Research Collection

Doctoral Thesis

Modification of titania-silica-based epoxidation catalysts and itseffect on surface processes

Author(s): Gisler, Andreas

Publication Date: 2003

Permanent Link: https://doi.org/10.3929/ethz-a-004674897

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.

ETH Library

Diss. ETH No 15382

Modification of Titania-Silica-Based

Epoxidation Catalysts and Its

Effect on Surface Processes

A dissertation submitted to the

Swiss Federal Institute ofTechnology Zurich (ETH)

for the degree of Doctor of Natural Sciences

presented by

Andreas Gisler

Dipl. Chem., University of Zürich

born 5 May 1970

citizen of Zürich

accepted on the recommendation of

Prof. Dr. A. Baiker, examiner

Prof. Dr. A. Togni, co-examiner

2003

Dedicated to my parents:

id Hans-Ueli Gisler-Röthli

for all the love, trust and support

tUrsi and Hans-Ueli Gisler-Röthlisberger

Acknowledgment

Firstly, I would like to express my sincere gratitude to Prof. Dr. A. Baiker for

his support, both personally and scientifically, and the opportunity to complete

my doctoral studies in his group. He provided me the freedom of managing my

work and making valuable experiences during the time being at the ETH.

Moreover, I would like to thank Prof. Dr. A. Togni for accepting the task

of co-examiner in this thesis.

A special thank is due to Prof. Dr. T. Bürgi for his support and contri¬

butions to the spectroscopic part of the thesis. I really appreciated the assis¬

tance, scientific discussions and efforts which were important for finishing this

work.

I furthermore thank Dr. Christian A. Müller for the introduction to the

world of aerogels and chemical engineering. His contributions were essential

for the first part of my thesis; Michael S. Schneider for all the measurements

and experiments performed during his diploma work; Dr. Tamas Mallat for his

help in drafting and reviewing the first publications and for interesting discus¬

sions.

Furthermore, I would like to thank for the contributions to the present

work: Dr. Marek. Maciejewski for measuring thermoanalysis and for always

having an open door for discussions - they were a valuable part of my time at

the ETH; Felix Bangerter for recording NMR spectra and his refreshing com¬

ments on scientific research; Dr. Frank Krumeich for TEM/SEM measure¬

ments; Dr. Charley Pickel for his technical support, beer and discussions;

Urs Krebs and Markus Kupfer for their help with fine mechanics; Max

Wohlwend for electronic support; Florian Eigenmann for Puls-TA experi¬

ments; Ronny Wirz for his support concerning ATR experiments and

Atsushi Urakawa for calculations.

A very big thank is dedicated to my office-mates Florian Eigenmann,

Dr. Reto Hess, Dr. Carsten Beck and Michael S. Schneider for sharing many

joyful moments in- and outside the ETH, for the splendid talks and discus¬

sions, and all the personal and scientific support which was essential to me.

Thanx guys!

Additionally I thank the whole Baiker-group for the unique atmosphere.

Specially I would like to mention the following members for sharing a lot of

good times and unforgettable moments: Dr. René Koppel, Dr. Steffen Auer,

Dr. Manuel Wildberger, Dr. Markus Schüren, Dr. Christian A. Müller,

Dr. Roland Wandeler - BMIG of all times, Dr. Clemens FX Wögerbauer,Dr. Reto Tschan, Dr. Nikiaus Künzle, Dr. Leo Schmid, Dr. Simon Frauchiger,Markus Rohr, Michael Ramin and Simon Diezi.

Besides comedy and sports, music was an important part of my time at the

ETH and was often an essential contrast to scientific research. Therefore a spe¬

cial thank is due to the members of the following bands for the time on stage

and elsewhere...: I.TA. Tribute Band, Papa's Blues Band, Crosswind,

59 Container Blues Band, Savage Blues Band, Funtonic, Polly Duster and

Ten4Soul.

Finally, I would like to thank my family, my friends and female company

for their love and support throughout all these years ofmy education.

Table of Contents

Acknowledgment v

Table of Contents ix

Summary xiü

Zusammenfassung xvü

1 Introduction 1

1.1 General Introduction on Epoxidation 1

1.2 Homogeneous Epoxidation Catalysts 2

1.3 Heterogenization of Homogeneous Catalyst Systems 3

1.4 Heterogeneous Epoxidation Catalyst 5

1.4.1 Supported Oxides 5

1.4.2 Ti-Substituted Molecular Sieves 5

1.4.3 Ti02-Si02 Mixed Oxides 7

1.5 Mechanistic Studies 8

1.5.1 Homogeneous Ti-Catalysts 8

1.5.2 Ti-Substituted Molecular Sieves 10

1.5.3 Epoxidation ofAllylic Alcohols 11

1.5.4 Ti02-Si02 Mixed Oxides 13

1.6 Attenuated Total Reflection Infrared Spectroscopy 16

1.6.1 In Situ Spectroscopy 16

1.6.2 Historical Development ofATR-IR Spectroscopy 17

1.6.3 Theory ofATR-IR Spectroscopy 17

1.6.4 ATR-IR Studies 24

1.7 Scope of Thesis 26

X

2 Experimental 29

2.1 Aerogel Preparation 29

2.2 Physicochemical Characterization 30

2.3 Epoxidation Reactions 30

2.4 Analysis 31

2.5 ATR-IR Spectroscopy 32

3 Titania-Silica Epoxidation Catalysts Modified by Mono- and

Bidentate Organic Functions 37

3.1 Introduction 37

3.2 Experimental 39

3.2.1 Synthesis of Sol-Gel Precursors 39

3.2.2 Aerogel Synthesis 41

3.2.3 Thermal Analysis 42

3.2.4 Nitrogen Physisorption 42

3.2.5 Nuclear Magnetic Resonance (NMR) 42

3.2.6 Electron Microscopy 43

3.2.7 Vibrational Circular Dichroism (VCD) 43

3.2.8 Epoxidation Procedure 43

3.3 Results AA

3.3.1 Structural Properties AA

3.3.2 Catalytic Properties 49

3.4 Discussion 53

3.5 Conclusions 55

4 Epoxidation on Titania-Silica Aerogel Catalysts Studied by

Attenuated Total Reflection Fourier Transform Infrared and

Modulation Spectroscopy 59

4.1 Introduction 59

4.2 Experimental 60

4.2.1 Preparation of Catalyst Layer 60

4.2.2 Nitrogen Physisorption 61

4.2.3 ATR Spectroscopy 62

4.2.4 Modulation Spectroscopy 62

XI

4.2.5 Adsorption Experiments 62

4.2.6 Epoxidation Experiments 63

4.2.7 Variable Temperature Experiments 63

4.2.8 Theoretical Calculations 6A

A3 Results 6A

A3.1 Adsorption Experiments 6A

4.3.2 Variable Temperature Experiments 69

4.3.3 Concentration Modulation Experiments of

Epoxidation Reaction 69

AA Discussion 76

A3 Conclusions 84

5 Epoxidation of Cyclic Allylic Alcohols on Titania-Silica Aerogels

Studied by Attenuated Total Reflection Fourier Transform

Infrared and Modulation Spectroscopy 87

5.1 Introduction 87

5.2 Experimental 88

5.2.1 Preparation of Catalyst Layer 88

5.2.2 Nitrogen Physisorption 88

5.2.3 ATR Spectroscopy 90

5.2.4 Modulation Spectroscopy 90

5.2.5 Adsorption and Epoxidation Experiments 90

5.3 Results 91

5.3.1 Adsorption Experiments 91

5.3.2 Concentration Modulation Experiments of

Epoxidation Reactions 95

5.4 Discussion 100

5.5 Conclusions 106

6 References 109

Outlook 123

List of Publications 127

Curriculum Vitae 131

Summary

The main aim of this thesis was to investigate the influence of surface modifica¬

tion on the structure and reactivity of titania-silica aerogels and to gain a deeper

understanding of the different surface processes occurring during epoxidation.

Two routes were applied to synthesize functionalized trimethoxysilanes as

precursors. Modifiers carrying acetoxy groups were obtained from terminal

allylic alcohols by using Pt-catalyzed hydrosilylation. In the case of modifica¬

tion with amino groups, (3-chloropropyl)trimethoxysilane was reacted with the

corresponding organic primary amine to afford the desired bidentate amino-

precursor. These precursors were incorporated in the titania-silica matrix by

addition during the sol-gel process. Aerogels with different Ti content were

synthesized and characterized by 13C and 29Si CP/MAS-NMR, N2-physisorp-

tion, electron microscopy, and thermoanalysis. The reactivity and selectivity of

the titania-silica mixed oxides were investigated in the epoxidation of cyclohex-

ene and cyclohexenol using tert-butyl hydroperoxide (TBHP) as oxidant.

The amorphous mesoporous structure of the aerogels markedly depended

on the nature of the modifier. All modified aerogels showed a lower BET

surface area and specific pore volume compared to the corresponding unmodi¬

fied aerogels. 29Si CP-MAS NMR spectroscopy revealed that the unmodified

aerogel has a higher degree of crosslinking as derived from the Q /Q values.

Organic modification had a remarkably positive influence on the rate of epoxi¬

dation of cyclohexene and cyclohexenol. Also the selectivities could be

enhanced by aerogel modification with a highest achieved selectivity of 91% at

80% TBHP conversion for a bidentate (diaminoalkyl)-modified aerogel. The

acid-catalyzed side reactions could be suppressed by organic modification of

the catalyst surface. These effects could be due to direct interaction of the

XIV

organic modifying group with the active Ti sites or due to H-bonding to the

acidic surface silanols which leads to reduction of the surface polarity.

Studying the epoxidation process by in situ ATR-IR spectroscopy com¬

bined with modulation excitation spectroscopy showed to be a powerful tool

for gaining insight into the phenomena occurring at the catalytic solid-liquid

interface during epoxidation on titania-silica aerogels. Modulation excitation

spectroscopy (MES) was applied by periodically changing the concentration of

cyclohexene and of the oxidant (TBHP), respectively, while the concentration

of the corresponding compound remained constant. Two different species of

TBHP were discernible from the adsorption experiments. One interacting

strongly with the surface silanol groups and the other coordinating to the active

Ti site. The latter could be traced to a strong blue shift of the C-O stretchingvibration and a red shift of the signal originating from Ti-O-Si vibration.

Modulation experiments under reaction conditions showed that the first one is

a spectator species, whereas the latter is actively involved in the epoxidation

process. Modulating the cyclohexene concentration showed that the surface

coverage ofTBHP remains constant, while depletion ofTBHP occurred at the

active site. Observation of the phase angle at which the product and modulated

reactant disappear revealed a clear time (phase) lag, which was smaller between

cyclohexene and cyclohexene oxide compared to the one between TBHP and

product. This behavior could also be monitored by GC analysis of the effluent

reaction solution of the ATR-cell. Pore diffusion limitation might be at the

origin of this phenomenon, the lower pore diffusion rate for TBHP could be

explained by the higher affinity ofTBHP to the catalyst surface. This indicated

that both chemical kinetics and diffusion rate are decisive factors for the epoxi¬

dation on titania-silica mixed oxides. Therefore proper design of the pore struc¬

ture and surface polarity is a key factor for efficient titania-silica aerogels. When

the catalyst was slowly heated in the presence of the reaction mixture, structural

changes could be observed which occurred simultaneously with the product

detected by GC analysis in the effluent solution.

Based on this work, adsorption and epoxidation of cyclohexenol and

cyclooctenol was investigated by analogous experiments as performed in the

case of cyclohexene in order to elucidate the role of the hydroxy group of the

allylic alcohol. Interestingly, a stronger and less reversible bonding to the cata-

Summary xv

lyst surface was observed for cyclohexenol compared to cyclooctenol. This

behavior led to catalyst deactivation which was observed in epoxidation experi¬

ments when modulating the concentration of cyclohexenol and TBHP, respec¬

tively. In the case of cyclohexenol modulation, deactivation was traced to the

spectra of the catalyst surface which was static, when the experiment was

repeated. In the case ofTBHP modulation, it was evident by the fact that no

displacement of cyclohexenol and no adsorption of oxidant was discernible.

On the other hand, in the study of cyclooctenol epoxidation, TBHP and the

allylic alcohols showed a similar affinity to the catalyst surface and active sites

and displacement of each reactant could be observed. In analogy to the experi¬

ments with cyclohexene, a time (phase) lag between the appearance of reactant

and product in the volume probed by the evanescent field was observed. As a

result of the strong interaction of the hydroxy group of cyclohexenol with the

active site mainly eis epoxide was formed. Due to steric hindrance, for the

cyclooctenol interaction with silanol groups close to the active Ti sites is

favored and therefore the ^raws-epoxide was the main product. Evidence for the

former was delivered by the occurrence of a framework vibration upon adsorp¬

tion of cyclohexenol, whereas the latter was supported by large negative bands

of the silanol groups in case of cyclooctenol epoxidation.

Zusammenfassung

Das Hauptziel dieser Arbeit war es, den Einfluss von Oberflächenmodifizie¬

rung auf die Struktur und Reaktivität von Titandioxid-Siliciumdioxid Aeroge-

len zu untersuchen, sowie einen tieferen Einblick in die Prozesse zu erlangen,die während der Epoxidation an der fest-flüssigen Grenzschicht stattfinden.

Zwei Methoden wurden evaluiert, um Trimethoxysilan mit funktionellen

Gruppen zu synthetisieren. Modifikatoren mit einer Acetoxygruppe wurden,

ausgehend von terminalen Allylalkoholen, durch Platin-katalysierte Hydro-

silylierung hergestellt. Bei der Modifizierung mit Aminogruppen wurde

(3-Chloropropyl)trimethoxysilan mit dem entsprechenden primären Amin zur

Reaktion gebracht, um den gesuchten bidentaten Amin-Vorläufer herzustellen.

Die Modifikatoren wurden durch Zugabe im Sol-Gel Prozess in die Titan¬

dioxid-Siliciumdioxid Matrix eingebaut. Aerogele mit einem unterschiedlichen

Titangehalt wurden hergestellt und mit Hilfe von N2-Physisorption, 13C und

29Si CP/MAS NMR, Elektronenmikroskopie und Thermoanalyse charakteri¬

siert. Die Reaktivitäten und Selektivitäten der Titandioxid-Siliciumdioxid

Mischoxide wurden anhand der Epoxidierung von Cyclohexen und Cyclohexe¬

nol mit £<?r£-Butylhydroperoxid (TBHP) untersucht.

Die amorphe, mesoporöse Struktur der Aerogele hing deutlich von der Art

der gewählten Modifikatoren ab. Alle modifizierten Aerogele wiesen im Ver¬

gleich zum unmodifizierten Aerogel kleinere BET-Oberflächen und spezifische

Porenvolumina auf. Untersuchungen mit 29Si CP-MAS NMR Spektroskopie

zeigten, dass das nicht modifizierte Aerogel einen höheren Vernetzungsgrad

hatte, wie aus den dementsprechenden Q /Q Werten abgeleitet werden

konnte. Die organische Modifikation hatte einen markanten positiven Einfluss

auf die Epoxidationsrate von Cyclohexen und Cyclohexenol. Auch die Selek¬

tivitäten konnten durch modifizierte Aerogele gesteigert werden mit einer

XVlll

erreichten Selektivität von 91% bei 80% TBHP-Umsatz für das bidentate Dia-

minoalkyl-modifizierte Aerogel. Die Säure-katalysierten Nebenreaktionen

konnten durch organische Modifizierung unterdrückt werden. Diese Effekte

könnten von direkten Wechselwirkungen der organisch-funktionellen Grup¬

pen mit den aktiven Titan-Zentren herrühren oder aufgrund von Wasserstoff-

Bindung mit den sauren Silanolgruppen an der Oberfläche entstehen, was zu

einer Reduktion der Polarität der Oberfläche führen würde.

Die Untersuchung des Epoxidationsprozesses mittels in situ ATR-IR Spek¬

troskopie erwies sich als eine sehr vielversprechende Methode, um Einblick in

die Phänomene zu gewinnen, die während der Epoxidierung an der fest¬

flüssigen Grenzfläche von Titandioxid-Siliciumdioxid Aerogelen auftreten.

Modulationsspektroskopie (engl. Modulation Excitation Spectroscopy; MES)

wurde angewandt, indem die Konzentration entweder von Cyclohexen oder

des Oxidationsmittels (TBHP) periodisch verändert wurde, während diejenigedes entsprechenden anderen Reaktanden konstant blieb. Zwei verschiedene

TBHP-Spezies waren aufgrund der Adsorptionsexperimente erkennbar: Eine

weist eine starke Wechselwirkung mit den Silanolgruppen an der Oberfläche

auf, während die andere an das aktive Titan-Zentrum koordiniert. Letzteres

konnte aufgrund einer Verschiebung der C-O Streckschwingung zu höheren

Wellenzahlen und einer Verschiebung der Ti-O-Si Schwingung zu tieferen

Wellenzahlen zugewiesen werden. Modulationsexperimente unter Reaktions¬

bedingungen zeigten, dass die erste ein Zuschauer-Spezies ist, während die

zweite die aktive Spezies im Epoxidationsprozess darstellt. Die Modulation der

Cyclohexenkonzentration zeigte, dass die Konzentration des auf der Katalysa¬

toroberfläche adsorbierten TBHP konstant blieb, während an den aktiven

Zentren die Konzentration des Oxidationsmittels verringert wurde. Beobachtet

man die Phasenwinkel, bei denen Produkt und Reaktanden verschwinden, so

entdeckt man eine Phasenverschiebung zwischen Cyclohexen und Epoxid, die

im Vergleich zu TBHP und dem Produkt kleiner war. Dies konnte auch durch

GC-Analyse der Reaktionslösung, die nach der ATR-Zelle gesammelt wurde,

verfolgt werden. Stofftransporthemmung durch Porendiffusion könnte ein

Grund für dieses Phänomen sein, die kleinere Porendiffusionsrate konnte

durch die höhere Affinität von TBHP zur Katalysatoroberfläche erklärt werden.

Dies zeigte, dass sowohl die chemische Kinetik als auch die Porendiffusion

Zusammenfassung xix

entscheidende Rolle für die Epoxidation an Ti02-Si02 Mischoxiden spielen.

Daher stellt die Gestaltung von Porengrösse und Oberflächenpolarität ein

wichtiger Faktor für die Effizienz von Ti02-Si02 Aerogelen dar. Beim langsa¬

men Erwärmen des Katalysators in Gegenwart der Reaktionslösung, konnten

strukturelle Veränderungen beobachtet werden, die zeitgleich mit dem gaschro-

matographisch nachgewiesenen Produkt in der Lösung auftraten.

Basierend auf dieser Arbeit wurde, analog zu den Experimenten mit Cyclo¬

hexen, die Adsorption und Epoxidierung von Cyclohexenol und Cyclooctenol

untersucht, um die Rolle der Hydroxygruppe des Allylalkohols näher zu be¬

leuchten. Interessanterweise konnte für Cyclohexenol eine stärkere und weni¬

ger reversible Bindung zur Katalysatoroberfläche nachgewiesen werden als dies

für Cyclooctenol der Fall war. Dieses Verhalten führte zu einer Deaktivierungdes Katalysators, was durch Epoxidierungsexperimente beobachtet werden

konnte, wenn die Konzentration von Cyclohexenol oder TBHP periodisch

verändert wurde. Im Fall von Cyclohexenol konnte die Deaktivierung der Tat¬

sache zugewiesen werden, dass bei der Wiederholung des Modulations-Experi¬

mentes das Spektrum der Katalysatoroberfläche nur noch statisch war. Wenn

die Konzentration von TBHP moduliert wurde, war dies erkennbar, da keine

Verdrängung von Cyclohexenol und keine Adsorption des Oxidationsmittels

beobachtet werden konnte. Die Untersuchung der Epoxidation von Cyclo-

octen hingegen zeigte eine ähnliche Affinität zur Oberfläche und zu den akti¬

ven Zentren für TBHP wie für den Allylalkohol und eine gegenseitige

Verdrängung beider Reaktanden war erkennbar. Analog zu den Experimenten

mit Cyclohexen, wurde eine Zeit- (Phasen)verschiebung zwischen Reaktand

und Produkt im Probevolumen des evaneszenten Feldes beobachtet. Aufgrundder starken Wechselwirkung der Alkoholgruppe von Cyclohexenol mit dem

aktiven Zentrum, wurde hauptsächlich das cis-Epoxid gebildet. Im Fall von

Cycloocten führte die sterische Hinderung zu einer bevorzugten Wechselwir¬

kung mit Silanolgruppen in der Umgebung des aktiven Titan-Zentrums, was

wiederum zu einer Bildung von trans-Epoxiden als Hauptprodukt führte. Der

Beweis für den ersten Fall wurde dem Auftauchen einer weiteren Gerüst¬

schwingung bei der Adsorption von Cyclohexenol zugeschrieben, während im

zweiten Fall die starken negativen Banden der Silanolgruppen bei der Epoxidie¬

rung von Cycloocten den entscheidenden Hinweis lieferten.

Chapter

Introduction

1.1 General Introduction on Epoxidation

Epoxidation is an important step in the syntheses of various organic com¬

pounds. The formation of an oxirane ring by epoxidation of an unsaturated

C-C bond is a crucial step and a powerful reaction in organic synthesis. This

reactive compound can undergo subsequent reactions and it is of great interest,

how the ring opening process takes place. Substituted alkenes present two

enantiotopic sides, from which the electrophilic attack of the oxygen can take

place. Therefore, enantioselectivity of the epoxidation reaction plays an impor¬

tant role since absolute configuration is essential in the synthesis of many

industrial products such as vitamins, pharmaceutical compounds, pheromones

and food additives.

The most common used oxidants in organic synthesis are peracids [1-3],

particularly m-chloroperbenzoic acid (mCPBA) is widely used for epoxidation

reactions [4]. The main reason for the application of peracids is their high reac¬

tivity; no drastic reaction conditions are necessary and reaction temperatures of

0°C allow kinetic resolution of different epoxides. Henbest and Wilson were

the first to establish that the epoxidation of allylic alcohols with peracids occurs

principally eis to the hydroxy group [5]. The authors proposed a transition state,

where the hydroxy group of the allylic alcohols coordinates via H-bonding to

the peroxy group (Scheme 1-1). In many investigations, mCPBA was often

used as reference for epoxidations by homogeneous and heterogeneous cata¬

lysts [6-8].

1

2 Chapter 1

////,

Ar

H

—A Y

H^;0; o "»•••r/ \-j...rt»tt^

Scheme 1 -1 : Suggested transition complex of olefin epoxidation by mCPBA as proposed

by Henbest and Wilson [5].

1.2 Homogeneous Epoxidation Catalysts

The origin of metal catalyzed epoxidation can be found in the work of Milas

[9-ii] and the consecutive investigations of Payne et al. [12]. The latter showed

that using H202 in the presence of transition metals like W, V, Mo leads to

selective epoxidation of olefins. Different metal acetylacetonates were tested for

their reactivity in epoxidation of different alkenes [13]. Cr, V and Mo catalysts

were found to be very active and lead to high selectivities. VO(acac)2 was used

as a reference for many studies on epoxidation of allylic alcohols and activity

and selectivity were compared with mCPBA and different homogeneous epoxi¬

dation catalysts [6,7,14]. Sheldon and van Doom proposed, that the highest oxi¬

dation state of the metals is necessary for the activation of the peroxide [15]. In

their function as Lewis acids, the withdrawal of electrons from the peroxidic

oxygen renders the hydroperoxides active for the nucleophilic attack and there¬

fore no change in the oxidation state is involved. The authors suggested the

M=0 group in the catalyst to react in a similar manner to the carbonyl group

of the organic peroxy acids.

A milestone in homogeneous catalysis for epoxidation was the discovery of

the Ti(01Pr)4 catalyzed selective and asymmetric epoxidation of allylic alcohols

at low temperature with tert-butyl hydroperoxide (TBHP) as oxidant in the

presence of chiral tartrate esters by Sharpless and Katsuki (Scheme 1-2) [16].

Introduction 3

D-(-) Diethyl tartrate

"O"

(CH3)3COOH, TiÇOWu

CH2C12 -20°C

RVR1o.

R3OH

L-(+) Diethyl tartrate

Scheme 1-2: First schematic presentation of asymmetric epoxidation of allylic alcohols

with TBHP catalyzed by Ti(0'Pr)4/tartrate [16].

Another successful example are manganese(III) complexes derived from chiral

tetradentate ^zV(salicylaldiminate) ligands which showed to be most powerful

for asymmetric epoxidation [17,18].

Disadvantages of homogeneous catalysts are the low thermal and mechani¬

cal stability and particularly the difficult separation from reactants and pro¬

ducts which makes them unsuitable for continuous application.

1.3 Heterogenization of Homogeneous Catalyst Systems

Immobilization of homogeneous catalyst has been the scope of extensive work

attempting to merge the high selectivity and activity of the homogeneous cata¬

lyst with the stability and easy separation of heterogeneous systems. Graftingresults in direct binding to the solid surface. This technique was used by

Maschmeyer et al. to anchor a [(C5H5)2TiCl2] complex in the piano stool

configuration on the surface of mesoporous silica MCM-41 [19]. Tethering is a

similar method but in this case, a spacer ligand is used for the attachement of

the metal complex [20,21]. To prevent leaching, the complex was covalently

bound via Si-C binding [22]. Another way to immobilize homogeneous catalyst

4 Chapter 1

is polymerization with a cross-linking agent. Immobilization of Mn(III)-salen

complexes was achieved by this method, however with lower selectivities than

the homogeneous species (Scheme 1-3) [23]. A successful example presents the

use of chiral polyaminoacids for the epoxidation of a,ß-unsaturated carbonyl

compounds [24]. The heterogenized polystyrene-bound Ti-tartrate complex or

the combination of a dialkyl tartrate and titanium-pillared montmorillonite

present another markable step in immobilization of homogeneous cata¬

lysts [25-29]. A further possibility presents the coordinative immobilization

through strong coordinating ligands which are anchored in the matrix [30].

R,R = (CH2)4

Scheme 1 -3: Immobilized Mn-salen complex via polymerization by Minutolo et al. [23].

Introduction 5

However, immobilization often results in loss of symmetry elements and the

immobilized complex becomes rigid and therefore less active and selective

[31,32]. Besides, metal leaching was often found to be the reason for the

observed selectivity and activity [33-36].

1.4 Heterogeneous Epoxidation Catalyst

1.4.1 Supported Oxides

The first truly heterogeneous epoxidation catalyst was presented by Shell work¬

ers in 1970 [37]. Impregnating a silica surface with a TiCl4 precursor lead to so

called Ti02-on-Si02 catalysts. The Ti-precursors were attached by free silanol

groups on the surface of the silica carrier. In this way, Ti(IV) site isolation can

be achieved and Lewis acidity is created by electron withdrawal through the

Si-O-ligands [38]. When using alkyl hydroperoxides as oxidants, this supported

oxide showed high activity. Today, this catalyst is still applied in half of the

worldwide propene oxide production performed by continuous liquid phase

epoxidation. When using H202, the hydrophilic catalyst is rapidly deactivated

by leaching of Ti. Alternatively, organic Ti(IV) precursors like Ti(01Pr)4 or

Ti(OiPr)x(acac)4,x [39,40] were used to prepare supported oxides.

Other metals like Ag were also applied to impregnate the surface to obtain

Ag/Si02 catalysts. This catalyst is used to produce ethylene oxide by vapor-

phase oxidation with molecular oxygen [41]. Also ZrCl4 was immobilized onto

the silica surface and used as epoxidation catalyst [37]. Alumina was used as car¬

rier to immobilize Mo precursors [42]. However, this active and selective catalyst

for the epoxidation of allylic alcohols was not stable during reaction.

1.4.2 Ti-Substituted Molecular Sieves

An important milestone in the history of heterogeneous epoxidation catalysts is

the discovery of the crystalline microporous TS-1 by Taramasso et al. [43]. The

name originates from the structural similarity with silicalite-1 (S-l). Part of the

6 Chapter 1

silicon atoms are isomorphously replaced by Ti(IV). This titanium-substituted

molecular sieve was found to be an active and selective oxidation catalyst using

H202 as oxidant and due to its hydrophobic surface, leaching of Ti can be

prevented. Epoxidation takes place already at low temperatures and is carried

out in a polar solvent. TS-1 has also been found to catalyze many other oxida¬

tions like alkane and alkene hydroxylations and alcohol oxidation under mild

conditions (Scheme 1-4) [44-46].

Based on this work Belussi et al. developed a Ti-substituted molecular sieve

based on the structure of silicalite-2 [47]. Also other metals such as AI, Ga, Fe,

Ge, Cr, Sn and V have been incorporated in the structure of S-l and S-2 [48-51].

O

Scheme 1 -4: Catalytic oxidations performed with titanium silicalite 1 (TS-1) and aqueous

hydrogen peroxide.

Introduction 7

However, due to the small diameter of the micropores (-0.6 nm for TS-1),

these materials are not suitable for epoxidation reaction with bulkier substrates.

These limitations led to the development of molecular sieves with bigger pores.

The Ti-beta, a material isomorphous to zeolite ß (containing Al), was synthe¬

sized by Camblor et al. [52]. Ti-beta shows bigger cavities consisting of 12 Si

tetrahedra compared to TS-1 (10 Si tetrahedra) and can activate tert-butyl

hydroperoxide (TBHP) as oxidant [53,54]. Another attempt to overcome the

limitations of microporous molecular sieves was the development ofTi-MCM-

41 [55,56] and Ti-MCM-48 [57]. These ultra large-pore catalysts were found to

be active for epoxidation of bulky olefins. Also these molecular sieves were syn¬

thesized with different metals. A Zr-substituted form ofMCM-41 for instance

showed to be active for epoxidation of cholesterol [58].

Molecular sieves based on alumo-phosphate (APO) were used to synthesize

a novel class of epoxidation catalysts. Additionally, the Al(III) centers were

observed to be crucial for ring opening reactions of the epoxide [59,60]. Besides

Ti (TAPO), V and Co metal ions were also used for substitution [61-63]. These

materials show low activity and often suffer from metal leaching.

1.4.3 Ti02-Si02 Mixed Oxides

Despite of their amorphous nature, Ti02-Si02 mixed oxides are very interest¬

ing materials for epoxidation of bulky olefins [39,45,64-71]. The most common

ways to produce titania-silica mixed oxides are the sol-gel method [69,70,72-76],

coprecipitation [77-80] and flame pyrolysis [81,82]. The conditions of the sol-gel

process have a major influence on the properties of titania-silica mixed oxides.

It was found that at least two types of Ti species are present: segregated Ti02

microdomains and isolated Ti species [39,76,83]. When the titania content was

higher than 15 wt% (nominal Ti02), the Ti02 microdomains became more

prominent in the mixed oxides [80,82]. Since Si and Ti precursors have different

condensation rates, a two-stage acid catalyzed hydrolysis was introduced to pre¬

pare atomically mixed Ti02-Si02 oxides with high homogeneity [80,82-85].

Another key factor for the physical properties of the mixed oxide is the drying

process. While evaporative drying and heating results in microporous xerogels,

8 Chapter 1

extraction with supercritical C02 preserves the mesoporous structure of the

aerogel because the capillary stresses can be reduced [75,86-88]. Since these

titania-silica mixed oxides are hydrophilic and therefore prone to Ti-leaching,

alkyl hydroperoxides are used as oxidants. Side reactions like dehydration, iso-

merization and oligomerization could be observed when using these cata¬

lysts [89]. To overcome these problems, organic modification by covalently bin¬

ding apolar surface functional groups via Si-C bonds were introduced [70,90-93].

This method was also applied for the modification ofTi-substituted molecular

sieves to enhance surface hydrophobicity and chemical stability [94-97]. Still,

despite the achievements in the past years to reduce hydrophilicity, it is still not

possible to perform epoxidations in aqueous medium as applied for TS-1 with

H202 as oxidant.

1.5 Mechanistic Studies

1.5.1 Homogeneous Ti-Catalysts

Since Ti-catalyzed epoxidation showed to be a promising and successful

method to produce epoxides in high yields and selectivities, extensive work has

been devoted to the investigation of the active sites, its coordination states and

reaction mechanism.

After the discovery of the homogeneous Sharpless catalyst for the epoxida¬

tion of allylic alcohols in 1980 [16], many studies have been carried out to shed

light on the mechanism and structure of the Ti-tartrate ester complex. In anal¬

ogy to the crystal structure of related V(IV) complexes with tartaric acid [98,99],

the complex was found to be dinuclear in a tricyclic structure where the free

carbonylic groups of the ester coordinate to the Ti centers (Scheme 1-5) [ioo].

Forming Ti(IV)-complexes with several tartrate derivatives, Potvin et al. found

all species to be dimeric, but for some complexes an acyclic structure was pro¬

posed [102]. The theoretical paper of Bach and Coddens proposed sterical

demands and the 3D-chirality of the whole complex to be the main reason for

the stereoselectivity [103]. Calculations using frontier orbital approach and

Introduction 9

PostScript error (typech

Scheme 1-5: Dimeric, active Ti(0'Pr)4/tartrate complex during Sharpless asymmetric

epoxidation of allylic alcohols [100,101].

extended Hiickel calculation highlighted two main electronic interactions for

the transition state in a spiro configuration: a) the peroxygen lone pair with the

3t*-orbital of the alkene b) the Ti-peroxygen antibonding orbital with the

Tt-orbital of the alkene [104]. The carbonyl groups of the tartrate esters were

considered to play an important role in stabilizing the dimeric complex by

interaction with the Ti centers. The allylic alcohol was found to coordinate via

dative bonding of the hydroxy group to the Ti(IV) site.

Based on the work of Henbest and Wilson [5], Itoh and coworkers investi¬

gated the stereochemistry of vanadyl acetylacetonate catalyzed epoxidation of

cyclic allylic alcohols with TBHP and compared the findings to the analogous

reactions with mCPBA [14]. While selectivity of the epoxidation with peroxy

acid changed from eis to trans for medium-rings, vanadium catalyzed reaction

preserved aV-stereoselectivity throughout. The authors claimed the position of

the hydroxyl group with respect to the double bond in the transition state to be

crucial for selectivities. This so called "dihedral" angle was found to be 150°, in

a quasi equatorial position for the epoxidation with peroxy acids while in the

case of vanadium, the ideal dihedral angle was determined to be 90°, in a quasi

axial position. The main reason for this difference was the direct coordination

of the hydroxy group to the vanadium atom and on the other hand a hydrogenbonded coordination to the peroxy-group in epoxidation with mCPBA

(Scheme 1-6).

/°VAr

a = 150°

Scheme 1 -6: Proposed transition states and dihedral angles a (C=C-C-0) for epoxidationof allylic alcohols presented by Itoh and coworkers. Catalyzed by VO(acac)2 (left) and epoxi¬dation with mCPBA (right).

1.5.2 Ti-Substituted Molecular Sieves

One of the most investigated heterogeneous catalysts is the Ti-substituted

molecular sieve TS-1. Due to its crystalline structure, the coordination state of

the Ti site was soon found to be tetrahedral. This has also been corroborated by

the spectroscopic data of UV-Vis [105-108], IR and Raman [105,106,109-112],

017-NMR [H3,ii4], EXAFS [106,115-118] andXANES [106,108,115,116,119] measure¬

ments. Despite these unambiguous spectroscopic data, complete understand¬

ing of the active species and the reaction mechanism is still obscure. Extensive

effort has been made to study the activated peroxo complex, which seemed to

be an initial step towards the transition state of the epoxidation. Breaking of at

least one of the Ti-O-Si bonds was proposed to be crucial for the formation of

the active peroxo-complex [120]. The active species was often described as a

stable five-membered ring, where the Ti center is bound to the bulk by three

Ti-O-Si bonds [45,121]. This was also confirmed by calculations which proposed

a Ti(r| -OOH) complex being formed and to be the active species for oxygen

transfer [122]. Under basic conditions, deprotonation of the hydroperoxide leads

to catalyst deactivation. Yet, isolation of an active peroxo-complex has only

succeded in a few cases. Isolation of a peroxo-silsesquioxane complex was

described by Roesky et al. and characterized by NMR [123]. The active site was

found to be in a tetrahedral coordination state and the peroxide was r\ -coordi¬

nated. When this complex and an excess of cyclohexene was stirred at room

10 Chapter 1

OH

fR-

Ri

n? Rs

Ra

0--.W

a = 90°

Introduction 11

temperature, cyclohexene oxide was formed without additional oxidant. A

titanium tert-butyl peroxo complex has already been isolated from titanium-tri-

ethanol-aminate as a five-coordinated titanium complex [124]. In both cases, a

so called tripodal Ti site was found to be the active site, where the titanium

center is linked by three Ti-O-Si bonds. This was also found for different silses-

quioxane analogues [125-127] and grafted titanocene dichloride on MCM-41

surface [19,128]. Studies using IR and NMR spectroscopy show that, in the

absence of olefins, putative alkyl peroxo complexes formed by the addition of

TBHP to tripodal complexes decompose rapidly at ambient temperature [125].

However, the rate of epoxidation was found to be significantly greater than that

of alkyl peroxo intermediate decomposition.

Adam and coworkers on the other hand proposed a transition state, where

two Ti-O-Si bonds are cleaved and substituted intermediately by a solvent mol¬

ecule [129].

Ab initio calculations and comparison with IR/Raman data demonstrated

a r| Na+-peroxo complex to be the most stable one. t| -analogues showed to be

35.4 kjmof higher in energy and therefore less stable. The theoretical findings

were in good agreement with the spectroscopic data [no].

It is known that during reactions the coordination geometry of titanium

may change from fourfold to five- or sixfold coordination as in the case of

epoxidation reactions [19] which was corroborated by the findings of UV-vis

[106,119,130] and XANES [130,131] measurements. Modelling the active Ti sites of

heterogeneous titanium catalysts with soluble silsesquioxane analogues corrob¬

orated this model [127]. Addition of methanol to a solution of these model cata¬

lyst led to fast ligand exchange and to formation of a six-coordinated dimer,

whose crystal structure could be determined. This dimeric complex showed

high activity in epoxidation of cyclohexene with high turnover frequency

(TOF) and high selectivity.

1.5.3 Epoxidation of Allylic Alcohols

Extensive work had been devoted to shed light on the coordination of allylic

alcohols to the active site during epoxidation reaction. Based on the work of

Sharpless for the homogeneous Ti(01Pr)4 tartrate complex [16] and Itoh for the

12 Chapter 1

VO(acac)2 catalyst [14], Adam and coworkers studied the epoxidation of several

allylic alcohols by TS-1 andTi-ß [8,129,132]. The threolerythro ratio of the desired

epoxide was determined and compared with the results found for epoxidation

with mCPBA, and for homogeneous systems such as methyltrioxorhenium

(MTO), VO(acac)2 and "Sharpless"-catalyst (Scheme 1-7). Substitution of the

R

OH

R*

threo

R

OH

P

erythro

R'

Scheme 1-7: Asymmetric epoxidation of substituted allylic alcohols and diastereomeric

products.

allylic alcohol resulting in a 1,2A strain (substitution at R1, and R ) and a

1,3A strain (substitution at R3 and R ) was found to play an important role for

the obtained threo/erythro ratios. A higher sensitivity of the homogeneousvanadium and titanium systems to 1,2-allylic strain was found and a high

sensitivity to a !'3A strain for mCPBA epoxidation and TS-1/Ti-ß. Due to these

findings the authors excluded a direct coordination of the hydroxy group to the

active Ti site for Ti-substituted molecular sieves. The authors claimed a transi¬

tion state similar to that of the epoxidation with mCPBA described by Henbest

and Wilson [5]. In comparison to the findings of Itoh et al. [14], the authors

described a different dihedral angle for the coordination via H-bonding to the

peroxide (140°) and dative O-bonding to the active metal center (90°)

(Scheme 1-8). The reason for the different coordination of the allylic alcohol

for TS-1 and homogeneous Ti(01Pr)4, respectively, was proposed to be a too

encumbered space around the Ti centers in the zeolite lattice, which disables a

closer coordination of the substrate and oxidant.

Introduction 13

a = 120'

R-o—-X.

a = 40°1,2-allyhc

strain

OH

erythro threo erythro threo

Scheme 1-8: Preferred dihedral angles a (C=C-C-0) in the allylic alcohol controlled by

1,2- or 1,3-allylic strain. Catalyzed by VO(acac)2 (right) and epoxidation with mCPBA (left).

Kumar and coworkers investigated regioselectivity of the epoxidation of geran-

iol and cyclic allylic alcohols catalyzed by TS-1 [133]. Due to the preferred

epoxidation of the allylic double bond they claimed a so called hydroxy-assisted

mechanism, where the hydroxy group coordinated directly to the Ti site in the

transition state. Based on the observation of the preferred cw-selectivities for

cyclic allylic alcohols, they claimed the enhanced reactivity and the hydroxy-assisted mechanism to be responsible for the formation of a reactive species with

a dative bond to the Ti site.

1.5.4Ti02-Si02 Mixed Oxides

One problem rendering the control of titanium coordination in the mixed

oxide difficult lies in the fact that titanium does not favor tetrahedral coordina¬

tion in an oxide matrix. Therefore, elucidation of the coordination state of the

Ti sites on the surface as well as in the bulk was of primary interest. A lot of

work has been carried out to collect IR [133], Raman [85,109], UV-Vis [I08,i3i,i34-

136], 29Si NMR [71,76,85,113] and Ti K-edge EXAFS/XANES [72,82,106,108] data on

titania-silica mixed oxides and to compare them with the data obtained for

TS-1. The most important results are summarized in Table 1-1.)

Comparison of the data with the corresponding ones for TS-1 indicate a

mainly tetrahedral coordination state of the Ti sites. However the higher values

14 Chapter 1

Table 1-1 : Physico-chemical properties of TS-1, Ti02-Si02 mixed oxides and Ti02-Si02

supported oxides.

Techniques TS-1 Ti02-Si02 Ti02/Si02

mixed oxides supported oxides

IR

Ti-O-Si vibration [cm"1] 960 910-960 930

Raman 960 935-960 950

Ti-O-Si vibration [cm"1] 1125 1100-1110 1080

UV-Vis

LMCT peak (dehydrated) [cm"1] 45000 - 50000 40900-45000 39000 - 47900

XANES (dehydrated)

Pre-edge peak intensitya 75% 58% 65%

Pre-edge peak position [eV] 4969.7 eV 4969.7 eV 4969.5 eV

EXAFS

Ti-O bond length (average) [Â] 1.80-1.81 1.81-1.82 1.81

Ti-O-Si bond angle 163° 159° -

a

Samples with low Ti contents (<2 wt% Ti02) are used for comparison.

Pre-edge peak position given using first inflection point of Ti foil at 4966.0 eV.

for UV-Vis in the case ofTS-1 suggested that isolated Ti04 sites in titania-silica

mixed oxides and supported oxides are not unique and even at low Ti content,

small amount of polymerized Ti species may also be present [136,137]. Direct

information about Ti-O-Si linkages could be obtained with 170-NMR spec¬

troscopy [113,114]. Additional to the significant pre-edge signal in XANES

experiments at 4969.7 eV, Greegor et al. obtained an average Ti-O bond lengthof 1.81 Â for tetrahedrally coordinated and 1.99 Â for octahedrally coordinated

Ti sites [82]. The intensity of tetrahedrally coordinated Ti centers was found to

be highest below 10 wt% nominal Ti02 and the amount of sixfold coordinated

titanium increased significantly with higher Ti content than 15 wt%. Experi¬

mental data show a strong correlation between catalytic activity of olefin epoxi¬

dation and the fraction of tetrahedrally coordinated Ti atoms [68,70,72].

Another important factor for the activity was found to be the location of

the Ti sites. While for supported oxides titanium sites are mostly located on the

surface of the catalyst, accessibility to the active sites in mixed oxides often

Introduction 15

presents a problem [138]. But it was also shown, that a large part ofTi atoms are

located on the surface for materials with low content. Still, diffusion in the

pores of the catalyst plays an important role for high activity. It was therefore of

primary interest to tune the preparation method of the catalyst for gaining high

accessibility of the Ti centers in the pores [70,74,139].

The preparation method has also a strong impact on the generation of

additional Brensted acid sites which are responsible for catalyzing undesired

side reactions [89,140]. The generation of new acid sites has been in the focus of

many investigations [78,141,142]. Kataoka and Dumesic proposed a model for the

generation of Brensted sites with two important criteria [142]: a) Brensted acid

sites are associated with Ti-O-Si bridges where the Ti atoms are not in tetrahe¬

dral coordination but form pentahedral or octahedral sites, regardless of

composition; b) the coordination change of the Ti atoms upon hydration will

generate weak Brensted acid sites. Additional acidity is also created by surface

hydroxylation as reported in several studies [77,143]. However, Brensted cata¬

lyzed side reactions could be suppressed in the case ofTS-1 by basic treatment

of the catalyst [91] or by ion exchange with metals like Li, Na, K, Ba and

Mg[i44]. Adding organic bases to the reaction mixture could remarkably

increase the selectivities of allylic cyclohexenols [89]. Another important factor

to reduce Brensted acidity is calcination and careful drying of the catalyst prior

to use in epoxidation reaction, because by losing the surface OH group, the Ti

site forms a coordinatively unsaturated state responsible for Lewis acidity

[19,145-147]. Hydrophobization of the surface by silylation [95,148,149] or modifi¬

cation by covalently bound organic groups [90,92,93] was found to have a posi¬

tive effect, reducing undesired side products.

Mechanistic studies with Ti02-Si02 mixed oxides are relatively sparse.

Beck et al studied the epoxidation of allylic alcohols over aerogels with low

Ti-content and discriminated between a hydroxy-assisted mechanism and a

silanol-assisted mechanism [150]. Moreover, the authors claimed a catalyst

restructuring by cleaving a Ti-O-Si bond as the initial step for epoxidation.

Ti(OSiMe3)4 was used as a homogeneous model for the active site of sily-

lated titania-silica mixed oxides [151]. The hydroxy group was found to play an

important role in the epoxidation process, since geraniol was exclusively

oxidized at the 2,3 double bond. Yet, cyclohexenol and cyclooctenol were both

16 Chapter 1

converted predominantly to cw-epoxides. Despite the high activity and struc¬

tural similarity of the homogeneous model, cyclohexene was only epoxidized in

small amounts. However, it is questionable, whether any clear conclusions for

the active Ti sites in titania-silica mixed oxides can be drawn from this model.

Therefore, the in situ investigation of the coordination geometry ofTi(IV)

atoms and adsorption of reactants during reactions is crucial for fully under¬

standing the reaction mechanism of titania-silica in epoxidation reactions.

1.6 Attenuated Total Reflection Infrared Spectroscopy

1.6.1 In Situ Spectroscopy

Many methods were used successfully to investigate chemical reactions which

occur at gas-solid interfaces, including X-ray photoelectron spectroscopy (XPS/

ESCA), Low-energy electron diffraction (LEED), Auger electron spectroscopy

(AES) and Fourier transform infrared (FTIR) spectroscopy just to name a few

[152,153]. Transmission and diffuse reflectance modes are the most often used in

vibrational spectroscopy. The main problem faced with transmission IR

measurements applied to solid-liquid interfaces are the strong absorptions of

the solvent, reactants and bulk catalyst in contrast to the relatively small signalsof the liquid-solid interface. Consequently, this technique is unsuitable for the

investigation of the surface processes such as heterogeneous catalysis. Monitor¬

ing surfaces of solid catalysts in the liquid phase by vibrational spectroscopy has

been applied in a few cases, mostly by reflection-absorption infrared spectros¬

copy (RAIRS) and surface-enhanced Raman spectroscopy (SERS) [154]. While

the latter can be applied under reaction conditions, it is restricted to the study

of polycrystalline metal surfaces [155]. ATR-IR spectroscopy has not been

widely used so far for the investigation of phenomena occurring at the liquid-

solid interface, only a few studies have been reported (vide infra). However, this

technique proved to be a powerful tool to gain insight into the processes taking

place at the catalyst surface.

Introduction 17

1.6.2 Historical Development of ATR-IR Spectroscopy

Newton described the phenomenon of total internal reflection of light in the

early seventeenth century. He observed an evanescent field in a medium with

lower refraction index in contact with a medium of a higher refraction index in

which a propagating wave of radiation undergoes total internal reflection. It

was not until the early thirties in the twentieth century when this phenomenon

was used for spectroscopic purpose by the studies of Taylor and cowor¬

kers [156-158]. However, it took another three decades until this remarkable

techniques were developed by Harrick who was investigating free charge carrier

distributions in semiconductors. He found that parallel polarized radiation had

low reflectivity while perpendicularly polarized radiation showed high reflecti¬

vity at the germanium-mercury interface [159]. The experiments showed to be

in good agreement with theoretical calculations. Discovering the work of

Eischens [160], Harrick concluded that with this technique it should be possible

to study the properties of the rarer medium. In the late fifties, Fahrenfort inde¬

pendently developed the use of total internal reflection to observe the spectra of

organic materials on silver chloride and called this technique attenuated total

reflection [161].

The studies of Harrick and Fahrenfort present the theoretical background for

the widely spread utilization ofATR-IR spectroscopy today [162].

1.6.3 Theory of ATR-IR Spectroscopy

For conventional transmission IR spectroscopy the detected intensity \ can be

described by the Lambert-Beer's law:

Where I0 stands for the incoming light intensity, K for the molar absorption

coefficient, c for the concentration and 1 for the thickness of the sample. In

contrary to transmission experiments, in ATR measurements the IR-beam

propagates inside an internal reflection element (IRE) and gets totally reflected

at the interface. Therefore it is necessary to have a closer look at the phenome-

18 Chapter 1

non of total reflection. The theory of reflection and transmission of an electro¬

magnetic wave was first derived by Fresnel (Figure 1-1). The incident (i) plane

wave consists of parallel (||) and perpendicular (_L) polarized electric field

components Eji and E1±, respectively. The corresponding components of the

reflected and refracted (transmitted) field components are denoted by ErM, Erl,

Eji, and Etl.Fresnel's equations relate the reflected and transmitted compo¬

nents to the corresponding incident components for non absorbing media:

Fig. 1-1 : Specular reflection and transmission. The angles of incidence (i), reflection (r)

and refraction (t) are denoted by 6P 0r and 6t, respectively. The corresponding electric field

components are denoted by E. They are split into orthogonal portions, one parallel to the

plane of incidence (x,z-plane) and the other perpendicular to this plane (parallel to y-axis).

Accordingly, electric fields are referred to as parallel (||) and perpendicular (_L) polarized, rij,

n2, kj and k2 denote the refractive and adsorption indices in the two media.

Introduction 19

Er„ n^cosd,-n,cosd,r» =

EM n2cosd[ + n1cosdt

r _

En_

nicosfl.-wacosfl, ^El± nicosdl + n2cosdt

In the case of internal reflection, the angle of the refracted beam 9t must be

larger than the one of the incident beam 9j and therefore, according to Snell's

law:

n^inSj = n2sin9t (4)

the refractive index of medium 2 must be smaller than the one of medium 1

(n2 < n:). When 9t reaches 90°, total reflection occurs and 9; is at the critical

angle of incidence 9C. It follows from Snell's law:

sin9c = i^/ii! = n21 (5)

Above the critical angle according to the equations (4)/(5) and to

sin9t = sin9i/sin9c > 1 (6)

the corresponding cosine is complex:

cos0, =±mlV\/sin2 0,

-

nl, (7)12-v^u ul ,«.21

Introducing these criteria in Fresnel's equation (Eq. 2 and 3) results in the fol¬

lowing relation between internally incident and reflected electric field compo¬

nents:

E„ n,, cos0, - jJsin 0, - n,,

n =-^

= — -—\' 21

(8)EM n\x cos 0, + i-y sin2 0( - n21

r, =

1

E

E^_=

cos0,-/Vsin 0,-n21 (9),1 cos0( + jJsin 0( - n21

20 Chapter 1

For the medium 2 the corresponding ratios for incident and transmitted elec¬

tric field components can be derived in the same way. However, if medium 2 is

absorbing, the complex refractive index has to be inserted. In this case one has

attenuated total reflection.

Fig. 1 -2: Evanescent/standig wave at the interface ofthe IRE and the probed medium.

With these findings, several conclusions can be made. Calculating the plane

wave in medium 2 results in an electromagnetic wave which is called evanes¬

cent wave whose electric field strength decreases with increasing distance (z)

from the interface (Figures 1-2 and 1-3):

z

Ex,y,z = E0x,y,ze"

(10)

Introduction 21

.E,

EX

Fig. 1 -3: ATR setup for an IRE with 5 total reflections. Optical and structural features are

related to the IRE fixed-coordinate system x,y,z. En and E^ denote the parallel and perpen¬

dicular polarized electric field components of the light incident to the IRE under the angle

6;. En results in the Ex and Ez components of the evanescent wave, while E^ results in the E

component.

Based on this equation, a so-called penetration depth (d ) can be determined,

where the initial electric field strength has decreased to 1/e of its value:

Kdp=— f—-

2jr^sin dt-n(ID

Where X1 = X/n1 denotes the wavelength in medium 1. Interestingly, d is

depending on the wavelength of the probing beam which results in higher

absorption intensities at lower frequencies. Typically, this distance is in the

order of 1/10 of the probing wavelength and varies between a micron and a few

microns, depending on the refractive indices of the two media.

22 Chapter 1

For bulk material, an effective thickness (de) can be defined, which is the

equivalent path in a hypothetical transmission experiment that results in the

same absorption signal as in the ATR experiment under identical conditions. de

is a function of the electric field at the interface En, the refractive index and the

angle of the incident beam:

0'

_

n2\EQdp2cos0

(12)

By choosing the appropriate IRE material and angle of the incident beam 0i5

the penetration depth can be adjusted to an optimal value with respect to the

investigated catalyst layer. Therefore, ATR-IR spectroscopy is a powerful tool

to investigate catalytic liquid-solid interfaces in situ. The most common used

IRE materials are zinc selenide and germanium. Several materials and their

refractive properties are listed in Table 1-2:

Table 1-2: Most common materials used for internal reflection elements (IRE) and their

physical properties.

Material Refractive Usable wavelength Critical

index (n)a range [cm"1] angle (6C)

ZnSe 2.4 20000 - 450 24,6°

KRS-5b 2.37 20000 - 250 24.6°

Si 3.4 10000-1500 15.6°

Ge 4.0 5500-600 14.5°

AMTIR (As/Se/Ge glass) 2.5 11000-750 23.6°

aAt 5000 cm'1.

Eutectic mixture of thallium bromide/i(adide.

Due to low signal intensities at the surface, improvement of the signal to noise

(S/N) ratio is an important factor. Moreover, most spectrometers are used in a

single beam mode, which makes it problematic to accumulate data over a longtime due to instrumental instabilities. A widely used technique is changing the

geometry and volume of the IRE in order to increase the number of total

Introduction 23

reflections, resulting in so called multiple internal reflection elements (MIRE).

The reflectivity (R) changes with N reflections (according to the Lambert-

Beer's law):

>NRiN = (l-ade)

N(13)

By this way, the effective thickness is amplified with each reflection and the

signal to noise (S/N) ratio is increased.

Another interesting method is changing the spectrometer into a pseudo-

double-beam mode, which is achieved by mechanically changing the vertical

position of the IRE by means of a lift, so that sample and reference can be

measured at the same time. This so called single-beam-sample-reference

(SBSR) technique will be highlighted in the experimental part (chapter 2).

A very promising technique, modulation excitation (ME) spectroscopy and

phase-sensitive detection (PSD) has been presented by Baurecht and

Fringeli [163]. If a system is disturbed by periodically varying an external para¬

meter such as temperature, pressure, or concentration of a reactant, then all the

species in the system, which are affected by this parameter will also change

Sample Demodulation

Demodulated

Spectra

Fig. 1 -4: Schematic setup for modulated excitation (ME) experiments. Periodic excitation

ofthe system perturbation of an external parameter is performed with frequency go. Detected

time-resolved response of the system is transformed by phase-sensitive detection (PSD) to a

phase-resolved spectrum where static signals are suppressed.

24 Chapter 1

periodically at the same frequency as the stimulation, or harmonics thereof

(Figure 1 -4). At the beginning of the modulation, the system relaxes to a new

quasi steady-state around which it is oscillating at angular frequency oo. As a

consequence, it is possible to separate the signals of the response from static

signals which are not affected by the periodic perturbation through a phase-

sensitive detection (PSD). The response of the system is followed by recordingtime-resolved spectra which are converted to phase-resolved spectra by a digital

PSD according to the equation:

Af\f) = ^JA(f,t)sm(kcot + fkSD)dt (U)1

0v /

k = l,2,...

A(f,t) is the time dependent absorbance at wavenumber tf, oo is the stimulation

frequency, T is the modulation period and </>fSD is the demodulation phase-

angle. Moreover, if the kinetics of the stimulated process is in the same range as

the excitation-period of the external parameter, phase lags and damped ampli¬

tudes will result. Using ME spectroscopy, even small signals arising from

changes at the catalyst surface can be detected with a good (S/N) ratio [164].

1.6.4 ATR-IR Studies

Attenuated total reflection spectroscopy has been used to analyze different

types of solid-liquid interfaces related to heterogeneous catalyst. McQuillan

and coworkers investigated adsorption processes of different organic

compounds on metal oxide films including Ti02, Zr02 and A1203 gel

layers [165-167]. This work was based on reported investigations of adsorption

experiments using metal oxide coated IREs [168-171]. A further development of

these studies was the introduction of the surface titration by internal reflection

spectroscopy (STIRS) to study water-solid interfaces [172]. Williams et al.

studied the adsorption of CO in aqueous and ethanolic solutions over thin

films (10 u,m) of Pt/Y-Al203 catalyst [173]. The authors found that CO resides

in both atop and bridged configurations on the catalyst surface. Adsorption of

butyronitrile from hexane was found to bind via ö-bond of the CN group with

the platinum. Rivera and Harris studied pyridine adsorption over bare and

cyano-derivatized silica sol-gel films and observed a strong interaction of the

Introduction 25

substrate with surface silanol groups [174]. Derivatizing the surface of the sol-geland therefore reducing free surface silanol groups reduced the total number of

adsorbed pyridine by 43%. In another publication, the same authors investi¬

gated the kinetics of transport and binding within silica sol-gel films for non-

binding molecules and species with a high affinity for the silica surface [175].

The rate of transport into the film was found to decrease drastically for mole¬

cules which showed a high interaction with the surface. Photocatalytic reduc¬

tion of 02 and intermediates on nanocrystalline Ti02 films in contact with

aqueous solutions were monitored by ATR-IR [176]. The observed signals upon

photocatalytic reaction were interpreted as Ti(02) peroxo species and

Ti(OOH) hydroperoxo species. Hydrogénation of C02 in CH2C12 was

followed by in situ ATR-spectroscopy over a thin Pt/Al203 film which was

prepared on the IRE by vapor deposition [177]. Based on this work, the adsorp¬

tion of cinchonidine (CD), a chiral auxiliary for asymmetric hydrogénation,

was studied over Pt/Al203 catalysts [178,179]. It was found, that the orientation

ofCD on the surface strongly depends on surface coverage and solvent effects.

The modulation excitation technique was used to investigate the enantio-

selective hydrogénation of a substituted pyrone over a Pd/Ti02 catalyst modi¬

fied by cinchonidine [164]. The formation of carboxylates by alcoholysis was

observed and compared with the kinetics of the appearance and disappearance

of reactant. The latter was found to be faster which was indicated by the

obtained phase lag. The adsorption of enantiomers at a chiral interface was

followed by modulating the absolute configuration of the admitted mole¬

cule [180]. It was possible to distinguish the adsorption of the two enantiomers

on the chiral silica used as stationary phase.

More recently, ATR spectroscopy was also successfully applied for in situ

studies at high pressure, which opens a new field for investigating liquid-solid

interfaces [181,182].

This selection of examples clearly show the potential ofATR spectroscopy

allowing the study of adsorption processes, kinetics, and nature of adsorbed

species truly in situ. Despite the so far rare utilization of this spectroscopic tech¬

nique in catalysis, it represents presently the most promising method to eluci¬

date the phenomena on solid-liquid interfaces.

26 Chapter 1

1.7 Scope of Thesis

In this work, a combination of different approaches like surface modifica¬

tion, model reactions and in situ spectroscopy should be used to shed light on

the performance of titania-silica mixed oxides as catalysts for epoxidation reac¬

tion.

In a first step the surface of titania-silica aerogels will be modified by intro¬

ducing organic functional groups which are covalently bound to silicon atoms.

An important target is the creation of a less hydrophilic catalyst surface. Adding

organic bases or alcohols to the reaction mixture was already found in previous

studies to have a positive influence on the selectivity of epoxidation reactions.

With this in mind, mono- and bidentate aminoalkyl and acetoxyalkyl precur¬

sors will be introduced in the sol-gel process. Physical characterization will be

performed to gain information on the modification - structure relation of the

different aerogels. By performing epoxidation of cyclohexene and cyclohexenol

as model reactions, the influence of organic modification on activity and selec¬

tivity will be investigated.

Only little work has been performed so far on the study ofprocesses taking

place at the liquid-solid interface during epoxidation over titania-silica mixed

oxides. Since unambiguous information is hard to obtain due to the amor¬

phous structure of mixed oxides, it is of primary interest to investigate the cata¬

lyst under working conditions during epoxidation. To gain further information

on this process, the reaction of cyclohexene with TBHP will be performed over

titania-silica aerogels and followed by attenuated total reflection (ATR) spec¬

troscopy. Modulating the concentration of reactants will be applied for these

studies.

Epoxidation of allylic alcohols was investigated in many previous studies

performed over Ti-substituted molecular sieves. However, in the case of mixed

oxides, evidence of the role of the allylic hydroxy group is still lacking. In anal¬

ogy to the experiments with cyclohexene, epoxidation of cyclohexenol and

cyclooctenol will be studied by ATR spectroscopy and compared to the data of

the previous investigations. The aim is to study the influence of the hydroxy

group of the allylic alcohols, the observed catalyst deactivation in the case of

Introduction 27

cyclohexenol and the different cis/trans-select'wiûes which were found in epoxi¬

dation reactions.

Chapter

Experimental

Detailed information on specific experimental procedures can be found in the

corresponding chapters. Here, the focus is on the description of the basic

experimental setups and procedures. In addition, some general remarks on the

applied methods and techniques are made.

2.1 Aerogel Preparation

Syntheses of the precursors were carried out under Ar-atmosphere using the

Schlenk-technique unless otherwise stated. All solvents were distilled prior to

use in the reactions, reactants were used as received. Acylation was monitored

by thin layer chromatography (TLC) using silica gel on aluminum foil (Mach-

erey-Nagel). Hydrosilylation was followed by transmission IR-spectroscopy

using a Perkin-Elmer 2000 FTIR spectrometer with a Perkin-Elmer transmis¬

sion cell. Hydration of (3-acetoxy-3,7-dimethyl-6-octenyl)trimethoxysilane

was performed in a home built steel autoclave and H2 was supplied by a Biichi

controller as described elsewhere [183].

Mesoporous titania-silica mixed oxides were prepared according to the well

known method which was previously described in detail [184].

The sol-gel process was carried out in a 250 ml round bottom flask,

equipped with a dropping funnel and magnetic stirrer. All reactants were

diluted in TrOH to prevent inhomogeneity. Since hydrolysis of tetramethoxy-

silane was found to be exothermic, the hydrolysing agent (HN03) was added

dropwise to the reaction mixture. Addition of trihexylamine (THA), to force

2

30 Chapter 2

gelation, was performed very quickly. After gelation, the sol-gels were aged for

6 days.

Drying of the aerogel by semicontinuous extraction with C02 under

supercritical conditions was performed in a steel autoclave, equipped with a

glass liner to prevent metal leaching from the autoclave. Prior to drying, the

aged gel was covered with 25 ml ofTrOH. Typically, extraction was completed

within 5 h at 20 MPa, 318K and a C02 flow of 15 g min'1 affording 60 ml of

extracted liquid. The as prepared aerogel clumps were ground in a mortar and

calcinated in a tubular reactor with an upward air stream.

2.2 Physicochemical Characterization

Trimethoxysilane precursors were described by H-, C- and Si-nuclear

magnetic resonance (NMR). Catalyst materials were investigated by transmis¬

sion and scanning electron microscopy (TEM, SEM), cross-polarization magic

angle spinning (CP-MAS) NMR, thermoanalytical analysis (TG, DTA) and

N2-physisorption.

2.3 Epoxidation Reactions

Epoxidations were performed under Ar-atmosphere in a 50 ml glass reactor

equipped with a reflux condenser, a thermometer, a septum for taking samples

by syringe and a magnetic stirrer.

Prior to adding reactants, the aerogel was dried under inert gas atmosphere at

373K. After cooling to ambient temperature, toluene (distilled over Na) and

dodecane (internal standard) were added and the mixture was heated to the

desired temperature. After addition of the substrate, the reaction was started by

adding tert-butyl hydroperoxide (TBHP) to the vigorously stirred reaction

mixture.

Experimental 31

2.4 Analysis

The samples taken during epoxidation reaction were analyzed using a HP-6890

gas Chromatograph, equipped with a HP-FFAP column (length: 30 m, diame¬

ter 320 mm, film thickness 0.25 um, 95 kPa He, column flow 1.9 ml min'1)

and a cool-on-column inlet (track oven mode) to prevent epoxide decomposi¬

tion during injection. Detection was performed by a TCD (523K, reference

flow 20 ml min'1, combined flow 7 ml min'1) and a FID (523K, H2 40 ml

min'1, air 450 ml min'1) simultaneously. Retention times of reactants and

products are listed in table 2.1. The epoxides were identified by comparison

with authentic samples.

Table 2-1 : Temperature profiles and retention times of the GC-methods used (only FID

retention times are listed).

Substrate Temperature Detected Retention time

profile compounds [min]

cyclohexene 40°C, 5 min, cyclohexene 2.7

lO-Cmin^to 130°C, cyclohexene oxide 9.4

25°C min'1 to 220°C, dodecane (IS)a 10.2

5 min TBHP 11.8

cyclohexenol 60°C, 5 min, dodecane (IS) 7.1

lO-Cmin^to 160°C, TBHP 9.3

5°C min1 to 180°C, Cyclohexenone 11.5

20°C min'1 to 220°C, Cyclohexenol 12.0

9 min Cyclohexenol oxide eis 15.9

Dicyclohex-2-enylether 16.1

Cyclohexenol oxide trans 17.9

cyclooctenol 60°C, 5 min, dodecane (IS) 6.7

25°C min'1 to 170°C, TBHP 8.1

50°C min'1 to 220°C, Cyclooctenol 11.8

1 min Cyclooctenol oxide eis 12.7

Cyclooctenol oxide trans 18.6

a Internal Standard.

32 Chapter 2

2.5 ATR-IR Spectroscopy

The experimental set up used for the investigations is schematically pre¬

sented in Figure 2-1. Reactant solutions were flown over the catalyst by means

Computer

ATR-Cell

Spectrometer

Analysis

Pump GC, UV-Vis

tu

Auto-Sampler

Fig. 2-1 : Schematic experimental setup. The measurement program controls the timing of

the data acquisition, the valve switching and sample collection for the off-line analysis (dotted

lines).

of a peristaltic pump (ISMATEC Regio 100) located after the cell. Liquid was

provided from two separate glass bubble tanks, where the liquid could be satu¬

rated with argon. The flow from the two tanks was controlled by a pneumati¬

cally actuated three way Teflon valve (PARKER PV-1-2324). Teflon tubing was

used throughout. All spectra are presented in absorbance units as A=-Log(I/I0),

where I and I0 are the reflected intensity of the sample and reference, respec¬

tively. The experiments were performed without polarizer.

Experimental 33

The spectroscopic experiments were carried out using a commercial

trapezoidal (45°, 50 x 20 x 2 mm, KOMLAS) ZnSe internal reflection element

(IRE). The catalyst layer was prepared by dropping a slurry of aerogel and

TrOH on the surface of the IRE which was subsequently dried under reduced

pressure. Since only one side of the ATR crystal was coated with a thin catalyst

layer and the cell did not cover the whole crystal, only ca. 8 reflections were

active. The IRE was mounted in a stainless steel flow through cell (s. Figure

2-2). The gap between the polished steel surface of the cell and the IRE is

Fig. 2-2: Pictures of the home-made flow-through cell, a: The open cell exposing the

sample chambers with inlet and outlet (A) sealed by the O-rings (B). The IRE is placed on

the O-rings. b: After mounting the cover and the heating jackets (C).

about 250 um and defined by a 30 x 1 mm viton O-ring (Johannsen AG) fit

into two precision electro-eroded ellipsoid nuts of the steel cell (Figure 2-2 a).

The length and width of the exposed areas of the IRE are 42.8 and 7.5 mm,

respectively. The edges of the exposed area are rounded off to avoid stagnating

34 Chapter 2

regions of the flowing liquid. The total volume of the cell is 0.077 mL. Nor¬

mally, only the upper part of the cell was used for in situ adsorption and epoxi¬

dation experiments, except in the case of the SBSR-method (s. page 35). A

thermostat was used to heat the cell. Two water-heated jackets were fixed on the

two sides of the ATR cell and the temperature was measured with a sensor posi¬

tioned at the bottom part of the cell (Figure 2-2 b). This stainless steel cell was

mounted onto a dedicated ATR attachment (OPTISPEC) within the sample

compartment of the Fourier transform IR spectrometer (Bruker IFS-66/S)

equipped with a liquid nitrogen cooled medium band MCT (HgCdTe) photo-

detector as depicted in Figure 2-3. The sample compartment of the spectro¬

meter was closed with a plexiglass cover and flushed with nitrogen in order to

suppress atmospheric C02 and water.

PM3 PM2

Fig. 2-3: Schematic top view of the setup ofthe ATR-sample compartment with detector

(DET), IRE, mirrors (M) and parabolic mirrors (PM). The mirrors are used for an appropri¬

ate focussing of the IR-beam.

Experimental 35

In the SBSR method the horizontal position of the ATR-cell is changed by

means of a lift (s. Figure 2-4). The blind attached in front and after the IRE

ensures that only adsorptions originating from the desired compartment are

detected. The upper one was used for the sample and the lower for the

reference. Sample and reference lift positions were determined by measure¬

ments in "dry" conditions prior to the experiments. Sample and reference spec¬

trum in "dry" conditions should have similar adsorption intensities.

R

Blind

Fig. 2-4: Alternating change from sample to reference is performed by computer-

controlled lifting and lowering of the ATR cell body.

Chapter

Titania-Silica Epoxidation Catalysts Modified

by Mono- and Bidentate Organic Functions

3.1 Introduction

It has been demonstrated in the past years that titania-silica mixed oxides are

not only good acid catalysts [185,186] but also active in various epoxidation reac¬

tions. Synthesis via the solution sol-gel route provided access to highly

dispersed (isolated) Lewis acidic Ti sites which can effectively activate the alkyl

hydroperoxide oxidant [39,40,66,187,188]. Restructuring the gel during drying can

be minimized by removal of the solvent in supercritical C02 [75]. This method

leads to mesoporous aerogels which are active and selective in the epoxidation

of bulky cyclic alkenes, alkenones and alkenols [39,40,66,140].

A drawback of titania-silica mixed oxides is their strongly hydrophilic char¬

acter due to the presence of surface silanol groups. This property hinders the

application of aqueous hydrogen peroxide as oxidant [66]. Leaching of Ti in

aqueous medium can be suppressed by hydrophobization of the surface. One

approach is the replacement of the surface silanol groups by methyl or phenyl

groups covalently bound to Si. Partial substitution of the tetraethoxysilane

precursor by methyltriethoxysilane or phenyltriethoxysilane affords stable

catalysts for the epoxidation of olefins with aqueous hydrogen peroxide [90,91],

and improves the selectivity when using tert-butyl hydroperoxide (TBHP)

[93,189]. Another successful strategy is the silylation of the gel or the final mixed

oxide [190,191].

Surface modification by covalently bound organic functional groups has a

much broader application range than tuning only the polarity. The potential of

organic modification of silica is reflected by the wealth of data demonstrating a

3

38 Chapter 3

multitude of applications [192-195]. Recently, the beneficial influence of various

polar organic functional groups which were built into the titania-silica matrix

with the aim of tuning the acidity of active sites and suppressing the acid-

catalyzed side reactions in demanding epoxidation reactions, has been

shown [196-198].

Esters and amines as organic functional groups can coordinate to Ti and

modify its acidity. In this work, various mono- and bidentate acetoxyalkyl and

aminoalkyl functions with similar structures were incorporated in the silica

matrix. Several functionalized trialkoxysilane precursors RSi(OMe)3 were

synthesized by varying the modifying group R. The performance of these

hybrid aerogels were tested in the epoxidation of cyclohexene and cyclo¬

hexenol.

Chiral surfaces are of great interest in the field of asymmetric catalysis, separa¬

tion of chiral compounds (e.g. chromatography), chemical sensor development

and nonlinear optical materials. One approach is to immobilize a chiral ligandwithin the catalyst framework [199-201]. Another promising way to create chiral

surfaces is the adsorption of chiral compounds onto a metal surface [178,202,203].

Chiral modification of titania-silica mixed oxides has not been reported yet and

represents an interesting challenge. As a mean for discriminating chiral surface

sites, vibrational circular dichroism (VCD)-IR spectroscopy was applied. This

technique measures the differential absorption of left- and right-circularly

polarized infrared light by chiral molecules [204,205]. Whereas enantiomers give

identical infrared spectra, their VCD spectra have different signs. Modulation

excitation (ME) attenuated total reflection (ATR)-IR spectroscopy was used in

our laboratory to investigate the adsorption behavior on a chiral silica sur¬

face [180]. By periodically changing the absolute configuration of the substrate,

specific signals could be detected, which was not the case for a correspondingachiral surface.

Organically Modified Aerogels 39

3.2 Experimental

3.2.1 Synthesis of Sol-Gel Precursors

3-Aminopropyltrimethoxysilane (Fluka, purum), A^,A^-dimethyl-3-aminopro-

pyltrimethoxysilane (Fluorochem. 95%), and 3-acetoxypropyltrimethoxysilane

(ABCR), were used for the sol-gel process without further purification. The

syntheses of other precursors were carried out under an argon atmosphere using

Schlenk-tube techniques. All chemicals were used as received unless otherwise

stated.

For the synthesis of (ethylenediaminopropyl)trimethoxysilane (EDAP-

TMOS), 30.5 ml (450 mmol) ethylenediamine (Fluka, >99.5%) and 16.5 ml

(90 mmol) (3-chloropropyl)trimethoxysilane (Aldrich 97%) were stirred in a

100 ml round bottom flask and refluxed at 100°C for 2 h. Two layers formed

after cooling to room temperature. The upper layer with the desired product

was distilled under reduced pressure (81°C/0.5 Torr) to yield 14.8 g

(66.5 mmol/73.8%) EDAP-TMOS.

(Propylenediaminopropyl)trimethoxysilane (PDAP-TMOS) was synthe¬

sized according to the procedure described above. Distillation (90°C/0.1 Torr)

yielded 14.2 g (60.4 mmol/67.1%) of the desired product. According to the

!H-, 13C- and 29Si-NMR spectra, two isomers of PDAP-TMOS could be

detected (Table 3-1). The ratio of the two isomers was 1:2.7, according to the

!H-NMR spectra.

Acylation of 3-buten-2-ol: 25.24 g (30.2 ml; 0.35 mol) 3-buten-2-ol

(Fluka, >97%) was dissolved in 50 ml CH2C12 (distilled over CaH2) and

degassed. Then 890 mg (7.28 mmol) A^A^-dimethyl-3-aminopyridine (Fluka,

>98%) and 42.5 g (0.42 mol) Et3N (Fluka, >99.5%) were added. The reaction

mixture was cooled to 0°C, and 43 g (0.42 mol) acetic anhydride (pract.,

distilled) was added dropwise. The solution was warmed to room temperature

and the reaction was monitored by TLC. After completion (16 h), purification

by distillation yielded 36.52 g (0.32 mol/91.4%) 3-buten-2-yl acetate.

Hydrosilylation of 3-buten-2-yl acetate: In a 50 ml Schlenk-tube 205 mg

(0.5 mmol) H2PtCl6 (Fluka) was degassed and the catalyst was covered with

11.4 g (0.1 mol) 3-buten-2-yl acetate. The reaction tube was protected against

40 Chapter 3

light by an aluminum foil. The reaction mixture was degassed and cooled to

-78°C in a dry ice/iPrOH bath. 14.6 g (0.12 mol) HSi(OMe)3 (Fluka, pract

-95%) was added dropwise with a syringe. The mixture was slowly allowed to

reach room temperature. The reaction was followed by IR spectroscopy. As

soon as the H-Si bond was not detectable, the reaction mixture was distilled

under reduced pressure. Distillation (60°C/0.3 Torr) yielded 14.9 g (63 mmol)

(3-acetoxybutyl)trimethoxysilane (AcOB-TMOS).

Acylation of 3-Buten-l,2-diol (Fluka, >99%) and subsequent hydrosilyla-

tion of the acylated precursor was performed as described above, yielding

18.07 g (61.4 mmol) (diacetoxybutyl)trimethoxysilane (DAcOB-TMOS).

Synthesis of Chiral Modifier:

Acylation of S-(-)-linalool (Fluka, purum) was performed as described on

page 39. Racemic (Acros, 95%) and S-linalyl acetate were hydrosilylated

according to the procedure described for 3-buten-2-yl acetate. Distillation in

vacuum (0.15 Torr; 72°C) yielded 20.41 g (64.1 mmol) (3-acetoxy-3,7-dime-

thyl-6-octenyl)trimethoxysilane as a colorless oil.

^-NMR: 5.08 - 5.02 (m, !H, C=CH-); 3.53 (s, 9H, -Si(OCH3)3); 1.93

(s, 3H, -OCOCH3); 1.93 - 1.63 (m, 4H); 1.63 (s, 3H, -CH(CH3)2); 1.56 (s,

3H, -CH(CH3)2); 1.35 (s, 3H, -C(OAc)CH3); 0.64 - 0.48 (m, 2H -CH2Si-);

13C-NMR: 170.2 (s, -OCO-); 131.6 (s, C=CH-); 123.9 (d, C=CH-); 85.1 (s,

-C(OAc)CH3); 50.5 (q, Si(OCH3)3); 37.4; 30.9; 25.6; 23.0; 22.3; 22.2; 17.5;

2.8 (td, V(29Si,13C) = 48, -CH2Si-);

29Si-NMR:-41.5;

Hydrogénation of (3-acetoxy-3,7-dimethyl-6-octenyl)trimethoxysilane:

10 g (31.4 mmol) of the substrate was dissolved in 15 ml hexane (Fluka; p.a.)

in a glass liner in the presence of 210 mg 5%Pd/C (Engelhard) catalyst. The

substrate was hydrogenated in a steel autoclave under H2 pressure (30 bar)

until completion of the hydrogénation could be monitored by H2 consump¬

tion. The suspension was filtered and the solvent was removed under reduced

pressure. Distillation in vacuum (0.15 Torr; 85°C) yielded 9.5 g (29.6 mmol/

94.4%) (3-acetoxy-3,7-dimethyloctanyl)trimethoxysilane as a colorless oil.

Organically Modified Aerogels 41

^-NMR: 3.54 (s, 9H, -Si(OCH3)3); 1.93 (s, 3H, -OCOCH3); 1.93 - 1.55

(m, 4H); 1.57 - 1.41 (m, 1H, -CH(CH3)2); 1.35 (s, 3H, -C(OAc)CH3); 1.28 -

1.10 (m, 4H); 0.83 (d, 3J = 6.6, 6H, -CH(CH3)2); 0.62 (m, 2H -CH2Si-);

13C-NMR: 170.2 (s, -OCO-); 85.1 (s, -C(OAc)CH3); 50.5 (q, Si(OCH3)3);

39.2; 37.7; 30.9; 27.8; 23.1; 22.5; 22.3; 21.3; 2.7 (td, 7(29Si,13C) = 48,

-CH2Si-);

29Si-NMR:-41.5;

3.2.2 Aerogel Synthesis

The aerogels were prepared according to procedures published previously [75].

Sol-gel processes were carried out in a glass reactor at room temperature under

an Ar atmosphere. For acetoxy-modified aerogels, prehydrolysis of the precur¬

sors in i-PrOH with aqueous HN03 as a hydrolyzing agent under vigorous

stirring (1000 rpm) lasted 6 h. The prehydrolysis was necessary to compensate

for the different sol-gel reactivities of the precursors [75,187]. Subsequently, tetra-

methoxysilane (TMOS; Fluka, puriss.) and titaniumbis(acetylacetonate)diiso-

propoxide (TIBADIP, 75% in iPrOH; Aldrich puriss.) in i-PrOH were added.

In all aerogels, the theoretical Ti02:Si02 mass ratio was 1:9, corresponding to a

Ti:Si =1:12 atomic ratio. After 24 h, trihexylamine (THA, Fluka >97%) in

i-PrOH was added and the stirring speed reduced (500 rpm). Gelation to an

opaque monolithic body occurred within 1 h. The total volume of the liquid

was ca. 170 ml and the corresponding molar ratios water : silicon dioxide : acid

: THA were 5 : 1 : 0.1 : 0.15. Different preparation conditions for the amine-

modified aerogels had to be chosen because the amine precursors themselves

act as base catalyst. A solution ofTMOS and TIBADIP in i-PrOH was mixed

with the acidic hydrolyzing agent. After 6 h, THA and the amine modifier in

i-PrOH were added and gelation occurred immediately. All gels were aged for

7 days.

Semicontinuous extraction with supercritical C02 was carried out at 40°C

and 230 bar. A glass liner was used to prevent contamination originating from

the steel autoclave. The as-prepared aerogel clumps were ground in a mortar

and calcined in a tubular reactor with upward flow at 100°C. All samples were

heated at a rate of 10°C min'1 in an air flow of 5 L min'1 and kept at the final

42 Chapter 3

temperature for 1 h. The calcination temperatures were chosen on the basis of

thermal analytical investigations. The composition of the samples with regard

to Si, Ti and Fe was determined by inductively coupled plasma atomic emission

spectroscopy (ICPAES). The Si to Ti ratio was nominal and the Fe-content was

below 0.01% (detection limit).

3.2.3 Thermal Analysis

Experiments were carried out on a Netzsch STA 409 thermoanalyzer. Gases

evolved during heating and/or injected into the system as a reference were

monitored on-line with a Balzers QMG 430 quadrupole mass spectrometer,

connected to the thermoanalyzer by a capillary heated to ca. 200°C. Details of

this method were described in a previous publication [93].

3.2.4 Nitrogen Physisorption

The specific surface area (SBET), mean cylindrical pore diameter (d ) and

specific desorption pore volume (V (N2)), assessed by the BJH method, were

determined by nitrogen physisorption at -196°C using a Micromeritics ASAP

2000 instrument. Prior to measurement, the sample was degassed at 100°C

until a final constant pressure below 0.1 Pa was achieved. BET surface area was

calculated in a relative pressure range between 0.05 and 0.2, assuming a cross

sectional area of 0.162 nm for the nitrogen molecule. Pore size distribution

was calculated applying the BJH method to the desorption branch of the

isotherm [206]. The fractal dimensions of the surface structure were calculated

according to Jarzebski et al. [207].

3.2.5 Nuclear Magnetic Resonance (NMR)

NMR experiments were performed on a Bruker AMX 400 WB spectrometer.

13C NMR (CP-MAS) were performed with Dl=3 s, P15=l ms and P3=5.5 ms.

For the deconvolution of the T2, T3, Q2, Q3, and Q4 CP/MAS signals, starting

values of-56 ppm, -64 ppm, -92 ppm, -100 ppm and -109 ppm, respectively,

were chosen. The peak positions and the width of the peaks were not fixed.

Organically Modified Aerogels 43

Gaussian function was chosen for the fitting and the least squares method for

optimization. To prove that the chiral recognition was still present in the

precursor, NMR experiments were performed using Eu(hfc)3 as a chiral shift

reagent.

3.2.6 Electron Microscopy

For TEM and SEM investigation, the sample was crushed and deposited on a

holey carbon foil supported by a copper grid. The Philips CM30 microscope,

operated at 300 kV, was equipped with a Supertwin lens (cs =1.2 mm, point

resolution < 0.2 nm).

3.2.7 Vibrational Circular Dichroism (VCD)

VCD spectra were measured using a Bruker PMA 37 accessory coupled to a

IFS/66 Fourier transform infrared spectrometer. The infrared beam from the

spectrometer is polarized by a wire grid polarizer. The linearly polarized light is

alternately switched at 50 kHz between left- and right-handed circular polari¬

zation by a photoelastic modulator (Hinds PEM 90) set at 1/4 retardation.

More detailed information can be found elsewhere [208].

3.2.8 Epoxidation Procedure

2-Cyclohexen-1-ol (Fluka, ca 97%), cyclohexene (Fluka, > 99.5%) and tert-

butyl hydroperoxide (TBHP, Fluka, ca. 5.5 M solution in nonane, stored over

molecular sieve 4 A) were used as received. Toluene (Riedel-de Haën, >99.7%)

was distilled from sodium and stored over molecular sieve 4 A.

In a 50 ml round bottom flask equipped with reflux condenser and ther¬

mometer, 70 mg catalyst was heated to 100°C under a Ar stream for 2 h. After

cooling to room temperature, a solution of 2 ml toluene, 20 mmol olefin and

0.4 g hexadecane as internal standard were added. At 90°C, the reaction was

started by addition of 5 mmol TBHP. The reaction was carried out under a Ar

atmosphere to prevent contamination with oxygen or moisture. Samples were

analyzed by a HP 6890 gas Chromatograph (cool on-column injection,

AA Chapter 3

HP-FFAP column). In all experiments, the reactant : peroxide molar ratio

was 4:1. Olefin and peroxide selectivities are defined as follows:

•selectivity of the epoxide related to the olefin converted,

S , p= 100%

[£P°xide]. .

olefin [olefin]0- [olefin] (1)

•selectivity of the epoxide related to the peroxide converted,

S •, = 100% [£P°xide],0n

peroxide [peroxide]0- [peroxide] (2)

•Initial rate was determined by measuring the epoxide yield after 5 min.

3.3 Results

3.3.1 Structural Properties

For the synthesis of organically modified titania-silica aerogels, the original

sol-gel process had to be adjusted to ensure the desired properties: (i) well

dispersed Ti in the silica matrix, (ii) mesoporous structure providing access for

bulky reactants to the active sites, and (iii) covalently anchored organic func¬

tional groups stable under the conditions of catalyst preparation and epoxida¬

tion reaction.

Some structural properties of the calcined aerogels derived from nitrogen

physisorption measurements are listed in Table 3-1. The characteristics of

unmodified aerogel (Ae) and three hybrid aerogels modified by monodentate

ligands (AP-Ae, DMAP-Ae, AcOP-Ae) have already been published [196-198].

These results are given here for comparison. All modified aerogels posses signi¬

ficantly lower surface areas and pore volumes than the unmodified titania-silica

aerogel. The acetoxy modified aerogels have lower surface areas than the amine

modified samples whose divergence may partly be due to the different modifier

concentration (10% or 5% of the Si precursors contained an aminoalkyl or an

Organically Modified Aerogels 45

acetoxyalkyl functional group, respectively). AcOB-Ae had the lowest BET-

surface area and pore volume and exhibited a bimodal pore size distribution.

The pore size distribution of all other aerogels were monomodal and indicated

a mesoporous structure.

Table 3-1 : Textural properties of aerogels.

Aerogel Modifier WmV] VpMcmV1] dmax>m]

unmodified (Ae) Si-OH (no modifier) 813 2.3 62

AP-Ae 347 1.38 31

DMAP-Ae 324 0.84 11

EDAP-Ae -Sk

~NH2372 1.58 80

PDAP-Ae

--Sk

--Sk

"NH2

NH,

436 1.38 32

AcOP-Ae „OAc 251 1.61 65

AcOB-Ae

DAcOB-Ae

-Sk172

OAc

"oac 251

OAc

0.52

2.02

2.4, 67

62

a Total volume of the pores with a diameter between 1.7 and 300 nm.

The graphically determined maximum of the pore size distribution.

A6 Chapter 3

Thermal stability of the mixed oxides was limited by the decomposition

behavior of the organic modifier. Thermal analysis indicated a significant

decomposition already at 150°C. For example, during heating PDAP-Ae up to

800°C several organic fragments evolved which were detected by MS. Interpre¬

tation of these fragments showed that carbon dioxide, water, solvent residue

from catalyst synthesis and fragments containing amino groups evolved.

Accordingly, the aerogels were calcined at 100°C and the in situ re-dryingbefore epoxidation was also performed at 100°C.

X-ray diffraction analysis confirmed the amorphous structure of the aero¬

gels. In one case (PDAP-Ae) the sample was heated up to 850°C, but still no

crystalline phase could be detected.

Transmission electron microscopy corroborated the amorphous structure

of organically modified titania-silica (Fig. 3-1, bottom). Scanning electron

microscopy (Fig. 3-1, top) revealed that the particle size of amino modified

aerogels is smaller than that of acetoxy modified catalysts [197]. This variation

may be due to different conditions in the gelation process.

Preservation of the structure of covalently bound modifying groups after

synthesis was confirmed by C-NMR. An example is shown in Fig. 3-2.

Although solid state NMR and solution NMR are not easy to compare, a

certain correlation between the observed signals in the aerogels and in the

precursor Si-compounds can be made. Despite the very broad signals of the

solid state NMR experiment, the peaks found for the precursor in solution

NMR can be traced to the 13C-NMR spectra of the aerogel.The bulk structure of aerogels was elucidated using Si-NMR, (no cross-

polarization applied). The NMR examination is based on the variation of the

pulse time for 29Si-CP-MAS experiments [76,209]. The quantitative values for

the deconvoluted signals, T2, T3, Q2, Q3 and Q are listed in Table 3-2. Qn

denotes a 29Si nucleus with a (Si(OSi)n(OX)4,n) local environment, i.e.,

n oxygen bridges to neighboring silicon nuclei. The remaining (4-n) coordina¬

tion sites are occupied either by hydroxyl groups (X=H) or bridges to Ti(IV)-

centers (X=Ti). Tn denotes a corresponding Si nucleus where one Si-O-Si is

substituted by an organic group R. All modified aerogels showed considerably

lower degree of crosslinking, expressed as Q /Q ,than the unmodified aerogel.

Organically Modified Aerogels A7

Fig. 3-1 : Scanning (a) and transmission (b) electron micrographs of PDAP-Ae. Inset in

(a): magnification of white box (horizontal length corresponds to 10 urn), inset in (b): elec¬

tron difftactogram.

48 Chapter 3

—i—i—i—i—i—|—i—i—i—i—|—i—i—i—i—|—i—i—i—i—|—i—i—i—i—|—i—i—i—i—|—i—i—i—i—|—i—i—i—i—|—i—i—i—i—|—i—i—i—i—|—r

100 80 60 40 20 0

ppm

Fig. 3-2: 13C-CPMAS NMR spectra of PDAP-Ae (top) and the 13C-NMR spectra of the

corresponding modifier precursor (propylenediaminopropyl)trimethoxysilane (bottom).

Again, the acetoxy modified aerogels, with the exception of DAcOB-Ae, show a

lower degree of crosslinking than the amino modified samples.

The presence of an organic function covalently bound to the Si nuclei has

been confirmed by the T2 and T3 signals in the spectra [T2 = R-Si(OX)(OSi)2,

T = R-Si(OSi)3]. The cumulative values for acetoxy-alkyl modified aerogelsvaried in the range 13-15% (Table 3-2), compared to the nominal value of

10%. In the case of aminoalkyl-modified aerogels the T sites were not discern¬

ible, probably due to the low degree of functionalization (5%).

Organically Modified Aerogels 49

Table 3-2 : iJSi-CP-MAS NMR signals of hybrid aerogels.

Aerogel T2 [%] T3 [%] Q2 [%] Q3 [%] Q4 [%] Q4/Q3

unmodified (Ae) - - 4 24 72 3.0

AP-Ae n.d.a n.d. 11 36 53 1.47

DMAP-Ae n.d. n.d. 5 37 A7 1.53

EDAP-Ae n.d. n.d. 4 36 60 1.67

PDAP-Ae n.d. n.d. 11 37 52 1.41

AcOP-Ae 3 10 4 43 40 0.93

AcOB-Ae 6 7 8 AA 35 0.79

DAcOB-Ae 3 12 6 28 51 1.83

a n.d. = not determined.

3.3.2 Catalytic Properties

All aerogels have been tested in the epoxidation of cyclohexene and cyclohexe¬

nol using TBHP as oxidant. Epoxidation of cyclohexene is a facile reaction

characterized by high rates and selectivities, as illustrated in Fig. 3-3 and

Table 3-3. Accordingly, the influence of organic modification on the epoxide

Table 3-3: Epoxidation of cyclohexene.

Aerogel Initial rate

[^mol m 2min :]

TBHP conversiona

[%]

° olefinsb

peroxide

unmodified (Ae) 7.3 88.3 96.5 94.2

AP-Ae 19.1 97.1 96.6 91.2

DMAP-Ae 18.2 96.4 97.0 ^100

EDAP-Ae 13.8 95.6 95.9 88.7

PDAP-Ae 18.3 96.8 97.4 97.3

AcOP-Ae 32.2 ^100 97.8 92.4

AcOB-Ae 21.2 92.7 97.3 89.1

DAcOB-Ae 31.3 98.6 97.5 94.6

aConversion at 60 mm.

Selectivities determined at 80% conversion.

50 Chapter 3

0 10 20 30 40 50 60 70

Time [min]

Fig. 3-3: Formation of cyclohexene oxide on various titania-silica aerogels as a function of

reaction time, (for conditions see page 39).

selectivities related to the olefin or peroxide consumed, is small and no clear

correlation with the structure of organic modification can be established. The

effect of organic modification is more substantial when considering the epoxi¬

dation activity of the aerogels. All hybrid aerogels, in particular the acetoxy-

alkyl-modified materials, are more active than the reference unmodified

aerogel. The initial rates related to unit surface area increases by a factor of 1.9

to 4.4 in the presence of polar organic functional groups (Table 3-3).

Epoxidation of 2-cyclohexen-1-ol is disturbed by acid-catalyzed side reac¬

tions including dimerization and oligomerization of the substrate as depicted

in Scheme 3-1 [210]. When using the unmodified aerogel (Ae) only about 70%

of the substrate was converted to the desired epoxide at 80% TBHP conversion

(Table 3-4). All hybrid aerogels were more selective, reaching up to 91% olefin

CD

•!-H

XOOhpq

5-

3-

2-

0

Organically Modified Aerogels 51

oligomers

Scheme 3-1 : Main and side reactions during epoxidation of cyclohexenol.

selectivity with the PDAP-Ae sample. It seems that acetoxy modified aerogelsshow slightly higher selectivities than amino modified materials, with the

exception of PDAP-Ae. The organically modified aerogels were also more

active than the unmodified aerogel: the initial rate increased by a factor of 2.2 -

5.7 in the presence of aminoalkyl or acetoxyalkyl functional groups. Note that

the kinetic curves in Figs. 3-3 and 3-4 show experiments carried out with the

same amount of catalyst and the differences in surface area are reflected in the

specific initial rates listed in Tables 3-3 and 3-4.

Table 3-4: Epoxidation of cyclohex-2-en-l-ol.

Aerogel Initial rate

[^mol m min ]

TBHP conversiona

[%]

^ olefinsb

peroxide

unmodified (Ae) 6.1 75.2 70.7 74.9

AP-Ae 17.2 93.2 76.7 73.4

DMAP-Ae 13.4 85.0 80.4 76.1

EDAP-Ae 17.0 83.0 80.9 74.5

PDAP-Ae 13.4 87.6 91.1 80.6

AcOP-Ae 32.6 97.9 90.4 80.1

AcOB-Ae 29.3 90.7 83.8 78.5

DAcOB-Ae 34.6 ^100 84.9 77.A

aConversion at 60 mm.

Selectivities determined at 80% conversion.

52 Chapter 3

0 10 20 30 40 50 60 70

Time [min]

Fig. 3-4: Formation of epoxide in the oxidation of 2-cyclohexen-l-ol on various titania-

silica aerogels as a function of reaction time, (for conditions see page 39).

Chiral Modification

NMR experiments with the chiral shift reagent Eu(hfc)3 revealed different

spectra for racemic and S-(-)-precursors. Typically, the two diastereotopic

methyl groups at the 6-position split up into two signals for the enantiomeri-

cally pure modifier, while for the racemic precursor four signals occurred upon

addition of shift reagent. The textural data for the two mixed oxides by

N2physisorption are listed in Table 3-5. In thermal analysis experiments, typi¬

cal fragments occurred upon heating above 150°C which were also detectable

when the corresponding experiments were performed with the precursor.

3-

CD

• i—t

Xo

2-

1-

0-

Organically Modified Aerogels 53

Table 3-5: Textural properties of aerogels.

Aerogel Sbet NY1] VpMcmV1] dmJM

S-Aerogel

rac-Aerogel

384

406

1.92

1.91

37 (19)

40 (21)

a Total volume of the pores with a diameter between 1.7 and 300 nm.

graphically determined maximum of the pore size distribution.

Unfortunately, no specific signals could be detected by VCD-IR transmission

measurements. Also, modulation experiments, where the absolute configura¬tion of ethyl lactate was periodically changed, revealed no difference in the

adsorption spectra for (S)-modified and racemic titania-silica aerogel.

3.4 Discussion

The results presented in Tables 3-3 and 3-4, and Figs. 3-3 and 3-4 demonstrate

that organic modification of titania-silica aerogels has a considerable positive

influence on the rate of epoxidation reactions, whereas the epoxide selectivity

related to the olefin consumed is enhanced only in the more demanding reac¬

tion, the epoxidation of cyclohexenol. In the epoxidation of cyclohexene and

cyclohex-2-en-1-ol the selectivities related to the peroxide consumed showed

no clear trend, except a slight enhancement with acetoxy modified aerogels in

the latter reaction. The lack of influence on the peroxide selectivity observed

for most modifiers suggests that the modifying group does not influence the

peroxide decomposition or the peroxide consumption by undesired oxidation

reactions.

The most important question to be answered is why and how organic

modification enhances the activity and selectivity of titania-silica epoxidation

catalysts. There are probably several reasons for the beneficial influence of

acetoxyalkyl and aminoalkyl functions. A likely explanation for the better

performance of hybrid aerogels is the interaction of the amino and acetoxy

54 Chapter 3

groups as electron donor ligands with some of the Ti sites. Another feasible

effect is the change in polarity of the aerogel by introducing an organic func¬

tional group in the silica matrix. This effect is understandable by assuming that

the strongly polar (but weakly acidic [211]) surface Si-OH groups are replaced

by acetoxyalkyl or aminoalkyl moieties. Furthermore, the organic functional

group can interact with another surface silanol group via hydrogen bonding

(Scheme 3-2). This interaction reduces the polarity of the hybrid aerogel and

eliminates the possible interaction of this silanol group with the substrate or

product. A recent evidence for this assumption is the remarkably reduced

polarity of silica after organic modification with aminopropyl groups [212]. For

comparison, organic modification by phenyl or aminopropyl groups decreased

the polarity expressed as E Tfrom 0.967 (unmodified silica) to 0.50 and 0.55,

respectively.

Scheme 3-2 : Reduction ofthe surface polarity via a hydrogen-bonding interaction between

an acetoxyalkyl functional group and a surface silanol group.

The good epoxidation activity and selectivity of titania-silica mixed oxides is

attributed to the presence of isolated Ti sites. To achieve a good distribution of

Ti in the silica matrix at the atomic level, it is crucial to compensate the strik¬

ingly different sol-gel reactivities ofTi and Si precursors. The considerable vari-

Organically Modified Aerogels 55

ation in the key structural properties of the aerogels (surface area, pore

structure, cross linking, Tables 3-1 and 3-2) indicates that this compensation

was only partially successful. Variation in the structural properties of the aero¬

gels with the chemical structure of the organic modifying group raises difficul¬

ties in answering the intriguing question whether modification of titania-silica

with bidentate ligands is a more efficient tool to improve the catalytic proper¬

ties than using monodentate aminoalkyl or acetoxyalkyl functions. The

epoxide selectivities related to cyclohexenol consumed seem to increase by

introducing a second amino group. The two bidentate-modified aerogels,EDAP-Ae and PDAP-Ae, show enhanced olefin selectivities compared to the

monodentate-modified aerogels AP-Ae and DMAP-Ae. The same effect could

not be observed for acetoxy modified aerogels. As for the activity of bidentate

modifiers, there is no clear trend observable.

Chiral Modification

NMR experiments using shift reagent revealed that the stereochemistry was

preserved during the synthesis steps of the chiral modifier. Despite the evidence

of the modifier covalently bound to the aerogel, traced by thermoanalysis, a

proof of chiral recognition on the catalyst surface could neither be obtained by

VCD-IR spectroscopy nor by adsorption experiments using in situ ME

ATR-IR spectroscopy. The reason for this behavior is not clear. It might be due

to the size of the chiral modifier, which allows the functional group to appear

in different conformations possibly resulting in cancellation the chiral informa¬

tion. Another explanation for the lack of evidence in the adsorption experi¬

ments could be the large hydrophobic substituent at the stereogenic center

which inhibits a proper interaction with the lactate.

3.5 Conclusions

Amino and acetoxy groups were incorporated into the silica matrix of titania-

silica aerogels using the corresponding modified R-Si(OMe)3 precursors,

tetramethoxysilane and titaniumbis(acetylacetonate)diisopropoxide in a sol-gel

56 Chapter 3

process. Supercritical extraction of gel afforded amorphous mesoporous aero¬

gels. Covalent incorporation of the organic functions was confirmed by NMR

analysis. All modified aerogels were considerably more active than the reference

purely inorganic mixed oxide. Organic functionalization of titania-silica

suppressed the acid-catalyzed side reactions in the epoxidation of cyclo-

hex-2-en-l-ol and in the best case 91% epoxide selectivity at 80% TBHP

conversion could be achieved. The positive effect of polar functional groups

may be attributed to a Lewis base - Lewis acid type interaction between the

"ligand" and a Ti active site, or to suppressed polarity of the hybrid aerogels.An unambiguous differentiation between the influence of mono- and bidentate

functional groups requires further effort.

Chapter

Epoxidation on Titania-Silica Aerogel Catalysts

Studied by Attenuated Total Reflection Fourier

Transform Infrared and Modulation Spectroscopy

4.1 Introduction

Catalysts based on Ti and Si belong to the most powerful heterogeneous epoxi¬

dation catalyst known today [185,213]. Various structural variants of these cata¬

lysts are applied, ranging from crystalline titanium-substituted molecular sieves

(TS-1 [43], TS-2 [214], Ti-ZSM-48 [215], Ti-Beta [52], TAPO-5 [216], Ti-MWW

[217]) to amorphous titania-silica aerogels. The latter catalysts possess relatively

large pores and have been designed to overcome the rigid structural constraints

of the molecular sieves which render them unsuitable for application in epoxi¬

dation of bulky olefins [40,129].

Extensive work has been devoted to the determination of the nature of

active sites in crystalline Ti-based catalyst. Elucidating the coordination state of

the Ti active sites was a primary goal in several investigations. The framework

vibrations observed in IR and Raman spectra [107,110,135,218], UV measure¬

ments [106,135], computational studies [110,112] and X-Ray absorption near-edge

structure (XANES) investigations [106-108] revealed a coordination number of

n=4 for the most intensively examined molecular sieve TS-1. For the titania-

silica aerogels the preparation method has a strong influence on the structure of

the Ti centers and the incorporation ofTi within the silica matrix [85]. Because

of the amorphous nature of these catalysts, it is more difficult to gain unam¬

biguous information on the structure of the active sites. Yet, for aerogels with a

low Ti content and well-dispersed Ti centers, spectroscopic measurements

revealed mainly tetrahedral coordination for Ti(IV)-ions, which replace a

4

60 Chapter 4

Si-atom in the matrix [68,70,85,186,219]. However, during the epoxidation reac¬

tion the Ti coordination may change from fourfold to five- or sixfold coordina¬

tion [19]. It is therefore of primary importance to characterize the catalyst under

working conditions. This requires the study of the catalytic solid-liquid inter¬

face, which is experimentally quite demanding because the applied method has

to be sensitive enough to probe the interface in the presence of a solvent and

many other species, including reactants, intermediates, products and spectator

species (not involved in the catalytic cycle).

Knowledge of the process occurring at the catalytic solid/liquid interface is

an important prerequisite for a rational design of improved epoxidation cata¬

lysts. With this in mind, we have investigated the dynamic processes occurring

at the solid/liquid interface during epoxidation of cyclohexene on titania-silica

aerogels using a combination of attenuated total reflection Fourier transform

infrared (ATR-IR) [162] and modulation [163,164] spectroscopy. Upon internal

reflection of an infrared beam in ATR an evanescent field is forming, which

probes a small volume near the internal reflection element. A heterogeneous

catalyst within this volume can thus be investigated, without exceedingly

strong absorptions from the bulk solvent. In our group this technique has been

used previously for the study of thin film model catalysts [177,178,220] as well as

supported metal catalysts [164,221]. Modulation spectroscopy takes advantage of

phase-sensitive detection (PSD) of periodically varying signals and drastically

increases the sensitivity [163,164].

4.2 Experimental

4.2.1 Preparation of Catalyst Layer

The aerogels were prepared according to reported procedures [93,222]. Detailed

information on aerogel preparation using the sol-gel process and semiconti-

nuous extraction with supercritical C02 can be found in chapter 3 on page 41.

The as-prepared aerogel clumps were ground in a mortar and calcined in a

tubular reactor with upward flow at 400°C. Titania-silica aerogels with 0, 10

ATR Studies on Epoxidation of Cyclohexene 61

and 20 wt% nominal Ti02 were prepared, designated OTi, 1 OTi and 20Ti,

respectively. Also, an aerogel modified with covalently bound methyl groups

was synthesized, designated 1 OTi-Me.

500 mg of calcined aerogel was added to 10 ml i-PrOH and vigorouslystirred for 2 h. The resulting slurry was dropped onto the surface of a ZnSe

internal reflection element (IRE, 45°, 50 x 20 x 2 mm, KOMLAS). The IRE

was subsequently kept at 500 Torr and 40°C for 3 h to allow evaporation of the

solvent. This procedure was repeated to gain a uniform catalyst layer on the

ZnSe IRE. Prior to the experiments, the IRE was kept at 5 Torr and 80°C for

2 h to remove adsorbed water and solvent.

4.2.2 Nitrogen Physisorption

The specific surface area (SBET), mean cylindrical pore diameter (<dp>) and

specific desorption pore volume (V (N2)), assessed by the BJH method, were

determined by nitrogen physisorption at -196°C using a Micromeritics ASAP

2000 instrument. Further details are described in chapter 3 on page 42. The

textural properties of the different aerogels are listed in Table 4-1.

Table 4-1 : Textural properties of aerogels.

Aerogel" SBET [m2gl] VPb [cm3gl] dmaxc [nm]

OTi (Si02)lOTi

20Ti

10Ti-Me

a

Acronyms denote OTi, pure silica aerogel, lOTi and 20Ti, titania-silica aerogels containing 10 and 20 wt%

nominal Ti02, 10Ti-Me, methylated aerogel lOTi, 10 mol% of the total Si content is modified with a

covalently bound methyl group

designates the BJH cumulative desorption pore volume of pores in the maximum range 1 7 - 300 nm diameter

<dp> 4V /SgET, in parentheses the graphically assessed maximum of the pore size distribution, derived from the

desorption branch, is given

1102 3.77 17 (35)

768 2.21 16 (36)

761 1.46 13 (96)

635 2.08 18 (60)

62 Chapter 4

4.2.3 ATR Spectroscopy

The heatable ATR flow-through cell used for the investigations was previously

reported [164]. Experiments were started after stable conditions were achieved at

the desired temperature. The experimental setup has already been described in

chapter 2 on page 32.

4.2.4 Modulation Spectroscopy

Modulation excitation (ME) spectroscopy has been applied to increase sensi¬

tivity and to help disentangle crowded spectra. The advantages of this method

have been described on chapter 1 on page 23. More detailed information of the

technique can be found elsewhere [163,164,180,223].

4.2.5 Adsorption Experiments

For adsorption and epoxidation experiments, 60 single beam spectra were

recorded during one modulation period by averaging several scans per singlebeam spectrum at a rate of about 10 spectra/s. Data acquisition was started

after two full modulation periods. During this initial time the system relaxes to

a quasi-stationary state, around which it oscillates at frequency ca. Data were

typically averaged over six modulation periods. Before demodulation the singlebeam spectra were transformed into absorbance spectra using the average of all

single beam spectra as the background. However the resulting demodulated

spectra are independent of the choice of the reference spectrum. Eighteen

phase-resolved absorbance spectra were calculated by varying the phase angle in

10° steps between 0 and 180°. Note that phase-resolved absorbance spectra

differing by 180° are identical but for the sign of the absorbance signals.

Adsorption experiments were carried out at room temperature. Neat cyclo-

hexane (Fluka, puriss p.a.) from the first tank was allowed to flow through the

cell for 20 - 30 min before collecting a background spectrum. Subsequently the

valve was switched manually and the solution from the second tank, containing

the substrate, was allowed to enter the cell. Spectra were collected by co-adding200 interferograms per spectrum until saturation of the catalyst surface was

ATR Studies on Epoxidation of Cyclohexene 63

observed. The valve was then switched to the former position and spectra were

recorded during removal of the substrate. Neat cyclohexane was subsequently

allowed to flow through the cell for 5 min before modulation experiments were

started. Measurement of the time-resolved spectra was synchronized with the

concentration modulation by switching the computer-controlled valve within

the data acquisition loop. One modulation period lasted T = 244 s.

4.2.6 Epoxidation Experiments

Epoxidation experiments were carried out at 70°C in cyclohexane (Fluka,

puriss. p.a.) solvent. Tank 1 contained a solution of cyclohexene (Fluka,

purum) in cyclohexane and £-butyl hydroperoxide (TBHP; Fluka, purum,

-5.5M in nonane) in cyclohexane, respectively, and tank 2 a mixture of the two

reactants. Note that in this way the concentration of one reactant (cyclohexene

and TBHB, respectively) was kept constant during modulation, whereas the

concentration of the other reactant was modulated. The catalyst used in these

experiments was lOTi-Me (Table 4-1) containing 10 wt% nominal Ti02 and

modified by 10mol% covalently bound methyl groups. In epoxidation

experiments [93] this catalyst showed the highest activity. Data acquisition was

performed as described above for the adsorption experiments with a modula¬

tion period T = 299 s. Within one modulation period typically twelve samples

of the reaction solution were collected after the cell. The samples were subse¬

quently analyzed using a HP 6890 gas Chromatograph (cool on-column injec¬

tion, HP-FFAP column, 30 m x 0.32 mm x 0.25 um).

4.2.7 Variable Temperature Experiments

Variable temperature measurements were carried out using the single-beam-

sample-reference (SBSR) technique [223,224]. In the SBSR method the horizon¬

tal position of the ATR-cell is changed by means of a lift. The ATR cell consists

of two compartments. The upper one is used for the sample and the lower for

the reference. In this way sample and reference are recorded quasi-simulta-

neously and have the same history, such as temperature program. Signals aris-

6A Chapter 4

ing due to dissimilar sample and reference history are thus efficiently

eliminated.

A cyclohexane solution of TBHP or the reaction mixture was allowed to

flow through the upper compartment of the ATR cell, while neat cyclohexane

was flown through the lower part, used as reference. Both on the sample and

reference side of the IRE, a layer of 1 OTi-Me catalyst was deposited. Several

SBSR spectra were recorded at the same temperature. The reported spectra

were obtained by subtraction of the SBSR spectrum, which was recorded just

before heating the cell to the next higher temperature. At each temperature,

samples of the reaction solution were collected after the cell and analyzed using

a HP 6890 gas Chromatograph (cool on-column injection, HP-FFAP column,

30 m x 0.32 mm x 0.25 um).

4.2.8 Theoretical Calculations

In order to better understand the infrared spectra of reactants (TBHP and

cyclohexene) and product (cyclohexene oxide) theoretical infrared spectra were

calculated using quantum chemical methods. All calculations were performed

at the b3pw91 level (hybrid density functional method) using a 6-31G(d,p)

basis set. Prior to the calculation of the infrared spectra based on a normal

modes analysis, the structure of the molecules was optimized by relaxing all

internal degrees of freedom. Frequencies were scaled by a factor of 0.96. All

calculations were performed using GAUSSIAN98 [225].

4.3 Results

4.3.1 Adsorption Experiments

Figure 4-1 shows ATR spectra of dissolved cyclohexene, TBHP (5.5M in

nonane), and cyclohexene oxide (-10 mmol/1 each) in cyclohexane measured

over the uncoated internal reflection element (IRE) in the absence of catalyst.

Also shown are the calculated spectra of the corresponding molecule. The

ATR Studies on Epoxidation of Cyclohexene 65

1300 1200 1100 1000 900 800 700

Wavenumber / cm1

Fig. 4-1 : Demodulated ATR spectra ofreactants and products on blank internal reflection

element (IRE) at room temperature (top) and calculated (bottom): a) f-butyl hydroperoxide

(TBHP); b) cyclohexene oxide; c) cyclohexene. The concentration was modulated between 0

and -10 mmol/1 in cyclohexane.

agreement between measured and calculated spectra is good enough to warrant

a reliable assignment of the bands.

Figure 4-2 shows phase-resolved ATR spectra recorded when modulatingthe cyclohexene concentration between 0 and 3 mmol/1 at room temperature

over the titania-silica aerogel (lOTi-Me). Prominent signals at 3024, 2945,

918, 876 and 719 cm' are revealed. These signals were also observed in the

modulation experiments at 70°C (see p. 69) and are due to cyclohexene, likely

dissolved. The bands are generally very weak. The spectra do not reveal any

66 Chapter 4

3600 3400 3200 3000 1100 900 700

Wavenumber / cm-1

Fig. 4-2: Demodulated ATR spectrum of adsorption of cyclohexene at room temperature

on methyl-modified titania-silica aerogel (lOTi-Me). The concentration of cyclohexene was

modulated between 0 and 3 mmol/1 in cyclohexane.

strong interaction between the aerogel and cyclohexene, however, a weak inter¬

action cannot be excluded.

On the other hand adsorption ofTBHP at room temperature on the same

catalyst is clearly indicated by the phase-resolved ATR spectra depicted in

Fig. 4-3. When TBHP is admitted relatively strong bands appear at 2983,

1389, 1367, 1250 (broad), 1191, 844 and 747 cm'1 which are assigned to

adsorbed TBHP. Correlated to the occurrence of these signals the silanol band

at 3700 cm' disappears and a broad band with a maximum at about 3348 cm'

appears. The latter is assigned to hydrogen-bonded silanol groups and the

hydrogen bonded O-H group of the peroxide. Note that in general the signals

arising upon TBHP adsorption are much larger than in the case of cyclohexene

(Fig. 4-2). The strong broad signals around 1020 and 944 cm'1 fall within the

aerogel framework vibration.

In the spectral range between 700 and 1300 cm' several absorptions are

expected for the aerogel: At about 800 cm'1 the symmetric v(Si-O-Si) stretch¬

ing vibration is expected, at 930-950 cm'1 the v(Ti-O-Si) stretching vibration

ATR Studies on Epoxidation of Cyclohexene 67

Fig. 4-3: Demodulated ATR spectra of adsorption of TBHP at room temperature on

methyl-modified titania-silica aerogel (lOTi-Me) (top) and corresponding non-modified cat¬

alyst lOTi (bottom). The concentration ofTBHP was modulated between 0 and 3 mmol/1

in cyclohexane.

and at 1080-1105 cm'1 and at about 1200 cm'1 the asymmetric v(Si-O-Si)

stretching vibrations. Furthermore, at 980 cm'1 the Si-OH vibrations absorb.

The ATR spectrum of the dry catalyst shown in Fig. 4-4 reveals the bands

discussed above and it additionally shows the band of the surface silanols at

3745 cm' as well as the CH stretching signals of the methyl groups incorpo¬

rated in the catalyst at 2981 cm'.The spectrum obtained for the skeletal vibra¬

tions is in good agreement with the results reported for The Ti02/Si02 mixed

oxides by Schraml et al. [85]. A maximum could be detected at 1075 cm' with

a shoulder at 1200 cm',which can be assigned to the v(Si-O-Si) asymmetric

68 Chapter 4

3600 3400 3200 3000 1300 1100

Wavenumber / cm1

900 700

Fig. 4-4: Spectrum ofthe dry modified titania-silica (lOTi-Me) recorded with SBSR tech¬

nique at room temperature. The spectrum is shown in the relevant regions for the catalyst: a)

between 3775 and 2825 cm"1; b) between 1375 and 625 cm"1.

stretching vibration. At 947 cm'1 a signal of the v(Ti-O-Si) asymmetric stretch¬

ing vibration could be detected. (Note that for TS-1 this vibration was found at

960 cm'1 [107,110,135,218]) Probably overlapping with this band is the Si-OH

signal, which is normally observed at 980 cm'1. The v(Si-O-Si) symmetric

stretching vibration was detected at 810 cm'.

Figure 4-5 shows phase-resolved ATR spectra in the 800-1300 cm'1 spec¬

tral range recorded during TBHP concentration modulation (between 0 and

3 mmol/1) experiments for three catalysts differing in the Ti content (OTi, 1 OTi

and 20Ti; Table 4-1). A negative band is observed at 980 cm'1 due to Si-OH

groups. This is consistent with Fig. 4-3, which shows the disappearance of

Si-OH groups upon adsorption of TBHP. On the other hand positive bands

are observed at about 1014 and 944 cm'.The latter bands reflect changes in

the Si-O vibrations upon adsorption of TBHP. Figure 4-5 reflects significant

changes in this region with increasing Ti content. An additional signal is

ATR Studies on Epoxidation of Cyclohexene 69

observed at about 930 cm',which increases with Ti content and is absent on

the pure Si aerogel (OTi). Between the band at 944 cm',which is also observed

on the Si aerogel and the signal at 930 cm',which has to be associated with Ti,

a significant phase lag is observed. In the region of the band at 1250 cm' asso¬

ciated with the C-C and C-O vibrations of the peroxide (not resolved clearly,

resulting in a broad signal) an additional band at 1260 cm' becomes more

significant with increasing Ti content. This band also shows a significant phase

lag with respect to the band at 1250 cm'.These phase lags could not be

detected in the adsorption experiments with the methyl-modified aerogel

(10Ti-Me) at room temperature (s. Fig. 4-3, top).

4.3.2 Variable Temperature Experiments

Figure 4-6 shows ATR spectra recorded at different temperatures using the

SBSR method. In the sample channel cyclohexene and TBHP (3.5 mmol/1

each) were flown over the methyl-modified aerogel (lOTi-Me). In the reference

channel neat solvent was flown over the sample. By increasing the temperature

of the ATR-cell, a change in the area of the framework vibrations could be

detected. Signals at 1112, 1030, 942, 902, 857, 805 and 703 cm'1 evolved

with temperature. Note that the shape of the evolving spectra is different from

the spectrum of the catalyst (Fig. 4-4) such that the observation cannot simply

be explained for example by the partial removal of the catalyst layer. The

changes in the framework spectrum become significant at 60-65°C. This is also

the temperature at which cyclohexane oxide product could be detected in the

samples of the reaction mixture collected at the outlet of the ATR cell. The

inset in Fig. 4-6 shows the relative amount of product at the different tempera¬

tures detected by GC analysis.

4.3.3 Concentration Modulation Experiments of Epoxidation Reaction

Reactant concentration modulation experiments were performed at 70°C using

the methyl-modified aerogel lOTi-Me. Two different types of experiments were

performed. In one the concentration of TBHP was kept constant (3 mmol/1)

and the concentration of cyclohexene was modulated between 0 and 3 mmol/1.

70 Chapter 4

1200 1100 1000 900 800

Wavenumber / cm1

Fig. 4-5: Demodulated ATR spectra of adsorption of TBHP at room temperature on

titania-silica aerogels with diffèrent Ti content: OTi (Si02) (bottom), lOTi (middle), 20Ti

(top). The concentration ofTBHP was modulated between 0 and 3 mmol/1 in cyclohexane.

In the other type of experiments the cyclohexane concentration was kept

constant and the one of TBHP was modulated between 0 and 3 mmol/1.

Figures 4-7 and 4-8 show some phase-resolved ATR spectra of these experi¬

ments. TBHP concentration modulation clearly affected the signal of surface

silanols at 3700 cm' (Figure 4-8), while this effect was not observed for the

corresponding experiment with cyclohexene. Note that the same interaction of

TBHP with the silanol groups of the catalyst was also observed at room

temperature and in the absence of cyclohexene. The binding of TBHP to the

silanol groups is furthermore indicated by the band at 980 cm' associated with

ATR Studies on Epoxidation of Cyclohexene 71

1200 1100 1000 900 800 700

Wavenumber / cm"1

Fig. 4-6: Spectra recorded at different temperatures using SBSR technique. Cyclohexeneand TBHP (3.5 mmol/1 each in cyclohexane) were heated up on methyl-modified aerogel

(lOTi-Me). The inset shows the detected peak area for cyclohexene oxide as analyzed by gas

chromatography of the effluent product solution.

silanol groups. The intensity of the latter is anti-correlated with respect to the

intensity of the TBHP bands, see for example the band ofTBHP at 844 cm'1

(Fig. 4-7, bottom).

The spectra of the modulation experiments reveal the formation of cyclo¬

hexene oxide. The signals indicated with an asterisk in Fig. 4-7 are associated

with the epoxidation product and could also be found in the spectrum of

dissolved cyclohexene oxide (s. Fig. 4-lb). Prominent cyclohexene oxide signals

were found at 840, 783 and 743 cm'. Cyclohexene oxide was observed in both

modulation experiments, i.e. when cyclohexene and TPHB, respectively, were

72 Chapter 4

1200 1100 1000 900 800 700

Wavenumber / cm-1

Fig. 4-7: Demodulated ATR spectra recorded at 70°C. The concentration of cyclohexene

(top) andTBHP (bottom) were periodically modulated between 0 and 3 mmol/1. The spectra

are shown in the region between 1275 and 690 cm"1.

modulated. The appearance of product demonstrates that the catalyst is active

under the applied conditions and the infrared spectra were recorded truly in

situ. Note, that the cyclohexene oxide signals at 840 and 743 cm" partially

overlap with TBHP signals (s. Figure 4-1) and therefore are hardly discernible

in the corresponding spectrum. In the cyclohexene modulation experiment

(Fig. 4-7, top) the cyclohexene oxide and the TBHP signals are out of phase

and therefore the signal around 743 cm' is partially cancelled. However in the

TBHP modulation experiment (Fig. 4-7, bottom), the signals are in phase and

the corresponding band around 840 cm' is enhanced.

Product formation was also observed by GC analysis of the reaction solu¬

tion collected in the cell effluent. Fig. 4-9 shows the GC signal of cyclohexene

oxide as a function of time for the experiments where the cyclohexene and

ATR Studies on Epoxidation of Cyclohexene 73

Cyclohexene

3700 3500 3300 3100

Wavenumber / cm"1

Fig. 4-8: Demodulated ATR spectra recorded at 70°C. The concentration of cyclohexene

(top) andTBHP (bottom) were periodically modulated between 0 and 3 mmol/1. The spectra

are shown in the region between 3750 and 2935 cm"1.

TBHP concentration, respectively, was modulated. Also shown are the inte¬

grated absorbances of the signal at 783 cm' associated with cyclohexene oxide

and the one at 3024 and 1367 cm' associated with cyclohexene and TBHP,

respectively. Note that the GC signals were shifted on the time axis in order to

correct for the time that is required for the solution to flow from the cell to the

fraction collector. The flow from the cell through the pump and to the fraction

collector results in some back-mixing. Still the modulated cyclohexene oxide

concentration profile is obvious from the GC analysis. On the other hand the

time-dependent signals of the cyclohexene oxide and of cyclohexene as

observed in the infrared can directly be compared. Even though the time-

7A Chapter 4

0 49.8 99.7 149.5 199.3 249.2 299

Time / s

Fig. 4-9: Time dependence of signals for the experiments shown in Figs. 5-7 and 5-8:

a) Modulation of cyclohexene concentration and b) modulation of TBHP concentration.

Change of the detected peak area for cyclohexene oxide analyzed with gas chromatography(columns) and change of intensity for the cyclohexene oxide signal at 783 cm"1 in the time-

resolved IR-spectrum (thin line) and corresponding reaaants (bold line, cyclohexene at

719 cm"1, TBHP at 1367 cm"1.

resolved absorbance traces are rather noisy, a distinctly different time behavior

is observed for the reactants and cyclohexene oxide product. For example the

appearance of cyclohexene oxide product is retarded with respect to the signalof TBHP, whereas at the negative edge of the modulation the decrease of the

signals associated with reactant and product shows the same behavior within

ATR Studies on Epoxidation of Cyclohexene 75

the noise. The different time behavior of the two signals is reflected in their dif¬

ferent phase lag in the phase-resolved spectra shown in Fig. 4-7 (bottom). The

signal at 783 cm" associated with cyclohexene oxide vanishes at a demodula¬

tion angle of 170°, whereas the one at 1367 cm"1 due to TBHP vanishes at 180

degree. Note that we do not report absolute phase angles, that is we did not

correct the phase angle for the time it takes for the solution to flow from the

valve to the cell. Absolute phase angles are however not of importance here.

In the cyclohexene concentration modulation experiment a different time

behavior of reactant and product is also observed (Fig. 4-7, top). The signal at

783 cm" associated with cyclohexene oxide vanishes at a demodulation angleof 140°, while the one at 719 cm" due to cyclohexene vanishes at 170°.

Interestingly the signals of TBHP were also observed in the experiment

where the cyclohexene concentration was modulated. This is best seen for the

signal at 844 cm"1, which does not interfere with other bands (Fig. 4-7, top).

Additional bands associated with TBHP are visible at 1191, 1260 and

1367 cm".The signals of the peroxide are anti-correlated with respect to the

ones of the cyclohexene oxide (phase difference of about 180°). The typical

signals of the silanol groups (at 3700 and 980 cm' ) are not observed in

contrast to the peroxide adsorption experiments (Fig. 4-3) and the peroxide

modulation experiment (Fig. 4-7, bottom).

The band of TBHP at about 1250 cm'1 deserves special attention. It is

rather broad in the ATR spectra of dissolved TBHP (Fig. 4-1). The calculations

show that this broad feature consists of two modes: The low frequency mode

has C-O stretching character, whereas the high frequency mode is an asym¬

metric C-C stretching mode. In the calculations the two modes have almost

equal intensity. The symmetry of the band in the ATR spectra in Fig. 4-1

strongly indicates that this is also the case for the dissolved molecule. The shape

of the band for the adsorbed TBHP however changes with the Ti content

(Fig. 4-5). Furthermore there is a phase lag between a low and a high frequency

component of the band. In addition, in the epoxidation experiment where the

cyclohexene concentration was modulated (Fig. 4-7, top) the band maximum

is shifted to 1260 cm'1.

76 Chapter 4

The bands observed in the cyclohexene modulation experiment (Fig. 4-7,

top) in the 920-1040 cm'1 spectral region (catalyst framework vibrations) are

likely due to the interaction of the peroxide with the catalyst. The broad band

above about 950 cm' is out of phase with the peroxide bands (phase shift of

180°), indicating that the peroxide interacts with the corresponding sites. Note

that in the peroxide adsorption experiments (Fig. 4-5) and the epoxidation

experiment, where the peroxide concentration was modulated (Fig. 4-7,

bottom), this band is buried in the framework bands arising due to adsorption

of the peroxide on Si-O-H groups. The band at about 930 cm' corresponds to

the band also observed in the adsorption of the peroxide (Figs. 4-3 and 4-5)

and in the peroxide modulation experiment (Fig. 4-7, bottom). In the latter

experiments it is however overlapped by signals, which arise due to TBHP

adsorption onto silanol groups. Investigation of the time-resolved signals shows

that the band at about 930 cm' is steadily increasing, which also results in a

signal in the demodulated spectra. This steadily increasing band at 930 cm' is

also seen in the TBHP adsorption experiments, whereas it is not seen for cyclo¬

hexene adsorption. This shows that TBHP is responsible for the observed spec¬

tral change. Figure 4-10a shows time-resolved absorbance spectra recorded

during an epoxidation experiment where the cyclohexene concentration was

modulated, revealing the strong signal at about 930 cm'.The position of the

band corresponds to the Ti-O-Si spectral region. Note that the background

spectrum was recorded in neat cyclohexane before the catalyst was contacted

with the reactants.

4.4 Discussion

The results presented above demonstrate the feasibility to study the solid-liquid

interface of a working aerogel catalyst by attenuated total reflection infrared

spectroscopy. The use of a flow-through cell allows disturbing the working

catalytic system by a forced concentration modulation of the reactants. In this

way species are selectively probed, which are affected by the perturbation.

ATR Studies on Epoxidation of Cyclohexene 77

The advantages of the digital phase-sensitive detection with respect to the

time-resolved measurements become apparent from the comparison of the

spectra in Fig. 4-10, which correspond to the epoxidation experiment, where

the cyclohexene concentration was modulated. Note that the three sets of spec¬

tra (a-c) correspond to the same set of data. Figure 4-10a shows a selection of

time-resolved spectra. The reference was recorded before admitting the reac¬

tants. Relatively large signals are observed, which are however mostly static.

The small changes can be revealed by taking difference spectra, i.e. by subtract¬

ing one (arbitrarily chosen) spectrum recorded during modulation from all

other spectra. A selection of such spectra is given in Fig. 4-10b. The signals in

these spectra are about one order of magnitude smaller than the signals in the

original time-resolved spectra. The fact that the species of interest give rise to

comparably small signals represents a challenge for in situ infrared spectroscopy

and hence noise is a concern. It is here that the digital phase-sensitive detection

drastically increases the quality of the spectra. The phase-resolved spectra

shown in Fig. 4-10c, are of much higher quality than the time-resolved coun¬

terparts (difference spectra). The digital phase-sensitive detection is a narrow¬

band technique, which efficiently eliminates noise at frequencies different from

the stimulation frequency.

The cyclohexene adsorption experiments (Fig. 4-2) reveal only very weak

bands, indicating that the interaction with the catalyst surface is weak at room

temperature. On the other hand TBHP adsorbs strongly at room temperature

and at 70°C (Figs. 4-3 and 4-5, room temperature adsorption and Figs. 4-7

and 4-8, 70°C) both in the absence and presence of cyclohexene. Note that

catalytic activity was observed above about 60°C (Fig. 4-6).

The ATR experiments reveal two differently adsorbed TBHP species on

the Ti-Si aerogels. One species is hydrogen-bonded to the Si-OH groups and

observed both on the pure Si aerogel and the Ti-Si aerogels. On the latter an

additional adsorbed THBP species is found: The most prominent spectral

differences with respect to the species hydrogen-bonded to the Si-OH groups

are a shift of the band at 1250 to about 1260 cm' and the appearance of a

band at 920-930 cm'1 (Fig. 4-5). The strengths of these bands and the relative

amount of the associated species increased with higher Ti content, which

suggests that the TBHP adsorbs on sites containing Ti. In the following these

78 Chapter 4

1300 1200 1100 1000 900 800 700

Wavenumber / cm"1

Fig. 4-10: ATR spectra for epoxidation experiment recorded under forced modulation of

the cyclohexene concentration: a) time-resolved absorbance spectra (reference recorded before

molulation); b) difference spectra by subtracting one (arbitrary chosen) spectrum; c) phase-resolved (demodulated) spectra. Note that the data set for spectra in a-c is the same.

sites are called Ti sites. The TBHP is more strongly adsorbed on the Ti sites

than on the Si-OH sites. In the adsorption experiments at room temperature,

where the TBHP concentration was modulated (Figs. 4-3 and 4-5), the differ¬

ent adsorption strengths results in a distinct phase lag between the signals of

ATR Studies on Epoxidation of Cyclohexene 79

TBHP adsorbed on Ti- and Si-OH sites. The band at 920-930 cm' indicates a

change in the Ti-O-Si vibration of the catalyst [72,85,186] and furthermore

corroborates the interaction of TBHP with a Ti site. The band at 1250-

1260 cm' is associated with two vibrations ofTBHP itself, as discussed in the

results section (C-C at high and C-O at low frequency). For TBHP adsorbed

on the Ti site the band is shifted to higher wavenumbers (1260 cm' ) compared

to TBHP hydrogen-bonded to Si-O-H groups (1250 cm'1). This shows that

the lower frequency C-O vibration has shifted or vanished. This in turn indi¬

cates an interaction between the peroxo group and the Ti site.

Comparison of the two concentration modulation experiments shown in

Figs. 4-7 and 4-8 allows the discrimination between spectator and reactive

TBHP species adsorbed on the catalyst surface. In the case where the TBHP

concentration was modulated the TBHP adsorbs and desorbs from the catalyst

surface both on the Si-OH- and the Ti sites. In the cyclohexene concentration

modulation experiment only those TBHP species give rise to signals in the

demodulated spectra, which are involved in the catalytic reaction. The

observed band at 1260 cm' and the absence of the negative Si-OH bands at

3700 and 980 cm'1 show that only the TBHP adsorbed on the Ti site is

involved in the catalysis, whereas the one adsorbed on the Si-OH is not. This

means that the concentration of the peroxide adsorbed on the silanol groups

remains constant during cyclohexene modulation, while the concentration of

the peroxide on the active sites is periodically changing, as can be seen from

Fig 4-7, top. This clearly shows that the peroxide adsorbed on the silanol

groups is a spectator species.

Unfortunately, not much is known about the interaction of TBHP with

Ti-Si catalysts. Fujiwara et al. isolated a titania-silsesquioxane-peroxide

complex, which was analyzed by low temperature NMR [123]. It was proposed

that in this complex the Ti is four coordinated with one Ti-O-O^Bu and three

Ti-O-Si and linkages. The spectral signatures of the adsorbed peroxide, which

is involved in the catalytic cycle are consistent with an analogous coordination

of the peroxide (shift of C-O vibration), but can by no means be taken as a

proof of it.

The Ti content of the aerogel has a significant influence on the adsorption-

desorption behavior of TBHP. Comparing the textural properties of the aero-

80 Chapter 4

gels (Table 4-1), we note that the BET surface area and the cumulative pore

volume decrease with higher Ti content. For a liquid, pore diffusivity Dpore can

be expressed as:

D =Df%- (I)

pore f — V A JX

Where Df stands for the diffusion coefficient in solution, % for the internal

porosity of the particles and t for tortuosity [226]. Higher Ti content also results

in more of the strongly adsorbed TBHP, which can be monitored by the genu¬

ine TBHP signals as well as by the signals originating from the changes in the

framework vibrations due to adsorption. Compared to the signal at 1014 cm'

the signal at 944 cm' becomes more prominent and broader. Interestingly, for

the aerogel 20Ti, with a Ti-content corresponding to 20 wt% nominal Ti02

the negative signal at 980 cm' originating from the Si-O-H groups does not

increase upon TBHP desorption, in contrast to the other aerogels (not shown).

Possibly a large fraction of silanol groups are located near the Ti sites in this

catalyst and are also involved in binding TBHP at the active sites.

The reactant concentration modulation experiments (Figs. 4-7 and 4-8)

reveal a periodic increase and decrease of the cyclohexene oxide product. In the

experiment with constant TBHP and varying cyclohexene concentration in

solution, signals associated with adsorbed TBHP were observed, whereas no

cyclohexene signals were observed in the TBHP concentration modulation

experiment. This indicates a vastly different adsorption behavior of the two

reactants on the catalyst surface. In the presence ofTBHP in solution the cata¬

lyst surface is covered with TBHP. By admitting the cyclohexene reactant some

of the adsorbed TBHP reacts. Dissolved TBHP can readsorb on the empty

catalytic sites. The TBHP coverage on the surface is determined by the relative

rate of reaction and readsorption of TBHP. For the latter the diffusion rate of

TBHP to the active site and the rate of adsorption is crucial. The observation

that the concentration of adsorbed TBHP is influenced by the cyclohexene

concentration modulation shows that diffusion and/or adsorption ofTBHP is

affecting the global epoxidation rate. No similar effect was observed for cyclo¬

hexene: During TBHP concentration modulation no cyclohexene signals were

observed showing that the cyclohexene concentration in the volume probed by

ATR Studies on Epoxidation of Cyclohexene 81

the evanescent electromagnetic field did not change significantly. This is in

agreement with the observation that cyclohexene does not adsorb strongly on

the catalyst surface (even at room temperature) and hence its surface concentra¬

tion is low under the experimental conditions.

Figure 4-9 shows significant phase lags between the reactant and product

signals. Phase lags occur when the relaxation time of the system is on the order

of the inverse modulation frequency [223]. The different time-behavior of the

reactant and product signals are also obvious from the traces shown in Fig. 4-9.

Notably in the TBHP modulation experiment the TBHP signal increases

significantly before the cyclohexene product signal.The observed phase lag between reactant and product could have different

origins. One feasible explanation is the different diffusion rate of the species.

TBHP, when it is first detected by ATR, has to diffuse to the active site before it

can react. The time lag between the TBHP and the cyclohexene oxide signal is

about 20 seconds (Fig. 4-9b). The root-mean square path travelled by a mole¬

cule in a given time t is given by

<x2>1/2=(2Dt)1/2. (2)

Assuming a diffusion constant of D=10 cm /s for the peroxide in solution it

would take about 0.05 s to travel the probed surface region of about 3 um

thickness. Obviously diffusion of TBHP in solution can not explain the

observed time lag. On the other hand, pore diffusion may be at its origin. The

larger time lag observed for the TBHP experiment than for the cyclohexene

experiment would be consistent with a slower pore diffusion of the former due

to its stronger interaction with the catalyst surface. Since the calculated Van

der Waals radii of cyclohexene and TBHP are very similar (-5.4 Â), sterical

factors are likely not the only reason for the different pore diffusion rate. An

observation, which supports this assumption is the comparison of typical reac¬

tant signals in the time-resolved spectra recorded on a blank IRE and an IRE

coated with aerogel. For cyclohexene these revealed only a small difference due

to the presence of the catalyst, whereas a large difference was observed for the

TBHP (Fig. 4-11). This points to a slower pore diffusion ofTBHP compared

to cyclohexene likely due to a stronger interaction of the former with the

82 Chapter 4

0 40.7 81.3 122 162.7 203.3 244

Time / s

Fig. 4-11 : Change of intensity as a function oftime for the two reactants on uncoated IRE

(thin line) and on IRE coated with methyl-modified aerogel (lOTi-Me) (bold line) for

a) cyclohexene (at 3024 cm"1) and b) TBHP (at 1367 cm"1). The concentration was modu¬

lated between 0 and 3 mmol/1 in cyclohexane.

surface. A similar observation was also made by Rivera and Harris, who investi¬

gated the rate of transport into thin sol-gel films for molecules with different

affinity to the surface [175]. They found for example that the diffusion of 2-pro-

panol was an order of magnitude slower than the diffusion of toluene.

An additional explanation for the observed time and phase lag between

reactant and product concerns the adsorption step itself, which may be slow on

the active site. An indication that the adsorption on the Ti sites is activated on

ATR Studies on Epoxidation of Cyclohexene 83

the catalyst modified with methyl groups (lOTi-Me) emerges from the TBHP

adsorption experiments. At room temperature no significant adsorption of

TBHP on the Ti sites could be found as indicated by the absence of the

930 cm"1 and the shift of the 1250 cm"1 band, in contrast to the unmodified

aerogels TilO and Ti20. Also, adsorption on the surface silanol groups seems to

be more reversible and weaker for the modified aerogel. Comparison of the

textural properties of the unmodified and modified aerogel (containing

10 wt% Ti02) shows that the modified aerogel has, besides the smaller BET-

surface area, bigger pores. This hints to a higher pore diffusion rate in the

modified aerogel, which may contribute to the different behavior. Modifyingthe surface with covalently bound methyl groups has a marked influence on the

interaction between peroxide and the active sites on catalyst due to (i) steric

hindrance and (ii) lower polarity of the surface. The modified aerogel catalyst

usually perform better than the corresponding unmodified counterparts. One

possible reason for this, which emerges from our study, is the higher pore diffu¬

sion ofTBHP and cyclohexene reactant. This may lead to pronounced changeof concentration gradients within the porous catalyst.

The variable temperature experiments, using the SBSR technique, indi¬

cated a change in the catalyst Si-O-Si and Ti-O-Si vibrations in the region

between 1200 and 700 cm".Online GC analysis showed epoxide formation at

about 60°C and above. This is the temperature at which strong changes in the

framework vibrations were observed in the ATR experiment. The above obser¬

vations suggest that the catalyst underwent some structural change once the

reaction was initiated. This supports recent kinetic studies by Beck et al., who

found that the initial turnover frequency declined after the first few turnovers.

They suggested a cleavage of one Ti-O-Si bond of the predried aerogels by

TBHP or water resulting in active Ti sites with remarkably different catalytic

properties [150].

84 Chapter 4

4.5 Conclusions

ATR infrared combined with modulation spectroscopy in a flow-through cell

was used to study the interaction of tert-butyl hydroperoxide (TBHP) and

cyclohexene with Ti-Si aerogel catalysts as well as the cyclohexene epoxidation

reaction. The phase-sensitive detection afforded an enhanced sensitivity, which

is required to study the relevant small changes due to the periodic perturbation

of this catalytic system by forced concentration modulation. Whereas no inter¬

action of cyclohexene with the catalyst surface was evident from the spectra,

two different adsorption modes of the peroxide were discernible. The modula¬

tion experiments revealed that peroxide, which adsorbs on Si-OH groups is a

spectator, whereas peroxide species adsorbing on Ti sites are involved in the

catalytic cycle. These species are more strongly adsorbed than those on the

Si-OH groups. The spectrum of this species is characterized by a shift of the

C-O frequency and a change of the Ti-O-Si framework vibrations of the cata¬

lyst.

The concentration modulation experiment revealed a time (phase) lagbetween the appearance of reactant and product in the volume probed by the

evanescent field. The main reason is likely the different rate of pore diffusion,

which is significantly slower for the peroxide than for cyclohexene. By modu¬

lating the cyclohexene concentration the TBHP coverage on the active sites can

be depleted, which shows that diffusion and/or adsorption of the peroxide on

the active site is affecting the observed epoxidation rate. Hence, concentration

gradients exist in the catalyst particles which are probably quite different for

peroxide and cyclohexene. The engineering of the pore structure and the

hydrophobicity of the surface are therefore key factors for the design of these

catalyst types. The spectroscopic studies furthermore indicate some structural

changes of the catalyst under conditions where epoxidation occurs.

Chapter

Epoxidation of Cyclic Allylic Alcohols

on Titania-Silica Aerogels

Studied by Attenuated Total Reflection Fourier

Transform Infrared and Modulation Spectroscopy

5.1 Introduction

Epoxidation of functionalized olefins using Ti and Si based catalysts has

received considerable attention in recent years [185,213]. Titania-silica aerogels

belong to the catalysts which exhibit interesting catalytic potential, particularly

for the epoxidation of bulky functionalized olefins, mainly due to their open

mesoporous structure combined with high abundance of isolated tetrahedrally

coordinated Ti sites. Most of the work aimed at elucidating the mechanism of

epoxidation has been carried out on crystalline well-defined molecular sieve

materials [122,129,133,227-230] or special homogeneous model catalysts

[100,102,104,151]. Various transition state structures have been proposed (see e.g.

[150] and references therein). However, the crystalline materials with their rigidstructural constraints are generally less suitable for the epoxidation of bulky

olefins [129,222]. Mechanistic studies on amorphous materials, such as titania-

silica aerogels are therefore demanded and such studies have to be performed

under working conditions due to the metastable nature of these materials.

In the past years considerable effort has been made to further develop

attenuated total reflection Fourier transform infrared (ATR-IR) spectroscopy

to a versatile tool for investigating catalytic solid-liquid interfaces [163/4,177/8].

Particularly, the combination of ATR-IR with modulation spectroscopy has

brought significant advantages concerning the sensitivity of the method and for

the discrimination of dynamic and static surface processes [164]. In chapter 4 we

5

88 Chapter 5

explored the potential of the technique to study differences in the adsorptive

interactions of the reactants in the epoxidation of cyclohexene [231].

Studies on the epoxidation of various allylic alcohols on titania-silica aero¬

gels revealed striking differences in the behavior of the epoxidation depending

on the allylic alcohol [150]. Particularly interesting is the difference observed

when comparing the epoxidation of cyclohexenol with that of cyclooctenol.

Based on the reaction rate and the greatly different stereoselectivity we specu¬

lated that the different catalytic behavior can be traced to a hydroxy-assistedmechanism in the case of cyclohexenol epoxidation, whereas for cyclooctenol a

silanol-assistedmechanism seems to prevail (Scheme 5-1).

In the present chapter we aim at unravelling the subtle differences in the

interactions occurring during epoxidation of cyclohexenol and cyclooctenol

using ATR-IR combined with modulation spectroscopy.

5.2 Experimental

5.2.1 Preparation of Catalyst Layer

The four aerogels used in the experiments were prepared as described in

chapter 3 on page 41 and reported elsewhere in detail [93,222]. The preparation

catalyst layer on surface of a ZnSe internal reflection element has been reported

in chapter 4 on page 61.

5.2.2 Nitrogen Physisorption

The textural properties of the different aerogels were determined by nitrogen

physisorption at -196°C using a Micromeritics ASAP 2000 instrument as

already described in chapter 3 on page 42. Maximum of pore size distribution

(dmax) were graphically assessed from desorption branch of isotherms.

ATR Studies on Epoxidation of Cyclic Allylic Alcohols 89

Textural properties of the aerogels (OTi, 1 OTi-Me, 1 OTi and 20Ti) were (same

sequence):

BET surface area: 1102, 635, 768 and 761 m2^1

dmax: 35, 60, 36 and 96 nm

Specific pore volume: 3.77, 2.08, 2.21 and 1.46 cm3g'!

Figure 5-1 shows the differential pore size distribution of the aerogels lOTi and

1 OTi-Me, derived from the desorption branch of the corresponding isotherms.

1 OTi-Me showed the highest performance for epoxidation of cyclohexenol [93].

i—i—i—i—i—i—i—| 1 1—i—i—i—i—i—i—|-

10 100

Pore Diameter (dp) / nm

Fig. 5-1 : Differential pore size distribution derived from the desorption branch of the

isotherm of methyl-modified titania-silica aerogel (1 OTi-Me) (bottom), and the correspond¬

ing unmodified aerogel lOTi (top). Note that the x-axis is logarithmic.

90 Chapter 5

5.2.3 ATR Spectroscopy

The experimental setup has been described in the experimental part on

page 32. Additional information can also be found elsewhere [164,232].

5.2.4 Modulation Spectroscopy

A short description of the modulation excitation (ME) spectroscopy and its

advantages can be found in chapter 1 on page 23. More detailed information of

the technique has been reported elsewhere [163,164,180,223].

5.2.5 Adsorption and Epoxidation Experiments

The procedure for adsorption and epoxidation experiments was essentially the

same as previously described on page 62.

Adsorption experiments were carried out at room temperature. Measure¬

ment of the time-resolved spectra was synchronized with the concentration

modulation by switching the computer-controlled valve within the data acqui¬

sition loop. One modulation period lasted T= 244 s.

Epoxidation experiments with cyclooct-2-en-l-ol (97%; prepared accord¬

ing the procedure described by Meier et al. [233]) were carried out at 70°C in

cyclohexane (Fluka, puriss p.a., dist. over CaH2) solvent, corresponding experi¬

ments with cyclohex-2-en-l-ol (Fluka, puriss. p.a.) were carried out at 80°C in

toluene solvent (Fluka, puriss, dist. over Na). For the latter experiment, toluene

was used as solvent since cyclohexane was unsuitable due to its lower boiling

point (81°C). For the concentration modulation, corresponding liquid feeds

were dosed from two different tanks. Tank 1 contained a solution of allylic

alcohol and £-butyl hydroperoxide (TBHP; Fluka, purum, -5.5M in nonane)

in the corresponding solvent, respectively, and tank 2 a mixture of the two reac¬

tants. Data acquisition was performed with a modulation period T = 299 s.

The samples collected after the cell were subsequently analyzed using a HP

6890 gas Chromatograph (cool on-column injection, HP-FFAP column, 30 m

x 0.32 mm x 0.25 um).

ATR Studies on Epoxidation of Cyclic Allylic Alcohols 91

5.3 Results

5.3.1 Adsorption Experiments

Figure 5-2 shows ATR spectra of dissolved cyclohex-2-en-1-ol and cyclohex¬

enol oxide (-10 mmol/1 each) in toluene as well as TBHP, cyclooct-2-en-l-ol

and cyclooct-2-en-l-ol oxide in cyclohexane measured over the uncoated inter¬

nal reflection element (IRE) in the absence of catalyst. Important vibrational

bands of reactants and products used in this study, are listed in Table 5-1.

Cyclohexenol oxide

Cyclooctenol oxide

1400 1300 1200 1100 1000 900 800 700

Wavenumber / cm1

Fig. 5-2 : Demodulated ATR spectra ofreactants and products on blank internal reflection

element (IRE) at room temperature: (top-down) cyclohex-2-en-l-ol, cyclohexenol oxide,

f-butyl hydroperoxide (TBHP), cyclooct-2-en-l-ol, cyclooctenol oxide. Note that in the

experiments of cyclohexenol and cyclohexenol oxide the spectral region between 730 -

700 cm"1 was obscured by strong absorptions of toluene used as solvent.

oCd

-eot/J

<

92 Chapter 5

Table 5-1 : Assignments of observed vibrational bands. Bands used in discussion are high¬

lighted.

Reactant Typical Reactant Framework Negativea)

Signals [cm"1] vibrations Framework

[cm"1] vibrations [cm"1]

905 - 875

Cyclohexenol 3026, 2960, 1064, 1053, 1035, 1027, 1020, 3700, 980,

1048,1041,953,804, 1014,1002,983, 890-810

727 958

Cyclohexenol oxide 1069, 1057, 1040, 1030, 1051, 999, 953

998, 944, 931, 864, 853,

846

Cyclooctenol 3019, 2964, 1131, 1049, 1050 - 1010,

1017,988,962,782,751, 990-910

3700, 980,

900 - 780

713

Cyclooctenol oxide 1053, 1024, 1017, 985, 1045, 1004, 954 850 - 830

978, 970, 912, 892, 881,

864,828,821,745

TBHP 2983, 1389, 1367, 1320, 1260, 1020, 944

1250, 1191,844,747

a

Signals decreasing upon exposure to concentration modulation of reactants.

Figure 5-3 shows phase-resolved ATR spectra recorded when modulating the

cyclohex-2-en-1-ol concentration between 0 and 3 mmol/1 at room tempera¬

ture over titania-silica aerogels (OTi, 1 OTi-Me, lOTi and 20Ti). Prominent

signals at 3026, 2960 (not shown), 1064, 1020, 1014, 1002, 953, 860, 848,

804 and 727 cm"1 were discernible. The negative bands at 3700 and 980 cm"1,

revealed a strong interaction with the surface silanol groups. The broad band at

3280 cm"1 indicated an interaction of the OH-group of the allylic alcohol with

the catalyst surface. With increasing Ti content, the area between 990 and

1080 cm' changed drastically. Note that cyclohexenol typically absorbed

between 1040 and 1065 cm"1 (Fig. 5-2). Whereas for the lOTi aerogel the

signals at 1019 and 1002 cm",

as well as the typical cyclohexenol signalsbecame more prominent with respect to OTi, these bands were weaker and

ATR Studies on Epoxidation of Cyclic Allylic Alcohols 93

1100 1000 900 800 700

Wavenumber / cm"1

Fig. 5-3: Demodulated ATR spectra of adsorption of cyclohex-2-en-l-ol at room tempe¬

rature on titania-silica aerogels with diffèrent Ti content and modification: (top-down): OTi

(Si02), methyl-modified titania-silica aerogel (1 OTi-Me), lOTi, 20Ti. The concentration of

cyclohex-2-en-l-ol was modulated between 0 and 3 mmol/1 in toluene.

saturated for 20Ti. Furthermore the bands at 1027 and 1035 cm"1 became less

intense on desorption of cyclohexenol. The broad negative band in the area

between 810 and 890 cm' increased with higher Ti content, as well as the

positive signals at 848 and 860 cm'1. Upon initial adsorption of cyclohexenol a

significant change in the framework vibration occurred giving rise to a band at

958 cm'1. Note that this band is close to the C-O stretching vibration of cyclo¬

hexenol at 953 cm'.

94 Chapter 5

The corresponding adsorption experiments with cyclooct-2-en-l-ol over

aerogels with different Ti content (OTi, 1 OTi-Me, lOTi and 20Ti) are shown in

Figure 5-4. Again, the concentration of the allylic alcohol was modulated

between 0 and 3 mmol/1 at room temperature. Typical signals could be

observed at 3019, 2964 (not shown), 1049, 1017, 988, 962, 782, 751 and

713 cm'.A broad negative framework band between 900 and 780 cm' was

detected, which indicates a perturbation of the symmetric Si-O-Si vibrations

1100 1000 900 800 700

Wavenumber / cm"1

Fig. 5-4: Demodulated ATR spectra of adsorption of cyclooct-2-en-l-ol at room tempe¬

rature on titania-silica aerogels with diffèrent Ti content and modification: (top-down): OTi

(Si02), methyl-modified titania-silica aerogel (1 OTi-Me), lOTi, 20Ti. The concentration of

cyclooct-2-en-l-ol was modulated between 0 and 3 mmol/1 in cyclohexane.

ATR Studies on Epoxidation of Cyclic Allylic Alcohols 95

(s. below). For all aerogels a negative band at 3700 cm' was observed. Due to

overlaps with the cyclooctenol signal at 988 cm',the corresponding negative

signal of the surface silanol groups at 980 cm' is obscured. A broad signal at

3280 cm' indicated an interaction of the hydroxy group of cyclooctenol with

the catalysts surface. As already observed for cyclohexenol, cyclooctenol showed

a similar adsorption behaviour on the methyl modified (1 OTi-Me) and the pure

silica aerogel (OTi). For these catalysts, the signals of the substrate (1048, 988

and 962 cm' ) were more dominant than the signals of the framework in this

area and a slight shift of the band at 1049 to 1043 cm' could be observed.

The spectra of adsorbed TBHP on the corresponding catalysts have been

discussed before [231]. Typical strong bands appear at 2983, 1389, 1367, 1320,

1250 (broad), 1191, 844 and 747 cm'1, which are assigned to adsorbed TBHP.

The strong broad signals around 1020 and 944 cm' originate from aerogelframework vibrations.

Also, the spectrum of the 1 OTi-Me catalyst was reported previously [231]

and discussed in comparison with the results reported for Ti02/Si02 mixed

oxides by Schraml et al. [85]. Most prominent, a broad and strong band with a

maximum at 1075 cm' and a shoulder at 1200 cm' could be detected, which

can be assigned to the v(Si-O-Si) asymmetric stretching vibration. The sym¬

metric v(Si-O-Si) stretching vibration can be found at 810 cm'1. At 947 cm'1 a

signal of the v(Ti-O-Si) asymmetric stretching vibration could be detected,

which is partially overlapped by the Si-OH vibration at 980 cm'1.

5.3.2 Concentration Modulation Experiments of Epoxidation Reactions

Reactant concentration modulation experiments were performed at 80°C and

70°C, respectively, using the methyl-modified aerogel 1 OTi-Me. Two different

types of experiments were performed. In one the concentration ofTBHP was

kept constant (3 mmol/1) and the concentration of the allylic alcohol was

modulated between 0 and 3 mmol/1. In the other type of experiments the

allylic alcohol concentration was kept constant and the one of TBHP was

modulated between 0 and 3 mmol/1. Figures 5-5 and 5-6 show some phase-

resolved ATR spectra of these experiments.

96 Chapter 5

1 ' 1 ' 1 ' 1 ' 1 ' 1 ' 1 ^

1400 1300 1200 1100 1000 900 800

Wavenumber / cm"1

Fig. 5-5: Demodulated ATR spectra of epoxidation experiments recorded in the regionbetween 1400 and 750 cm"1. The concentration of cyclohex-2-en-l-ol (top) and TBHP

(bottom) were periodically modulated between 0 and 3 mmol/1 in toluene at 80°C. Top most

spectrum shows adsorption of cyclohexenol oxide over 1 OTi-Me in the range between 1070

-965 cm"1.

When cyclohexenol was used as reactant (Fig. 5-5, top), the cyclohexenol

concentration modulation experiment affected the signals of surface silanols

980 cm",while this effect was not observed for the corresponding experiment

with TBHP (Fig. 5-5, bottom). For both experiments, no clear signals of the

epoxidation product were discernible.

Only relatively weak signals could be detected between 1100 and

900 cm"1, when the concentration of TBHP was modulated. While typical

peroxide signals (1367, 1326, 1250, 1191 and 844 cm'1) could be observed,

the framework vibrations in the region between 950 and 1070 cm' were

hardly influenced and exhibited a phase shift of 90° with respect to the TBHP

ATR Studies on Epoxidation of Cyclic Allylic Alcohols 97

signals. The changes in the framework vibrations typical for TBHP adsorption

on the methyl-modified aerogel (1 OTi-Me) at 1020 and 944 cm'1 could not be

detected [231]. GC analysis of the effluent reaction solution showed only traces

of the desired epoxide and its intensity did not seem to be periodically modu¬

lated.

In the case of cyclohexenol modulation, signals of the allylic alcohol (1064,

1053, 1048, 1041, 953 and 804 cm'1) and the catalyst framework (1035,

1027, 1020, 1014, 983 and 958 cm'1) could be detected. The spectrum

showed a high similarity to the one of cyclohexenol adsorption on the unmodi¬

fied aerogel (lOTi) at room temperature (see Fig. 5-3). Furthermore, negative

bands at 1367, 1250, 1191, 844 and 747 cm'1 indicate a displacement of

TBHP. Analysis of the cell effluent by GC revealed a small concentration of

cyclohexenol oxide, which was however higher than in the TBHP modulation

experiment. Yet, the concentration showed only slight periodic time depen¬

dence.

Modulation experiments using cyclooctenol as reactant revealed an interac¬

tion with surface silanol groups for TBHP as well as for the allylic alcohol. In

contrast to the behaviour observed with cyclohexenol, the spectra revealed neg¬

ative signals at 980 cm' and most prominently at 3700 cm',

as shown in

Figure 5-7. The spectra of both experiments showed the formation of cyclooc¬

tenol oxide. The signals indicated with an asterisk in Fig. 5-6 are associated

with the epoxidation product and could also be found in the spectrum of dis¬

solved cyclooctenol oxide. Prominent cyclooctenol oxide signals were found at

1045, 1024, 1004, 985, 970, 912, 892, 881, 864, 821 and 745 cm'1. The

appearance ofproduct demonstrates that the catalyst is active under the applied

conditions and the infrared spectra were recorded truly in situ. Note, that the

cyclooctenol oxide signal at 745 cm'1 partially overlaps with TBHP (747 cm'1)

and cyclooctenol (751 cm' ) signals. Simultaneous gaschromatographic analy¬

ses of the samples taken from the cell effluents at different phase angles, as pre¬

sented in Figure 5-8, confirmed the formation of cyclooctenol oxide.

When the TBHP concentration was modulated (Fig. 5-6, bottom), the typical

signals for the adsorption of the peroxide on the catalyst could be observed

(1367, 1320, 1250, 1191, 1020, 944 and 844 cm'1). Interestingly, additional

signals at 1260 and at 936 cm' were observed, which hinted to adsorption on

98 Chapter 5

1400 1300 1200 1100 1000 900 800 700

Wavenumber / cm"1

Fig. 5-6: Demodulated ATR spectra of epoxidation experiments recorded in the regionbetween 1400 and 700 cm"1. The concentration of cyclooct-2-en-l-ol (top) and TBHP

(bottom) were periodically modulated between 0 and 3 mmol/1 in cyclohexene at 70°C.

the active sites [231]. The negative bands at 753 and 713 cm" are indicative for

cyclooctenol desorption from the catalyst. Reactants, activated species and

product showed different phase lags. The cyclooctenol signal at 713 cm"

reached its minimum at 70° demodulation phase angle, while the TBHP signal

at 1367 cm'1 had its maximum at 100°. The signal originating from the perox¬

ide adsorbed on the Ti site (1260 cm'1) reached its maximum at 30° and the

one for the cyclooctenol oxide signal at 881 cm' was observed at 60°. Spectra

recorded before starting the modulation experiment showed a slight increase of

the silanol bands at 3700 and 980 cm' before they decreased again when

TBHP concentration was raised. The scans were recorded starting at the

steady-state, when the allylic alcohol in cyclohexane was flown through the cell

and before the peroxide was applied for the first time.

O

Öa

-eoT

<

ATR Studies on Epoxidation of Cyclic Allylic Alcohols 99

1 x Iff3

3800 3600 3400 3200 3000

Wavenumber / cm1

Fig. 5-7: Demodulated ATR spectra in the region between 3935 and 2935 cm"1 ofepoxi¬dation experiments of cyclohexenol (cf. Fig 5-5) and cyclooctenol (cf. Fig. 5-6). Conditions

are given in legends ofFigs. 5-5 and 5-6, respectively. (Modulation ofreactant concentration:

a) cyclooctenol; b) TBHP in cyclooctenol epoxidation; c) cyclohexenol; d) TBHP in cyclo¬hexenol epoxidation).

For the corresponding epoxidation experiment with cyclooctenol concen¬

tration modulation (Fig. 5-6, top), the characteristic framework bands in the

area between 1050 and 940 cm" and the signals of the allylic alcohol at 753

and 713 cm" were observed, as for the adsorption experiments. The negative

signals at 2983, 1367, 1250, 1191 and 844 cm'1 indicate a desorption of

TBHP. Note that for both epoxidation experiments, the signal at 844 cm'1

overlaps with the broad negative framework band between 900 and 780 cm'.

Also in this case, the various compounds exhibit different phase lag. The maxi¬

mum of the cyclooctenol signal at 713 cm' appears at 80° demodulation phase

angle while the one of the TBHP signal (1367 cm'1) can be observed at 50°.

The epoxide signal at 881 cm' reaches its maximum at 180°.

100 Chapter 5

O 49 8 99 7 149 5 199 3 249 2 299

Time / s

Fig. 5-8: Time dependence of signals for the experiments shown in Fig. 5-6. a) Modula¬

tion of cyclooctenol concentration and b) modulation ofTBHP concentration. Change of

£ram-cyclooctenol oxide concentration analyzed by GC (pillars). Intensities ofthe cycloocte¬nol oxide signal at 912 cm"1 (thin line) and corresponding reactants (bold line - cyclooctenol

at 713 cm"1; dotted line -TBHP at 1396 cm"1) in the time-resolved IR-spectra.

5.4 Discussion

As mentioned in the introduction, comparative studies of the epoxidation of

various allylic alcohols on aerogels with a Ti content of 1 and 5 wt% using

TBHP as oxidant showed remarkable differences in reactivity, regio- and

stereoselectivity depending on the allylic alcohol [150]. In particular, vastly

ATR Studies on Epoxidation of Cyclic Allylic Alcohols 101

different rates, chemoselectivity and stereoselectivity were observed for the

epoxidation of cyclooctenol and cyclohexenol. The turnover numbers (TON)

for 2 min (7.2 and 1.4, respectively) and 2 h (39.5 and 10.2, respectively) as

well as the times necessary for 50% TBHP conversion (8 and 56 min, respec¬

tively) showed a higher reactivity for cyclooctenol. On the other hand, the

diastereoselectivity for aV-epoxide was much higher in the case of cyclohexenol

(82.5 and 14.5%, respectively). The present spectroscopic investigations

provide substantial evidence for the speculations made earlier that in the case of

cyclohexenol a hydroxy-assisted interaction prevails, whereas for the cyclo¬

octenol epoxidation a silanol-assisted interaction is dominant, due to steric

hindrance (Scheme 5-1).

Scheme 5-1 : Proposed transition states for the epoxidations of cyclohex-2-en-l-ol via

hydroxy-assisted mechanism (a) and cyclooct-2-en-l-ol by a silanol-assisted mechanism (b).

Note that different coordination for cyclooctenol may be feasible depending on the positionof the neighboring silanol group.

The cyclohexenol adsorption experiments (Fig. 5-3) reveal a strong adsorp¬

tion on the surface of the different catalysts, which can clearly be seen by the

negative signals at 3700 and 980 cm"1, indicating a strong binding to the

102 Chapter 5

surface silanol groups. Since the prominent signal at 1020 cm" is absent for the

adsorption on the pure silica aerogel (OTi), this band seems to be originating

from framework vibrations which are influenced by the Ti sites. The saturation

of typical cyclohexenol signals (1064 - 1041 cm"1) for the aerogel with 20 wt%

nominal Ti02 (2OTi), indicates an irreversible adsorption on the catalyst. On

the other hand, the appearance of a stronger negative band between 890 and

810 cm" for this catalyst also points to a strong but reversible interaction with

the catalyst framework, thus revealing different adsorption behaviour on diffe¬

rent sites of the catalyst. A second prominent signal next to the typical

cyclohexenol band at 953 cm" appears at 958 cm" for the adsorption on the

catalysts lOTi and 20Ti. This corroborates the assumption of an interaction

with the active site, since a similar behaviour could be found for adsorption of

TBHP [231].

When cyclooctenol is adsorbed on the catalysts, the spectra reveal an inter¬

action of the surface silanols with the hydroxy group of the allylic alcohol

(Fig. 5-4). For the lOTi and 20Ti aerogels, these interactions seem to be stron¬

ger than for the other two catalysts, which can be seen by the higher intensity

of the negative band at 3700 cm'. Also, the framework vibrations seem to have

bigger influence on the ATR spectra which underlines the assumption of a

stronger interaction between cyclooctenol and these catalysts. For 1 OTi-Me and

OTi aerogels on the other hand, the signals of cyclooctenol seems to be stronger,

for example, a weak signal at 1131 cm' (not shown) can only be observed for

these two catalysts. This hints to more reversible adsorption and desorption of

cyclooctenol on these catalysts with consequently less saturation of the signals.The molecules appearing at the interface therefore dominate the spectra.

Compared to the adsorption experiments with cyclohexenol, cyclooctenol

seems to have a more reversible and weaker interaction with all catalysts.

In the epoxidation experiments the difference of the adsorption becomes

more obvious. When the TBHP concentration is modulated in the epoxidation

of cyclohexenol (Fig. 5-5, bottom), the observations reveal a stronger adsorp¬

tion of cyclohexenol on the surface and the active sites compared to TBHP. The

lack of negative signals originating from the allylic alcohol (e.g. 953 or

727 cm'1) and typical framework bands observed for TBHP adsorption (1020,

944 cm' ) indicate that the catalyst remains mainly static when TBHP concen-

ATR Studies on Epoxidation of Cyclic Allylic Alcohols 103

tration is modulated. Hence, no significant displacement of cyclohexenol by

TBHP could be detected, except when the peroxide was flown through the cell

for a very long period. Even interaction with surface silanol groups could not

be detected. Interestingly, the framework vibrations seem to react with a delay

to the appearance ofTBHP on the surface, which corroborates the hindrance

of TBHP adsorption by cyclohexenol. Therefore it is not surprising that no

cyclohexenol oxide could be detected in the spectra and only traces were

discernible in the GC analysis. When modulating the concentration of

cyclohexenol (Fig. 5-5, top), the typical spectrum of the allylic alcohol can

clearly be observed. Furthermore the negative bands of TBHP indicate a

displacement of the peroxide. This clearly shows a stronger adsorption on the

catalyst for cyclohexenol with respect to TBHP. Since this experiment showed a

higher concentration of cyclohexenol oxide in the effluent solution and its

intensity even marked a slight dependence on the cyclohexenol concentration,

the shoulder at 999 cm'1 and the weak signals at 944, 931 and 853 cm'1 may

indicate cyclohexenol oxide formation. This is strongly supported by the top

most spectrum (bold) in the range between 1070 - 965 cm' of cyclohexenol

oxide over 1 OTi-Me (Fig. 5-5, top). Interestingly, when the experiment was

repeated 5 min later, the ATR spectrum revealed a clear deactivation since both

the cyclohexenol signals as well as the negative TBHP bands were much weaker

(Fig. 5-9). Figure 5-9 shows that in the second experiment the bands are in

general much weaker. As only dynamic changes are visible in the demodulated

spectra the comparison in Fig. 5-9 shows that in the second experiments the

catalyst was less active. This corroborates the findings of Beck et al. [150] who

observed a fast catalyst deactivation for the epoxidation of cyclohexenol. They

also suggested, that the allylic alcohol blocks the active sites and therefore deac¬

tivates the catalyst. This is fully supported by our experiments described above

and also by variable temperature experiments where the catalyst was slowly

heated. GC analysis of the effluent reaction solution showed the highest con¬

centration of cyclohexenol oxide at 75°C, although best performances were

reported for 90°C [93].

In the case of cyclooctenol epoxidation, the adsorption behaviour of the

allylic alcohol plays a different role. When the TBHP concentration was modu¬

lated (Fig 5-6, bottom), a displacement of cyclooctenol could clearly be moni-

104 Chapter 5

1300 1200 1100 1000 900 800

Wavenumber / cm"1

Fig. 5-9: Demodulated ATR spectra of epoxidation experiments recorded in the regionbetween 1380 and 780 cm"1. The concentration of cyclohex-2-en-l-ol was periodicallymodulated between 0 and 3 mmol/1 in toluene at 80°C. The second experiment (bottom)

was started 5 min after the first experiment (top) was finished.

tored by the two negative bands at 751 and 713 cm".The negative bands of

the silanol groups at 3700 and 980 cm"1 and the signal at 1260 cm"1 originat¬

ing from TBHP coordinating to the active Ti site [231] indicates a weaker

adsorption of the allylic alcohol on the catalyst surface and active sites than

observed for cyclohexenol. When investigating the scans taken before starting

the epoxidation experiment, it can clearly be observed that the silanol signal at

3700 cm" increased before decreasing again when TBHP concentration is

raised (vide infra). This indicates a desorption of cyclooctenol before TBHP

adsorbs on the surface silanols. Therefore coordination and activation of the

peroxide at the Ti active sites is possible and the formation of the epoxide is

O

Ö

oT

<

ATR Studies on Epoxidation of Cyclic Allylic Alcohols 105

evident. GC analysis showed a clear periodic increase and decrease of cyclo¬

octenol oxide concentration. The phase lag of 110° for the two TBHP signals

at 1250 and 1260 cm"1 reveals the rate determining factor. Since the first band

indicates the adsorption on the surface silanol groups and the latter the adsorp¬

tion at the Ti centers, the phase lag reflects the time needed for the pore diffu¬

sion and/or coordination to the active site. The phase lag for the TBHP signal

at 1250 cm"1 and the cyclooctenol oxide signal at 881 cm"1 is 140°, which is

smaller than was found for the epoxidation with cyclohexene (180°) [231].

Modulation of the cyclooctenol concentration reveals a strong adsorption

on the catalyst and a displacement ofTBHP not only by consumption of the

latter on the active sites (Fig. 5-6, top), but also on the residual surface. The

strong negative band at 3700 cm' indicates adsorption of cyclooctenol on the

catalyst surface via silanol groups. Note that in the corresponding experiment

using cyclohexene as substrate, this band was not discernible due to the weak

interaction with the catalyst surface and the blocking of the silanol sites byTBHP [231]. The phase lag for the cyclooctenol signal at 713 cm'1 and the one

of the epoxide at 881 cm'1 is 90° and therefore smaller than the corresponding

phase lag observed for the TBHP concentration modulation. This is probably

due to the fact, that TBHP is already coordinated to the Ti site and therefore

can react with the allylic alcohol as soon as the latter is adsorbed on the active

site.

Compared to the experiment with cyclohexene, where the observed phase

lag was 150° [231], the cyclooctenol oxide seems to be formed faster once the

allylic alcohol is detected on the surface. This may be surprising at first glance,since cyclohexene has a smaller Van der Waals radius than cyclooctenol and

most important, does not reveal any strong interaction with the catalysts sur¬

face. Hence the olefin should diffuse more rapidly into the pores, resulting in

shorter travelling time from the catalyst surface to the active site [175,231]. Using

an aerogel with 5 wt% nominal Ti02, Beck et al. found higher TONs for

cyclooctenol than for cyclohexene, for a reaction time of 2 min as well as for

2 hrs [150]. The faster formation of the cyclooctenol oxide may be traced to

the dominance of a silanol-assisted epoxidation mechanism where the allylic

alcohol adsorbs on a surface silanol group adjacent to the Ti site or vicinal to it

(s. Scheme 5-1) [125-127,230].

106 Chapter 5

In the case of cyclohexenol, the dative bond of the oxygen of the allylic

alcohol to the active site as described by Kumar et al. for TS-1 [133] is much

stronger mainly due to the smaller steric demand of the molecule compared to

cyclooctenol.

Our experiments revealed catalyst deactivation and reduced accessibility

for TBHP on the surface and the active site in the case of cyclohexenol, which

explains the slower formation of cyclohexenol oxide. The prevalence of the

hydroxy-assisted mechanism for cyclohexenol epoxidation is also indicated by

the stereoselectivity of the formed epoxide. GC analysis of the effluent solution

revealed formation of aV-cyclohexenol oxide as the main product, whereas with

cyclooctenol mainly the ^raws-epoxide was observed. A close and dative bond¬

ing to the Ti site leads to a hydroxy-assisted epoxidation mechanism which

favours the formation of a aV-epoxide.

5.5 Conclusions

ATR infrared combined with modulation spectroscopy in a flow-through cell

was used to study the interaction of cyclohex-2-en-l-ol and cyclooct-2-en-l-ol

with Ti-Si aerogel catalysts as well as the epoxidation of the two allylic alcohols

by tert-butyl hydroperoxide (TBHP). The phase-sensitive detection afforded an

enhanced sensitivity which is required to study the relevant small changes due

to the periodic perturbation of this catalytic system by forced concentration

modulation. For both allylic alcohols interaction with surface silanol groups

and Ti sites could be observed. Cyclohexenol revealed a stronger and less revers¬

ible adsorption on aerogels compared to cyclooctenol. No displacement of

cyclohexenol was observed when the peroxide concentration was modulated,

whereas TBHP desorption could be observed when the concentration of the

allylic alcohol was changed. The strong and irreversible adsorption of cyclo¬

hexenol leads to catalyst deactivation by blocking the active Ti sites. When

cyclooctenol was used as substrate, epoxide formation could clearly be

observed. Modulating the TBHP concentration revealed a clear displacement

of the allylic alcohol. Even activation of the peroxide by coordination at the

ATR Studies on Epoxidation of Cyclic Allylic Alcohols 107

active site was discernible. A time (phase) lag between the appearance of reac¬

tant and product in the volume probed by the evanescent field was observed.

The main part of this phase lag is caused by diffusion and/or adsorption at the

active site of TBHP. When the cyclooctenol concentration was modulated, a

displacement of peroxide was discernible. Also, a phase lag between the appear¬

ance of substrate and epoxide was observed, which was smaller than in the

experiment where the TBHP concentration was modulated.

Interaction of the hydroxy group of cyclohexenol with the active site is very

strong and therefore mainly aV-epoxide is formed. For the cyclooctenol, the

interaction of the hydroxy group with Ti sites plays a less important role in the

epoxidation process, due to steric hindrance. Interaction with silanol groups

adjacent to the active Ti sites is favored which leads to the trans-epoxlde as

main product.

Chapter

References

[I] C. L. Stevens, J. Tazuma,/. Am. Chem. Soc. 76, 715 (1954).

[2] E. E. Royals, L. L. Harrell,/ Am. Chem. Soc. 77, 3405 (1955).

[3] W Dittmann, F. Stiirzenhofecker, liebigs Ann. Chem. 688, 57 (1965).

[4] F. Camps, J. Coll, A. Messeguer, M. A. Pericas, Tetrahedron lett. 22,

3895 (1981).

[5] H. B. Henbest, R. A. L. Wilson,/ Chem. Soc. 1958 (1957).

[6] S. Tanaka, H. Yamamoto, H. Nozaki, K. B. Sharpies, R. C. Michaels,

J. D. Cutting,/ Am. Chem. Soc. 96, 5254 (1974).

[7] K. B. Sharpless, R. C. Michaels,/ Am. Chem. Soc. 95, 6136 (1973).

[8] W Adam, A. Corma, A. Martinez, C. M. Mitchell, T I. Reddy, M.

Renz, A. K. Smerz,/ Mol Catal A: Chem. 117, 357 (1997).

[9] N. A. Milas, / Am. Ceram. Soc. 59, 342 (1937).

[10] N. A. Milas, L. S. Maloney,/ Am. Ceram. Soc. 62, 841 (1940).

[II] N. A. Milas, S. Sussman,/ Am. Ceram. Soc. 58, 302 (1936).

[12] G. B. Payne, C. W Smith,/ Org Chem. 22, 1682 (1957).

[13] N. Indictor, W. F. Brill,/ Org. Chem. 30, 2074 (1965).

[14] T Itoh, K. Jitsukawa, K. Kaneda, S. Teranishi,/ Am. Chem. Soc. 101,

159 (1979).

[15] R. A. Sheldon, J. A. van Doom,/ Catal 31, 427 (1973).

[16] T. Katsuki, K. B. Sharpless,/ Am. Chem. Soc. 102, 5974 (1980).

[17] T. Katsuki,/ Mol Catal A: Chem. 113, 87 (1996).

[18] E. N. Jacobsen, in Comprehensive Organometallic Chemistry II, Vol. 12

(Eds.: E. W Abel, F. G. A. Stone, E. Wilkinson), Pergamon, New York,

(1995), p. 1097.

[19] T Maschmeyer, F. Rey, G. Sankar, J. M. Thomas, Nature 378, 159

(1995).

110 Chapter 6

[20] A. Corma, A. Fuerte, M. Iglesias, F. Sanchez, / Mol. Catal. A: Chem.

107,225(1996).

[21] A. Heckel, D. Seebach, Angew. Chem., Int. Ed. Engl 39, 163 (2000).

[22] R. Neumann, H. Miller,/ Chem. Soc, Chem. Commun. 2277 (1995).

[23] F. Minutolo, D. Pini, A. Petri, P. Salvadori, Tetrahedron: Asymmetry 7,

2293 (1996).

[24] S. Julia, J. Masana, J. C. Vega, Angew. Chem. Int. Ed. Engl. 19, 929

(1980).

[25] B. Altava, M. I. Burguete, B. Escuder, S. V. Luis, R. V. Salvador, J. M.

Fraile, J. A. Mayoral, A. J. Royo,/ Org Chem. 62, 3126 (1997).

[26] H. Sellner, D. Seebach, Angew. Chem., Int. Ed. Engl 38, 1918 (1999).

[27] B. E. Rossiter, m Asymmetric Synthesis (Ed.: J. D. Morrison), Academic

Press, Orlando, (1985), p. 193.

[28] M. J. Farral, M. Alexis, M. Tracarten, Nouv. J. de Chim. 7, 449 (1983).

[29] B. M. Choudary, V. L. K. Valli, A. D. Prasad, / Chem. Soc., Chem.

Commun. 1186(1990).

[30] R. I. Kureshy, N. H. Khan, S. H. R. Abdi, P. Iyer, React. Funct. Polym.

34, 153 (1997).

[31] B. B. De, B. B. Lohray, S. Sivaram, P. K. Dhal, Tetrahedron: Asymmetry

6, 2105 (1995).

[32] E. Breysse, C. Pinel, M. Lemaire, Tetrahedron: Asymmetry 9, 897

(1998).

[33] T. Yokoyama, M. Nishizawa, T. Kimura, T. M. Suzuki, Chem. lett.

1703 (1983).

[34] M. M. Dell'Anna, P. Mastrorilli, C. F. Nobile, G. P. Suranna, / Mol

Catal 103, 17 (1995).

[35] I. Arends, R. A. Sheldon, M. Wallau, U. Schuchardt, Angew. Chem.,

Int. Ed. Engl 36, 1144 (1997).

[36] G. Olason, D. C. Sherrington, Macromol Symp. 131, 127 (1998).

[37] F. Wattimena, H. P. Wulff, British Patent 1249'079, 1971, Shell Oil

Company.

[38] R. A. Sheldon, I. W. C. E. Arends, H. E. B. Lempers, in Supported

Reagents and Catalysts in Chemistry (Eds.: B. K. Hodnett, A. P. Kybett,

References 111

H. H. Clark, K. Smith), Royal Society of Chemistry, Limerick, (1997),

p. 37.

[39] R. Hutter, T Mallat, A. Baiker,/ Catal 153, 111 (1995).

[40] R. Hutter, T Mallat, D. Dutoit, A. Baiker, Top. Catal 3, 421 (1996).

[41] R. A. van Santen, H. P. C. E. Kuipers, Adv. Catal 35, 265 (1987).

[42] S. Imamura, H. Sasaki, M. Shono, H. Kanai,/ Catal 177, 72 (1998).

[43] M. Taramasso, Perego, G., Notari, B., USPatent 4410501, 1983,

[44] R. A. Sheldon, Top. Curr. Chem. 164, 21 (1993).

[45] C. B. Khouw, C. B. Dartt, J. A. Labinger, M. E. Davis,/ Catal 149,

195 (1994).

[46] E. Höft, H. Kosslick, R. Fricke, H. J. Hamann,/ Prakt. Chem. 338, 1

(1996).

[47] G. Bellussi, A. Carati, M. G. Clerici, A. Esposito, R. MiUini,

F. Buonomo, Belgian Patent 1,001,038, 1989,

[48] G. Bellussi, A. Carati, M. G. Clerici, G. Maddinelli, R. MiUini,

/ Catal 133,220(1992).

[49] T H. Chang, F. C. Leu, Zeolites 15, 496 (1995).

[50] J. Kornatowski, M. Sychev, S. Kuzenkov, K. Strnadova, W Pilz,

D. Kassner, G. Pieper, W H. Baur, / Chem. Soc., Faraday Trans. 91,

2217(1995).

[51] R. Joseph, M. Sasidharan, R. Kumar, A. Sudalai, T Ravindranathan,

/ Chem. Soc., Chem. Commun. 1341 (1995).

[52] M. A. Camblor, A. Corma, A. Martinez, J. Perezpariente, / Chem. Soc.,

Chem. Commun. 589 (1992).

[53] A. Corma, P. Esteve, A. Martinez, S. Valencia,/ Catal 152, 18 (1995).

[54] T Sato, J. Dakka, R. A. Sheldon,/ Chem. Soc., Chem. Commun. 1887

(1994).

[55] A. Corma, M. T Navarro, J. P. Pariente, / Chem. Soc., Chem.

Commun. 141 (1994).

[56] P. T Tanev, M. Chibwe, T J. Pinnavaia, Nature 368, 321 (1994).

[57] K. A. Koyano, T. Tatsumi, Chem. Commun. 145 (1996).

[58] K. Vercruysse, D. M. Klingeleers, T Colling, P. A. Jacobs, Stud. Surf.

Sei. Catal 117,469 (1998).

112 Chapter 6

[59] A. Corma, M. Iglesias, F. Sanchez, / Chem. Soc., Chem. Commun.

1635 (1995).

[60] M. van Klaveren, R. A. Sheldon, Stud. Surf Sei. Catal. 110, 561

(1997).

[61] M. S. Rigutto, H. Vanbekkum,/ Mol Catal 81,11 (1993).

[62] M. J. Haanepen, J. H. C. vanHooff, Appl Catal, A 152, 183 (1997).

[63] J. A. Martens, P. A. Jacobs, m AdvancedZeolite Science andApplications,Vol 85, (1994), p. 653.

[64] W F. Maier, J. A. Martens, S. Klein, J. Heilmann, R. Parton, K.

Vercruysse, P. A. Jacobs, Angew. Chem., Int. Ed. Engl. 35, 180 (1996).

[65] R. Hutter, T Mallat, A. Baiker, / Chem. Soc., Chem. Commun. 2487

(1995).

[66] R. Hutter, T Mallat, A. Baiker,/ Catal 157, 665 (1995).

[67] S. Imamura, T Nakai, H. Kanai, T Ito, Catal lett. 28, 277 (1994).

[68] S. Imamura, T Nakai, H. Kanai, T Ito, / Chem. Soc., Faraday Trans.

91, 1261 (1995).

[69] S. Imamura, T Nakai, H. Kanai, T Shiono, K. Utani, Catal. lett. 39,

79 (1996).

[70] S. Klein, S. Thorimbert, W F. Maier,/ Catal 163, 416 (1996).

[71] Z. F. Liu, G. M. Crumbaugh, R. J. Davis,/ Catal 159, 83 (1996).

[72] Z. F. Liu, R. J. Davis,/ Phys. Chem. 98, 1253 (1994).

[73] G. Dagan, S. Sampath, O. Lev, Chem. Mater. 7, 446 (1995).

[74] D. C. M. Dutoit, M. Schneider, R. Hutter, A. Baiker, / Catal. 161,

651 (1996).

[75] D. C. M. Dutoit, M. Schneider, A. Baiker,/ Catal 153, 165 (1995).

[76] K. L. Walther, A. Wokaun, B. E. Handy, A. Baiker,/ Non-Cryst. Solids

134,41(1992).

[11] P. K. Doolin, S. Alerasool, D. J. Zalewski, J. F. Hoffman, Catal. lett.

25,209(1994).

[78] H. Nakabayashi, Bull Chem. Soc.Jpn. 65, 914 (1992).

[79] B. M. Reddy, E. P. Reddy, B. Manohar, Appl Catal, A 96, LI (1993).

[80] A. Y Stakheev, E. S. Shpiro, J. Apijok,/ Phys. Chem. 97, 5668 (1993).

[81] C. H. Hung, J. L. Katz,/ Mater. Res. 7, 1861 (1992).

References 113

[82] R. B. Greegor, F. W. Lytle, D. R. Sandstrom, J. Wong, P. Schultz,

/ Non-Cryst. Solids 55, 27 (1983).

[83] J. B. Miller, S. T. Johnston, E. I. Ko,/ Catal 150, 311 (1994).

[84] M. Aizawa, Y. Nosaka, N. Fujii, / Non-Cryst. Solids 128, 77 ( 1991 ).

[85] M. Schraml-Marth, K. L. Walther, A. Wokaun, B. E. Handy, A. Baiker,

/ Non-Cryst. Solids 143, 93 (1992).

[86] C. J. Brodsky, E. I. Ko,/ Non-Cryst. Solids 186, 88 (1995).

[87] M. Schneider, A. Baiker, Catal Rev.-Sci. Eng 37,515 (1995).

[88] J. B. Miller, E. I. Ko, Catal Today 35, 269 (1997).

[89] M. Dusi, T. Mallat, A. Baiker,/ Catal 187, 191 (1999).

[90] S. Klein, W. F. Maier, Angew. Chem., Int. Ed. Engl 35, 2230 (1996).

[91] H. Kochkar, F. Figueras,/ Catal 171, 420 (1997).

[92] E. Lotero, D. Vu, C. Nguyen, J. Wagner, G. Larsen, Chem. Mater. 10,

3756 (1998).

[93] C. A. Müller, M. Maciejewski, T. Mallat, A. Baiker,/ Catal 184, 280

(1999).

[94] B. Notari, Adv. Catal 41, 253 (1996).

[95] A. Corma, M. Domine, J. A. Gaona, J. L. Jorda, M. T. Navarro, F. Rey,

J. Perez-Pariente, J. Tsuji, B. McCulloch, L. T. Nemeth, Chem. Com¬

mun. 2211 (1998).

[96] A. Corma, J. L. Jorda, M. T. Navarro, F. Rey, Chem. Commun. 1899

(1998).

[97] M. A. Camblor, A. Corma, P. Esteve, A. Martinez, S. Valencia, Chem.

Commun. 795 (1997).

[98] G. L. Robbins, R. E. Tapscott, Inorg Chem. 15, 154 (1976).

[99] H. Mimoun, P. Chaumette, M. Mignard, L. Saussine, J. Fischer,

R. Weiss, Nouv. J. de Chim. 7, 461 (1983).

[100] K. B. Sharpless, S. S. Woodard, M. G. Finn, Pure Appl Chem. 55,

1823 (1983).

[101] C. J. Burns, C. A. Martin, K. B. Sharpless, / Org Chem. 54, 2826

(1989).

[102] P. G. Potvin, P. C. C. Kwong, M. A. Brook,/ Chem. Soc., Chem. Com¬

mun. 113 (1988).

114 Chapter 6

103] R. D. Bach, B. A. Coddens, G. J. Wolber, / Org Chem. 51, 1030

(1986).

104] K. A. j0rgensen, R. A. Wheeler, R. Hoffmann,/ Am. Chem. Soc. 109,

3240 (1987).

105] A. Zecchina, S. Bordiga, G. Spoto, L. Marchese, G. Petrini,

G. Leofanti, M. Padovan,/ Phys. Chem. 96, 4985 (1992).

106] S. Bordiga, S. Coluccia, C. Lamberti, L. Marchese, A. Zecchina,

F. Boscherini, F. Buffa, F. Genoni, G. Leofanti, G. Petrini, G. Vlaic,

/ Phys. Chem. 98, 4125 (1994).

107] C. Lamberti, S. Bordiga, A. Zecchina, A. Carati, A. N. Fitch,

G. Artioli, G. Petrini, M. Salvalaggio, G. L. Marra, / Catal. 183, 222

(1999).

108] D. T. On, L. LeNoc, L. Bonneviot, Chem. Commun. 299 (1996).

109] G. Deo, A. M. Turek, I. E. Wachs, D. R. C. Huybrechts, P. A. Jacobs,

Zeolites 13,365 (1993).

110] G. Tozzola, M. A. Mantegazza, G. Ranghino, G. Petrini, S. Bordiga,G. Ricchiardi, C. Lamberti, R. Zulian, A. Zecchina, / Catal. 179, 64

(1998).

111] A. Zecchina, S. Bordiga, G. Spoto, L. Marchese, G. Petrini,

G. Leofanti, M. Padovan,/ Phys. Chem. 96, 4991 (1992).

112] G. Ricchiardi, A. Damin, S. Bordiga, C. Lamberti, G. Spano,

F. Rivetti, A. Zecchina,/ Am. Chem. Soc. 123, 11409 (2001).

113] P. J. Dirken, M. E. Smith, H. J. Whitfield, / Phys. Chem. 99, 395

(1995).

114] M. E. Smith, H. J. Whitfield,/ Chem. Soc., Chem. Commun. 723

(1994).

115] V Bolis, S. Bordiga, C. Lamberti, A. Zecchina, A. Carati, F. Rivetti,

G. Spano, G. Petrini, langmuir 15, 5753 (1999).

116] V Bolis, S. Bordiga, C. Lamberti, A. Zecchina, A. Carati, F. Rivetti,

G. Spano, G. Petrini, Microporous Mesoporous Mater. 30, 61 (1999).

117] R. MiUini, E. P. Massara, G. Perego, G. Bellussi,/ Catal 137, 491

(1992).

118] C. Lamberti, G. T. Palomino, S. Bordiga, D. Arduino, A. Zecchina,

G. Vlaic, Jpn.J. Appl Phys., Part 1 38, 55 (1999).

References 115

119] A. Zecchina, S. Bordiga, C. Lamberti, G. Ricchiardi, D. Scarano,

G. Petrini, G. Leofanti, M. Mantegazza, Catal. Today 32, 97 (1996).

120] A. Fernandez, J. Leyrer, A. R. Gonzalezelipe, G. Munuera,

H. Knozinger,/ Catal 112, 489 (1988).

121] M. G. Clerici, P. Ingallina,/ Catal 140, 71 (1993).

122] P. E. Sinclair, C. R. A. Catlow,/ Phys. Chem. B 103, 1084 (1999).

123] M. Fujiwara, H. Wessel, P. Hyung-Suh, H. W Roesky, Tetrahedron 58,

239 (2002).

124] G. Boche, K. Mobus, K. Harms, M. Marsch,/ Am. Chem. Soc. 118,

2770 (1996).

125] M. Crocker, R. H. M. Herold, A. G. Orpen, M. T A. Overgaag, Chem.

Soc., Dalton Trans. 3791 (1999).

126] M. C. Klunduk, T Maschmeyer, J. M. Thomas, B. F. G. Johnson,

Chem. Eur. J. 5, 1481 (1999).

127] T Maschmeyer, M. C. Klunduk, C. M. Martin, D. S. Shephard,

J. M. Thomas, B. F. G. Johnson, Chem. Commun. 1841 (1997).

128] R. D. Oldroyd, J. M. Thomas, T Maschmeyer, P. A. McFaul,

D. W Snelgrove, K. U. Ingold, D. D. M. Wayner, Angew. Chem. 108,

2966 (1996).

129] W Adam, A. Corma, T I. Reddy, M. Renz, / Org Chem. 62, 3631

(1997).

130] C. Lamberti, S. Bordiga, M. Salvalaggio, G. Spoto, A. Zecchina, F.

Geobaldo, G. Vlaic, M. Bellatreccia,/ Phys. Chem. B 101, 344 (1997).

131] L. LeNoc, D. T. On, S. Solomykina, B. Echchahed, F. Beland, C. C. D.

Moulin, L. Bonneviot, in 11th International Congress on Catalysis - 40th

Anniversary, Pts a andB, Vol. 101, (1996), p. 611.

132] W Adam, C. M. Mitchell, C. R. Saha-Moller, Eur. J. Org Chem. 785

(1999).

133] R. Kumar, G. C. G. Pais, B. Pandey, P. Kumar,/ Chem. Soc, Chem.

Commun. 1315 (1995).

134] J. Klaas, G. SchulzEkloff, N. I. Jaeger, / Phys. Chem. B 101, 1305

(1997).

116 Chapter 6

135] C. Li, G. Xiong, Q. Xin, J. K. Liu, P. L. Ying, Z. C. Feng, J. Li, W B.

Yang, Y Z. Wang, G. R. Wang, X. Y Liu, M. Lin, X. Q. Wang,

E. Z. Min, Angew. Chem., Int. Ed. Engl 38, 2220 (1999).

136] S. Klein, B. M. Weckhuysen, J. A. Martens, W F. Maier, P. A. Jacobs,

/ Catal 163,489(1996).

137] X. T Gao, S. R. Bare, J. L. G. Fierro, M. A. Banares, I. E. Wachs,

/ Phys. Chem. B 102, 5653 (1998).

138] Z. F. Liu, J. Tabora, R. J. Davis,/ Catal 149, 117 (1994).

139] R. J. P. Corriu, D. Leclercq, Angew. Chem., Int. Ed. Engl. 35, 1420

(1996).

140] M. Dusi, T Mallat, A. Baiker,/ Mol Catal A: Chem. 138, 15 (1999).

141] A. Molnàr, M. Bartok, M. Schneider, A. Baiker, Catal. lett. 43, 123

(1997).

142] T. Kataoka, J. A. Dumesic,/ Catal 112, 66 (1988).

143] C. Contescu, V T. Popa, J. B. Miller, E. I. Ko, J. A. Schwarz,/ Catal.

157,244(1995).

144] R. Hutter, T. Mallat, A. Peterhans, A. Baiker,/ Mol Catal A: Chem.

138,241 (1999).

145] M. Primer, P. Pichat, M. V Mathieu,/ Phys. Chem. 75, 1216 (1971).

146] M. Primer, P. Pichat, M. V Mathieu,/ Phys. Chem. 75, 1221 (1971).

147] M. Dusi, T. Mallat, A. Baiker,/ Catal 173, 423 (1998).

148] J. Bu, H. K. Rhee, Catal lett. 66, 245 (2000).

149] M. L. Pena, V Dellarocca, F. Rey, A. Corma, S. Coluccia, L. Marchese,

Microporous Mesoporous Mater. 44, 345 (2001).

150] C. Beck, T. Mallat, A. Baiker, Catal lett. 88, 203 (2003).

151] C. Beck, T. Mallat, A. Baiker, New]. Chem. 27, 1284 (2003).

152] J. Ryczkowski, Catal. Today 68, 263 (2001).

153] C. J. Hirschmugl, Surf Sei. 500, 577 (2002).

154] M. J. Weaver, S. Zou, m Advances in Spectroscopy, Vol. 26 (Eds.: R. J. H.

Clark, R. E. Hester), Wiley, Chichester, U.K., (1998), p. 219.

155] A. Campion, P. Kambhampati, Chem. Soc. Rev. 27, 241 (1998).

156] A. M. Taylor, A. M. Glover,/ Opt. Soc Am. 23, 206 (1933).

157] A. M. Taylor, D. A. Durfee,/ Opt. Soc Am. 23, 263 (1933).

158] A. M. Taylor, A. King,/ Opt. Soc Am. 23, 308 (1933).

References 117

159] N. J. Harrick,/ Opt. Soc Am. 49, 316 (1959).

160] R. P. Eischens,/ Phys. Chem. Solids 14, 56 (1960).

161] J. Fahrenfort, Spectrochim. Acta 17, 698 (1961).

162] N. J. Harrick, Internal reflection spectroscopy, Interscience, New York,

(1967).

163] D. Baurecht, U. P. Fringeli, Rev. Sei. Instrum. 72, 3782 (2001).

164] T. Bürgi, A. Baiker,/ Phys. Chem. B 106, 10649 (2002).

165] K. D. Dobson, A. J. McQuillan, Phys. Chem. Chem. Phys. 2, 5180

(2000).

166] K. D. Dobson, A. J. McQuillan, Spectrochim. Acta, Part A 56, 557

(2000).

167] K. D. Dobson, A. D. Roddick-Lanzilotta, A. J. McQuillan,

Vib. Spectrosc 24, 287 (2000).

168] R. P. Sperline, Y Song, H. Freiser, langmuir 8, 2183 (1992).

169] R. P Sperline, Y Song, H. Freiser, langmuir 10, 37 (1994).

170] S. J. Hug, B. Sulzberger, langmuir 10, 3587 (1994).

171] A. Couzis, E. Gulari, langmuir 9, 3414 (1993).

172] K. D. Dobson, P. A. Connor, A. J. McQuillan, langmuir 13, 2614

(1997).

173] I. Ortiz-Hernandez, C. T Williams, langmuir 19, 2956 (2003).

174] D. Rivera, J. M. Harris, langmuir 17, 5527 (2001).

175] D. Rivera, J. M. Harris, Anal Chem. 73,411 (2001).

176] R. Nakamura, A. Imanishi, K. Murakoshi, Y Nakato, / Am. Chem.

Soc 125,1443 (2003).

177] D. Ferri, T Bürgi, A. Baiker, / Phys. Chem. B 105, 3187 (2001 ).

178] D. Ferri, T Bürgi,/ Am. Chem. Soc 123, 12014 (2001).

179] D. Ferri, T Bürgi, A. Baiker, Phys. Chem. Chem. Phys. 4, 2661 (2002).

180] R. Wirz, T Bürgi, A. Baiker, langmuir 19, 785 (2003).

181] N. M. B. Flichy, S. G. Kazarian, C. J. Lawrence, B. J. Briscoe,/ Phys.Chem. B 106, 154 (2002).

[182] M. S. Schneider, J. D. Grunwaldt, T Bürgi, A. Baiker, Rev. Sei.

Instrum. 74,4121 (2003).

118 Chapter 6

[183] M. P. Schürch, "Heterogene enantioselektive Hydrierung von akti¬

vierten Carbonyl-Verbindungen", Doctoral dissertation #12799,

ETH (Zürich), 1998.

[184] D. C. M. Dutoit, "Silica Based Mixed Oxide Aerogel Catalysts",

Doctoral dissertation #11674, ETH (Zürich), 1996.

[185] X. T. Gao, I. E. Wachs, Catal Today 51, 233 (1999).

[186] R. J. Davis, Z. F. Liu, Chem. Mater. 9, 2311 (1997).

[187] C. J. Brinker, G. W Scherer, Sol-Gel Science, Academic Press, Inc.,

Boston, (1990).

[188] K. Kosuge, P. S. Singh,/ Phys. Chem. B 103, 3563 (1999).

[189] F. Figueras, H. Kochkar, Catal lett. 59, 79 (1999).

[190] M. B. D'Amore, S. Schwarz, Chem. Commun. 121 (1999).

[191] J. L. Sotelo, R. Van Grieken, C. Martos, Chem. Commun. 549 (1999).

[192] M. Chiari, M. Nesi, P. G. Righetti, Surface Modification ofSilica Walls:

A Review ofDifferent Methodologies, CRC, Boca Raton, (1996).

[193] C. Gründling, G. EderMirth, J. A. Lercher,/ Catal 160, 299 (1996).

[194] S. V Slavov, K. T Chuang, A. R. Sanger,/ Phys. Chem. 100, 16285

(1996).

[195] J. Grobe, Organosilicon Chem. II591 (1994).

[196] C. A. Müller, M. Schneider, T. Mallat, A. Baiker, Appl Catal, A 201,

253 (2000).

[197] C. A. Müller, M. Schneider, A. Gisler, T Mallat, A. Baiker, Catal. lett.

64, 9 (2000).

[198] C. A. Müller, M. Schneider, T. Mallat, A. Baiker, / Catal 189, 221

(2000).

[199] L. Canali, D. C. Sherrington, Chem. Soc Rev. 28, 85 (1999).

[200] P. M. Price, J. H. Clark, D. J. Macquarrie, Chem. Soc, Dalton Trans.

101 (2000).

[201] C. Bianchini, P. Barbara, Top. Catal 19, 17 (2002).

[202] M. O. Lorenzo, V Humblot, P. Murray, C. J. Baddeley, S. Haq,

R. Raval,/ Catal 205, 123 (2002).

[203] M. Schunack, E. Laegsgaard, I. Stensgaard, I. Johannsen,

F. Besenbacher, Angew. Chem., Int. Ed. Engl. 40, 2623 (2001).

References 119

[204] L. A. Nafie, T. A. Keiderling, P. J. Stephens, / Am. Chem. Soc 98, 2715

(1976).

[205] L. A. Nafie, Annu. Rev. Phys. Chem. 48, 357 (1997).

[206] J. C. P. Broekhoff, Preparation of Heterogenous Catalysts II, Elsevier,

Amsterdam, (1979).

[207] A. B. Jarzebski, L. Pajak,/ Non-Cryst. Solids204, 172 (1996).

[208] T Bürgi, Chimia 57, 623 (2003).

[209] K. L. Walther, A. Wokaun, A. Baiker, Mol Phys. 71, 169 (1990).

[210] M. Dusi, C. Beck, T Mallat, A. Baiker, in Catalysis ofOrganic Reactions

(Ed.: M. E. Ford), Marcel Dekker Inc., New York, (2000), ch. 7.

[211] J. N. Kondo, E. Yoda, H. Ishikawa, F. Wakabayashi, K. Domen,

/ Catal 191, 275 (2000).

[212] S. J. Tavener, J. H. Clark, G. W Gray, P. A. Heath, D. J. Macquarrie,

Chem. Commun. 1147 (1997).

[213] M. Dusi, T. Mallat, A. Baiker, Catal Rev.-Sci. Eng 42,213 (2000).

[214] A. Corma, M. A. Camblor, P. Esteve, A. Martinez, J. Perezpariente,

/ Catal 145, 151 (1994).

[215] D. P. Serrano, H. X. Li, M. E. Davis,/ Chem. Soc, Chem. Commun.

145 (1992).

[216] N. Ulagappan, V Krishnasamy, / Chem. Soc, Chem. Commun. 373

(1995).

[217] P. Wu, T Tatsumi, T Komatsu, T Yashima, Chem. lett. 114 (2000).

[218] D. Scarano, A. Zecchina, S. Bordiga, F. Geobaldo, G. Spoto,

G. Petrini, G. Leofanti, M. Padovan, G. Tozzola,/ Chem. Soc, FaradayTrans. 89,4123 (1993).

[219] A. S. Soult, D. F. Carter, H. D. Schreiber, L. J. van de Burgt,

A. E. Stiegman,/ Phys. Chem. B 106, 9266 (2002).

[220] D. Ferri, T Bürgi, A. Baiker, Chem. Commun. 1172 (2001).

[221] C. Keresszegi, T Bürgi, T. Mallat, A. Baiker,/ Catal 211, 244 (2002).

[222] R. Hutter, D. C. M. Dutoit, T Mallat, M. Schneider, A. Baiker,

/ Chem. Soc, Chem. Commun. 163 (1995).

[223] U. P. Fringeli, D. Baurecht, M. Siam, G. Reiter, M. Schwarzott,

T. Bürgi, P. Brüesch, in Handbook of Thin Film Materials, Vol. 2 (Ed.:

H. S. Nalwa), Academic Press, New York, (2001), p. 191.

120 Chapter 6

[224] U. P. Fringeli, J. Goette, G. Reiter, M. Siam, D. Baurecht, in Fourier

Transform Spectroscopy: 11th International, Vol. ATP Conference Proceed¬

ings 430 (Ed.: J. A. deHaseth), Am. Inst, of Phys., (1998).

[225] M. J. Frisch, G. W Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb,

J. R. Cheeseman, V G. Zakrzewski, J. A. Montgomery, R. E.

Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels,

K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V Barone, M. Cossi,

R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford,

J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma,

D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman,

J. Cioslowski, J. V Ortiz, A. G. Baboul, B. B. Stefanov, G. Liu,

A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin,

D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara,

C. Gonzalez, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen,

M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon,

E. S. Replogle, J. A. Pople, A.7 ed., Gaussian Inc., Pittsburgh, PA,

(1998).

[226] R. H. Perry, D. Green, in Perry's chemical engineers' handbook, 6th ed.,

McGraw-Hill Book Co, New York, (1984), p. 16_24.

[227] H. C. L. Abbenhuis, S. Krijnen, R. A. van Santen, Chem. Commun.

331 (1997).

[228] B. Notari, Stud. Surf Sei. Catal 37,413 (1988).

[229] G. L. Marra, G. Artioli, A. N. Fitch, M. Milanesio, C. Lamberti,

Microporous Mesoporous Mater. 40, 85 (2000).

[230] D. Gleeson, G. Sankar, C. R. A. Catlow, J. M. Thomas, G. Spano, S.

Bordiga, A. Zecchina, C. Lamberti, Phys. Chem. Chem. Phys. 2, 4812

(2000).

[231] A. Gisler, T Bürgi, A. Baiker, Phys. Chem. Chem. Phys. 5, 3539 (2003).

[232] D. Ferri, T Bürgi, A. Baiker, Helv. Chim. Acta 85, 3639 (2002).

[233] H. Meier, C. Antony-Mayer, C. Schulz-Popitz, G. Zerban, liebigs Ann.

Chem. 1087 (1987).

Outlook

Organic modification of titania-silica aerogels by covalently bound aminoalkyl

and acetoxyalkyl groups was found to be a promising method to enhance reac¬

tivity and selectivity in epoxidation of cyclic alkenes and allylic alcohols with

tert-butyl hydroperoxide (TBHP). The change of surface polarity and inter¬

action of the organic functional groups with surface silanol groups are at the

origin of these effects. An interaction of the functional groups with the active

Ti site - and therefore a direct influence on the epoxidation process - can nei¬

ther be excluded nor proved yet. All modifiers reveal a strong influence on the

pore structure of the synthesized aerogels. Besides the conditions chosen for the

sol-gel process, addition of the desired modifier can be a powerful tool for

designing a catalyst suitable to the demands of the desired reaction.

Unfortunately no evidence was found for chirality on a surface of the aero¬

gel modified with (3-acetoxy-3,7-dimethyloctanyl)trimethoxysilane. Neither

vibrational circular dichroism (VCD) spectroscopy nor studying adsorption

with modulation experiments by changing the absolute configuration could

show that the catalyst surface was optically active. Probably choosing a chiral

modifier which is smaller and less flexible might be more promising. Neverthe¬

less introducing chiral recognition into titania-silica mixed oxides remains a

great challenge.

Studying the adsorption of reactants and the epoxidation process over

titania-silica aerogels by in situ attenuated total reflection (ATR) spectroscopy

gave useful insight into the phenomena taking place at the catalytic solid-liquid

interface. It was possible to distinguish between a spectator species of TBHP

adsorbed at the surface silanol sites and an activated species coordinating to the

Ti site. Besides, observed phase lags between product and reactant appearance

revealed that the rate of pore diffusion is a crucial step for the epoxidation over

124

titania-silica mixed oxides. The diffusion was found to be strongly depending

on the chemical affinity of the reactant to the catalyst surface.

A comparative study of the epoxidation of cyclohexenol and cyclooctenol

revealed a much stronger and irreversible adsorption of cyclohexenol on the

surface of the aerogel which leads to blocking of the active Ti sites and thus

catalyst deactivation. For cyclooctenol oxidation a phase lag could be observed

for reactant and product appearance, which was found to be smaller for

cyclooctenol compared to TBHP. Because of the high affinity for the Ti centers,

cyclohexenol was found to be epoxidized by a hydroxy-assisted mechanism,

whereas cyclooctenol preferred a silanol-assisted mechanism due to steric

hindrance.

In situ ATR-IR spectroscopy combined with modulation spectroscopy was

shown to be a powerful tool to gain information on the processes taking place

during epoxidation over titania-silica aerogels. It was possible to detect and dis¬

criminate spectator and active species and even kinetic aspects could be studied

by phase-resolved spectra. This technique provides an extremely powerful

means to further the knowledge on the crucial processes occurring at catalytic

liquid-solid interfaces, which can be extended to many catalytic processes

awaiting exploration.

List of Publications

List of publications related to the thesis

The following list summarizes chronologically the publications which are based

on this thesis. The pertinent chapters of the thesis are given in brackets.

"Titania-silica epoxidation catalysts modified by mono- and bidentate organic

functional groups"

A. Gisler, CA. Müller, M. Schneider, T. Mallat, and A. Baiker, Top. Catal. 15,

247-255 (2001).

(Chapter 3)

"Epoxidation on Titania-Silica Aerogel Catalysts Studied by Attenuated Total

Reflection Fourier Transform Infrared and Modulation Spectroscopy"

A. Gisler, T Bürgi, and A. Baiker, Phys. Chem. Chem. Phys., 5, 3539 (2003).

(Chapter 4)

"Epoxidation of Cyclic Allylic Alcohols on Titania-Silica Aerogels Studied by

Attenuated Total Reflection Fourier Transform Infrared and Modulation Spec¬

troscopy"

A. Gisler, T Bürgi, and A. Baiker,/ Catal, (inpress).

(Chapter 5)

128

List of other publications

It follows a chronological list of publications where contributions have been

made.

"Synthesis of Organically Modified Titania-Silica Aerogels: Application for

Epoxidation of Cyclohexenol"

A. Gisler, CA. Müller, M. Schneider, T. Mallat, and A. Baiker, Stud. Surf Sei.

Catal 130, 1637-1642 (2000).

"Titania-silica epoxidation catalysts modified by acetoxypropyl groups"

CA. Müller, M. Schneider, A. Gisler, T Mallat, A. Baiker, Catal. lett. 64 9-14

(2000).

List of Conference Contributions

It follows a list of conference contributions, where the author was first author

(posters) or lecturer (oral presentations).

A. Gisler, CA. Müller, M. Schneider, T. Mallat, and A. Baiker: "Synthesis of

Organically Modified Titania-Silica Aerogels - Application for Epoxidation of

2-Cyclohexen-l-ol", 12th Int. Congress on Catalysis 2000, Granada (Spain),

poster.

A. Gisler, CA. Müller, M. Schneider, T. Mallat, and A. Baiker: "Synthesis of

Organically Modified Titania-Silica Aerogels: Application for Epoxidation of

Olefins and Allylic Alcohols", 12th Int. Congress on Catalysis 2000, Granada

(Spain), conference proceedings.

A. Gisler, CA. Müller, M. Schneider, T. Mallat, and A. Baiker: "Epoxidation

with Titania-Silica Aerogels Modified by Polar Organic Functional Groups",

Fall Meeting of the Swiss Chemical Society 2000, Lausanne (Switzerland),

poster and oral presentation.

Curriculum Vitae

Name Andreas Gisler

Date of Birth 5 May 1970

City Zürich

Citizen of Zürich

Nationality Swiss

Education

1983-1990

1991-1997

1997-2003

Gymnasium Urdorf

Graduation with Matura Type B

University of Zürich, Department of Organic Chemistry

Organic Chemistry Studies

ETH Zürich, Institute for Chemical and Bioengineering

Doctor Thesis under the Supervision of

Prof. Dr. A. Baiker

Finally..,

WIR HftS&t pHMO« BNEWOCHE ZBT F(JR Dil BATTER -

smMWM&.msem taRfMN

MiINT» WIR MÜSSTWSIEAISO

ZUR tÖÜJfffi FERTIG HABEN.