1598 1572 - C-Qu.<br />
strch., H-<br />
Qu. b.<br />
1534<br />
(med)<br />
1500<br />
(st)<br />
1538 - Unharmonic<br />
vibr.<br />
1498 1542 1501 1495 C-Qu.<br />
strch., H-<br />
Qu. b.<br />
quiphos<br />
-<br />
Quinoline<br />
flat<br />
Phenyl flat<br />
Phenyl flat<br />
-<br />
Phenyl flat<br />
1468 1468 1464 1502 1471 - H-Qu. Phenyl flat<br />
(med) (st)<br />
asym. ip. b.<br />
1422 1427 1422 1477 1432 1479 H-Qu. sym. Phenyl flat<br />
(wk) (sh)<br />
ip. b.<br />
1390 1388 1383 1434 1394 -<br />
37,42 H-Q. Phenyl flat<br />
82
(wk) (med) sym. ip. b.,<br />
C,<br />
C 16 N-<br />
Qu. strch.<br />
1379<br />
(sh)<br />
1375<br />
(med)<br />
1369 1408 1371 - H,C-Qu.<br />
comp.,<br />
30,31,32 H b.<br />
- 1324 1361 1314 1304 C-Qu.<br />
strch., H-<br />
Qu. ip. b.,<br />
C-A. strch.<br />
1309<br />
(st)<br />
1259<br />
(sm)<br />
1319<br />
(st)<br />
1258<br />
(sm)<br />
1311 1337 - 1273<br />
1247 1292 - -<br />
3 C 5 N strch.,<br />
C,H-A.<br />
comp.<br />
H 11,13,14 2 C<br />
b.<br />
1 O 23 C<br />
strch.,<br />
40,41,42 H ip.<br />
b., H 14 2 C b.<br />
- 1225 1258 - - H 11,12,13,14 2 C<br />
b.<br />
- 1202 1240 - 1171<br />
26,16 H-A,<br />
28,29 H-A<br />
sci.,<br />
H 11,12,13,14 2 C<br />
b.<br />
- 1087 1113 1118 - H,C-Qu.<br />
comp.,<br />
11,12 C strch.,<br />
- 1027 1027 - 992 H-Qu. oop.<br />
wag.,<br />
C 7 C,<br />
8,10 C 9 C ip.<br />
b.<br />
- 988 1014 - - C-A. sym.<br />
expan.,<br />
H 11,12,13,14 2 C<br />
b.<br />
- 971 989 - -<br />
- 954 951 - -<br />
13,14 C,<br />
12 C 4 N –<br />
syn. strch.,<br />
H 11,12,14 2 C<br />
b.<br />
13,14 C,<br />
12 C 4 N –<br />
asyn. strch.,<br />
H 11,12,14 2 C<br />
b.<br />
- 901 926 - - H,C-A., -<br />
Phenyl flat<br />
-<br />
Quinoline<br />
flat<br />
Phenyl flat<br />
-<br />
-<br />
-<br />
-<br />
-<br />
-<br />
-<br />
83
comp.,<br />
H 11,12,13,14 2 C<br />
b., 2 P 11 O 3 N,<br />
N 14 C 13 C b.<br />
- 887 912 - 871 C-Qu.<br />
comp. def.,<br />
H-A. oop.<br />
wag.<br />
- 864 884 - -<br />
12,13 C 14 C<br />
comp.,<br />
11,12 C strch.,<br />
3,4 NP asym.<br />
strch., C-A.<br />
comp.<br />
833 823 829 849 805 - C,H-Qu.<br />
(med) (med)<br />
oop. wag.<br />
-<br />
-<br />
Quinoline<br />
flat<br />
- 807 835 - - C-Qu. ip. b. -<br />
- 803<br />
(sm)<br />
793 812 786 - C,H-Qu.<br />
oop. wag.<br />
Quinoline<br />
flat<br />
- 760 780 739 - C,H-Qu. -<br />
oop. wag.<br />
758 (st) 755 (st) 746 782 - 745 C,H-A. oop. Phenyl flat<br />
wag.<br />
696 (st) 697 (st) 690 712 - 688 C,H-A. oop. Phenyl flat<br />
wag.<br />
- 669 682 - -<br />
7,9 C 8 C, -<br />
6,10 C 5 C ip.<br />
b. , 3 NP<br />
strch.<br />
Comments: The same as for the Table 4-3, plus the following n.o.- not observed, expan.<br />
– expansion, asym. – asymmetrical, syn. – synchronously, asyn. – asynchronously, A-<br />
aniline.<br />
The results <strong>of</strong> ethyl pyruvate hydrogenation with quiphos modified Pt, Pd nanoparticles<br />
<strong>and</strong> activated conventional Pt/Al 2 O 3 are summarized in the Table 7-3. As can be seen<br />
from the Table 7-3 the addition <strong>of</strong> free quiphos stops the reaction completely. In other<br />
words quiphos can be considered as a poison for this reaction with Pt or Pd based<br />
catalysts. Since conversion was not observed at all, ee can not be defined.<br />
Table 7-3. Results <strong>of</strong> ethyl pyruvate hydrogenation with quiphos modified Pt <strong>and</strong> Pd<br />
nanoclusters.<br />
Sample<br />
name<br />
Catalyst<br />
loading, mg.<br />
Reaction rate,<br />
mmol·(min·g) -<br />
Initial rate<br />
(min) -1 Conversion A , [Free<br />
% after time, quiphos],<br />
1<br />
min<br />
mM<br />
B<br />
Pt/Al 2 O 3 20 0.6 16.5 12.5 (345) C 0<br />
Pt/Al 2 O 3 20 ~0 ~0 ~0 (90) 1<br />
Quiphos on<br />
Pt<br />
nanoparticles<br />
20 ~0 ~0 ~0 (60) 0<br />
84
Quiphos on 20 ~0 ~0 ~0 (90) 0<br />
Pd<br />
nanoparticles<br />
Comments: A - initial amount <strong>of</strong> racemic ethyl lactate (as impurity in commercial ethyl<br />
pyruvate) was taken into account. B- racemic hydrogenation without a modifier, C –<br />
according to the curve development (conversion vs. time) reaction can go further.<br />
7.4 Summary<br />
It has been found that quiphos adsorbs on Pt nanoclusters in a similar mode as<br />
cinchonidine through the quinoline moiety. An additional adsorption mode <strong>of</strong> the<br />
quiphos is via the phenyl ring with possibly contribution from the 3,4 N <strong>and</strong>/or P atoms.<br />
The presence <strong>of</strong> quiphos in the reaction mixture induces strong poison effect that stops<br />
reaction completely, probably because <strong>of</strong> occupation <strong>of</strong> active surface sites by quiphos<br />
molecule.<br />
85
Chapter 8<br />
“Quiphos-spider” modified Pt colloidal<br />
nanoparticles<br />
8.1 Introduction<br />
The “quiphos-spider” (4-isopropyl-2 methyl cyclohexyl-1methyl-phosphonite -2-2'-1,5<br />
binaphtyl) compound is the second <strong>of</strong> three lig<strong>and</strong>s which is provided by our<br />
collaborators (Marseille). The chirality <strong>of</strong> this lig<strong>and</strong> is conditioned by the fact that both<br />
couple <strong>of</strong> aromatic rings do not lie in the one plane, making ~32° <strong>and</strong> the image mirror<br />
can not be superimposed with the original molecule.<br />
P O<br />
Fig. 8-1. Quiphos-spider.<br />
Here we investigate catalytic activity <strong>of</strong> Pt nanoparticles modified with quiphos-spider<br />
lig<strong>and</strong> in the ethyl pyruvate hydrogenation <strong>and</strong> explain obtained results based on<br />
proposed models <strong>of</strong> the molecule adsorption geometry.<br />
8.2 Experimental<br />
Materials<br />
Quiphos-spider was provided by Pr<strong>of</strong>. Gerard Buono <strong>and</strong> Dr. Didier Nuel <strong>of</strong> Université<br />
Aix-Marseille III, France. Other chemicals were purchased <strong>and</strong> used as received.<br />
Samples preparation<br />
Preparation <strong>of</strong> quiphos-spider modified Pt nanoclusters had been done according to the<br />
following procedure. 321 mg (0.73 mmol) quiphos-spider was dissolved in argon<br />
purged 70 ml THF with 400 mg Pt 2 (DBA) 3 (0.37 mmol Pt) the mixture was transferred<br />
in to the glass reactor <strong>and</strong> hydrogen pressure <strong>of</strong> 2 bar was applied under constant<br />
stirring at room temperature for 19 hours. Then the mixture was transferred into 400 ml<br />
cyclohexene, after 24 hours black-brown precipitate was collected <strong>and</strong> washed with<br />
(THF/cyclohexene mixture 1/3) 200 ml <strong>and</strong> then with 200 ml cyclohexene, finally<br />
sample was dried at 40°C over night.<br />
Due to the limited amount <strong>of</strong> provided quiphos-spider lig<strong>and</strong> Pt modification <strong>of</strong> Pt<br />
nanoclusters was done only.<br />
Characterization<br />
FTIR spectra were recorded on “Thermo Nicolet AVATAR – 370” instrument, with<br />
DRIFTS accessories, measuring KBr background first.<br />
TEM images were recorded according to the already described in the chapter 4<br />
procedure.<br />
86
Molecular modelling consisting <strong>of</strong> geometry optimization <strong>and</strong> IR spectra calculation<br />
has been done with help <strong>of</strong> G03 program [175] for quiphos-spider <strong>and</strong> naphthalene.<br />
Ethyl pyruvate hydrogenation (including activation <strong>of</strong> conventional Pt/Al 2 O 3 ) was<br />
performed according to the procedure described in the chapter 6. Important parameters<br />
are summarized in the Table 8-1.<br />
8.3 Results <strong>and</strong> discussion<br />
The average size <strong>of</strong> Pt nanoclusters was found to be 1.9±0.3 nm from TEM image (Fig.<br />
8-2). Particles were found to be not uniformly distributed; however no significant<br />
agglomeration was found.<br />
Fig. 8-2. TEM image <strong>of</strong> quiphos-spider modified Pt nanoclusters.<br />
As can be seen from comparison <strong>of</strong> DRIFT spectra <strong>of</strong> free <strong>and</strong> adsorbed on Pt<br />
nanoclusters quiphos-spider lig<strong>and</strong> relative intensities <strong>and</strong> positions <strong>of</strong> some peaks have<br />
changed. Since quiphos-spider molecule consists <strong>of</strong> two (not identical, due to their<br />
orientation within skeleton geometry <strong>of</strong> the lig<strong>and</strong>) naphthalene rings, assignment <strong>of</strong><br />
the main adsorption peaks can be done through comparison IR spectra <strong>of</strong> naphthalene<br />
<strong>and</strong> quiphos-spider. In fact in plane deformation <strong>of</strong> naphthalene ring(s) at 1595 cm -1 ,<br />
1389 cm -1 , 1267 cm -1 having dipole moment orientated along short axis <strong>of</strong> naphthalene<br />
<strong>and</strong> at 1509 cm -1 , 1209 cm -1 <strong>and</strong> 1013 cm -1 with dipole moment orientated along long<br />
its axis, as well as out <strong>of</strong> plane vibrations at 956 cm -1 , 790 cm -1 <strong>and</strong> 617 cm -1 have<br />
dipole moment along the normal vector <strong>of</strong> molecule’s plane. Since quiphos-spider<br />
molecule have two naphthalene planes <strong>and</strong> reduced symmetry, the one peak in the<br />
spectrum <strong>of</strong> naphthalene corresponds to two or more peaks (shoulders) in the spectrum<br />
<strong>of</strong> quiphos-spider. Due to the fact that vibrations <strong>of</strong> free lig<strong>and</strong>s having dipole moment<br />
orientated along short naphthalene axis <strong>and</strong> along long naphthalene axis <strong>and</strong> along<br />
normal vector <strong>of</strong> naphthalene planes are present as strong corresponding peaks in the<br />
spectrum <strong>of</strong> adsorbed quiphos-spider, we would like to propose the following<br />
adsorption models <strong>of</strong> the molecule (Fig. 8-5). In fact, both these models allow having<br />
non zero projection <strong>of</strong> every mentioned orientation <strong>of</strong> dipole moment along normal<br />
vector <strong>of</strong> the metal surface. It has to be mentioned that we do not exclude adsorption on<br />
corners or edges <strong>of</strong> a nanocluster however we think that molecules adsorb mostly on a<br />
87
flat crystal surface due to the higher number <strong>of</strong> atoms (making adsorption sites) on<br />
planes <strong>of</strong> nanocluster than on corners.<br />
Relative units<br />
1<br />
2<br />
1600 1400 1200 1000 800 600<br />
Wavenumber, cm -1<br />
Fig. 8-3. Drift spectra <strong>of</strong> quiphos-spider on Pt nanoclusters (1) <strong>and</strong> (2) free quiphosspider.<br />
1<br />
2<br />
3<br />
1600 1400 1200 1000 800 600 400<br />
Wavenumber, cm -1<br />
Fig. 8-4. Comparison <strong>of</strong> experimentally obtained IR spectra <strong>of</strong> free quiphos-spider (1),<br />
naphthalene (2) <strong>and</strong> calculated IR spectra (correlation coefficient 0.96) <strong>of</strong> free<br />
naphthalene (3). The huge peak at ~800 cm -1 is not shown fully. The similar vibration<br />
modes are marked.<br />
88
Fig. 8-5. Proposed adsorption models <strong>of</strong> quiphos-spider on Pt nanoclusters.<br />
The results <strong>of</strong> ethyl pyruvate hydrogenation with quiphos-spider modified Pt<br />
nanoparticles <strong>and</strong> activated conventional (unmodified) Pt/Al 2 O 3 are summarized in the<br />
Table 8-1. Quiphos-spider modified Pt nanoclusters sample demonstrates slow racemic<br />
hydrogenation <strong>of</strong> ethyl pyruvate, but higher normalized on gram <strong>of</strong> a catalyst reaction<br />
rate than unmodified activated with hydrogen (see experimental section) conventional<br />
Pt/Al 2 O 3 , however it shows lower reaction rate normalized on number <strong>of</strong> surface metal<br />
atoms, probably because some <strong>of</strong> them are “poisoned” by quiphos-spider lig<strong>and</strong>.<br />
Zero enantiomeric excess was found probably due to the inability <strong>of</strong> lone pair <strong>of</strong><br />
phosphorus to “catch” the adsorbed ethyl pyruvate (as it is required for cinchonidineethyl<br />
pyruvate system) because lone pair <strong>of</strong> phosphorus might be involved into the<br />
adsorption binding.<br />
Despite <strong>of</strong> the fact that the reaction rate normalized on gram <strong>of</strong> a catalyst was higher for<br />
colloidal sample, reaction rate normalized on estimated mole <strong>of</strong> Pt surface atoms was<br />
higher for conventional Pt/Al 2 O 3 , probably because the latter one was thermally<br />
activated under hydrogen.<br />
Initial rate A Conversion B , %<br />
Table 8-1. Results <strong>of</strong> ethyl pyruvate hydrogenation with quiphos-spider modified Pt<br />
nanoclusters.<br />
Sample name Catalyst Reaction rate,<br />
loading, mg. mmol·(min·g) -1 (min) -1 after time, min<br />
C<br />
Pt/Al 2 O 3 20 0.6 16.5 12.5 (345) D<br />
Quiphos-spider 17 2 3.3 17 (160) D<br />
on Pt<br />
nanoparticles<br />
Comments: A – values were estimated based on chemisorption data (Pt/Al 2 O 3 ), size <strong>of</strong><br />
Pt nanoclusters <strong>and</strong> metal content (30%), B - initial amount <strong>of</strong> racemic ethyl lactate (as<br />
impurity in commercial ethyl pyruvate) was taken into account. C- racemic<br />
hydrogenation without a modifier, D – according to the curve pr<strong>of</strong>ile (conversion vs.<br />
time) reaction can go further.<br />
8.4 Summary<br />
The adsorption models <strong>of</strong> quiphos spider on Pt nanocluster based on DRIFTS <strong>and</strong><br />
molecular modeling investigations has been proposed. According to these models<br />
quiphos-spider adsorbs via its P atom <strong>and</strong> <strong>of</strong> one naphthalene ring. This system was not<br />
found to be able to induce enantioselectivity in the hydrogenation <strong>of</strong> ethyl pyruvate<br />
(racemic reaction was observed) probably due to the inability <strong>of</strong> adsorbed quiphosspider<br />
to make semi-hydrogenated transition state by analogy with adsorbed<br />
cinchonidine-hydrogen-ethyl pyruvate system.<br />
89
Chapter 9<br />
Diphos <strong>and</strong> diop modified Pt <strong>and</strong> Pd<br />
colloidal nanoparticles<br />
9.1 Introduction<br />
In this chapter we used diphos <strong>and</strong> diop organophosphorus lig<strong>and</strong>s as modifiers for Pt<br />
<strong>and</strong> Pd nanoclusters. The chiral lig<strong>and</strong> diop attracts attention since it has been<br />
successfully used in enantioselective homogeneous catalysis [24], whereas non chiral<br />
diphos molecule has rather similar <strong>and</strong> simple structure that facilitates DRIFTS<br />
investigation <strong>of</strong> lig<strong>and</strong>s adsorbed on nanoclusters <strong>and</strong> relative comparisons e.g.<br />
catalytic properties <strong>of</strong> diphos <strong>and</strong> diop modified Pt <strong>and</strong> Pd nanoclusters.<br />
P<br />
P<br />
P<br />
P<br />
O<br />
O<br />
9.2 Experimental<br />
Diphos<br />
Diop<br />
Fig. 9-1. Diphos <strong>and</strong> Diop molecules.<br />
Materials<br />
Ethylene bis(diphenylphosphine) (diphos) 96 % (Aldrich), (4S,5S)-4,5-<br />
Bis(diphenylphosphinomethyl)-2,2-dimethyl-1,3-dioxolane ( or (S,S)-Diop) <strong>and</strong><br />
(4R,5R)-4,5-Bis(diphenylphosphinomethyl)-2,2-dimethyl-1,3-dioxolane ( or (R,R)-<br />
Diop) 98 % (Aldrich) <strong>and</strong> Pd(DBA) 2 (Aldrich) were used as received, Pt 2 (DBA) 3 was<br />
synthesize as described in the chapter 4.<br />
Samples preparation<br />
In order to prepare diphos on Pt nanoclusters, 560 mg Pt 2 (DBA) 3 (1 mmol Pt) in 40 ml<br />
THF was added to 400 mg diphos (1 mmol) in 30 ml THF. The mixture was purged<br />
with Ar for 10 min in metal reactor <strong>and</strong> hydrogen pressure <strong>of</strong> 4.5 bar was applied for<br />
2.5 hours under constant stirring. After that time, hydrogen pressure was realized <strong>and</strong><br />
the mixture was transferred into 400 ml cyclohexene, after 30 hour supernatant was<br />
decanted <strong>and</strong> precipitated black substance was transferred into 400 ml<br />
THF:cyclohexene mixture (1:7), stirred for 10 min <strong>and</strong> after 4 hour precipitate was<br />
collected <strong>and</strong> this procedure has been repeated until no trace <strong>of</strong> yellow DBA solution<br />
left. Finally black powder was washed with 400 ml cyclohexene <strong>and</strong> dried in air at 65<br />
°C.<br />
The diphos on Pd nanoclusters sample was prepared according to the following<br />
procedure. 1084 mg Pd(DBA) 2 (1.9 mmol Pd) in 40 ml THF was added to 752 mg (1.9<br />
mmol) diphos in 20 ml THF. The mixture was purged with Ar for 10 min in the metal<br />
reactor <strong>and</strong> then hydrogen (5 bar) was applied for 20 hours under constant stirring.<br />
90
After that time the mixture was transferred to 450 ml cyclohexene <strong>and</strong> after 10 hour<br />
black-brown precipitated was collected by G4 glass filter. 400 ml THF: cyclohexene<br />
(1:1) mixture <strong>and</strong> then 200 ml cyclohexene were used for washing. Finally sample was<br />
dried under vacuum at 40°C.<br />
In order to prepare diop on Pd nanoclusters 400 mg (0.8 mmol) (S, S) diop in 40 ml<br />
THF was mixed with 460 mg Pd(DBA) 2 (0.8 mmol Pd) in 40 ml THF in metal reactor.<br />
The mixture was purged with Ar for 10 min <strong>and</strong> then 5 bar <strong>of</strong> hydrogen pressure was<br />
applied for 6.5 hours. After that time, reaction mixture was transferred into 400 ml<br />
cyclohexene, after 16 hours black precipitated was collected <strong>and</strong> washed with 200 ml<br />
THF: cyclohexene mixture (1:4) <strong>and</strong> with 200 ml cyclohexene. Finally it was dried<br />
under vacuum for during 10 hour at 40 °C.<br />
(R, R) <strong>and</strong> (S, S) enantiomers <strong>of</strong> diop were used as modifiers for Pt nanocluster. 250<br />
mg (0.5 mmol) e.g. (S, S) diop in 35 ml THF was transferred into the metal reactor with<br />
274 mg Pt 2 (DBA) 3 in 35 ml THF. After 10 min for Ar purging hydrogen pressure <strong>of</strong> 4<br />
bar has been set for 5.5 hours. After that time the mixture was transferred into 400 ml<br />
cyclohexene, collected after 20 hours precipitate was washed with 400 ml THF:<br />
cyclohexene mixture (1:4) <strong>and</strong> with 200 ml cyclohexene. Finally sample was dried at<br />
40 °C under vacuum.<br />
Characterization <strong>and</strong> experiments<br />
Molecular modelling including geometry optimisation <strong>and</strong> IR spectra calculation has<br />
been performed with Gaussian 03 s<strong>of</strong>tware [175] for diop <strong>and</strong> diphos molecules by<br />
using the DFT method (b3lyp) <strong>and</strong> (6-31G) basis set. Visualisation <strong>of</strong> the geometry <strong>and</strong><br />
IR vibrations was performed using Gaussview s<strong>of</strong>tware. In order to facilitate assigment<br />
<strong>of</strong> the peaks <strong>of</strong> the experimental IR spectrum for the free molecule, its IR spectrum was<br />
compared (in terms <strong>of</strong> intensities <strong>of</strong> the signals <strong>and</strong> their vibrational frequencies) with<br />
calculated IR spectra using MS Excel <strong>and</strong> Origin (intensities <strong>of</strong> both spectra were<br />
normalized, frequencies <strong>of</strong> the calculated spectra were multiplied by a correlation<br />
coefficient). The correlation c<strong>of</strong>ficient dωtheory/dωexp <strong>of</strong> 1.02 was found to provide<br />
the best fit.<br />
XRD spectra were measured in the way described in the chapter 4.<br />
Hydrogenation <strong>of</strong> ethyl pyruvate <strong>and</strong> DRIFTS investigations have been performed<br />
according to the procedure described in details in the chapter 7 <strong>and</strong> 4 correspondingly.<br />
Due to the absence <strong>of</strong> TEM instrument in the IUB <strong>and</strong> limited access to the instrument<br />
outside the university, TEM investigation <strong>of</strong> diop modified Pt nanoclusters has been<br />
done only.<br />
9.3 Results <strong>and</strong> discussion<br />
XRD analysis <strong>of</strong> diop <strong>and</strong> diphos modified Pt <strong>and</strong> Pd nanoclusters clearly demonstrate<br />
Pt <strong>and</strong> Pd crystalline nature <strong>of</strong> the samples. In fact, all spectra have a strong peak at 39°<br />
corresponding (111) crystal face (Fig. 9-2). The small peak at 46° (diop on Pt, Fig. 9-2,<br />
curve 1) corresponds to (200) crystal face [221], where as in case <strong>of</strong> diphos on Pt<br />
sample (same Fig. curve 4 this peak is not seen. The asymmetrical peak at ~35° (diphos<br />
on Pd, curve 3) is difficult to assign unambiguously it probably might be assigned to<br />
(311) Pd crystal face base on work <strong>of</strong> Hull [222].<br />
Based on Scherrer equation [177] the average size <strong>of</strong> metal crystals was estimated<br />
considering width on half height <strong>of</strong> the strongest peak at 39°. The corresponding sizes<br />
<strong>of</strong> Pt <strong>and</strong> Pd nanoclusters are summarized in the Table 9-1 together with data obtained<br />
from the TEM image (Fig. 9-3 ).<br />
91
Table 9-1. Average crystal size <strong>of</strong> diphos <strong>and</strong> diop modified Pt <strong>and</strong> Pd nanoclusters<br />
estimated according to the Scherrer model.<br />
Sample<br />
, nm (Scherrer , nm (TEM)<br />
model)<br />
Diop on Pt 1.7 2<br />
Diop on Pd 1.4 -<br />
Diphos on Pt 1.2 -<br />
Diphos on Pd 1.8 -<br />
Estimated values <strong>of</strong> the size <strong>of</strong> crystals lie within 1-2 nm range. We will use these<br />
values for estimation <strong>of</strong> initial rate in further. Metal content in all samples was found to<br />
be 43±5 % from TGA.<br />
1<br />
2<br />
3<br />
4<br />
30 40 50 2Θ 60 70<br />
Fig. 9-2. XRD spectra <strong>of</strong> diop modified Pd (1), Pt (2), diphos modified Pd (3) <strong>and</strong> Pt (4)<br />
nanoclusters.<br />
92
Fig. 9-3. TEM image <strong>of</strong> diop modified Pt nanoclusters.<br />
Conformation analysis<br />
As was found from molecular modelling the diphos molecule can exist in several<br />
geometrical conformations (Fig. 9-4) at room temperature. The presence <strong>of</strong> different<br />
conformations is conditioned by free rotation around Ph-P <strong>and</strong> P-C bonds.<br />
93
Fig. 9-4. Conformation <strong>of</strong> diphos molecules.<br />
At least three conformations <strong>of</strong> diphos were found with the energy gaps ΔE 2,1 =0.014<br />
eV, ΔE 2,3 =0.003 eV that are less that energy corresponding to the heat motion <strong>of</strong><br />
molecules (K b T=0.023 eV, T = 300 °K), thus these three (<strong>and</strong> probably more)<br />
conformation coexist at room temperature.<br />
The very similar situation was found with diop molecule, in fact, the energy gap<br />
ΔE 2,1 =7.5 meV results in coexisting <strong>of</strong> both conformations (Fig. 9-5) at room<br />
temperature.<br />
Fig. 9-5. Conformation <strong>of</strong> diop molecule. For classification <strong>of</strong> dipole moment<br />
orientation please the text below.<br />
DRIFTS investigation<br />
94
Since the orientation <strong>of</strong> phenyl rings <strong>of</strong> diphos <strong>and</strong> diop molecules is not rigidly fixed<br />
due to the free rotation around mentioned bonds, we have to do not take into account<br />
the orientations <strong>of</strong> dipole moment induced by any vibration <strong>of</strong> any phenyl ring <strong>of</strong><br />
diphos <strong>and</strong> diop in determination <strong>of</strong> adsorption mode from analysis <strong>of</strong> DRIFT spectra<br />
<strong>of</strong> corresponding samples. In fact, for example, in the case <strong>of</strong> conformation 1 <strong>of</strong> diop<br />
the out <strong>of</strong> plane vibration <strong>of</strong> right upper phenyl ring (Fig. 9-5 left) induces dipole<br />
moment orientated almost along the axis connecting two phosphorous atoms in the<br />
molecule, whereas in the case <strong>of</strong> conformation 2 (Fig. 9-5 right) the dipole moment<br />
induced by the same phenyl ring is orientated in almost perpendicular direction (out <strong>of</strong><br />
paper). However the vibrations produced by phenyl ring(s) are very strong <strong>and</strong> can be<br />
use for comparison between DRIFT spectra <strong>of</strong> diop <strong>and</strong> diphos based samples.<br />
Since diop <strong>and</strong> diphos have similar structure, their DRIFT spectra have very similar<br />
groups <strong>of</strong> peaks (Fig. 9-6). In fact, for example, two strong peaks at 1479 cm -1 , 1432<br />
cm -1 can be found in spectra <strong>of</strong> diop <strong>and</strong> diphos. It is important to note, that each <strong>of</strong><br />
them is a observed superposition <strong>of</strong> low resolved (due to the very close positions) group<br />
<strong>of</strong> single peaks (red sticks) corresponding to in plane vibration <strong>of</strong> every four phenyl<br />
ring, as is demonstrated from theoretical IR modelling <strong>of</strong> diphos molecule (Fig. 9-7).<br />
Taking this fact into account, we would like to say, that it is not possible to answer on<br />
the following question exactly. Which one <strong>of</strong> single peaks (red sticks, Fig. 9-7) does<br />
correspond to the observed peak (1484 cm -1 ) in the spectra <strong>of</strong> adsorbed (Fig. 9-8,<br />
curves 2 <strong>and</strong> 3) molecule diphos (as well as diop)?<br />
2<br />
1400 1200 1000 800<br />
Wavenumber, cm -1 1<br />
Fig. 9-6. DRIFT spectra <strong>of</strong> free diphos (1) <strong>and</strong> diop (2).<br />
It is also possible that the observed peak (1484 cm -1 ) in the spectrum <strong>of</strong> adsorbed<br />
molecule consists <strong>of</strong> several (or all) singe peaks (some <strong>of</strong> them present as shoulders)<br />
<strong>and</strong> it is not clear which one <strong>of</strong> them dominates. At the same time the geometrical<br />
orientations <strong>of</strong> phenyl rings are not fixed in the space due to the conversion between<br />
conformations. The same objections are valid for other in plane vibrations (in the range<br />
1160-1000 cm -1 ) <strong>and</strong> out <strong>of</strong> plane vibrations (in the range 846-675 cm -1 ) <strong>of</strong> phenyl<br />
95
ings. Due to these reasons we can not successfully apply the metal adsorption selection<br />
rule for determining the geometry <strong>of</strong> adsorption. This is why only some vibrations <strong>of</strong><br />
molecule which do not affect vibrations <strong>of</strong> phenyl rings were taken into the account.<br />
Among these vibrations (at 1276 cm -1 , 1099 cm -1 <strong>and</strong> 907 cm -1 ) the vibration at 1099<br />
cm -1 inducing dipole moment orientated (for all three conformations) along the line<br />
connecting two phosphor atoms (Fig. 9-4 or 9-9) is the most strongest one. Since this<br />
vibration presences in the spectrum <strong>of</strong> diphos adsorbed on Pt <strong>and</strong> Pd nanoclusters as a<br />
strong peak, (both at 1105 cm -1 ) (Fig. 9-8) we propose that the axis connecting two<br />
phosphor atoms is not parallel to the metal surface, i.e. it either parallel to the surface<br />
normal vector, either has a significant angle <strong>of</strong> tilt, as schematically shown in Fig. 9-9.<br />
We think that the adsorption on corners <strong>of</strong> Pt <strong>and</strong> Pd nanocluster can not be excluded,<br />
however, in order to facilitate analysis <strong>and</strong> larger number <strong>of</strong> atoms (making adsorptions<br />
sites) are on the plane <strong>of</strong> crystals we consider adsorption occurring on a flat surface<br />
only.<br />
Since no principal differences were found between spectra <strong>of</strong> adsorbed diphos on Pt<br />
<strong>and</strong> Pd nanoclusters except that some shoulders in the spectrum <strong>of</strong> diphos on Pt are<br />
represented as peaks in the spectrum <strong>of</strong> diphos on Pd <strong>and</strong> vice versa we think that there<br />
are no principal differences in adsorption geometry. However some degree <strong>of</strong> tilt <strong>of</strong> the<br />
adsorbed diphos is possible that might be a reason for slight difference in wavenumbers<br />
<strong>of</strong> the corresponding peaks e.g. (out <strong>of</strong> plane vibration <strong>of</strong> phenyl ring at 754 cm -1 in<br />
case <strong>of</strong> Pt <strong>and</strong> 747 cm -1 in case <strong>of</strong> Pd nanoparticles), that can also be explained as a<br />
different adsorption strength.<br />
Fig. 9-7. Fragment <strong>of</strong> calculated (low frequency axis) IR spectrum <strong>of</strong> diphos.<br />
Corresponding frequencies <strong>of</strong> experimental observed peaks are written above, see also<br />
text for details.<br />
96
1106<br />
1106<br />
3<br />
2<br />
1099<br />
1<br />
1400 1200 1000 800<br />
Wavenumber, cm -1<br />
Fig. 9-8. Drift spectra <strong>of</strong> free (1) <strong>and</strong> adsorbed on Pt (2) <strong>and</strong> Pd (3) nanoparticles<br />
diphos. Intensity <strong>of</strong> spectrum <strong>of</strong> free diphos is decreased for better performance.<br />
We think that the lone pair <strong>of</strong> phosphor atom does not bind to the metal surface because<br />
<strong>of</strong> steric hinder induced by both neighbour phenyl rings. The possible distance between<br />
metal surface <strong>and</strong> phosphor atom can be roughly (without considering changes in the<br />
geometry <strong>of</strong> adsorbed molecule) estimated as a half size <strong>of</strong> phenyl ring (~1.4 Å,<br />
hydrogens were ignored).<br />
The same logic was applied in the analysis <strong>of</strong> DRIFT spectra <strong>of</strong> adsorbed diop on Pt<br />
<strong>and</strong> Pd nanoclusters (Fig. 9-10), thus only those vibrations which do not affect<br />
vibrations <strong>of</strong> phenyl rings were considered. Among them the strongest vibrations<br />
(marked in the Fig. 9-10) were chosen <strong>and</strong> analyzed as shown in the Table 9-2. As can<br />
be seen from orientation <strong>of</strong> corresponding dipole moment the main axis (line<br />
connecting both P-atoms) <strong>of</strong> the diop molecule can not have only exactly perpendicular<br />
or parallel to the normal <strong>of</strong> surface orientations, otherwise all these vibrations<br />
(absorption peaks) will not be observable simultaneously, in one spectrum. This means,<br />
that the molecule has more then one adsorption mode, or what we think is the most<br />
probable, the axis <strong>of</strong> the molecule has a significant tilt with respect to the all three<br />
coordinate axis’s (one normal <strong>and</strong> two tangent vectors <strong>of</strong> the plane), schematically<br />
adsorption mode is shown in the Fig. 9-11. Here again we would like to note that the<br />
adsorption on the edges <strong>and</strong> corners <strong>of</strong> nanoparticle is also not excluded, however due<br />
to the mentioned above reasons we consider adsorption on a plane.<br />
97
Fig. 9-9. Proposed orientation <strong>of</strong> adsorbed diphos on Pt <strong>and</strong> Pd nanoclusters.<br />
1087<br />
3<br />
1385<br />
1371<br />
1251<br />
1216<br />
891<br />
2<br />
1400 1200 1000<br />
Wavenumber, cm -1 800<br />
1<br />
Fig. 9-10. DRIFT spectra <strong>of</strong> free (1) <strong>and</strong> adsorbed diop on Pt (2) <strong>and</strong> Pd (3)<br />
nanoclusters.<br />
Table 9-2. Assignment <strong>of</strong> some frequencies (cm -1 ) in DRIFT spectra <strong>of</strong> free, adsorbed<br />
on Pt <strong>and</strong> Pd nanoclusters diop.<br />
Free diop Diop on Pt Diop on Pd Diop theoretical A<br />
1385 (st) 1376 B (st) 1381 (st)<br />
C<br />
1461 ↕ 1,5<br />
1371 (st) 1376 B (st) 1372 (sh) 1447 ● 1,2<br />
1251 (st) C 1236 (st) 1240 (st)<br />
B<br />
1282 ↕ 1,6<br />
B<br />
1275 ↕ 1,5<br />
98
1216 (st) 1217 (sh) -<br />
B<br />
1210 ↔ 1,4<br />
1087 (st) 1101 (st) 1104 (med)<br />
B<br />
1016 ↕ 1,5<br />
891 (st) 885 (st) 889 (st)<br />
B<br />
874 ↔ 1,4<br />
Comments: A – Type <strong>of</strong> the vibration summarized in this table <strong>and</strong> corresponding<br />
orientation <strong>of</strong> the dipole moment are the same for all conformations <strong>of</strong> diop, B -<br />
unambiguous assignment is difficult, C – orientation <strong>of</strong> the dipole moment was<br />
characterized as “↔” - horizontal (in plane <strong>of</strong> the figure),“↕” – vertical <strong>and</strong> “●” -<br />
horizontal (perpendicular to the figure’s plane) with numerical subscript according to<br />
scheme on Fig. 9-5.<br />
No principal differences between DRIFT spectra <strong>of</strong> diop adsorbed on Pt <strong>and</strong> Pd<br />
nanoparticles were found. However the relative intensity <strong>of</strong> some peaks is slightly<br />
different, some peaks have their analogs as shoulders, small shifts in position <strong>of</strong> some<br />
peaks probably indicate on some space deviation <strong>of</strong> main axis <strong>of</strong> diop being adsorbed<br />
on Pd with respect to Pt nanoclusters or/<strong>and</strong> different adsorption strength. The peaks at<br />
721 cm -1 (out <strong>of</strong> plane vibration <strong>of</strong> phenyl ring) in the spectrum <strong>of</strong> diop on Pd has no<br />
corresponding peak in case <strong>of</strong> Pt, probably due to the mentioned above reasons: some<br />
degree <strong>of</strong> tilt <strong>of</strong> molecular axis <strong>and</strong> different bonding strength.<br />
Here we again propose that the lone pair <strong>of</strong> closest to the metal surface phosphorous<br />
atom does not take part in binding to metal (at least in case <strong>of</strong> adsorption on flat<br />
surface), because <strong>of</strong> “big” distance (~1.4 Å, estimated in the same way as have been<br />
done for diphos) to surface that is caused by space shielding <strong>of</strong> phenyl rings.<br />
Fig. 9-11. Proposed adsorption model <strong>of</strong> diop on Pt <strong>and</strong> Pd nanoclusters, hydrogens are<br />
not shown.<br />
99
The model <strong>of</strong> adsorption <strong>of</strong> diop on Pt <strong>and</strong> Pd nanoclusters is very similar to model <strong>of</strong><br />
diphos adsorption. This is not surprising because both molecules have the same (Ph) 2 P-<br />
X-P(Ph) 2 structure.<br />
Results <strong>of</strong> ethyl pyruvate hydrogenation with diop <strong>and</strong> diphos modified Pt <strong>and</strong> Pd<br />
catalysts are summarized in the Table 9-3.<br />
Table 9-3. Results <strong>of</strong> ethyl pyruvate hydrogenation with diop <strong>and</strong> diphos modifiers <strong>of</strong><br />
Pt nanoclusters.<br />
Sample<br />
name<br />
Catalyst<br />
loading,<br />
mg.<br />
Reaction rate,<br />
mmol·(min·g) -1<br />
Conversion B ,<br />
Initial<br />
(min) -1 min<br />
rate A % after time,<br />
Ee, %<br />
(dominated<br />
enantiomer)<br />
Diphos on 20 2 1.2 11 C (90) 0<br />
Pt<br />
Diphos on 20 ≈0 ≈0 0 D (75) -<br />
Pd<br />
(S,S)Diop 20 13 10 80 C (300) 4.4 (S)<br />
on Pt<br />
(R,R)Diop 20 10 7 50 C (90) 4.4 (R)<br />
on Pt<br />
Diop on Pd 20 ≈0 ≈0 0 D (90) -<br />
E<br />
Pt/Al 2 O 3 20 ≈0 ≈0 0 D (240) -<br />
F<br />
Pt/Al 2 O 3 20 0.6 16 2.3 C (90) 0<br />
Comments: A – Values were estimated based on average crystal size (XRD) <strong>and</strong> metal<br />
content (TGA) <strong>of</strong> the samples, B - initial amount <strong>of</strong> racemic ethyl lactate (as impurity<br />
in commercial ethyl pyruvate) was taken into account, C – according to the curve<br />
pr<strong>of</strong>ile (conversion vs. time) reaction can go further, D - according to the curve pr<strong>of</strong>ile<br />
(conversion vs. time) reaction can not go further E- reaction was performed with<br />
presence <strong>of</strong> 34 mg (S,S)diop (1 mM), F- racemic hydrogenation without a modifier.<br />
As can be seen from the Table 9-3, surprisingly, diop modified Pt nanoclusters were<br />
found to be able to induce low (4.4 %) but reproducible enantioselectivity in ethyl<br />
pyruvate hydrogenation reaction. Moreover this type <strong>of</strong> catalyst demonstrates<br />
enhancement <strong>of</strong> specific reaction rate <strong>and</strong> thus enhancement <strong>of</strong> the initial rate values in<br />
factor <strong>of</strong> 7-8 with respect to Pt nanoclusters with similar average size modified by non<br />
chiral diphos.<br />
No conversion was found in ethyl pyruvate hydrogenation with activated (400°C,<br />
H 2 /N 2 , 2 hours) conventional Pt/Al 2 O 3 with presence <strong>of</strong> free diop (1 mM), probably due<br />
to the differences in surface chemistry (surface contaminations <strong>and</strong> external impurities)<br />
between diop modified nanoclusters <strong>and</strong> conventional Pt/Al 2 O 3 .<br />
Interestingly, that (S,S)diop leads to excess <strong>of</strong> S lactate <strong>and</strong> (R,R)diop to R lactate <strong>and</strong><br />
corresponding catalysts demonstrate comparable reaction rate. These facts clearly<br />
indicate the importance <strong>of</strong> chiral nature <strong>of</strong> modifier on yielded direction <strong>of</strong><br />
enantioselectivity.<br />
Taking into the account the model <strong>of</strong> enantioselectivity developed for cinchonidine on<br />
Pt system we would like to propose the following model explaining the sense <strong>of</strong><br />
enantioselectivity. Author does not pretend to absolute rightfulness <strong>of</strong> the model,<br />
however finds it to be sufficient for qualitative explanation <strong>of</strong> enantioselectivity, its<br />
inverse with using opposite enantiomers <strong>of</strong> diop <strong>and</strong> enhancement <strong>of</strong> the reaction rate.<br />
100
According to the determined adsorption geometry <strong>of</strong> diop, the lone pair <strong>of</strong> phosphorous<br />
is not involved into binding to the metal surface (at least for adsorption on a flat<br />
surface), thus it is possible that the phosphorous atom (blue marked) (Fig. 9-10 right)<br />
can “catch” adsorbed on Pt ethyl pyruvate molecule via hydrogen bond (rate<br />
enhancement), like unproronated cinchonidine does [50] in non-ionic solvents (toluene,<br />
THF, etc.). In both cases α-carbonyl group <strong>of</strong> ethyl pyruvate molecule forms halfhydrogenated<br />
state with modifier, is in asymmetric environment produced by red<br />
marked OH <strong>and</strong> H groups <strong>of</strong> a modifier. It is known for cinchona alkaloids on Pt that<br />
reverse <strong>of</strong> positions <strong>of</strong> OH <strong>and</strong> H groups leads to inverse <strong>of</strong> enantioselectivity<br />
(cinchonidine “became” cinchonine) [87]. The same (or very similar) fact was found<br />
for diop obtaining S lactate by using (S,S)diop <strong>and</strong> R lactate with (R,R) diop.<br />
O<br />
H<br />
N<br />
OH<br />
H<br />
H<br />
P<br />
H<br />
O<br />
O<br />
O<br />
O<br />
*<br />
= X<br />
C<br />
X<br />
C<br />
X<br />
P<br />
O<br />
X<br />
N<br />
HHO<br />
O H<br />
N<br />
O<br />
P<br />
H<br />
H<br />
O<br />
X<br />
Pt surface<br />
Fig. 9-12. Models <strong>of</strong> half hydrogenated state in case <strong>of</strong> literature known [84-86] [ethyl<br />
pyruvate-cinchonidine] ad complex (left bottom) on Pt <strong>and</strong> proposed [ethyl pyruvatediop]<br />
ad complex (right bottom) on Pt. Schematic positions <strong>of</strong> OH <strong>and</strong> H groups <strong>of</strong><br />
modifiers with respect to α-carbonyl C=O bond <strong>of</strong> ethyl pyruvate for cinchonidine (left<br />
up) <strong>and</strong> diop (right up) modifier.<br />
It is interesting to note, that due to the geometry <strong>of</strong> modifiers OH <strong>and</strong> H groups are<br />
located significantly closer to “caught” C=O group <strong>of</strong> ethyl pyruvate in case <strong>of</strong><br />
cinchonidine than for diop. Probably because <strong>of</strong> remoteness <strong>of</strong> OH <strong>and</strong> H groups their<br />
asymmetrical influence on C=O bond is not as strong as it is in the case <strong>of</strong> cinchona<br />
system, that results in low ee. It is also possible that the different nature <strong>of</strong> H-N <strong>and</strong> H-<br />
P hydrogen bonds also influences on ee. The electric (electrostatic) influence <strong>of</strong> closed<br />
to the metal surface phenyl rings, in principal, can also cause asymmetrical influence <strong>of</strong><br />
C=O bond, since one <strong>of</strong> them in tilted. However we do not think their influence <strong>and</strong><br />
resulted asymmetric environment are significant, because they have very similar nature<br />
(with respect to OH, H pair), in fact, by keeping (Ph) 2 P group in the same state <strong>and</strong><br />
reversing positions <strong>of</strong> OH <strong>and</strong> H groups (using another enantiomer <strong>of</strong> diop) only, we<br />
change direction <strong>of</strong> enantioselectivity <strong>of</strong> the system.<br />
101
The calculation <strong>of</strong> energy <strong>of</strong> forming pro-R <strong>and</strong> pro-S half-hydrogenated diop-ethyl<br />
pyruvate complexes can confirm or disprove this model; however we leave this out <strong>of</strong><br />
the subject <strong>of</strong> the current thesis.<br />
9.4 Summary<br />
The models <strong>of</strong> adsorption geometry <strong>of</strong> diop <strong>and</strong> diphos molecules on Pt <strong>and</strong> Pd<br />
nanoclusters have been proposed. According to these models both molecules are<br />
adsorbed via their phenyl rings pointing molecular axis (a line connecting two<br />
phosphorous atoms) out from the surface, some degree <strong>of</strong> tilt <strong>of</strong> the axis is also<br />
possible. In this model lone pair <strong>of</strong> phosphorous does not take part in adsorption,<br />
probably due to this fact <strong>and</strong> asymmetrical (with respect to C=O bond) location <strong>of</strong> OH<br />
<strong>and</strong> H groups <strong>of</strong> diop result in observed enantioselectivity in hydrogenation <strong>of</strong> ethyl<br />
pyruvate. Diop modified Pt nanoclusters demonstrate enhancement <strong>of</strong> the reaction rate<br />
with respect to Pt nanoclusters modified with non chiral lig<strong>and</strong> in a factor <strong>of</strong> 7-8. The<br />
absence enhanced reaction rate <strong>and</strong> enantioselectivity in ethyl pyruvate hydrogenation<br />
with conventional Pt/Al 2 O 3 with presence <strong>of</strong> diop shows on differences between<br />
catalysts, probably in surface modification.<br />
102
Chapter 10<br />
Binap <strong>and</strong> synphos modified Pt<br />
colloidal nanoparticles<br />
10.1 Introduction<br />
In the current chapter we use chiral binap <strong>and</strong> synphos molecules (Fig. 10-1) as<br />
modifiers for Pt (synphos, binap) <strong>and</strong> Pd (binap) nanoclusters.<br />
O<br />
PPh 2 O PPh 2<br />
PPh 2 O PPh 2<br />
Binap<br />
Synphos<br />
Fig. 10-1. Binap <strong>and</strong> synphos molecules.<br />
As it was already mentioned binap attracts attention due to the numerous successful<br />
applications <strong>of</strong> it in enantioselective homogeneous catalysis [223]. The Synphos<br />
molecule is similar to binap structure <strong>and</strong> synphos based organometalic complex as a<br />
homogeneous catalyst has been found to be even more efficient in some <strong>of</strong> reactions<br />
[33]. Immobilization <strong>of</strong> this type <strong>of</strong> chiral lig<strong>and</strong> on a surface <strong>of</strong> metal nanoparticles is<br />
one strategy for heterogenation <strong>of</strong> enantioselective homogeneous catalysts. Recently,<br />
binap was immobilized on Au <strong>and</strong> Pd nanoclusters with size 1.7 <strong>and</strong> 2 nm,<br />
correspondingly, interestingly, that the latter system was found to be able to induce 95<br />
% enantiomeric excess in hydrosilylation <strong>of</strong> styrene [224]. Another example is the<br />
immobilization <strong>of</strong> Ru complex with phosphonic acid substituted binap on Fe 3 O 4<br />
nanoparticles with 10 nm size [225]. The obtained quasi-homogeneous catalyst<br />
demonstrated even higher ee’s in hydrogenation <strong>of</strong> aromatic ketones. It is very<br />
important to mention that magnetic properties <strong>of</strong> the nanoparticles allow their<br />
separation from reaction products by permanent magnet.<br />
In the current chapter we test binap <strong>and</strong> synphos modified Pt nanocluster in<br />
hydrogenation <strong>of</strong> ethyl pyruvate only, whereas the further test <strong>of</strong> these (<strong>and</strong> all other)<br />
samples is supposed to be conducted by our collaborators from Paris <strong>and</strong> Bucharest,<br />
according to the mentioned in the chapter 3 EU-COST project.<br />
10.2 Experimental<br />
Materials<br />
Synphos lig<strong>and</strong> was provided by Virginie Ratovelomanana-Vidal from the Laboratoire<br />
de Synthese Selective Organiquie et Produits Naturels, UMR (Ecole Nationale<br />
Superieure de Chemie de Paris), binap was purchased from Aldrich.<br />
Samples preparation<br />
O<br />
103
For the preparation <strong>of</strong> synphos modified Pt nanoclusters 273 mg Pt 2 (DBA) 3 (0.51 mmol<br />
<strong>of</strong> Pt) was dissolved in nitrogen flushed THF (40 ml), 300 mg (0.47 mmol) synphos<br />
dissolved in 20 ml nitrogen flushed THF. Both mixtures were transferred into the metal<br />
reactor, purged with argon for 10 min <strong>and</strong> kept under 3 bar <strong>of</strong> hydrogen pressure. After<br />
five hours the pressure was released <strong>and</strong> the mixture was transferred into 300 ml<br />
cyclohexene, then additional 250 ml cyclohexene was added. The brown-back<br />
precipitate was collected by G4 filter. The precipitate in the filter was washed three<br />
times with 100 ml THF/cyclohexene (2/3 by volume) mixture, then with 300 ml<br />
cyclohexene. The brown-black powder was collected mechanically from the filter <strong>and</strong><br />
finally fried on air at 35°C for 24 hours. The sample was found to be redispersible in<br />
e.g. chlor<strong>of</strong>orm or THF.<br />
Due to the limited amount <strong>of</strong> synphos, immobilization <strong>of</strong> it on Pd nanoclusters has not<br />
been done.<br />
In order to prepare binap modified Pt nanoclusters 560 mg Pt 2 (DBA) 3 (0.51 mmol Pt)<br />
in nitrogen purged for 10 min 40 ml THF was added to nitrogen purged 30 ml THF<br />
with 635 mg (1.02 mmol) binap. The mixture was transferred into the metal reactor <strong>and</strong><br />
after purging with Ar for 10 min hydrogen pressure <strong>of</strong> 4 bar has been applied. After 5<br />
hours the mixture was transferred into 500 ml cyclohexene. After 24 hours black<br />
precipitate was collected <strong>and</strong> washed with 400 ml THF: cyclohexene mixture (1:4) <strong>and</strong><br />
with 200 ml cyclohexene. Finally sample was dried at 40 °C under vacuum.<br />
Characterization<br />
The free (not adsorbed) lig<strong>and</strong> <strong>and</strong> quasi-homogeneous synphos modified Pt colloidal<br />
sample (3 mg) were dissolved (separately) in 1.3 ml <strong>of</strong> CDCl 3 <strong>and</strong> 31 P-NMR<br />
measurements were performed on JEOL ECX 400 MHz instrument making 20 <strong>and</strong><br />
20000 scans, respectively.<br />
Molecular modelling, XRD spectra measuring, DRIFTS investigations <strong>and</strong><br />
hydrogenation <strong>of</strong> ethyl pyruvate have been performed as it is described e.g. in the<br />
chapter 9.<br />
Due to the absence <strong>of</strong> TEM instrument in the IUB <strong>and</strong> limited access to the instrument<br />
outside the university, TEM investigation <strong>of</strong> prepared samples is missing.<br />
10.3 Results <strong>and</strong> discussion<br />
31 P-NMR investigation<br />
NMR investigation <strong>of</strong> free lig<strong>and</strong> <strong>and</strong> lig<strong>and</strong> modified colloidal Pt nanoclusters is<br />
illustrated on example <strong>of</strong> synphos.<br />
Synphos molecule has two phosphorous atoms; however 31 P-NMR spectrum (Fig. 10-2<br />
up) has only one peak at -14.8 ppm due to the specific structure <strong>of</strong> the synphos<br />
molecule (identical surrounding <strong>of</strong> P atoms). The absolute absence <strong>of</strong> this peak in the<br />
spectrum <strong>of</strong> colloidal sample unambiguously says that the obtained after written above<br />
purification procedure sample contains no trace <strong>of</strong> free (not adsorbed) synphos.<br />
Whereas, in case <strong>of</strong> adsorbed synphos, significantly shifted peak to 8.2 ppm says about<br />
changes in chemical environment <strong>of</strong> phosphorous atoms.<br />
The obtain spectrum is very similar to the spectra obtained by Hyeon <strong>and</strong> co workers<br />
[226] for binap (<strong>and</strong> other phosphor organic compounds) adsorbed on colloidal Pd<br />
nanoclusters.<br />
104
-14.8<br />
Free synphos<br />
8.2<br />
Synphos on Pt<br />
40 30 20 10 0 -10 -20 -30 -40<br />
PPM<br />
Fig. 10-2. 31 P-NMR spectra <strong>of</strong> free synphos (up) <strong>and</strong> synphos modified colloidal Pt<br />
nanoclusters (bottom). Both samples were dissolved in CDCl 3 .<br />
It is important to note, that due to the small size <strong>of</strong> colloidal Pt nanoclusters estimated<br />
as 1 nm based on XRD (see below) <strong>and</strong> 2 nm based on dynamic light scattering (data<br />
are not included) it was estimated that r<strong>and</strong>om rotation <strong>of</strong> single nanocluster (in<br />
solution) is fast enough to do not consider anisotropic NMR effects <strong>and</strong> the system can<br />
be considered within liquid phase NMR theory, i.e. free r<strong>and</strong>om rotation <strong>of</strong> nanocluster<br />
corresponds to free r<strong>and</strong>om rotation <strong>of</strong> a molecule. In fact, rotation correlation time <strong>of</strong><br />
an object with radius R (2 nm) in a solvent with viscosity η (0.58 cp or 5.8·10 -4<br />
kg/(m·sec)) at room temperature T (300 K) can be determined from the following<br />
equation, based on Debye-Stokes-Einstein formula.<br />
4 3 η<br />
−<br />
τ = ≈ 5 ⋅10<br />
9<br />
c<br />
πR<br />
sec<br />
3 KT<br />
From another h<strong>and</strong> the chemical shift anisotropy (CSA) effect can be roughly estimated<br />
as a 10-15 ppm, that is ≈6 kHz in terms <strong>of</strong> frequencies or τ SCA = 0.2·10 -3 sec. Thus, the<br />
rotation correlation time is much less than typical CSA time that verifies our mentioned<br />
above assumption.<br />
The absent <strong>of</strong> the peak corresponding to the free state <strong>of</strong> binap in the spectrum <strong>of</strong> the<br />
binap (as well as e.g. diop) modified Pt colloidal sample again clearly demonstrate that<br />
used preparation <strong>and</strong> purification procedures allow having lig<strong>and</strong> in adsorbed state only<br />
<strong>and</strong> no trace <strong>of</strong> free lig<strong>and</strong>. However due to the technical <strong>and</strong> organisation problems<br />
with NMR instrument it was not possible to perform required continuous measurement<br />
(≈ 20000 scans, ~3 days) for obtaining sufficient signal/noise ratio for resolving peak<br />
corresponding to the adsorbed lig<strong>and</strong> for all samples having P atoms.<br />
XRD spectra <strong>of</strong> binap <strong>and</strong> synphos modified Pt nanoclusters demonstrate crystalline<br />
nature <strong>of</strong> the samples.<br />
105
1<br />
2<br />
20 40 60 80<br />
2Θ<br />
Fig. 10-3. XRD spectra <strong>of</strong> binap (1) <strong>and</strong> synphos (2) modified Pt nanoclusters. Intensity<br />
<strong>of</strong> (1) was decreased for better performance.<br />
In fact, all spectra have a strong peak at 39° corresponding (111) crystal face (Fig. 10-<br />
3). In case <strong>of</strong> binap on Pt sample peaks at 36°, 46° <strong>and</strong> 67° correspond to (222), (200)<br />
<strong>and</strong> (220) faces, according to work <strong>of</strong> Swanson et al. [221]. The last tree peaks are not<br />
resolved in the spectrum <strong>of</strong> synphos on Pt sample, probably, due to the low sample<br />
loading. Based on the width <strong>of</strong> the peak at 39° the average crystal size was estimated<br />
with help <strong>of</strong> Scherrer model, data are summarized in the Table 10-1.<br />
Table 10-1. Average crystal size <strong>of</strong> binap <strong>and</strong> synphos modified Pt estimated according<br />
to the Scherrer model.<br />
Sample<br />
, nm<br />
Binap on Pt 1.6<br />
Synphos on Pt 1.0<br />
Estimated values <strong>of</strong> the size <strong>of</strong> crystals lie within 1-2 nm range. We will use these<br />
values for estimation <strong>of</strong> initial rate in further. Metal content in all samples was found to<br />
be 43±5 % from TGA.<br />
Conformation analysis<br />
It was found from molecular modelling that at least two conformations <strong>of</strong> synphos<br />
molecule are possible due two rotation around P-Ph <strong>and</strong> C-P(Ph) 2 bonds, like it was<br />
found for diphos <strong>and</strong> diop molecules.<br />
106
Fig. 10-4. Conformations <strong>of</strong> synphos.<br />
The energy difference between conformations was found to be 0.027 eV, that is almost<br />
the same value as energy <strong>of</strong> thermal motion (KT=0.026 eV) is, thus both conformations<br />
can coexist, especially in a solution at room temperature. Because <strong>of</strong> this fact, we do<br />
not take into the consideration molecular vibrations induced by phenyl rings.<br />
Regardless <strong>of</strong> the fact that such type <strong>of</strong> conformations were not found for binap<br />
molecule, we ignore any vibration <strong>of</strong> a phenyl ring in interpretation <strong>of</strong> DRIFT spectrum<br />
<strong>of</strong> adsorbed binap as well.<br />
DRIFTS investigation<br />
From analysis <strong>and</strong> comparison <strong>of</strong> DRIFT spectra <strong>of</strong> free synphos, binap <strong>and</strong> diphos, all<br />
(in <strong>and</strong> out <strong>of</strong> plane) vibrations <strong>of</strong> phenyl rings were excluded from analysis due to the<br />
mentioned above reasons.<br />
1<br />
2<br />
1600 1400 1200 1000 800<br />
Wavenumber, cm -1<br />
Fig. 10-5. DRIFT spectra <strong>of</strong> free binap (1) <strong>and</strong> synphos (2).<br />
107
The DRIFT spectra <strong>of</strong> synphos (Fig. 10-6) <strong>and</strong> modified Pt nanoclusters clearly show<br />
shifts in wavenumbers <strong>of</strong> 2-10 cm -1 that indicates to slight changes in skeleton<br />
geometry or/<strong>and</strong> changes <strong>of</strong> force constants <strong>of</strong> bonds. The relative intensity <strong>of</strong> some<br />
corresponding peaks is also changed. In order to facilitate spectra analysis, frequencies<br />
<strong>of</strong> some peaks allowing unambiguous assignment in terms <strong>of</strong> orientation <strong>of</strong> induced<br />
dipole moments are summarized in the Table 10-2. Left peaks were not considered due<br />
to the fact that some peaks are superposition <strong>of</strong> two or more vibrations, as was<br />
illustrated in Fig. 9-5.<br />
2<br />
1600 1400 1200 1000 800 600<br />
Wavenumber, cm -1 1<br />
Fig. 10-6. DRIFT spectra <strong>of</strong> free (1) <strong>and</strong> adsorbed on Pt nanoclusters (2) synphos.<br />
Intensity <strong>of</strong> free synphos was decreased for better performance.<br />
Table 10-2. Assignment <strong>of</strong> some frequencies (cm -1 ) <strong>of</strong> DRIFT spectra <strong>of</strong> free <strong>and</strong><br />
adsorbed on Pt nanoclusters synphos.<br />
Free synphos<br />
(experimental)<br />
Synphos on Pt Calculated synphos A Dipole moment<br />
orientation B<br />
1452 (str) 1463 (wk. sh.) 1501 (str) ↔ 1,4<br />
1433 (str) 1441 (str) 1485 (str) ↕ 1,5<br />
1401 (med) 1404 (med) 1449 (wk) ↔ 1,3<br />
1289 (str) 1301 (str) 1316 (str) ↔ 1,4 or ↕ 1,5<br />
1255 (med) 1257 (med) 1281 (med) ↔ 1,4<br />
1150 (str) 1159 (med) 1172 (med) ↔ 2,4<br />
938 (med) 939 (med) 975 (med) ↕ 1,5<br />
873 (med) 877 (med) 906 (med) C<br />
902 (wk) C ↔ 1,4<br />
↕ 1,5<br />
Comments: A - conformation from Fig. 10-4 (left) was used, calculated frequencies for<br />
other conformation were just slightly (± max. 10cm -1 ) different, B - orientation <strong>of</strong> the<br />
dipole moment was the same for both conformations <strong>and</strong> was characterized as “↔” -<br />
horizontal (in plane <strong>of</strong> the figure),“↕” – vertical <strong>and</strong> with numerical subscript according<br />
to scheme in Fig. 10-4, C – unambiguous assignment is difficult.<br />
108
Below in the text we assume that only one adsorption mode <strong>of</strong> synphos exists on Pt<br />
surface <strong>and</strong> adsorption takes place on a flat surface.<br />
Based on the data from the Table 10-2 it follows, that the synphos molecule can not<br />
adsorb (see denied examples N1, 2 <strong>and</strong> 3 in Fig. 10-7) on the flat surface without<br />
significant tilt (with respect to depicted in Fig. 10-4 orientation, assuming that the metal<br />
surface has a normal parallel to ↕ 1,5 direction). At the same time it allows a coexistence<br />
<strong>of</strong> non zero projection to the surface’s normal <strong>of</strong> all orientations <strong>of</strong> dipole moment from<br />
this table. However “denied” orientations become possible if they coexist on the Pt<br />
surface, e.g. 1 coexist with 2 or 3.<br />
Due to the isotropic shape <strong>of</strong> the molecule it is not possible to determine the orientation<br />
<strong>of</strong> the adsorbed molecule unambiguously; in fact, at least two adsorption species (see<br />
allowed examples (4 <strong>and</strong> 5) in Fig. 10-7-bottom) can be proposed, they can also coexist<br />
on the surface together.<br />
Fig. 10-7. Schematic illustration <strong>of</strong> denied mono modes (up) <strong>and</strong> allowed mono <strong>and</strong><br />
multi modes (bottom) <strong>of</strong> adsorption <strong>of</strong> synphos molecule on Pt nanoclusters. Hydrogen<br />
atoms are not shown.<br />
It is interesting to note that proposed adsorption mode N5 (Fig. 10-7) is similar to<br />
adsorption modes found for diphos <strong>and</strong> diop molecules. In fact, all three molecules<br />
have P(Ph) 2 functional group which can be considered as a common anchor for them, in<br />
the similar way as quinoline is an anchor for cinchonidine <strong>and</strong> quiphos molecules.<br />
In spite <strong>of</strong> similarity in the structure <strong>of</strong> synphos <strong>and</strong> binap, IR spectrum <strong>of</strong> the latter is<br />
difficult for full analysis because it has a lot <strong>of</strong> peaks which have no analogs in the IR<br />
spectrum <strong>of</strong> synphos (probably due to the enhanced inharmonic vibrations for binap).<br />
Unfortunately, almost every strong peak in the spectrum <strong>of</strong> binap which can be<br />
unambiguously assigned has to be excluded from consideration due the mentioned<br />
above reasons (different conformations <strong>and</strong>/or contribution from vibration <strong>of</strong> the<br />
similar parts <strong>of</strong> the molecule, e.g. another phenyl rings). From another h<strong>and</strong>, those<br />
109
peaks which can be used for analysis <strong>of</strong> orientation <strong>of</strong> adsorbed binap can not be<br />
assigned unambiguously.<br />
However ambiguity in assignment <strong>and</strong> interpretation <strong>of</strong> spectra very <strong>of</strong>ten results in<br />
difficulties (impossibilities) in discrimination between ↕ 1,5 <strong>and</strong> ↔ 1,4 orientations <strong>of</strong><br />
induced dipole moment. For example the strong peaks at 816 cm -1 in the spectrum <strong>of</strong><br />
binap on Pt (Fig. 10-8) corresponds to the peak at 818 cm -1 in the spectrum <strong>of</strong> free<br />
binap which consists <strong>of</strong> two components (due to the symmetrical <strong>and</strong> asymmetrical out<br />
<strong>of</strong> plane bending <strong>of</strong> hydrogens in naphthalene rings, both are not shown). The<br />
symmetrical vibration has dipole moment along the ↕ 1,5 direction, whereas<br />
asymmetrical one has it along the ↔ 1,4 direction. The similar situation is with all other<br />
strong peaks.<br />
816<br />
2<br />
1600 1400 1200 1000 800<br />
Wavenumber, cm -1 1<br />
Fig. 10-8. DRIFT spectra <strong>of</strong> free (1) <strong>and</strong> adsorbed on Pt nanoclusters (2) binap.<br />
Intensity <strong>of</strong> free binap was decreased for better performance.<br />
Since spectra <strong>of</strong> adsorbed binap on Pt nanoclusters allows interpretation in terms <strong>of</strong><br />
orientation <strong>of</strong> dipole moment in ↕ 1,5 or/<strong>and</strong> ↔ 1,4 directions, we would like to use the<br />
same logic for determination <strong>of</strong> orientations <strong>of</strong> adsorbed binap. This results to the same<br />
adsorption modes like for synphos molecule (Fig. 10-7).<br />
In hydrogenation <strong>of</strong> ethyl pyruvate both catalysts: binap <strong>and</strong> synphos did not show<br />
detectable enantioselectivity. The specific reaction rate <strong>and</strong> initial rate (normalized on<br />
the number <strong>of</strong> metal surface atoms) were similar for both samples. Reaction rate<br />
normalized on number <strong>of</strong> surface Pt atoms was higher for conventional Pt/Al 2 O 3 ,<br />
probably due to the thermal activation <strong>of</strong> the latter under hydrogen. The reason <strong>of</strong><br />
absence <strong>of</strong> enantioselectivity with synphos <strong>and</strong> binap modified Pt nanoclusters<br />
probably lies in principal difference in their structure, i.e. both molecules do not meet<br />
the specific requirements developed for cinchona on Pt system, whereas, we think, diop<br />
molecule (see chapter 9) fits partly, that results in low ee.<br />
Table 10-3. Results <strong>of</strong> ethyl pyruvate hydrogenation with binap <strong>and</strong> synphos modifiers<br />
<strong>of</strong> Pt nanoclusters<br />
110
Initial rate A Conversion B , %<br />
Sample name Catalyst Reaction rate,<br />
loading, mg. mmol·(min·g) -1 (min) -1 after time, min<br />
Binap on Pt 20 5 3.7 30 (120) D<br />
Synphos on Pt 10 4.5 2.8 16 (120) D<br />
C<br />
Pt/Al 2 O 3 20 0.6 16.5 12.5 (345) D<br />
Comments: A – values were estimated based on chemisorption (Pt/Al 2 O 3 ), size <strong>of</strong> Pt<br />
nanoclusters <strong>and</strong> metal content (43%) data, B - initial amount <strong>of</strong> racemic ethyl lactate<br />
(as impurity in commercial ethyl pyruvate) was taken into account. C- hydrogenation<br />
without a modifier, D – according to the curve pr<strong>of</strong>ile (conversion vs. time) reaction<br />
can go further.<br />
10.4 Summary<br />
The models <strong>of</strong> adsorption geometry <strong>of</strong> synphos <strong>and</strong> binap molecules on 1 <strong>and</strong> 1.6 nm Pt<br />
nanoclusters have been proposed. According to these models both molecules are<br />
adsorbed via their phenyl rings, however existence <strong>of</strong> other adsorption modes is also<br />
possible. In any model <strong>of</strong> adsorption on a flat surface, the lone pair <strong>of</strong> phosphorous<br />
does not take part in adsorption, due to steric hinder <strong>of</strong> induced by e.g. phenyl rings. No<br />
enantioselectivity was found with binap <strong>and</strong> synphos modified Pt nanoclusters in the<br />
hydrogenation <strong>of</strong> ethyl pyruvate, probably due the fact that both molecules do not meet<br />
specific requirements which were developed based on cinchona on Pt system.<br />
111
References<br />
[1] G.E. Tranter, J. Chem. Soc., Chem. Commun. (1986) 60-61.<br />
[2] A. Streitwieser, W.D. Schaeffer, J. Am. Chem. Soc. 78 (1956) 5597-5599.<br />
[3] J. March, Advanced Organic Chemistry, John Wiley & Sons, New Your, 1992.<br />
[4] Corriu, Guerin, Moreau, The Chemistry <strong>of</strong> Organic Silicon Compounds, John<br />
Wiley & Sons, New York, 1998.<br />
[5] M. Gielen, Top. Curr. Chem. 104 (1982) 57-105.<br />
[6] Stackhou.J, R.D. Baechler, K. Mislow, Tetrahedron Lett. (1971) 3437.<br />
[7] Stackhou.J, R.D. Baechler, K. Mislow, Tetrahedron Lett. (1971) 3441.<br />
[8] J.D. Andose, J.M. Lehn, K. Mislow, J. Wagner, J. Am. Chem. Soc. 92 (1970)<br />
4050.<br />
[9] G. Delapierre, J.M. Brunel, T. Constantieux, G. Buono, Tetrahedron:<br />
Asymmetry 12 (2001) 1345-1352.<br />
[10] G. Delapierre, T. Constantieux, J.M. Brunel, G.R. Buono, Eur. J. Org. Chem.<br />
(2000) 2507.<br />
[11] J.M. Brunel, T. Constantieux, G. Buono, J. Org. Chem. 64 (1999) 8940-8942.<br />
[12] J.M. Brunel, T. Constantieux, A. Lab<strong>and</strong>e, F. Lubatti, G. Buono, Tetrahedron<br />
Lett. 38 (1997) 5971-5974.<br />
[13] S.J. Brois, J. Am. Chem. Soc. 90 (1968) 506.<br />
[14] S.J. Brois, J. Am. Chem. Soc. 90 (1968) 508.<br />
[15] H. Hamill, M.A. McKervey, J. Chem. Soc., Chem. Commun. (1969) 864.<br />
[16] A. Vargos, Phd Thesis, Zurich, 2004.<br />
[17] V. Prelog, M. Kovacevic, M. Egli, Angew. Chem., Int. Ed. Engl. 28 (1989)<br />
1147-1152.<br />
[18] V.S. Martin, S.S. Woodard, T. Katsuki, Y. Yamada, M. Ikeda, K.B. Sharpless,<br />
J. Am. Chem. Soc. 103 (1981) 6237-6240.<br />
[19] J.D. Morrison, H.S. Mosher, Asymmetric Organic Reactions, Englewood Cliffs,<br />
New Jersey, 1971.<br />
[20] I. Ojima, Catalytic Asymmetric Synthesis, Willey-VCH, 1999.<br />
[21] D.E. De Vos, I.F.J. Vankelecom, P.A. <strong>Jacobs</strong>, Chiral catalyst immobilization<br />
<strong>and</strong> recycling, Weinheim, 2000.<br />
[22] H.-U. Blaser, M. Eissen, P. Fauquex, K. Hungerbühler, E. Schmidt, G.<br />
Sedelmeir, M. Studer, in: H.U. Blaser, E. Schmidt (Eds.), Asymmetric Catalysis<br />
on Industrial Scale: Challenges, Approaches <strong>and</strong> Solutions, WILEY-VCH,<br />
Weinheim, 2004, pp. 91-103.<br />
[23] R. Noyori, M. Koizumi, D. Ishii, T. Ohkuma, Pure Appl. Chem. 73 (2001) 227-<br />
232.<br />
[24] T.P. Dang, H.B. Kagan, J. Chem. Soc., Chem. Commun. (1971) 481-481.<br />
[25] T. Hayashi, M. Kumada, Acc. Chem. Res. 15 (1982) 395-401.<br />
[26] S. Jeulin, N. Champion, P. Dellis, V. Ratovelomanana-Vidal, J.P. Genet,<br />
Synthesis-Stuttgart (2005) 3666-3671.<br />
[27] R. Touati, T. Gmiza, S. Jeulin, C. Deport, V. Ratovelamanana-Vidal, C. Ben<br />
Hassine, J.P. Genet, Synlett (2005) 2478-2482.<br />
[28] R. Le Roux, N. Desroy, P. Phansavath, J.P. Genet, Synlett (2005) 429-432.<br />
[29] C. Mordant, P. Dunkelmann, V. Ratovelomanana-Vidal, J.P. Genet, Eur J Org<br />
Chem (2004) 3017-3026.<br />
[30] C. Mordant, P. Dunkelmann, V. Ratovelomanana-Vidal, J.P. Genet, Chem<br />
Commun (2004) 1296-1297.<br />
112
[31] S. Jeulin, S.D. de Paule, V. Ratovelomanana-Vidal, J.P. Genet, N. Champion, P.<br />
Dellis, P Natl Acad Sci USA 101 (2004) 5799-5804.<br />
[32] J.P. Genet, Accounts Chem Res 36 (2003) 908-918.<br />
[33] S.D. de Paule, S. Jeulin, V. Ratovelomanana-Vidal, J.P. Genet, N. Champion, P.<br />
Dellis, Eur. J. Org. Chem. (2003) 1931-1941.<br />
[34] S.D. de Paule, S. Jeulin, V. Ratovelomanana-Vidal, J.P. Genet, N. Champion,<br />
G. Deschaux, P. Dellis, Org Process Res Dev 7 (2003) 399-406.<br />
[35] S.D. de Paule, S. Jeulin, V. Ratovelomanana-Vidal, J.P. Genet, N. Champion, P.<br />
Dellis, Tetrahedron Lett 44 (2003) 823-826.<br />
[36] K. Inoguchi, K. Achiwa, Synlett (1991) 49-51.<br />
[37] K. Achiwa, J. Am. Chem. Soc. 98 (1976) 8265-8266.<br />
[38] G.M. Schwab, F. Rost, L. Rudolph, Kolloid Z 68 (1934) 157-159.<br />
[39] G.M. Schwab, L. Rudolph, Naturwisserschaft 20 (1932) 363-364.<br />
[40] G.J. Hutchings, Annu. Rev. Mater. Res. 35 (2005) 143-166.<br />
[41] A. Tai, T. Kikukawa, T. Sugimura, Y. Inoue, S. Abe, T. Osawa, T. Harada, Bull.<br />
Chem. Soc. Japan 67 (1994) 2473-2477.<br />
[42] T. Harada, T. Osawa, in: G. Jannes, V. Dubois (Eds.), Chiral Reactions in<br />
Heterogeneous Catalysis, Plenum Press, New York, 1995, p. 83.<br />
[43] T. Osawa, A. Tai, Y. Imachi, S. Takasaki, in: G. Jannes, V. Dubois (Eds.),<br />
Chiral Reactions in Heterogeneous Catalysis, Plenum Press, New York, 1995, p.<br />
75.<br />
[44] M. Studer, H.U. Blaser, C. Exner, Adv. Synth. Catal. 345 (2003) 45-65.<br />
[45] Y. Orito, S. Imai, S. Niwa, J. Chem. Soc. Jpn. (1979) 1118.<br />
[46] Y. Orito, S. Imai, S. Niwa, J. Chem. Soc. Jpn. (1980) 670.<br />
[47] Y. Orito, S. Imai, S. Niwa, Nippon Kagaku Kaishi (1979) 1118-1120.<br />
[48] N. Kunzle, R. Hess, T. Mallat, A. Baiker, J. Catal. 186 (1999) 239-241.<br />
[49] P.B. Wells, A.G. Wilkinson, Top. Catal. 5 (1998) 39-50.<br />
[50] A. Baiker, J. Mol. Catal. A: Chem. 163 (2000) 205-220.<br />
[51] B. Törok, K. Felfoldi, G. Szakonyi, K. Balazsik, M. Bartok, Catal. Lett. 52<br />
(1998) 81-84.<br />
[52] M. Schürch, N. Künzle, T. Mallat, A. Baiker, J. Catal. 176 (1998) 569-571.<br />
[53] O.J. Sonderegger, T. Burgi, A. Baiker, J. Catal. 215 (2003) 116-121.<br />
[54] E. Toukoniitty, P. Maki-Arvela, J. Kuusisto, V. Nieminen, J. Paivarinta, M.<br />
Hotokka, T. Salmi, D.Y. Murzin, J. Mol. Catal. A: Chem. 192 (2003) 135-151.<br />
[55] E. Toukoniitty, P. Maki-Arvela, M. Kuzma, A. Villela, A.K. Neyestanaki, T.<br />
Salmi, R. Sjoholm, R. Leino, E. Laine, D.Y. Murzin, J. Catal. 204 (2001) 281-<br />
291.<br />
[56] J.A. Slipszenko, S.P. Griffiths, P. Johnston, K.E. Simons, W.A.H. Vermeer,<br />
P.B. Wells, J. Catal. 179 (1998) 267-276.<br />
[57] W.A.H. Vermeer, A. Fulford, P. Johnston, P.B. Wells, J. Chem. Soc., Chem.<br />
Commun. (1993) 1053-1054.<br />
[58] B. Torok, K. Felfoldi, K. Balazsik, M. Bartok, Chem. Commun. (1999) 1725-<br />
1726.<br />
[59] M. Studer, S. Burkhardt, H.-U. Blaser, Chem. Commun. (1999) 1727-1728.<br />
[60] M. von Arx, T. Mallat, A. Baiker, Catal. Lett. 78 (2002) 267-271.<br />
[61] M. von Arx, T. Mallat, A. Baiker, Tetrahedron: Asymmetry 12 (2001) 3089-<br />
3094.<br />
[62] M. von Arx, T. Mallat, A. Baiker, J. Catal. 193 (2000) 161-164.<br />
[63] K. Balazsik, K. Szori, K. Felfoldi, B. Torok, M. Bartok, Chem. Commun.<br />
(2000) 555-556.<br />
113
[64] A. Szabo, N. Kunzle, T. Mallat, A. Baiker, Tetrahedron: Asymmetry 10 (1999)<br />
61-76.<br />
[65] N. Kunzle, A. Szabo, M. Schurch, G.Z. Wang, T. Mallat, A. Baiker, Chem.<br />
Commun. (1998) 1377-1378.<br />
[66] G.Z. Wang, T. Mallat, A. Baiker, Tetrahedron: Asymmetry 8 (1997) 2133-2140.<br />
[67] W.R. Huck, T. Mallet, A. Baiker, J. Catal. 193 (2000) 1-4.<br />
[68] M. von Arx, T. Mallat, A. Baiker, Top. Catal. 19 (2002) 75-87.<br />
[69] D.Y. Murzin, P. Maki-Arvela, E. Toukoniitty, T. Salmi, Catal. Rev. - Sci. Eng.<br />
47 (2005) 175-256.<br />
[70] T. Bürgi, A. Baiker, J. Am. Chem. Soc. 120 (1998) 12920-12926.<br />
[71] M. Schürch, O. Schwalm, T. Mallat, J. Weber, A. Baiker, J. Catal. 169 (1997)<br />
275-286.<br />
[72] A.F. Carley, M.K. Rajumon, M.W. Roberts, P.B. Wells, Journal Of The<br />
Chemical Society-Faraday Transactions 91 (1995) 2167-2172.<br />
[73] K.E. Simons, P.A. Meheux, S.P. Griffiths, I.M. Sutherl<strong>and</strong>, P. Johnston, P.B.<br />
Wells, A.F. Carley, M.K. Rajumon, M.W. Roberts, A. Ibbotson, Recueil Des<br />
Travaux Chimiques Des Pays-Bas-Journal Of The Royal Netherl<strong>and</strong>s Chemical<br />
Society 113 (1994) 465-474.<br />
[74] T. Evans, A.P. Woodhead, A. Gutierrez-Sosa, G. Thornton, T.J. Hall, A.A.<br />
Davis, N.A. Young, P.B. Wells, R.J. Oldman, O. Plashkevych, O. Vahtras, H.<br />
Agren, V. Carravetta, Surf. Sci. 436 (1999) L691.<br />
[75] G. Bond, P.B. Wells, J. Catal. 150 (1994) 329.<br />
[76] D. Ferri, T. Burgi, J. Am. Chem. Soc. 123 (2001) 12074-12084.<br />
[77] W. Chu, R.J. LeBlanc, C.T. Williams, J. Kubota, F. Zaera, J. Phys. Chem. B 107<br />
(2003) 14365-14373.<br />
[78] C. LeBlond, J. Wang, J. Liu, A.T. Andrews, Y.K. Sun, J. Am. Chem. Soc. 121<br />
(1999) 4920.<br />
[79] M. Bartok, B. Torok, K. Balazsik, T. Bartok, Catal. Lett. 73 (2001) 127-131.<br />
[80] X.B. Li, N. Dummer, R. Jenkins, R.P.K. Wells, P.B. Wells, D.J. Willock, S.H.<br />
Taylor, P. Johnston, G.J. Hutchings, Catal. Lett. 96 (2004) 147-151.<br />
[81] M. Garl<strong>and</strong>, H.-U. Blaser, J. Am. Chem. Soc. 112 (1990) 7048-7050.<br />
[82] D. Ferri, T. Burgi, A. Baiker, J. Catal. 210 (2002) 160-170.<br />
[83] H.-U. Blaser, H.P. Jalett, D.M. Monti, A. Baiker, J.T. Wehrli, Stud. Surf. Sci.<br />
Catal. 67 (1991) 147.<br />
[84] O. Schwalm, B. Minder, J. Weber, A. Baiker, Catal. Lett. 23 (1994) 271-279.<br />
[85] O. Schwalm, J. Weber, B. Minder, A. Baiker, Theochem-Journal Of Molecular<br />
Structure 330 (1995) 353-357.<br />
[86] O. Schwalm, J. Weber, B. Minder, A. Baiker, Int. J. Quantum Chem. 52 (1994)<br />
191-197.<br />
[87] P.B. Wells, R.P.K. Wells, in: D.E. De Vos, I.F.J. Vankelecom, P.A. <strong>Jacobs</strong><br />
(Eds.), Chiral Catalyst Immobilization <strong>and</strong> Recycling, Wiley-VCH, Weinheim,<br />
2000, pp. 123-138.<br />
[88] G. Bond, P.A. Meheux, A. Ibbotson, P.B. Wells, Catal. Today 10 (1991) 371.<br />
[89] H.-U. Blaser, H.P. Jalett, M. Garl<strong>and</strong>, M. Studer, H. Thies, A. Wirth-Tijani, J.<br />
Catal. 173 (1998) 282.<br />
[90] J.L. Margitfalvi, M. Hegedus, J. Mol. Catal. A: Chem. 107 (1996) 281-289.<br />
[91] J.L. Margitfalvi, M. Hegedus, E. Tfirst, 11th International Congress On<br />
Catalysis - 40th Anniversary, Pts A And B, 1996, pp. 241-250.<br />
[92] J.L. Margitfalvi, M. Hegedus, E. Tfirst, Tetrahedron: Asymmetry 7 (1996) 571-<br />
580.<br />
114
[93] J.L. Margitfalvi, E. Talas, E. Tfirst, C.V. Kumar, A. Gergely, Appl. Catal., A<br />
191 (2000) 177.<br />
[94] H.-U. Blaser, H.P. Jalett, D.M. Monti, J.F. Reber, J.T. Wehrli, Stud. Surf. Sci.<br />
Catal. 41 (1988) 153.<br />
[95] K.E. Simons, A. Ibbotson, P. Johnston, H. Plum, P.B. Wells, J. Catal. 150<br />
(1994) 321-328.<br />
[96] E. Toukoniitty, P. Maki-Arvela, A.N. Villela, A.K. Neyestanaki, T. Salmi, R.<br />
Leino, R. Sjoholm, E. Laine, J. Vayrynen, T. Ollonqvist, P.J. Kooyman, Catal.<br />
Today 60 (2000) 175-184.<br />
[97] H.-U. Blaser, H.P. Jalett, M. Müller, M. Studer, Catal. Today 37 (1997) 441-<br />
463.<br />
[98] Y. Nitta, K. Kobiro, Chem. Lett. (1996) 897-898.<br />
[99] Y. Nitta, Y. Okamoto, Chem. Lett. (1998) 1115-1116.<br />
[100] P.A. Meheux, A. Ibbotson, P.B. Wells, J. Catal. 128 (1991) 387-396.<br />
[101] Y. Nitta, T. Imanaka, S. Teranishi, J. Catal. 80 (1983) 31-39.<br />
[102] A. Baiker, J. Mol. Catal. A: Chem. 115 (1997) 473-493.<br />
[103] M. Schürch, T. Heinz, R. Aeschimann, T. Mallat, A. Pfaltz, A. Baiker, J. Catal.<br />
173 (1998) 187.<br />
[104] T. Heinz, Phd Thesis, Universität Basel, Basel, 1997.<br />
[105] H.A. Wierenga, L. Soethout, J.W. Gerritsen, B.E.C. V<strong>and</strong>eleemput, H.<br />
Vankempen, G. Schmid, Advanced Materials 2 (1990) 482-484.<br />
[106] X.B. Zuo, H.F. Liu, D.W. Guo, X.Z. Yang, Tetrahedron 55 (1999) 7787-7804.<br />
[107] A. Kockritz, S. Bisch<strong>of</strong>f, V. Morawsky, U. Prusse, K.D. Vorlop, J. Mol. Catal.<br />
A: Chem. 180 (2002) 231-243.<br />
[108] A. Gamez, J. Kohler, J. Bradley, Catal. Lett. 55 (1998) 73-77.<br />
[109] J.U. Kohler, J.S. Bradley, Langmuir 14 (1998) 2730-2735.<br />
[110] J.U. Kohler, J.S. Bradley, Catal. Lett. 45 (1997) 203-208.<br />
[111] J.S. Bradley, J. Kohler, D. deCaro, Abstracts Of Papers Of The American<br />
Chemical Society 212 (1996) 14-CATL.<br />
[112] Y.L. Huang, J.R. Chen, H. Chen, R.X. Li, Y.Z. Li, L.E. Min, X.J. Li, J. Mol.<br />
Catal. A: Chem. 170 (2001) 143-146.<br />
[113] J. Zhang, X.P. Yan, H.F. Liu, J. Mol. Catal. A: Chem. 175 (2001) 125-130.<br />
[114] H. Bönnemann, G.A. Braun, Angew. Chem., Int. Ed. 35 (1996) 1992-1995.<br />
[115] H. Bönnemann, G.A. Braun, Chem. Eur. J. 3 (1997) 1200-1202.<br />
[116] X.B. Zuo, H.F. Liu, M.H. Liu, Tetrahedron Lett. 39 (1998) 1941-1944.<br />
[117] X.B. Zuo, H.F. Liu, J.F. Tian, J. Mol. Catal. A: Chem. 157 (2000) 217-224.<br />
[118] S. Jansat, M. Gomez, K. Philippot, G. Muller, E. Guiu, C. Claver, S. Castillon,<br />
B. Chaudret, J. Am. Chem. Soc. 126 (2004) 1592-1593.<br />
[119] H.H. Helcher, J. Herbord Klossen, 1718.<br />
[120] M. Faraday, Philos. Trans. R. Soc. London 147 (1857) 145.<br />
[121] U. Kolb, S.A. Quaiser, M. Winter, M.T. Reetz, Chem. Mater. 8 (1996) 1889-<br />
1894.<br />
[122] N. Toshima, Y. Wang, Langmuir 10 (1994) 4574-4580.<br />
[123] G. Schmid, Endeavour 14 (1990) 172-178.<br />
[124] DEFINITION AND CLASSIFICATION OF COLLOIDS, IUPAC,<br />
http://www.iupac.org/reports/2001/colloid_2001/manual_<strong>of</strong>_s_<strong>and</strong>_t/node33.ht<br />
ml<br />
[125] Classification <strong>of</strong> colloids, Wikipedia, http://en.wikipedia.org/wiki/Colloid<br />
[126] M.A. Watzky, R.G. Finke, Chem. Mater. 9 (1997) 3083-3095.<br />
[127] M.A. Watzky, R.G. Finke, J. Am. Chem. Soc. 119 (1997) 10382-10400.<br />
115
[128] J.D. Aiken III, R.G. Finke, J. Mol. Catal. A: Chem. 145 (1999) 1-44.<br />
[129] B.K. Teo, N.J.A. Sloane, Inorg. Chem. 24 (1985) 4545-4558.<br />
[130] N.J.A. Sloane, B.K. Teo, J. Chem. Phys. 83 (1985) 6520-6534.<br />
[131] A.J. Bard, L.R. Faulkner, Electrochemical Methods, Wiley, New York, 1980.<br />
[132] A. Roucoux, J. Schulz, H. Patin, Chemical Reviews 102 (2002) 3757-3778.<br />
[133] A. Amulyavichus, A. Daugvila, R. Davidonis, C. Sipavichus, Fiz. Met.<br />
Metalloved. 85 (1998) 111-117.<br />
[134] E. Gaffet, M. Tachikart, O. ElKedim, R. Rahouadj, Mater. Charact. 36 (1996)<br />
185-190.<br />
[135] H. Bönnemann, W. Brijoux, R. Brinkmann, R. Fretzen, T. Joussen, R. Koppler,<br />
B. Korall, P. Neiteler, J. Richter, J. Mol. Catal. 86 (1994) 129-177.<br />
[136] R. Houbertz, T. Feigenspan, F. Mielke, U. Memmert, U. Hartmann, U. Simon,<br />
G. Schon, G. Schmid, Europhys. Lett. 28 (1994) 641-646.<br />
[137] G. Schmid, S. Peschel, New J. Chem. 22 (1998) 669-675.<br />
[138] U. Simon, R. Flesch, H. Wiggers, G. Schon, G. Schmid, J. Mater. Chem. 8<br />
(1998) 517-518.<br />
[139] M. Bardaji, O. Vidoni, A. Rodriguez, C. Amiens, B. Chaudret, M.J. Casanove,<br />
P. Lecante, New J. Chem. 21 (1997) 1243-1249.<br />
[140] G. Schmid, H. West, J.O. Malm, J.O. Bovin, C. Grenthe, Chem. Eur. J. 2 (1996)<br />
1099-1103.<br />
[141] M.N. Vargaftik, V.P. Zagorodnikov, I.P. Stolarov, Moiseev, II, D.I. Kochubey,<br />
V.A. Likholobov, A.L. Chuvilin, K.I. Zamaraev, J. Mol. Catal. 53 (1989) 315-<br />
348.<br />
[142] G. Schmid, B. Morun, J.O. Malm, Angew. Chem., Int. Ed. Engl. 28 (1989) 778-<br />
780.<br />
[143] G. Schmid, A. Lehnert, Angew. Chem., Int. Ed. Engl. 28 (1989) 780-781.<br />
[144] G. Schmid, Polyhedron 7 (1988) 2321-2329.<br />
[145] G. Schmid, Applied Homogeneous Catalysis with Organometallic Compounds,<br />
Wiley-VCH, 1996, p. 677.<br />
[146] R. Franke, J. Rothe, J. Pollmann, J. Hormes, H. Bönnemann, W. Brijoux, T.<br />
Hindenburg, J. Am. Chem. Soc. 118 (1996) 12090-12097.<br />
[147] O. Vidoni, K. Philippot, C. Amiens, B. Chaudret, O. Balmes, J.O. Malm, J.O.<br />
Bovin, F. Senocq, M.J. Casanove, Angew. Chem., Int. Ed. 38 (1999) 3736-<br />
3738.<br />
[148] K. Patel, S. Kapoor, D.P. Dave, T. Mukherjee, J. Chem. Sci. 117 (2005) 311-<br />
316.<br />
[149] S.H. Sun, C.B. Murray, D. Weller, L. Folks, A. Moser, <strong>Science</strong> 287 (2000)<br />
1989-1992.<br />
[150] P.H. Hess, P.H. Parker, J. Appl. Polym. Sci. 10 (1966) 1915.<br />
[151] Y. Kato, S. Sugimoto, K. Shinohara, N. Tezuka, T. Kagotani, K. Inomata,<br />
Mater. Trans. 43 (2002) 406-409.<br />
[152] M.T. Reetz, W. Helbig, J. Am. Chem. Soc. 116 (1994) 7401-7402.<br />
[153] S.I. Stoeva, B.L.V. Prasad, S. Uma, P.K. Stoimenov, V. Zaikovski, C.M.<br />
Sorensen, K.J. Klabunde, J. Phys. Chem. B 107 (2003) 7441-7448.<br />
[154] W.T. Nichols, T. Sasaki, N. Koshizaki, Euromat, Prague, 2005.<br />
[155] K.M. Beck, T. Sasaki, N. Koshizaki, Chem. Phys. Lett. 301 (1999) 336-342.<br />
[156] X.Y. Zeng, N. Koshizaki, T. Sasaki, A. Yabe, F. Rossignol, H. Nagai, Y.<br />
Nakata, T. Okutani, M. Suzuki, Appl. Surf. Sci. 140 (1999) 90-98.<br />
[157] Turkevic.J, G. Kim, <strong>Science</strong> 169 (1970) 873.<br />
116
[158] H. Bönnemann, G. Braun, W. Brijoux, R. Brinkmann, A.S. Tilling, K. Seevogel,<br />
K. Siepen, J. Organomet. Chem. 520 (1996) 143.<br />
[159] H. Bönnemann, W. Brijoux, Active Metals, VCH, Weinheim, 1996.<br />
[160] H. Bönnemann, R.M. Richards, Eur. J. Inorg. Chem. (2001) 2455-2480.<br />
[161] H. Bönnemann, W. Brijoux, R. Brinkmann, E. Dinjus, T. Joussen, B. Korall,<br />
Angew<strong>and</strong>te Chemie-International Edition In English 30 (1991) 1312-1314.<br />
[162] L. D'Souza, Phd Thesis, International <strong>University</strong> Bremen, Bremen, 2005.<br />
[163] COST Chemistry Action D24, Sustainable Chemical Processes: Stereoselective<br />
Transition Metal-Catalysed Reactions,<br />
http://costchemistry.epfl.ch/load/d24_mid_term_report.pdf<br />
[164] T. Mallat, A. Baiker, J. Catal. 176 (1998) 271-274.<br />
[165] D.G. Blackmond, J. Catal. 176 (1998) 267-270.<br />
[166] T. Mallat, Z. Bodnar, B. Minder, K. Borszeky, A. Baiker, J. Catal. 168 (1997)<br />
183-193.<br />
[167] U.K. Singh, R.N. L<strong>and</strong>au, Y.K. Sun, C. Leblond, D.G. Blackmond, S.K.<br />
Tanielyan, R.L. Augustine, J. Catal. 154 (1995) 91-97.<br />
[168] Y.K. Sun, J. Wang, C. Leblond, R.N. L<strong>and</strong>au, D.G. Blackmond, J. Catal. 161<br />
(1996) 759-765.<br />
[169] K. Balazsik, M. Bartok, J. Catal. 224 (2004) 463-472.<br />
[170] X.B. Li, R.P.K. Wells, P.B. Wells, G.J. Hutchings, Catal. Lett. 89 (2003) 163.<br />
[171] V. Morawsky, Phd Thesis, Braunschweig, 2002.<br />
[172] T.O. Ely, C. Pan, C. Amiens, B. Chaudret, F. Dassenoy, P. Lecante, M.J.<br />
Casanove, A. Mosset, M. Respaud, J.M. Broto, J. Phys. Chem. B 104 (2000)<br />
695-702.<br />
[173] K. Moseley, P.M. Maitlis, J. Chem. Soc. D (1971) 982-983.<br />
[174] P. Yuan, D.Q. Wu, H.P. He, Z.Y. Lin, Appl. Surf. Sci. 227 (2004) 30-39.<br />
[175] M. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R.<br />
Cheeseman, V.G. Zakrzewski, J.A. Montgomery, Jr., R.E. Sratmann, J.C.<br />
Burant, S. Dapprich, J.M. Millam, A.D. Daniels, R. Cammi, B. Mennucci, C.<br />
Pomelli, C. Adamo, S. Clifford, J. Ochterski, G.A. Petersson, P.Y. Ayala, Q.<br />
Ciu, K. Morokuma, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B.<br />
Foresman, J.P. Cioslowski, I. Komaromi, R. Gompets, R.L. Martin, D.J. Fox, T.<br />
Keith, M. Al-Laham, P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, J.L.<br />
Andres, C. Gonzalez, M. Head-Gordon, E.S. Replogle, J.A. Pople, Gaussian<br />
Inc., Pittsburgh, 2003.<br />
[176] W.P. Davey, Physical Review 25 (1925) 763.<br />
[177] P. Scherrer, Göttingen Nachrichten 2 (1918) 98.<br />
[178] J. Kubota, Z. Ma, F. Zaera, Langmuir 19 (2003) 3371-3376.<br />
[179] R.G. Greenler, J. Chem. Phys. 44 (1966) 310.<br />
[180] H.-U. Blaser, M. Garl<strong>and</strong>, H.P. Jallet, J. Catal. 144 (1993) 569-578.<br />
[181] H.-U. Blaser, H.P. Jalett, J. Wiehl, J. Mol. Catal. 68 (1991) 215-222.<br />
[182] E. Orglmeister, T. Mallat, A. Baiker, J. Catal. 233 (2005) 333-341.<br />
[183] B. Minder, T. Mallat, A. Baiker, G. Wang, T. Heinz, A. Pfaltz, J. Catal. 154<br />
(1995) 371.<br />
[184] D. Ferri, T. Burgi, K. Borszeky, T. Mallat, A. Baiker, J. Catal. 193 (2000) 139-<br />
144.<br />
[185] V. Morawsky, U. Pruße, L. Witte, K.D. Vorlop, Catal. Commun. 1 (2000) 15.<br />
[186] K. Balazsik, M. Bartok, J. Catal. 224 (2004) 463.<br />
[187] J. Kubota, F. Zaera, J. Am. Chem. Soc. 123 (2001) 11115.<br />
[188] B.E. Spiewak, R.D. Cortright, J.A. Dumesic, J. Catal. 176 (1998) 405-414.<br />
117
[189] K. Umezawa, T. Ito, M. Asada, S. Nakanishi, P. Ding, W.A. Lanford, B.<br />
Hjorvarsson, Surf. Sci. 387 (1997) 320-327.<br />
[190] M.R. Voss, H. Busse, B.E. Koel, Surf. Sci. 414 (1998) 330-340.<br />
[191] J. Fearon, G.W. Watson, J. Mater. Chem. 16 (2006) 1989-1996.<br />
[192] Asymmetric Hydrogenation <strong>of</strong> Activated Ketones with catASium® F214, R.<br />
Hartung, J.G.E. Krauter, S. Seebald, P. Kukula, U. Nettekoven, M. Studer,<br />
http://www.degussa-catalysts.com/asp/documents/79_Document.pdf<br />
[193] CatASium® F214 Data Sheet heterogeneous catalyst for the enantioselective<br />
hydrogenation <strong>of</strong> -functionalised ketones, http://www.degussacatalysts.com/asp/documents/125_Document.pdf<br />
[194] K. Balazsik, M. Bartok, J. Mol. Catal. A: Chem. 219 (2004) 383-389.<br />
[195] Z. Ma, J. Kubota, F. Zaera, J. Catal. 219 (2003) 404-416.<br />
[196] C.P. Hwang, C.T. Yeh, J. Mol. Catal. A: Chem. 112 (1996) 295-302.<br />
[197] C.E. Smith, J.P. Biberian, G.A. Somorjai, J. Catal. 57 (1979) 426-443.<br />
[198] T. Matsushima, D.B. Almy, J.M. White, Surf. Sci. 67 (1977) 89-108.<br />
[199] H.H. Rotermund, J. Lauterbach, G. Haas, Appl. Phys. A: Mater. Sci. Process. 57<br />
(1993) 507-511.<br />
[200] H. Niehus, G. Comsa, Surf. Sci. 93 (1980) L147-L150.<br />
[201] P.R. Norton, R.L. Tapping, J.W. Goodale, J. Vac. Sci. Technol. 14 (1977) 446-<br />
451.<br />
[202] D. Ferri, T. Burgi, A. Baiker, J. Phys. Chem. B 105 (2001) 3187-3195.<br />
[203] B. Minder, M. Schurch, T. Mallat, A. Baiker, T.P.A. Heinz, J. Catal. 160<br />
(1996) 261.<br />
[204] C. Keresszegi, T. Burgi, T. Mallat, A. Baiker, J. Catal. 211 (2002) 244-251.<br />
[205] R.L. Augustine, S.K. Tanielyan, J. Mol. Catal. A: Chem. 118 (1997) 79-87.<br />
[206] U. Bohmer, F. Franke, K. Morgenschweis, T. Bieber, W. Reschetilowski, Catal.<br />
Today 60 (2000) 167-173.<br />
[207] X.H. Li, X. You, P.L. Ying, J.L. Xiao, C. Li, Top. Catal. 25 (2003) 63-70.<br />
[208] J.R.G. Perez, J. Malthete, J. Jacques, Comptes Rendus De L Academie Des<br />
<strong>Science</strong>s Serie Ii 300 (1985) 169-172.<br />
[209] Y. Nitta, A. Shibata, Chem. Lett. (1998) 161-162.<br />
[210] Y. Nitta, K. Kobiro, Y. Okamoto, Stud. Surf. Sci. Catal. 108 (1997) 191-198.<br />
[211] Y. Nitta, K. Kobiro, Chem. Lett. (1995) 165-165.<br />
[212] Y. Nitta, Y. Ueda, T. Imanaka, Chem. Lett. (1994) 1095-1098.<br />
[213] Y. Nitta, K. Kobiro, Y. Okamoto, Proc. 70th Ann. Meeting Chem. Soc. Japan,<br />
1996, p. 573.<br />
[214] H.-U. Blaser, M. Müller, Stud. Surf. Sci. Catal. 59 (1991) 73.<br />
[215] Y. Nakamura, Bull. Chem. Soc. Japan 16 (1941) 367.<br />
[216] T.J. Hall, P. Johnston, W.A.H. Vermeer, S.R. Watson, P.B. Wells, 11th<br />
International Congress On Catalysis - 40th Anniversary, Pts A And B, 1996, pp.<br />
221-230.<br />
[217] P.J. Collier, T.J. Hall, J.A. Iggo, P. Johnston, J.A. Slipszenko, P.B. Wells, R.<br />
Whyman, Chem. Commun. (1998) 1451-1452.<br />
[218] J. Haglund, A.F. Guillermet, G. Grimvall, M. Korling, Physical Review B 48<br />
(1993) 11685-11691.<br />
[219] T.J. Hall, P. Johnston, W.A.H. Vermeer, S.R. Watson, P.B. Wells, Stud. Surf.<br />
Sci. Catal. 101 (1996) 221.<br />
[220] F. Gao, L. Chen, M. Garl<strong>and</strong>, J. Catal. 238 (2006) 402-411.<br />
[221] Swanson, Tatge, Natl. Bur. St<strong>and</strong>. (U.S.), Circ. 539 (1953) 31.<br />
[222] Hull, Physical Review 17 (1921) 571.<br />
118
[223] R. Noyori, H. Takaya, Acc. Chem. Res. 23 (1990) 345-350.<br />
[224] M. Tamura, H. Fujihara, J. Am. Chem. Soc. 125 (2003) 15742-15743.<br />
[225] A.G. Hu, G.T. Yee, W.B. Lin, J. Am. Chem. Soc. 127 (2005) 12486-12487.<br />
[226] S.U. Son, Y. Jang, K.Y. Yoon, E. Kang, T. Hyeon, Nano Lett. 4 (2004) 1147-<br />
1151.<br />
119
I declare that the current Ph.D. thesis has been done by myself.<br />
Alex<strong>and</strong>er Kraynov,<br />
Date: 24.07.2006<br />
120