an exploration of silsesquioxanes and zeolites using high-speed experimentation

Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/27351603

AnExplorationofSilsesquioxanesandZeolitesusingHigh-SpeedExperimentation

ARTICLE

Source:OAI

CITATIONS

2

READS

30

1AUTHOR:

PaoloPescarmona

UniversityofGroningen

72PUBLICATIONS1,362CITATIONS

SEEPROFILE

Availablefrom:PaoloPescarmona

Retrievedon:04February2016

https://www.researchgate.net/publication/27351603_An_Exploration_of_Silsesquioxanes_and_Zeolites_using_High-Speed_Experimentation?enrichId=rgreq-a12b2390-ca2b-4f49-b14f-3a00f78a7aa6&enrichSource=Y292ZXJQYWdlOzI3MzUxNjAzO0FTOjEzNDQ1NjU4MTAzODA4MUAxNDA5MDY4MzU1MzUy&el=1_x_2

https://www.researchgate.net/publication/27351603_An_Exploration_of_Silsesquioxanes_and_Zeolites_using_High-Speed_Experimentation?enrichId=rgreq-a12b2390-ca2b-4f49-b14f-3a00f78a7aa6&enrichSource=Y292ZXJQYWdlOzI3MzUxNjAzO0FTOjEzNDQ1NjU4MTAzODA4MUAxNDA5MDY4MzU1MzUy&el=1_x_3

https://www.researchgate.net/?enrichId=rgreq-a12b2390-ca2b-4f49-b14f-3a00f78a7aa6&enrichSource=Y292ZXJQYWdlOzI3MzUxNjAzO0FTOjEzNDQ1NjU4MTAzODA4MUAxNDA5MDY4MzU1MzUy&el=1_x_1

https://www.researchgate.net/profile/Paolo_Pescarmona?enrichId=rgreq-a12b2390-ca2b-4f49-b14f-3a00f78a7aa6&enrichSource=Y292ZXJQYWdlOzI3MzUxNjAzO0FTOjEzNDQ1NjU4MTAzODA4MUAxNDA5MDY4MzU1MzUy&el=1_x_4


https://www.researchgate.net/institution/University_of_Groningen?enrichId=rgreq-a12b2390-ca2b-4f49-b14f-3a00f78a7aa6&enrichSource=Y292ZXJQYWdlOzI3MzUxNjAzO0FTOjEzNDQ1NjU4MTAzODA4MUAxNDA5MDY4MzU1MzUy&el=1_x_6


An Exploration of Silsesquioxanes and Zeolites using High-Speed Experimentation

Paolo Prospero Pescarmona

Cover: drawing by Gianpiero Pescarmona.

An Exploration of Silsesquioxanes and Zeolites using High-Speed Experimentation

Proefschrift

ter verkrijging van de graad van doctor

aan de Technische Universiteit Delft,

op gezag van de Rector Magnificus prof. dr. ir. J.T. Fokkema,

voorzitter van het College voor Promoties,

in het openbaar te verdedigen

op vrijdag 19 september 2003 om 10.30 uur

door

Paolo Prospero PESCARMONA

Dottore in Chimica, Università di Torino (Italië)

geboren te Torino (Italië)

Dit proefschrift is goedgekeurd door de promotor:

Prof. dr. Th. Maschmeyer

Adviseur: Dr. ir. J.C. van der Waal

Samenstelling promotiecomissie:

Rector Magnificus voorzitter

Prof. dr. Th. Maschmeyer Technische Universiteit Delft, promotor

Prof. dr. ir. H. van Bekkum Technische Universiteit Delft

Prof. dr. ir. D.E. De Vos Katholieke Universiteit Leuven, België

Prof. L. Marchese Università del Piemonte Orientale, Italië

Prof. A.F. Masters University of Sydney, Australië

Prof. Sir J.M. Thomas University of Cambridge, Verenigd Koninkrijk

Dr. ir. J.C. van der Waal Technische Universiteit Delft - Avantium

Technologies Amsterdam, adviseur

ISBN 90-9017266-1

Better being strange than stranger.

to all the persons in whose eyes I met a smile

Preface __________________________________________________________________________________________________________

I

Preface and thesis outline

This thesis describes the study of silsesquioxanes and zeolites by means of

Combinatorial Chemistry and High-Speed Experimentation. As it often happens with

specialised work, this dissertation includes topics that probably are not familiar to most

of the readers. In order to overcome this hurdle, introductory chapters providing general

information over the subjects of this thesis have been included. The information

supplied in these chapters will help readers in possession of chemical knowledge to

understand the content of the rest of the thesis.

Thesis outline:

• Chapter 1 gives an introduction to Combinatorial Chemistry and High-Speed

Experimentation. These recently developed methods for chemical research have been

widely applied in this thesis.

• Chapter 2 presents an overview about silsesquioxanes. The investigation of this

family of compounds constitutes the major topic of this thesis.

• Chapter 3 describes a first optimisation of the synthesis of titanium silsesquioxanes as

epoxidation catalysts by means of High-Speed Experimentation techniques.

• Chapter 4 is a continuation of Chapter 3 and describes a fine-tuning of the synthesis

of titanium silsesquioxanes using High-Speed Experimentation techniques.

• Chapter 5 reports new synthetic methods for cyclopentyl and cyclohexyl

silsesquioxanes based on the High-Speed Experimentation work described in

Chapters 3 and 4.

• Chapter 6 presents an application of the cyclopentyl silsesquioxane discussed in

Chapter 5 as ligand for an osmium catalyst for the dihydroxylation of alkenes.

• Chapter 7 describes new tert-butyl and phenyl silsesquioxane precursors for

titanium-based epoxidation catalysts identified by exploring the synthesis of silsesquioxanes

in highly polar solvents by means of High-Speed Experimentation techniques.

• Chapter 8, after a concise introduction to zeolites, reports the application of

High-Speed Experimentation techniques to the study of the synthesis of zeolite beta.

• Appendix A describes the High-Speed Experimentation equipment.

• Appendix B consists of a glossary of concepts and abbreviations used in this thesis.

Thesis outline __________________________________________________________________________________________________________

II

Chapter 2

Silsesquioxanes Chapter 8

Synthesis of zeolite beta

by means of HSE

Chapter 1

Combinatorial Chemistry and

High-Speed Experimentation (HSE)

Chapter 3

A new route to Ti-silsesquioxane

catalysts discovered using HSE

Chapter 4

Fine-tuning of the synthesis of

Ti-silsesquioxanes by means of HSE

Chapter 5

Syntheses of cyclopentyl and

cyclohexyl silsesquioxanes

Chapter 7

New Ti-silsesquioxane catalysts

discovered using HSE

Chapter 6

Os-silsesquioxane as

dihydroxylation catalyst

Contents __________________________________________________________________________________________________________

III

Contents

Preface and thesis outline

Contents

1. Combinatorial Chemistry and High-Speed Experimentation techniques applied to

catalysis.

1.1. Introduction.

1.2. Methods.

1.1.1. Mix&split synthesis

1.1.2. Masking strategies

1.1.3. Parallel synthesis and High-Speed Screening (HSE techniques).

1.3. Applications.

1.4. High-Speed Experimentation techniques applied to catalysis.

2. Silsesquioxanes.

2.1. Introduction and nomenclature.

2.2. Synthesis of oligomeric silsesquioxanes.

2.2.1. Hydrolytic condensation of RSiX3.

2.2.2. Cleavage of Si-O-Si bonds.

2.2.3. Reaction of the R-group.

2.2.4. Corner-capping reactions.

2.3.Characterisation.

2.3.1. NMR spectroscopy.

2.3.2. Mass spectrometry.

2.3.3. IR spectroscopy.

2.3.4. X-ray diffraction.

2.4. Completely condensed silsesquioxanes (RSiO1.5)a.

2.4.1. R4Si4O6 (a4b0).

2.4.2. R6Si6O9 (a6b0).

2.4.3. R8Si8O12 (a8b0).

I

III

1

2

3

4

6

7

8

10

13

14

17

17

20

20

21

22

22

23

24

24

24

25

25

26

Contents __________________________________________________________________________________________________________

IV

2.4.4. R10Si10O15 (a10b0).

2.4.5. R12Si12O18 (a12b0)

2.4.6. Higher completely condensed silsesquioxanes (a > 12).

2.5. Incompletely condensed silsesquioxanes

2.5.1. RSi(OH)3 (a1b3).

2.5.2. R2Si2O(OH)4 (a2b4).

2.5.3. R4Si4O4(OH)4 (a4b4).

2.5.4. R6Si6O7(OH)4 (a6b4).

2.5.5. R7Si7O9(OH)3 (a7b3).

2.5.5.1 Metallasilsesquioxane catalysts for the epoxidation of

alkenes.

2.5.5.2 Metallasilsesquioxane catalysts for the polymerisation of

alkenes.

2.5.5.3 Other metallasilsesquioxanes.

2.5.6. R8Si8O11(OH)2 (a8b2).

2.5.7. Other incompletely condensed structures (a6b2, a7b1, a8b4).

2.6. Conclusions.

3. A new, efficient route to titanium-silsesquioxane epoxidation catalysts

developed by using High-Speed Experimentation.

3.1. Introduction.

3.2. The High-Speed Experimentation approach.

3.3. Results and discussion.

3.3.1. Characterisation of the HSE lead.

3.4. Conclusions.

3.5. Experimental.

4. Fine-tuning of the synthesis of titanium-silsesquioxane epoxidation catalysts by

means of High-Speed Experimentation.

4.1. Introduction.



4.3.1. Synergetic effect of mixtures of silanes.

28

29

29

29

31

31

31

32

33

34

36

37

38

39

39

45

46

48

49

52

56

57

61

62

62

63

63

Contents __________________________________________________________________________________________________________

V

4.3.2. Effect of the reaction conditions.

4.3.3. Effect of the nature of the titanium centre.

4.4. Conclusions.

4.5. Experimental.

5. Fast and high-yield syntheses of cyclopentyl and cyclohexyl

silsesquioxanes using acetonitrile as reactive solvent.

5.1. Introduction.


5.2.1. Synthesis of cyclopentyl silsesquioxane a7b3.

2.5.5.4 Monitoring by means of mass spectrometry.

2.5.5.5 Monitoring by means of infrared spectroscopy.

5.2.2. Synthesis of cyclohexyl silsesquioxanes.

5.3. Conclusions.

5.4. Experimental.

5.4.1. Synthesis of cyclopentyl silsesquioxane a7b3.

5.4.2. Synthesis of cyclohexyl silsesquioxanes.

6. Osmium silsesquioxane as model compound and homogeneous catalyst for

the dihydroxylation of alkenes.

6.1. Introduction.


6.3. Conclusions.

6.4. Experimental.

7. New tert-butyl and phenyl silsesquioxane precursors for epoxidation

titanium catalysts discovered by means of High-Speed Experimentation.

7.1. Introduction.


7.2.1. The High-Speed Experimentation screening.

7.2.2. Phenyl silsesquioxanes synthesised in H2O.

7.2.3. tert-butyl silsesquioxanes synthesised in H2O.

7.2.4. tert-butyl silsesquioxanes synthesised in DMSO.

64

69

73

73

77

78

79

79

82

88

93

98

99

100

101

105

106

107

112

112

117

118

118

118

120

122

132

Contents __________________________________________________________________________________________________________

VI

7.3. Conclusions.

7.4. Experimental.

7.4.1. The High-Speed Experimentation screening.

7.4.2. Phenyl silsesquioxanes synthesised in H2O.

7.4.3. tert-butyl silsesquioxanes synthesised in H2O.

7.4.4. tert-butyl silsesquioxanes synthesised in DMSO.

8. Study of the synthesis of zeolite beta using High-Speed Experimentation.

8.1. Introduction.

8.1.1. Zeolite beta.



8.3.1. Up-scaling of the HSE lead.

8.4. Conclusions.

8.5. Experimental.

Appendix A. High-Speed Experimentation equipment.

A.1. The HSE automated workstation.

A.2. The heating blocks.

A.3. The vacuum centrifuge.

Appendix B. Glossary.

B.1. Concepts.

B.2. Abbreviations.

Summary.

Samenvatting.

Acknowledgements.

Publications and oral presentations.

Curriculum vitae.

135

136

136

136

137

139

143

144

145

146

147

150

151

152

155

155

157

158

159

159

160

161

165

169

171

174

1

1 Combinatorial Chemistry and High-Speed Experimentation techniques applied to catalysis

Abstract

Combinatorial Chemistry and High-Speed Experimentation techniques are methods that

have been developed recently and that allow the fast preparation and analysis of large

numbers of samples. These techniques, initially developed for the discovery of

pharmaceuticals, are being increasingly applied to various fields of chemical research,

particularly to catalysis and materials science.

____________________ The contents of this chapter have been published in:

P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, Catal. Lett., 1999, 63, 1.

P.P. Pescarmona, J.C. van der Waal, T. Maschmeyer, Catal. Today, 2003, 81, 347.

Chapter 1 __________________________________________________________________________________________________________

2

1.1 Introduction

Combinatorial Chemistry (CC) and High-Speed Experimentation (HSE)

techniques are research methods that allow the preparation and the testing of large

numbers of samples in a short time.1 The fascinating idea behind the first applications of

Combinatorial Chemistry methods was to mimic nature’s evolutionary approach in the

search of pharmaceutical active compounds. The concept was to combine many

synthetic parameters to produce a vast number of different compounds and to test them

for pharmaceutical activity.2 Just the active species would ‘survive’ this screening and

would be further studied as candidates for active drugs. Starting at the end of the

nineteen-eighties, the field of Combinatorial Chemistry rapidly developed both from the

methodological and technical point of view, in such a way that these methods are

becoming a standard tool for the discovery of novel drugs. The success in the

pharmaceutical area has stimulated application of these techniques to other fields of

chemical research and particularly to catalysis1,3-8 (both homogeneous9-11 and

heterogeneous12-14) and to materials science (e.g. superconducting, optical, magnetic,

dielectric, ferroelectric or polymeric materials).4,15-17 Combinatorial methods were

developed and modified in different ways to adapt them to each field of application and

started to be often associated with High-Speed Experimentation techniques. The term

combinatorial indicates the methodology of combining different experimental

parameters and to test the so-obtained combinations for selected properties, while the

term High-Speed Experimentation, alternatively known as High-Throughput

Experimentation (HTE), refers to the automated equipment that, allowing fast

preparation and screening of samples, is often necessary to perform a combinatorial

approach. HSE automated workstations can perform operations rapidly (and for 24

hours a day) and can also cope with very small amounts of reactants with high

precision. Besides being in many cases faster and cheaper per experiment,

Combinatorial Chemistry and High-Speed Experimentation methods are also safer and

have a lower environmental impact, since they use only small quantities of reactants. A

further advantage of the experiments carried out with these techniques is the high

reproducibility: since the operations are largely performed by robotic equipment, the

experimental errors due to different preparation conditions or different operators are

reduced to systematic ones and the reproducibility of the experiments is notably increased.

Combinatorial Chemistry and High-Speed Experimentation __________________________________________________________________________________________________________

3

The incentive to adopt Combinatorial Chemistry and High-Speed

Experimentation is provided by the fact that by these techniques it is possible to

synthesise and screen tens or hundreds of samples in a much shorter time compared to

traditional methods. Clearly, this approach has the potential for substantial time-savings

in research and development, allowing faster testing of hypotheses as well as reducing

considerably the time-to-market of new commercially relevant developments. Industrial

interest in CC and HSE technology is demonstrated by the increasing number of small,

dedicated new companies and joint ventures in this field. Besides their use in applied

research for the identification of new leads, CC and HSE techniques can result in a

powerful tool for academic scientific research. A correct application of these techniques

can provide a new approach for fundamental research: by means of a rational and

scientific design of the experiments these techniques can lead to the identification of

trends that give a much better first understanding/appreciation of the system under

study. Moreover, by performing a large number of experiments in a short time, more

opportunities may be offered to serendipity.

1.2 Methods

In order better to understand what Combinatorial Chemistry is, it is instructive to

consider, as an example, the reaction of a compound of class A (e.g. a Lewis acid) with

a compound of class B (e.g. a Lewis base) in a solvent S to give a product P. Combining

n different compounds of class A with m different compounds of class B and l different

solvents S, a parameter space constituted of n×m×l different combinations would be

defined, where each combination could lead to a different product or to a different yield

for a particular product. Such an approach can be useful if one were trying, e.g., to

obtain product P with the highest yield. Once the collection of products (usually

referred to as a library18) based on the selected parameter space has been synthesised, it

is necessary to subject it to a screening test regarding the performance or physical

properties of the products. The mode of screening depends on the type of product and

on its properties (e.g. catalyst selectivity or product yield).

When performing a synthetic Combinatorial Chemistry experiment, a number of

basically different strategies may be followed to create a library of compounds. The

Chapter 1 __________________________________________________________________________________________________________

4

most commonly used are: mix&split (or split and pool) synthesis, masking strategies

and parallel synthesis.

1.2.1 Mix&split synthesis

Mix&split synthesis is generally related to the use of polymer resin beads as

support for the reaction and used to synthesise bioactive compounds,2 typically

polypeptides. In the first step of the synthesis, the resin is divided into n equal parts and

each of them is treated with one of the n different reagents. Next, the n portions of resin

are washed, combined and mixed in a single pot, then divided again into n equal parts.

In the second step each portion is again treated with one of the n reagents and then the

washing, mixing and splitting steps are repeated. The procedure is iterated m times. At

the end a library of nm products is obtained in just m steps (Figure 1.1).

Figure 1.1. Mix&split synthesis on a solid-phase carrier; in this example, a library of 27

elements is generated in 3 steps by combining 3 different reagents.

-X -Y -Z

-XX -YX -ZX -XY -YY -ZY -XZ -YZ -ZZ

mix +X +Z

+Y

-YXX -YYX -YZX

-XXX -XYX -XZX

-ZXX -ZYX -ZZX

-YXY -YYY -YZY

-XXY -XYY -XZY

-ZXY -ZYY -ZZY

-YXZ -YYZ -YZZ

-XXZ -XYZ -XZZ

-ZXZ -ZYZ -ZZZ

mix +X +Z

+Y


5

Once such a library has been created, it has to be tested to determine, e.g., the bioactive

compound(s). A problem that arises when the mix&split synthesis method is used, is

how to identify the compound(s) giving positive results during the screening among the

set of the components of the library.19 One way to overcome this problem is by a

process known as deconvolution.2 In the first step, the elements of the library are

divided into the n vessels and screened for activity: the active vessel is identified and

the others are eliminated. The compounds of this vessel are resynthesised in smaller

libraries and screened again for activity. The process of elimination goes on until the

active compound is established (Figure 1.2).

Figure 1.2. Example of deconvolution process for a library of 27 elements.

- YXX - YYX - YZX - XXX - XYX - XZX

- ZXX - ZYX - ZZX -YXY -YYY - YZY -XXY -XYY - XZY

-ZXY -ZYY - ZZY -YXZ - YYZ - YZZ -XXZ - XYZ - XZZ

-ZXZ - ZYZ - ZZZ

- XX - YX - ZX -XY -YY - ZY - XZ - YZ -ZZ

- XXY - YXY - ZXY -XYY -YYY - ZYY - XZY -YZY - ZZY

Active vessel

Step 1

Step 2

- XZ -YZ - ZZ

- XZY -YZY -ZZY

Step 3

Active compound

+Y +Y +Y

+Y +Y +Y

Active vessel

Chapter 1 __________________________________________________________________________________________________________

6

Although deconvolution has proved to be useful, it presents some drawbacks: besides

being a quite time-consuming process, when the active vessel is identified, there is no

certainty that the activity is due to the presence of a single active product rather than

resulting from the presence of many weakly active compounds.

Another technique used to identify the compounds that were active during screening is

that of encoding. The method consists in tagging each compound of the library during

the synthesis with a different tag; suitable tags show particular chemical or physical

properties that can be easily detected.20,21 Common chemical tagging methods make use

of tags that can be cleaved to liberate compounds (e.g. amines), which can then be

detected by standard analytical techniques (e.g. HPLC, MS).19

1.2.2 Masking strategies

Masking strategies are used for the combinatorial synthesis of thin films of solid

materials.15,16 These strategies are based on the successive deposition of thin films of solids

over a support. By means of a set of masks, the deposition of each solid precursor is limited

to a specific region of the support, creating a library of solids with different compositions.

Figure 1.3. Examples of binary and quaternary masks.

quaternary masking

binary masking


7

Mainly two types of masking are reported in the literature: binary and quaternary

(Figure 1.3). When using binary masking, 2n compositions can be obtained with n

masks in n steps; with quaternary masking, 4n compositions are produced using n masks

in 4n steps (Figure 1.4). Binary masking allows the synthesis of the highest number of

different solids with a given number of deposition steps; quaternary masking allows

greater flexibility in the design of the library, since elements added in the four steps

used for each mask will not overlap.

The analytical technique employed for the screening of the library depends on the

desired property of the materials (e.g. resistance measurements for superconductors,

photoluminescence measurements for phosphors).4

Figure 1.4. Binary and quaternary masking strategies for the synthesis of materials libraries.

1.2.3 Parallel synthesis and High-Speed Screening (HSE techniques)

In the parallel synthesis approach the various reactions take place in separate

vessels; typically, robotic equipment is used to pick and mix the reactants in different

miniature vessels or wells, so that an array of distinct products is obtained. The library

of products has to be screened to determine the active compound(s). In contrast to the

mix&split approach, the identification of the active species is not a problem, since

A A B

D C

AE BE

CE DE

90° rotation + B90° rotation + C90° rotation + Dmask 1 + A mask 2 + E

AB A

B

A ABC AC AB A

C BCB

+A +B +C

binary masking

quaternary masking

Chapter 1 __________________________________________________________________________________________________________

8

separate wells are used. However, finding a suitable way to test all the isolated products

at high-throughput speeds presents a principal challenge. If the screening process of the

library is performed one well at a time, it can be very expensive and time-consuming

and would become a bottleneck nullifying some of the speed advantages of

combinatorial synthesis. To avoid this, fast and affordable analytical methods, usually

referred to as High-Speed Screening (HSS) or High-Throughput Screening (HTS)

techniques,22 are being developed by many groups. Current HSS technology is based on

the use of miniaturised, automated and parallel versions of tools conventionally

employed to screen the compounds under study, such as chromatographic,

spectroscopic or thermographic techniques, and in some cases of new specifically

developed methods.1,4,11,13,23 (e.g. a mass spectrometer equipped with an autosampler

has been used to assay the enantioselectivity of catalytic reactions of isotopically

labelled substrates, with a throughput of about 1000 samples per day).24 The choice of

the type of HSS is related to the particular feature that has to be detected and, therefore,

has to be tailored for any system under investigation.

The use of parallel synthesis associated to High-Speed Screening is presently referred to

as High-Speed Experimentation (or High-Throughput Experimentation).

1.3 Applications

To illustrate how and in which cases Combinatorial Chemistry and High-Speed

Experimentation techniques can be applied advantageously, a general case of the

discovery/optimisation of a catalytic system will be considered. Given a chemical reaction

for which a catalyst is wanted, the first step in primary screening is conventional, i.e.

consists of gathering all the knowledge already available in the literature about the

subject. In many cases this search provides basic information about the characteristics of

the most suitable catalytic system for the chosen chemical reaction (e.g. an acid rather

than a basic catalyst, a homogeneous rather than a heterogeneous catalyst). If this

information is scarce or absent, the issue becomes how to find a way to get basic knowledge

of the system under investigation. This could be achieved with an experimental approach

and/or by means of computational modelling. If an experimental approach is chosen, this

second step in primary screening should cover a broad range of parameters in order to gain


9

very general information about the system. This usually requires an extremely large

number of experiments to be performed, for which purpose Combinatorial Chemistry and

High-Speed Experimentation techniques are particularly helpful.

Comparing the three synthetic strategies described in Paragraph 1.2, it is clear

that the mix&split and the masking approaches might be advantageous for primary

screening since they allow the fast synthesis and screening of very large libraries of

compounds (thousands per week). High-Speed Experimentation parallel methods have a

lower throughput (hundreds of samples per week) than mix&split and masking

methods, but one that is still very high when compared with traditional, serial

techniques. The lower throughput of HSE parallel techniques might be a drawback for a

primary screening, which requires gathering a broad set of information in a short time.

This problem can be overcome by using Design of Experiments (DOE) methods and

search algorithms, that allow decreasing the number of experiments to be performed by

means of statistical data-handling.12,13,17,25 Moreover, parallel methods are more

versatile than mix&split and masking methods: with the latter, only the composition of

a target compound can be optimised, while parallel methods also allow the variation of

other synthetic parameters (solvent, pH, concentrations, temperature etc.). Besides, HSE

parallel equipment can be designed to handle solid, liquid and gas phases, while

mix&split and masking strategies are limited to liquid and solid phases, respectively.

Finally, parallel and masking approaches allow for the correlation of each sample with

its properties (i.e. its activity as a catalyst), while this is not possible, prior to

deconvolution, using mix&split methods where the samples are screened as a whole. In

this sense, parallel and masking techniques are more suitable when the goal of

identifying new leads is coupled with that of a fundamental understanding of the system

under study. Given all these considerations it is clear why HSE parallel techniques are

becoming the most common approach for primary screening.

Typically, the primary screening provides a set of lead compounds and some

basic correlation between the properties of the leads and their structural and synthetic

characteristics. In the secondary screening one or more of these leads are further

optimised: the information obtained from the primary screening is the basis to focus

further research on a smaller experimental parameter space. Dealing with a more

focussed parameter space means that more detailed and specific information can be

obtained in this screening phase. The fact that the parameter space screened is smaller in

Chapter 1 __________________________________________________________________________________________________________

10

the secondary than in the primary screening implies that the number of experiments

performed in the secondary screening is lower while the information gathered from each

experiment should be more detailed. Therefore, HSE parallel techniques are particularly

useful to speed up the secondary screening. The applicability of HSE techniques to

secondary screening is determined by the technical limitations of the automated

workstations employed. These limitations may range from the preparation of the

samples (e.g. inefficient mixing of reaction mixtures in the small HSE vessels) to their

analysis (e.g. equipment for high-throughput screening of the samples is not yet

available for every spectroscopic technique). The quality of the HSE equipment is

continuously improving, nevertheless at this stage not every chemical synthesis can be

studied by means of HSE techniques.

At the end of the secondary screening a restricted number of leads will be

identified and the correlation between their properties and their structural and synthetic

characteristics will be much more defined than at the end of the primary screening. It is

at this point that the lead(s) can be studied in detail at a conventional laboratory scale:

this typically includes the characterisation of the lead(s) with different spectroscopic

and/or diffractive techniques. If Combinatorial Chemistry and High-Speed

Experimentation techniques have been used during the earlier screening stages,

repeating the experiment in a conventional manner is an important check of the

reproducibility of the results on a larger-volume scale. Results should never be

published without such a check, since the ability to independently validate a particular

result should not be dependent on the HSE-equipment used.26

1.4 High-Speed Experimentation techniques applied to catalysis

The development of new catalysts is a challenging task. In many cases the

correlation between their features (structural, electronic) and their performance

(activity, selectivity, lifetime) is not easily established, especially for supported

heterogeneous catalysts. Therefore, an iterative process of ‘design’, synthesis and

testing is usually followed to improve catalyst performance. HSE techniques can

accelerate this process considerably, allowing for the simultaneous evaluation of a large

number of candidates (Figure 1.5). All operations involved in the development of a


11

catalyst lend themselves to the application of HSE techniques. The synthesis of both

homogeneous and heterogeneous catalysts, their screening for activity and selectivity in

test reactions, and also the determination of the optimal process parameters for a

specific reaction, can all be conducted with much greater speed and efficiency

employing the miniaturised, automated combinatorial/HSE procedures than by the use

of conventional methods.

It is important to realise that HSE techniques have to be considered as a powerful tool to

increase the number of samples to be studied, rather than as an alternative to rational

methods for the development of catalysts: a multi-dimensional and scientific approach –

based, e.g., on literature data, computational modelling, personal chemical knowledge

and intuition - is essential to determine which parameter space has to be investigated to

gain the desired information from the experiment.

Figure 1.5. Comparison between classic and High-Speed Experimentation approach.

In this sense, HSE techniques have to be seen as a new and useful tool for chemical

research in the same way as, say, spectroscopic techniques or quantomechanical

calculations. Therefore, they can be very suitable to study some systems and inefficient

for others. The range of chemical systems that can be studied using HSE techniques will

probably get broader with the technical improvement of the automated equipment.

Testing

Design

Synthesis Testing

Design

Synthesis

Classic HSE

Chapter 1 __________________________________________________________________________________________________________

12

References

1 P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, Catal. Lett., 1999, 63, 1. 2 N.K. Terret, M. Gardner, D.W. Gordon, R.J. Kobylecki, J. Steele Tetrahedron, 1995, 30, 8135. 3 J.M. Newsam, F. Schüth, Biotechnol. Bioeng. (Comb. Chem.), 1998/1999, 61, 203. 4 B. Jandeleit, D.J. Schaefer, T.S. Powers, H.W. Turner, W.H. Weinberg, Angew. Chem. Int. Ed., 1999,

38, 2494. 5 R. Schlögl, Angew,. Chem. Int. Ed., 1998, 37, 2333. 6 T. Bein, Angew. Chem. Int. Ed., 1999, 38, 323. 7 W.F. Maier, Angew. Chem. Int. Ed., 1999, 38, 1216. 8 J.M. Thomas, Angew. Chem. Int. Ed., 1999, 38, 3589. 9 Shimizu, K.D., Snapper, M.L., Hoveyda, A.H., Chem. Eur. J., 1998, 4, 1885. 10 R.H. Crabtree, Chem. Commun., 1999, 1611. 11 M.T. Reetz, Angew. Chem. Int. Ed., 2001, 40, 284. 12 A. Holzwarth, P. Denton, H. Zanthoff, C. Mirodatos, Catal. Today, 2001, 67, 309. 13 S. Senkan, Angew. Chem. Int. Ed., 2001, 40, 312. 14 J.M. Newsam, T. Bein, J. Klein, W.F. Maier, W. Stichert, Micropor. Mesopor. Mater., 2001, 48, 355. 15 T.X. Sun, Biotechnol. Bioeng. (Comb. Chem.), 1998/1999, 61, 193. 16 X.-D. Xiang, Biotechnol. Bioeng. (Comb. Chem.), 1998/1999, 61, 227. 17 J.M Cawse, Acc. Chem. Res., 2001, 34, 213. 18 See Appendix B for a general definition of library in CC and HSE. 19 C. Barnes, S. Balasubramanian, Curr. Opin. Chem. Biol., 2000, 4, 346. 20 J.J. Baldwin, J.J. Burbaum, I. Henderson, M.H.J. Ohlmeyer, J. Am. Chem. Soc., 1995, 117, 5588. 21 E.J. Moran, S. Sarshar, J.F. Cargill, M.M. Shahbaz, A. Lio, A.M.M. Mjalli, W.W. Armstrong, J. Am.

Chem. Soc., 1995, 117, 10787. 22 M.F. Asaro, R.B. Wilson, Chemistry & Industry, 1998, 19, 777. 23 M.T. Reetz, Angew. Chem. Int. Ed., 2002, 41, 1335. 24 M.T. Reetz, M.H. Becker, H.-W. Klein, D. Stöckigt, Angew. Chem. Int. Ed., 1999, 38, 1758. 25 D.C. Montgomery, Design and Analysis of Experiments, J. Wiley & Sons, 1996, 4th edition. 26 M. Baerns, C. Mirodatos, NATO Science Series, 2002, Ser. II Vol. 69, 469.

13

2

Silsesquioxanes

Abstract

Silsesquioxanes are a family of inorganic-organic hybrid compounds with applications

in the fields of catalysis, materials science and coordination chemistry. This Chapter

presents a review of their synthesis and characterisation as well as selected applications. ____________________

The contents of this chapter have been published in:

P.P. Pescarmona, T. Maschmeyer, Aust. J. Chem., 2001, 54, 583.

Chapter 2 __________________________________________________________________________________________________________

14

2.1 Introduction and nomenclature

The importance of silsesquioxanes as chemical structures of interest to a large

variety of chemists has grown steadily since their discovery as a ‘curious white

precipitate’ during silane polymerisations by Sprung and Guenther in 1955. Their

structures are manifold and usually crystallographically characterised. Their synthetic

accessibility, once, a major drawback due to synthesis times of months, has much

improved and many types of silsesquioxanes can now be synthesised in a matter of

hours or days. Applications of silsesquioxanes can be found in the areas of catalysis,

materials science (polymers, photo-resists) and coordination chemistry.

The etymology of the term silsesquioxane indicates a family of compounds

characterised by a ratio of 1.5 (sesqui, from the Latin semisque, ‘and a half’) between

silicon and oxygen atoms. Silsesquioxanes are compounds of the general formula

(RSiO1.5)a(H2O)0.5b, or, rearranging it, RaSiaO(1.5a-0.5b)(OH)b, where R is an hydrogen

atom or an organic group and a and b are integer numbers (a = 1,2,3…; b = 0,1,2,3,…)

related according to:

a + b = 2n, where n is an integer (n = 1,2,3,…)

b ≤ a + 2

Given this formula, a silsesquioxane with a particular degree of condensation can be

described by the values of a and b, e.g. a7b3 refers to a silsesquioxane containing seven

Si atoms and three -OH groups (Figure 2.2, compound 13). However, it is important to

notice that, for any set of a and b values, more than one structure might be drawn: for

instance, two a8b2 silsesquioxane structures are reported in the literature (Figure 2.2,

compounds 14 and 15).1,2

Structurally, silsesquioxane frameworks consist of tetrahedral units in which a silicon is

bound to three oxygens and one R-group. The oxygen atoms can act as bridges between

two silicon atoms, belonging to different tetrahedral units, or between a silicon and a

hydrogen atom. A unit where a silicon is connected to three oxygens is commonly

referred to as T unit (an M unit describes silicon bound to one oxygen, D when bound to

two oxygens and Q when bound to four oxygens, as is the case in a silicate).3 An

Silsesquioxanes __________________________________________________________________________________________________________

15

alternative nomenclature for silsesquioxanes may be based on the number of T units

they contain (Ta).

Figure 2.1. Completely condensed silsesquioxanes (these structures are simplified

representations as to, e.g., SiOSi angles).

Silsesquioxanes can be synthesised either as oligomeric species, i.e. discrete structures

constituted of a limited number of tetrahedral units, or as polymeric species. The

properties and applications of these two types of silsesquioxanes are rather different and

their fields of research separated. Here, the attention will be focussed on oligomeric

silsesquioxanes (oligosilsesquioxanes).

On the basis of the general formula reported above, silsesquioxanes can be

subdivided in two main groups: completely condensed and incompletely condensed

silsesquioxanes, according to b = 0 or b ≠ 0, respectively. In the case of completely

condensed silsesquioxanes, the general formula can be reduced to (RSiO1.5)a, where a is

now an even number greater than 2. In these compounds, all oxygen atoms act as

SiSi

Si Si

O

O

O

O R

R

R

R R

RO

Si Si

Si Si

O

O

OO

R

OO

O

R

Si

Si

Si

Si

Si

OO

O

OO

O

R

R

R

R R

O

O

R

O

Si

SiSiOO

O O

OO

Si

Si

RR

R

R

Si SiO

O

R

RO

R

O

O

R R

O

O

O

R

Si

Si

O

R

R

Si

Si

O

R

SiSi

Si Si

SiSi

RR R

O O

OO

O

O O

OO

O

SiOOO

Si

R

O

O

R R

O

OR

O

R

SiSi

Si Si

Si

R

O

SiO

O

O

Si

O

SiO

R R

R

R

ROSi

R

RO

R

OR R

O

O

O

R

Si

Si

O

R

RSi

SiSi

SiSi

SiSi

R

R

OO

O

O

OO

Si

O

O

R

Si

O

O

O

21 3

54 6

SiSi

Si Si

O

O

O

O R

R

R

R R

RO

Si Si

Si Si

O

O

OO

R

OO

O

RSiSi

Si Si

O

O

O

O R

R

R

R R

RO

Si Si

Si Si

O

O

OO

R

OO

O

R

Si

Si

Si

Si

Si

OO

O

OO

O

R

R

R

R R

O

O

R

O

Si

Si

Si

Si

Si

Si

OO

O

OO

O

R

R

R

R R

O

O

R

O

Si

SiSiOO

O O

OO

Si

Si

RR

R

R

SiSiOO

O O

OO

Si

Si

RR

R

R

Si SiO

O

R

RO

R

O

O

R R

O

O

O

R

Si

Si

O

R

R

Si

Si

O

R

SiSi

Si Si

SiSi

RR R

O O

OO

O

O O

O

Si SiO

O

R

RO

R

O

O

R R

O

O

O

R

Si

Si

O

R

R

Si

Si

O

R

SiSi

Si Si

SiSi

RR R

O O

OO

O

O O

OO

O

SiOOO

Si

R

O

O

R R

O

OR

O

R

SiSi

Si Si

Si

R

O

SiO

O

O

Si

O

SiO

R R

R

R

O

O

SiOOO

Si

R

O

O

R R

O

OR

O

R

SiSi

Si Si

Si

R

O

SiO

O

O

Si

O

SiO

R R

R

R

ROSi

R

RO

R

OR R

O

O

O

R

Si

Si

O

R

RSi

SiSi

SiSi

SiSi

R

R

OO

O

O

OO

Si

O

O

R

Si

O

O

ORO

Si

R

RO

R

OR R

O

O

O

R

Si

Si

O

R

RSi

SiSi

SiSi

SiSi

R

R

OO

O

O

OO

Si

O

O

R

Si

O

O

O

21 3

54 6

Chapter 2 __________________________________________________________________________________________________________

16

bridges between the silicon atoms and no silanol group (Si-OH) is present (Figure 2.1).

On the other hand, incompletely condensed silsesquioxanes contain silanol groups that

confer onto them utility as ligands for metal coordination and as model compounds for

silica surfaces (Figure 2.2).

Figure 2.2. Incompletely condensed silsesquioxanes (these structures are simplified

representations).

Si Si

SiSi

Si

Si

Si

Si

OH

OO

O

O

OO

O

OO

O

R

R

R

R

R

R R

RO

HOSi Si

Si

OH

O O

O

O

R

R

R

R

RR

Si O

O

RO

HO

Si Si

SiSi

O

O

OOH

OH

R

R

O

Si Si

SiSi

Si

Si

Si

Si

O O O

O

OO

O

O

O

OH

R

R

R

HO

R

RO

R

R

Si

SiSi

Si

Si

Si

Si

OHO

O

OH

O

OO

OH

OO

O

R

R

R

R

R R

RO

Si

SiSi

Si

Si

Si

Si

OO

O

OO

OH

OO

O

R

R

R

R

R R

R

O

O

Si Si

Si

Si Si

OO

OHHO

OH

O

O

OHO

O

R

R

R

R

R

RO

Si

O

Si Si

Si

Si Si

OO HO

OH

O

O

O

R

R

R

R

R

RO

Si

OSi

OHHO

R

OH

Si

OHHOR

O

SiOHHO

R

SiSiO

Si SiO

R

OH

OO

OHHO

R

HO

RR

87 10

1211 13

9

1514 16

Si Si

SiSi

Si

Si

Si

Si

OH

OO

O

O

OO

O

OO

O

R

R

R

R

R

R R

RO

HOSi Si

SiSi

Si

Si

Si

Si

OH

OO

O

O

OO

O

OO

O

R

R

R

R

R

R R

RO

HOSi Si

Si

OH

O O

O

O

R

R

R

R

RR

Si O

O

RO

HO

Si Si

SiSi

O

O

OOH

OH

RSi Si

Si

OH

O O

O

O

R

R

R

R

RR

Si O

O

RO

HO

Si Si

SiSi

O

O

OOH

OH

R

R

O

Si Si

SiSi

Si

Si

Si

Si

O O O

O

OO

O

O

O

OH

R

R

R

HO

R

RO

R

R

R

O

Si Si

SiSi

Si

Si

Si

Si

O O O

O

OO

O

O

O

OH

R

R

R

HO

R

RO

R

R

Si

SiSi

Si

Si

Si

Si

OHO

O

OH

O

OO

OH

OO

O

R

R

R

R

R R

RO

Si

SiSi

Si

Si

Si

Si

OHO

O

OH

O

OO

OH

OO

O

R

R

R

R

R R

RO

Si

SiSi

Si

Si

Si

Si

OO

O

OO

OH

OO

O

R

R

R

R

R R

R

O

O

Si

SiSi

Si

Si

Si

Si

OO

O

OO

OH

OO

O

R

R

R

R

R R

R

O

O

Si Si

Si

Si Si

OO

OHHO

OH

O

O

OHO

O

R

R

R

R

R

RO

Si

Si Si

Si

Si Si

OO

OHHO

OH

O

O

OHO

O

R

R

R

R

R

RO

Si

O

Si Si

Si

Si Si

OO HO

OH

O

O

O

R

R

R

R

R

RO

Si

O

O

Si Si

Si

Si Si

OO HO

OH

O

O

O

R

R

R

R

R

RO

Si

OSi

OHHO

R

OH

Si

OHHO

R

OH

Si

OHHOR

O

SiOHHO

R

Si

OHHOR

O

SiOHHO

R

SiSiO

Si SiO

R

OH

OO

OHHO

R

HO

RR

SiSiO

Si SiO

R

OH

OO

OHHO

R

HO

RR

87 10

1211 13

9

1514 16


17

2.2 Synthesis of oligomeric silsesquioxanes

A number of different ways are currently used to synthesise oligomeric silsesquioxanes.

2.2.1 Hydrolytic condensation of RSiX3

The most common synthetic route to obtain oligomeric silsesquioxanes is the

hydrolytic condensation of monosilanes RSiX3 (X = Cl, OMe, OEt,…):4,5

RSiX3 + 3 H2O RSi(OH)3 + 3 HX [1]

a RSi(OH)3 (RSiO1.5)a(H2O)0.5b + (1.5a – 0.5b) H2O [2]

The first step is the hydrolysis of the monosilane to give the corresponding trisilanol.

This reaction is generally very fast.6 The second step is the condensation of the trisilanol

yielding different silsesquioxane species (see Figures 2.1 and 2.2 for some structures

reported in literature). The trisilanol itself is very reactive and in most of the cases

cannot be isolated; the only reported exceptions are in the case of rather large organic

groups.7-10 The condensation reaction is a multistep process involving the formation of

many different intermediate structures. Sprung and Guenther11,12 and later Brown and

Vogt6,13 studied the hydrolytic condensation of different organomonosilanes RSiX3

(R = methyl, ethyl, phenyl, cyclohexyl). They hypothesised various mechanisms of

formation, involving the consecutive condensation of the monomeric trisilanol to give

linear, cyclic and, finally, polycyclic and polyhedral silsesquioxanes. For instance, in

the case of the synthesis of completely condensed phenyl silsesquioxanes 3 (Ph8Si8O12,

a8b0 or T8), they proposed the reaction to proceed by the consecutive formation of the

dimer, of the cyclic tetramer and, finally, of the cubic silsesquioxane (Figure 2.3).13

Lavrent’yev et al. followed the hydrolytic condensation of ethyltrichlorosilane in

aqueous butanol by GC-MS: the identification of intermediates and products allowed

them to outline the complex mechanism of formation.4 Recently, Kudo and Gordon

performed a theoretical study of the mechanism for the synthesis of

hydrosilsesquioxanes.14,15 Using ab initio quantum mechanical methods, they calculated

the energy barriers for the initial hydrolysis of trichlorosilane HSiCl3 to HSi(OH)3 and

Chapter 2 __________________________________________________________________________________________________________

18

the successive formation of the dimer H2Si2O(OH)4 [a2b4], cyclic trimer H3Si3O3(OH)3

[a3b3] and cyclic tetramer H4Si4O4(OH)4 [a4b4]. They found that the presence of a

water molecule strongly reduced the energy barriers that have to be overcome to form

the selected species, suggesting that the solvent may play a relevant role in influencing

the synthesis.

Figure 2.3. Proposed mechanism of formation for the completely condensed

silsesquioxane 3.

In fact, many factors are known to influence the hydrolytic condensation, determining

which silsesquioxanes structures are formed and in which ratios,4,5 they are namely:

- the nature of the R-group

- the nature of the X-group

- the solvent

- the concentration of the monosilane RSiX3

- the rate of addition and quantity of H2O

- the temperature

- the pH

- the reaction time

All these factors influence the hydrolytic condensation mutually and, therefore, they

cannot be studied independently. Nevertheless, some general conclusions about the

effect of any of them can be drawn.

The nature of the R-group might influence the thermodynamics and kinetics of

formation of silsesquioxanes via both steric and electronic effects. For example, for

bulky R-groups like cyclohexyl1 or cyclopentyl,16 the formation of incompletely

condensed silsesquioxanes is known to be favoured, while the hydrolytic condensation

SiSi

Si Si

O

O

O

O R

R

R

R R

RO

Si Si

Si Si

O

O

OO

R

OO

O

R

SiOHHO

R

OH

Si

OHHOR

O

SiOHHO

R

Si

SiO R

R

Si

Si

O

R

O

O

RHO

HO

HO

HO

SiSi

Si Si

O

O

O

O R

R

R

R R

RO

Si Si

Si Si

O

O

OO

R

OO

O

R

SiOHHO

R

OH

SiOHHO

R

OH

Si

OHHOR

O

SiOHHO

R

Si

OHHOR

O

SiOHHO

R

Si

SiO R

R

Si

Si

O

R

O

O

RHO

HO

HO

HO

Si

SiO R

R

Si

Si

O

R

O

O

RHO

HO

HO

HO


19

of monosilanes with smaller groups like methyl11,17 or hydrogen18 yields more

frequently completely condensed structures. This trend indicates that a steric effect of

the R-group determines to a significant extent the level of condensation of the products.

Moreover, the nature of the R-group determines, together with the nature of the solvent,

the solubility of the silsesquioxanes species, thereby influencing the equilibria and the

rate of the condensation reactions. In this context, it is interesting to note that the

solubility of the incompletely condensed silsesquioxane a7b3 in many organic solvents

is substantially lower when R is cyclopentyl or cycloheptyl than when it is cyclohexyl.16

This difference could account for the higher selectivity as well as for the higher yields

reported for cyclopentyl and cycloheptyl silsesquioxane a7b3.19

The nature of the X-group does not have a major influence on the synthesis of

silsesquioxanes, since the group reacts during the first, usually fast, step of the process.

When X is a halide, the hydrolysis is faster than when X is an alkoxy group. Moreover,

the hydrolysis of trihalidesilanes produces acids (HX) that may catalyse successive

condensation reactions. Trichlorosilanes are usually chosen as starting material for the

synthesis of silsesquioxanes.

The solvent plays a strong role in influencing the synthesis of silsesquioxanes, since

solvent molecules interact with the silsesquioxanes species present in the reaction

solution. Polar molecules form hydrogen bonds with the silanol groups (Si-OH) and,

therefore, stabilise incompletely condensed species. Recently, the favourable effect of

an high polarity solvent on the synthesis of incompletely condensed silsesquioxanes was

reported.20 As mentioned above, the solvent also determines the solubility of the

silsesquioxane species, influencing the rate and the products of the reaction. Finally, the

solvent can affect the kinetics of the synthesis by interacting with the transition-state

intermediates of the condensation reactions: theoretical calculations14,15 showed that the

interaction of a transition-state intermediate with a polar solvent molecule can stabilise

the system and, therefore, reduce the activation barrier for the formation of higher

silsesquioxane structures.

The starting concentration of the monosilane RSiX3 influences the reaction kinetics.

Since it has been impossible, so far, to obtain a satisfactory kinetic equation for the

synthesis of silsesquioxanes, it is not possible to preview which can be the effect of the

starting concentration on the reaction. It has been suggested that a high concentration

facilitates the formation of polymeric silsesquioxanes.4

Chapter 2 __________________________________________________________________________________________________________

20

The quantity and the rate of addition of H2O influence the kinetics of the hydrolytic

condensation (water acts both as a reactive compound and as a solvent). However, for the

reasons mentioned previously, the precise effect of this parameter has not yet been elucidated.

The temperature of the reaction influences the reaction kinetics and the solubility of the

silsesquioxanes species present in the reaction solution. It has been reported that a high

reaction temperature favours the formation of highly condensed polymeric species.5

The hydrolysis and condensation reactions of silsesquioxanes are catalysed by both acid

and basic medium. Low pH values have been reported to favour the formation of

oligosilsesquioxanes, while at high pH values polymeric species are favoured.12

The synthesis of silsesquioxanes species can be very slow and periods of several

months, even years, can be required before the reaction goes to completion.1 If the

reaction is stopped after a set reaction time and the silsesquioxane species forced to

condense by drying the reaction solution, the silsesquioxane structures that are formed

are different from those that would have been formed by letting the synthesis proceed.

Usually, the silsesquioxanes so obtained are of lower molecular weight, since they are

generated by the rapid condensation of intermediate species.

2.2.2 Cleavage of Si-O-Si bonds

Recently, Feher and coworkers developed new methods for the synthesis of

incompletely condensed silsesquioxanes by the cleavage of Si-O-Si bonds of

completely condensed silsesquioxanes.21 The reaction of the readily available R6Si6O9

and R8Si8O12 (Figure 2.1, compounds 2 and 3) with a strong acid (HBF4/BF3, CF3SO3H

or CH3SO3H)2,22,23 or with a base (NEt4OH)24,25 produces various incompletely

condensed structures (e.g. compounds 10, 11, 13, 14, 15, 16) in good yields and often

with good selectivity. These methods offer a fast, though somewhat complex,

alternative to the hydrolytic condensation of organomonosilanes to obtain various

incompletely condensed silsesquioxanes in synthetically useful amounts.

2.2.3 Reaction of the R-group

Different types of reaction involving the conversion of the R-group of

silsesquioxanes to another group are known. These reactions can be useful if the desired


21

silsesquioxane were difficult to obtain via the hydrolytic condensation. These types of

synthesis methods have been applied to completely condensed silsesquioxanes and

particularly to the cubic structure R8Si8O12 (Figure 2.1, compound 3). The

octahydrosilsesquioxane H8Si8O12 has been reacted with a number of 1,2-unsaturated

hydrocarbons in the presence of H2PtCl6 as the catalyst to get the addition of the organic

group on the Si atom (Figure 2.4).26 This hydrosilylation reaction is rather general and

can be used to produce a variety of octasilsesquioxanes a8b0, from mono- to completely

substituted.27 The cyclohexyl silsesquioxane (c-C6H11)8Si8O12 can be synthesised via the

catalytic hydrogenation of the corresponding phenyl silsesquioxane (C6H5)8Si8O12.28

(2-C4H3S)8Si8O12 can be brominated with Br2 in the presence of HBr to give

(2-C4Br3S)8Si8O12.29 Modification of the R-group with various functionalities has been

reported fo R = C6H530 and p-ClCH2C6H4.31

Figure 2.4. Hydrosilylation reaction of 1,2-unsaturated hydrocarbons with H8Si8O12.

2.2.4 Corner-capping reactions

This method has been used by Feher and coworkers to obtain the cubic

silsesquioxane a8b0 from the incompletely condensed structure a7b3 (Figure 2.5).1 The

reaction of the cyclohexyl silsesquioxane (c-C6H11)7Si7O9(OH)3 with an

organotrichlorosilane RSiCl3 in the presence of an amine offers a straightforward route

to various monosubstituted octasilsesquioxanes.

Figure 2.5. Corner-capping reaction (Chex = c-C6H11).

Si H + H2C CH Si CH2CH2

H2PtCl6Si H + H2C CH Si CH2CH2

H2PtCl6

SIOR

OSi

SiSi

Si

Si

Si

Si

OO

O

OOO

O

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

ORSiCl3

Si

SiSi

Si

Si

Si

Si

OO

O

OOO

O

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

OHOH

OH

SIOR

OSi

SiSi

Si

Si

Si

Si

OO

O

OOO

O

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

O

SIOR

OSi

SiSi

Si

Si

Si

Si

OO

O

OOO

O

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

ORSiCl3

Si

SiSi

Si

Si

Si

Si

OO

O

OOO

O

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

OHOH

OH

Si

SiSi

Si

Si

Si

Si

OO

O

OOO

O

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

OHOH

OH

Chapter 2 __________________________________________________________________________________________________________

22

2.3 Characterisation

2.3.1 NMR spectroscopy

Silsesquioxane structures have at least three groups of nuclei that can be

analysed by NMR techniques: 1H, 13C and 29Si. The NMR spectrum of each of these

three provides useful information for the characterisation of silsesquioxane species. 1H NMR spectroscopy is used to determine the number of silanol groups present in

incompletely condensed silsesquioxanes and to establish if they are hydrogen bonded to

other silanol groups (δ 5 / 7.5 ppm, e.g. the peak marked with × in Figure 2.6) or if they

are isolated (δ 2 / 4 ppm). 13C NMR spectroscopy often allows to distinguish the ipso carbon atoms, i.e. those

bound to a silicon atom, from the other carbons of the R-group. For instance, in the

a7b3 silsesquioxane (c-C6H11)7Si7O9(OH)3 the ipso carbons are in a different region

(δ 22 / 25 ppm) from the remaining CH2 groups (δ 26.5 / 30 ppm) (Figure 2.6).32 29Si NMR spectroscopy is the most useful, though also the most time-consuming, NMR

technique to characterise silsesquioxanes. The 29Si NMR spectrum of a silsesquioxane

provides information about the symmetry of the structure and about the coordination

sphere of the Si atoms. For example, the 29Si NMR spectrum of the a7b3 silsesquioxane

(c-C6H11)7Si7O9(OH)3 displays three peaks with integral ratios 3:1:3, in agreement with

the C3v symmetry of the molecule (Figure 2.6).32 The position of the peaks accounts for

the presence of silanol groups: peaks at lower shifts are due to Si exclusively connected

to other Si atoms through oxygen bridges (δ -68 / -70 ppm), while peaks of Si atoms

connected to an -OH group are at higher shifts (around -60 ppm). More generally, the

position of the peaks also depends on the nature of the R-group connected to the silicon,

while the shift between peaks from silicons with a different number of -OH groups is

approximately constant (~10 ppm per -OH group).33 The presence of geometrically

strained structures also causes the Si peak position to be moved towards higher shifts:

for example, the completely condensed silsesquioxane (c-C6H11)6Si6O9 gives a single

peak at ~-56 ppm, while the less strained (c-C6H11)8Si8O12 gives a single peak at

~-69 ppm.1


23

Figure 2.6. 1H, 13C and 29Si NMR spectra of cyclohexyl silsesquioxane 13.

2.3.2 Mass spectrometry

Mass spectrometry (MS) is slowly becoming a standard technique to analyse

silsesquioxane species. Various ionisation techniques have been used to study

silsesquioxanes with mass spectrometry: electronspray ionisation (ESI),34-36

atmospheric pressure chemical ionisation (APCI),37 turbo ion-spray (TISP)37 and

matrix-assisted laser desorption ionisation (MALDI).36,38,39 In principle, it is not

possible to predict which ionisation technique is the most convenient to analyse a

Chapter 2 __________________________________________________________________________________________________________

24

certain silsesquioxane sample. Advantages of mass spectrometry are its rapidity and the

low amount of sample needed for the analysis (10-50 µg/ml). A drawback is the

impossibility to have an absolute intensity scale, limiting the results to a qualitative

level. For this reason MS is often coupled to other analytic techniques, amongst which

is usually NMR. The applications of MS range from the assignment of structures to

unknown silsesquioxane species1,37 to the study of the species formed during the

synthesis of silsesquioxanes.34,35

2.3.3 IR spectroscopy

Infrared spectroscopy has been applied to study many silsesquioxane structures.

Characteristic vibrations present in the IR spectra of silsesquioxanes are:4 the Si-O-Si

asymmetric stretching, at 1100-1140 cm-1 or at 1057-1085 cm-1 in case of strained

geometry; deformational vibrations of the silicon-oxygen framework, in the region

between 360 and 600 cm-1; the Si-R vibrations, which are strongly influenced by the

nature of the R-group. 6,17,18,40 Recently, the information provided by IR spectroscopy

was used as the basis of a theoretical study of the structure of an organometallic

silsesquioxane complex.41

2.3.4 X-ray diffraction

X-ray single-crystal diffraction is used to obtain a structural characterisation of

silsesquioxane species that can be synthesised in crystalline form, providing information

about bond distances and angles on the basis of which plots of the silsesquioxane

structures can be drawn. Crystallisation of silsesquioxanes is usually difficult and even

apparently suitable crystals often show pronounced disorder.

2.4 Completely condensed silsesquioxanes (RSiO1.5)a

These small, discrete cage compounds are usually referred to as polyhedral

oligomeric silsesquioxanes (POSS). No structure is known with a = 2, probably because

the compound would be geometrically too strained. For a ≥ 4, completely condensed


25

silsesquioxanes with various R-groups have been synthesised and characterised

(Figure 2.1).5

2.4.1 R4Si4O6 (a4b0)

The structure of this silsesquioxane can be represented as a distorted triangular

pyramid (compound 1 in Figure 2.1); the Si-O framework has a Td symmetry. This

structure is very geometrically strained and this makes the silsesquioxane very unlikely

to be formed. In fact, R4Si4O6 silsesquioxane has only been claimed in a very early

publication42 in the case of R = tert-butyl.

2.4.2 R6Si6O9 (a6b0)

The completely condensed silsesquioxanes R6Si6O9 have the shape of a prism,

with two distorted-triangular faces and three distorted-square faces (structure 2, Figure

2.1). The Si-O framework presents a D3h symmetry. The distorted triangular faces

contain three Si atoms, each bridged by three O atoms: this arrangement is somewhat

strained and this accounts for the fact that this silsesquioxane structure is usually found

just as a by-product. This structure has been reported for R = methyl,11,17 ethyl,12 heptyl,

octyl, isononyl,43 phenyl44 and cyclohexyl.6,1.

Figure 2.7. X-ray structure of methyl silsesquioxane a8b0.

Chapter 2 __________________________________________________________________________________________________________

26

2.4.3 R8Si8O12 (a8b0)

The completely condensed silsesquioxanes R8Si8O12 are commonly represented

as cubic structures (compound 3, Figure 2.1). However, the structure is rather distorted

from that of a regular cube, as it has been determined by X-ray analysis:45,46 the Si atoms

tend to keep a tetrahedral geometry, causing the O atoms to point outwards from the cube

edges (Figure 2.7). The Si-O framework presents an Oh symmetry. This silsesquioxane

structure gained interest as its similarity to the Si8O12 building block in zeolite LTA and

zeolite AFY was recognised (Figure 2.8).47

Figure 2.8. Comparison of zeolite LTA (top) and zeolite FAU (bottom) structures with

silsesquioxanes T8 (a8b0) and T12 (a12b0), respectively.

T8

T12Si SiO

O

R

RO

R

O

O

R R

O

O

O

R

Si

Si

O

R

R

Si

Si

O

R

SiSi

Si Si

SiSi

RR R

O O

OO

O

O O

O

SiSi

Si Si

O

O

O

O R

R

R

R R

RO

Si Si

Si Si

O

O

OO

R

OO

O

R

T8

T12Si SiO

O

R

RO

R

O

O

R R

O

O

O

R

Si

Si

O

R

R

Si

Si

O

R

SiSi

Si Si

SiSi

RR R

O O

OO

O

O O

O

Si SiO

O

R

RO

R

O

O

R R

O

O

O

R

Si

Si

O

R

R

Si

Si

O

R

SiSi

Si Si

SiSi

RR R

O O

OO

O

O O

O

SiSi

Si Si

O

O

O

O R

R

R

R R

RO

Si Si

Si Si

O

O

OO

R

OO

O

RSiSi

Si Si

O

O

O

O R

R

R

R R

RO

Si Si

Si Si

O

O

OO

R

OO

O

R


27

The fact that silsesquioxanes R8Si8O12 could be seen as model compounds for zeolites

encouraged many theoretical calculations of the energies and structural features of these

silsesquioxanes and particularly of H8Si8O12.47-50 Due to the increase in computational

power, the most recent calculations could be performed on an ab initio quantum

mechanical level and were, therefore, more rigorous than the previous ones.49,50 The

total energy of H8Si8O12 was compared to that of other completely condensed

hydrosilsesquioxanes (HSiO1.5)a, showing that hydrosilsesquioxanes with lower values

of a are less stable than H8Si8O12, while those with higher a values have a similar

energy. This same trend was found using different calculation methods (HF49 and

DFT50).

The completely condensed silsesquioxane R8Si8O12 has been obtained either as a main

or as a by-product from the hydrolytic condensation of various monosilanes:

R = hydrogen,18,51 linear alkyl groups (methyl,11,52,53 ethyl,12,52 n-propyl,52 n-butyl,52

n-pentyl,54 n-hexyl43), vinyl,55 cyclohexyl,52 phenyl,13,40,54 benzyl,28 m-tolyl,28

3,5-dimethylphenyl28 and para-substituted phenyl.31 Alternative synthetic methods are

reported in Paragraphs 2.2.3 and 2.2.4. Monosubstituted octasilsesquioxanes RIR7Si8O12

with different R- and RI-groups have also been synthesised.27 Octasilsesquioxanes with

R being a saturated alkyl group are rather stable. More reactive is the Si-H group in the

octahydrosilsesquioxane H8Si8O12, even if still not so reactive when compared to Si-H

groups in non-cyclic silanes.5 H8Si8O12 can be obtained in relatively good yields

(17.5%) by the hydrolytic condensation of HSiCl3 in a biphasic aqueous HCl/hexane

system, using partially hydrated FeCl3 as water source.51 H8Si8O12 can react with

different substrates to give a complete substitution of the H atoms with other groups, i.e.

Cl,56 OCH3,56 OSi(CH3)3,57 OSn(CH3)3,58 OSb(CH3)458 and a number of organic groups

by hydrosilylation reactions (see Paragraph 2.2.3). Another reactive octasilsesquioxane

is (CH2=CH)8Si8O12, where the reactive moiety is the double bond of the vinyl groups.

H8Si8O12 and (CH2=CH)8Si8O12 have been used as precursors for hybrid

organic-inorganic polymers: a detailed description of these polymeric species is outside

the purposes of this dissertation and can be found elsewhere.5,59 H8Si8O12,60-62

(CH2=CH)8Si8O1263-65 and (p-ICH2C6H4)8Si8O12

66 have been used as precursors for

dendrimers with silsesquioxane cores (Figure 2.9). Among these highly-branched

molecules, particularly interesting for their application to catalysis are those containing

phosphine groups that can act as ligands for metal centres.64-68 The dendrimer 17 with

Chapter 2 __________________________________________________________________________________________________________

28

16 PPh2 arms shown in Figure 2.9 was used as a ligand for Rh-complexes, producing a

much more selective catalyst for the hydroformylation of 1-octene than those obtained

using smaller phosphine ligands.65 The improved selectivity was ascribed to the bidentate

coordination of the Rh-centre, favoured by the high steric crowding of the ligand.

Figure 2.9. a8b0-based silsesquioxane dendrimer.

2.4.4 R10Si10O15 (a10b0)

The completely condensed silsesquioxanes R10Si10O15 have the shape of a prism,

with two distorted-pentagonal faces and four distorted-square faces (structure 4,

Figure 2.1). The Si-O framework presents a D5h symmetry. This silsesquioxane

structure has been synthesised and characterised for a number of R-groups, including:

H,18,51 CH3,17,52 C2H5,69 and C6H5.40 R10Si10O15 is never found as the only product, but

usually as a by-product in the synthesis of R8Si8O12. Few reactions are known for

silsesquioxanes R10Si10O15 and all involve the substitution of the H atoms in

silsesquioxane H10Si10O15.70

Si

PPh2PPh2

Si

PPh2

Ph2P

Si

Ph2P

Ph2P

SiSi

Si

Si

Si

PPh2

Ph2P

PPh2 PPh2PPh2

PPh2

PPh2

PPh2

PPh2

PPh2

17

Si

PPh2PPh2

Si

PPh2

Ph2P

Si

Ph2P

Ph2P

SiSi

Si

Si

Si

PPh2

Ph2P

PPh2 PPh2PPh2

PPh2

PPh2

PPh2

PPh2

PPh2Si

PPh2PPh2

Si

PPh2

Ph2P

Si

Ph2P

Ph2P

SiSi

Si

Si

Si

PPh2

Ph2P

PPh2 PPh2PPh2

PPh2

PPh2

PPh2

PPh2

PPh2

17


29

2.4.5 R12Si12O18 (a12b0)

For the completely condensed silsesquioxanes R12Si12O18, two unstrained

structures can be drawn: one has the shape of a prism, with two distorted-hexagonal

faces and six distorted-square faces (structure 5, Figure 2.1), the other consists of four

distorted-pentagonal faces and four distorted-square faces (structure 6, Figure 2.1). The

Si-O framework of 5 has a D6h symmetry, while that of 6 has a D2d symmetry.

Theoretical calculations showed that the two isomeric structures 5 and 6 have very

similar energies.49,50 R12Si12O18 silsesquioxanes have been obtained in small amounts

from the hydrolytic condensation of monosilanes, for R = H18,71, CH317 and C6H5.40

R12Si12O18 structure 5 is interesting since it can be seen as a model compound for a

building block present in zeolite FAU (Figure 2.8).

2.4.6 Higher completely condensed silsesquioxanes (a > 12)

Completely condensed structures with a > 12 are uncommon, though, some

examples have been reported.4,71

2.5 Incompletely condensed silsesquioxanes

Incompletely condensed silsesquioxanes (RSiO1.5)a(H2O)0.5b contain Si-OH

groups (the number of silanol groups is given by the value of b). The presence of these

silanol groups makes these silsesquioxanes suitable as soluble model compounds for

surfaces of silica and silicate-based materials.72,73 Particularly, incompletely condensed

silsesquioxanes with an extensive silicon-oxygen framework (Figure 2.2, structures with

a ≥ 6) should have electronic properties similar enough to those of silica.32 As their

homologues on silica surfaces, the silanol groups on incompletely condensed

silsesquioxanes such as a7b3 can bind metal centres (Figure 2.10). The so-obtained

complexes can be used as model compounds to study active metal sites at the surface of

heterogeneous catalysts at a molecular level.74 Additionally, they may themselves result

in homogeneous catalysts with comparable or even higher activity than their

heterogeneous counterparts.75,76 Analogously to what happens on silica surfaces, silanol

Chapter 2 __________________________________________________________________________________________________________

30

groups on incompletely condensed silsesquioxanes with a rigid framework will retain

their geometry after the insertion of a metal centre, thereby determining the

coordination geometry around the metal atom.32 It can be concluded that incompletely

condensed silsesquioxanes as ligands for metal centres can be excellent models for

silica-supported systems, given their chemical, electronic and geometric similarity.

These silsesquioxanes are usually obtained by the hydrolytic condensation of

monosilanes: the factors that favour the formation of incompletely rather than

completely condensed species are manifold, as discussed in Paragraph 2.2.1. An

alternative synthetic path is the selective cleavage of Si-O-Si bonds described in

Paragraph 2.2.2.

Figure 2.10. Reaction of silsesquioxane 13 and its trimethylsilylated derivatives with

metal centres.

MOL

OL

Si

SiSi

Si

Si

Si

Si

OHO

O

OH

O

OO

OH

OO

O

R

R

R

R

R R

RO

Si

SiSi

Si

Si

Si

Si

OHO

O

OH

O

OO

OH

OO

O

R

R

R

R

R R

RO

Si

SiSi

Si

Si

Si

Si

OHO

O

OH

O

OO

OH

OO

O

R

R

R

R

R R

RO

SiSi

Si Si

O

O

O

O R

R

R

R R

RO

Si M

Si Si

O

O

OO

L

OO

O

RML4

Si

SiSi

Si

Si

Si

Si

OHO

O

OH

O

OO

OSiMe3

OO

O

R

R

R

R

R R

RO

Si

SiSi

Si

Si

Si

Si

OHO

O

OSiMe3

O

OO

OSiMe3

OO

O

R

R

R

R

R R

RO

Me3SiCl

Me3SiCl

5% Et3N

5% Et3N

ML4

ML4

Si

SiSi

Si

Si

Si

Si

OO

O

OO

OSiMe3

OO

O

R

R

R

R

R R

RO

Si

SiSi

Si

Si

Si

Si

OO

O

OSiMe3

O

OO

OSiMe3

OO

O

R

R

R

R

R R

RO

ML3

MOL

OLMO

L

OL

Si

SiSi

Si

Si

Si

Si

OHO

O

OH

O

OO

OH

OO

O

R

R

R

R

R R

RO

Si

SiSi

Si

Si

Si

Si

OHO

O

OH

O

OO

OH

OO

O

R

R

R

R

R R

RO

Si

SiSi

Si

Si

Si

Si

OHO

O

OH

O

OO

OH

OO

O

R

R

R

R

R R

RO

Si

SiSi

Si

Si

Si

Si

OHO

O

OH

O

OO

OH

OO

O

R

R

R

R

R R

RO

Si

SiSi

Si

Si

Si

Si

OHO

O

OH

O

OO

OH

OO

O

R

R

R

R

R R

RO

Si

SiSi

Si

Si

Si

Si

OHO

O

OH

O

OO

OH

OO

O

R

R

R

R

R R

RO

SiSi

Si Si

O

O

O

O R

R

R

R R

RO

Si M

Si Si

O

O

OO

L

OO

O

RML4

Si

SiSi

Si

Si

Si

Si

OHO

O

OH

O

OO

OSiMe3

OO

O

R

R

R

R

R R

RO

Si

SiSi

Si

Si

Si

Si

OHO

O

OSiMe3

O

OO

OSiMe3

OO

O

R

R

R

R

R R

RO

Me3SiCl

Me3SiCl

5% Et3N

5% Et3N

ML4

ML4

Si

SiSi

Si

Si

Si

Si

OO

O

OO

OSiMe3

OO

O

R

R

R

R

R R

RO

Si

SiSi

Si

Si

Si

Si

OO

O

OSiMe3

O

OO

OSiMe3

OO

O

R

R

R

R

R R

RO

ML3


31

2.5.1 RSi(OH)3 (a1b3)

This incompletely condensed silsesquioxane (Figure 2.2, compound 7) is

extremely reactive and tends to condense to give higher silsesquioxane species.

Nevertheless, in the case of some rather bulky R-groups, the isolation and

characterisation of the trisilanol monomer has been reported (R = tert-butyl,7 phenyl,8

cyclohexyl,9,77 tris(trimethylsilyl)silyl and tris(trimethylsilyl)methyl10). Typically, the

synthesis of RSi(OH)3 is performed by careful hydrolysis of the corresponding

trichlorosilane RSiCl3 in the presence of aniline as hydrochloric acid acceptor.8

2.5.2 R2Si2O(OH)4 (a2b4)

When silsesquioxanes are synthesised via the hydrolytic condensation of

monosilanes, the first step after the formation of the trisilanol monomer RSi(OH)3 is

supposed to be its condensation to give the dimer R2Si2O(OH)4 (Figure 2.2, compound 8).

As the monomer, in general, this compound is not very stable and tends to condense

further to give higher silsesquioxanes. Just two reports of R2Si2O(OH)4 are known, for

R = tert-butyl78 and cyclohexyl,6 two bulky substituents. In the case of R = tert-butyl, just

two other silsesquioxanes have been reported: the monomer ButSi(OH)37 and the completely

condensed But4Si4O6.42 In the case of R = cyclohexyl, many incompletely and completely

condensed silsesquioxanes have been reported, 1,6,9,52 ranging from the monomer

(c-C6H11)Si(OH)3 to the cubic (c-C6H11)8Si8O12: the preferential formation of any of them

is strongly dependent on the reaction conditions, as described in Paragraph 2.2.1.

2.5.3 R4Si4O4(OH)4 (a4b4)

This incompletely condensed silsesquioxane has been synthesised and isolated

for R = phenyl,13,79 cyclopentyl80 and isopropyl81 as a square planar structure

(Figure 2.2, compound 9). (C6H5)4Si4O4(OH)4 and (c-C5H9)4Si4O4(OH)4 are obtained

via the hydrolytic condensation of the respective trichlorosilanes, while the method for

the synthesis of (i-C3H7)4Si4O4(OH)4 is somewhat more complex:

(isopropyl)diphenylchlorosilane first undergoes a hydrolytic condensation, followed by

the substitution of the phenyl groups with chlorine atoms and, finally, by a second

Chapter 2 __________________________________________________________________________________________________________

32

hydrolytic condensation. In principle, different isomers of the R4Si4O4(OH)4 structure

are possible, depending on the relative position of the four OH groups. A detailed

characterisation by NMR and X-ray allowed to determine that the actual structure is the

cis-cis-cis one, with all the OH groups on the same side of the plane of the

molecule.79,81 This result is supported by quantum mechanical calculations which

suggest that such structure is stabilised by the hydrogen bonding among the four OH

groups.15 Both (C6H5)4Si4O4(OH)4 and (i-C3H7)4Si4O4(OH)4 have been used as

precursors to synthesise other silsesquioxanes, including completely condensed and

polymeric species.79,81

2.5.4 R6Si6O7(OH)4 (a6b4)

Three incompletely condensed silsesquioxanes corresponding to the formula

R6Si6O7(OH)4 have been isolated and characterised, with R = cyclohexyl,24

cycloheptyl16 and norbornyl.82 Like for all the other mentioned incompletely condensed

silsesquioxanes, the organic groups are rather bulky. The framework of these

silsesquioxane structures presents a C2v symmetry and can be described as two Si4O4

planes sharing an edge or as an R6Si6O9 structure with two hydrolysed Si-O-Si bonds

(Figure 2.2, compound 11). (c-C6H11)6Si6O7(OH)4 is obtained by the base-catalysed

cleavage of silsesquioxane (c-C6H11)Si6O9.24 Both the cycloheptyl and the norbornyl

structures are obtained via the hydrolytic condensation of the corresponding

trichlorosilanes. R6Si6O7(OH)4 is not the only silsesquioxane product: in the case of

cycloheptyl, (c-C7H13)6Si6O7(OH)4 [~40%] is obtained together with

(c-C7H13)7Si7O9(OH)3 [~60%]; in the case of norbornyl, many other silsesquioxanes

besides (C7H11)6Si6O7(OH)4 are synthesised. (c-C7H13)6Si6O7(OH)4 has recently found

application as ligand for a water-stable titanium-catalyst that is claimed to be active in

the epoxidation of alkenes with both TBHP and H2O2.72 The catalyst is obtained by

reacting TiCl4 with (c-C7H13)6Si6O7(OH)4; the resulting cluster contains four titanium

atoms and three silsesquioxane units, as determined by X-ray analysis.


33

2.5.5 R7Si7O9(OH)3 (a7b3)

This incompletely condensed silsesquioxane has been synthesised in significant

quantities, isolated and characterised for three bulky, cyclic R-groups: cyclopentyl,16

cyclohexyl1,6 and cycloheptyl.16 The framework of this silsesquioxane is characterised

by a C3v symmetry and can be described as a cubic R8Si8O12 structure with a missing

corner (Figure 2.2, compound 13).

R7Si7O9(OH)3 is the silsesquioxane structure with the largest number of applications as

a ligand for metal centres and as a model compound for silica and silicate-based

materials surfaces. The compound with R = cyclohexyl, (c-C6H11)7Si7O9(OH)3, was first

synthesised and studied by Brown and Vogt.6 It was prepared by the hydrolytic

condensation of the trichlorosilane (c-C6H11)SiCl3 in aqueous acetone. The reaction was

extremely slow and proceeded for approximately 3 years to give a final yield in

(c-C6H11)7Si7O9(OH)3 of 60-70%. Given the interest of this silsesquioxane structure,

Feher et al. studied in detail the same reaction for cyclohexyl as well as for cyclopentyl

and cycloheptyl as R-groups,1,16 and were able to isolate and identify some other

silsesquioxane structures formed in smaller amounts together with the main product

R7Si7O9(OH)3. In the case of R = cyclopentyl, the a7b3 silsesquioxane was almost the

only product and the synthesis was optimised by refluxing the reaction mixture to give a

yield of 29% in pure (c-C5H9)7Si7O9(OH)3 after 3 days.16 A further optimisation of the

synthesis of (c-C5H9)7Si7O9(OH)3 has been realised on the basis of High-Speed

Experimentation results20 by using acetonitrile instead of acetone as a solvent, thus

getting to a yield of 64% in 18 hours.19

An alternative way to synthesise R7Si7O9(OH)3 is by the base-mediated cleavage of

R8Si8O12.25

Silsesquioxane R7Si7O9(OH)3 contains three adjacent silanol groups that can react as a

tridentate ligand with a metal centre to give a complex (Figure 2.10). In some cases, it is

preferred to have the silsesquioxane doubly or even just singly coordinated to the metal

centre. For this purpose, it can be useful to let one or two of the silanol groups react

beforehand with trimethylchlorosilane [TMSCl, (CH3)3SiCl] (Figure 2.10). The

silylation of R7Si7O9(OH)3 was studied by Feher and Newman, providing also a

mechanistic study for the silylation of silica surfaces.83 It was found that silylation is

favoured for mutually hydrogen-bonded silanol groups with respect to isolated, not

Chapter 2 __________________________________________________________________________________________________________

34

hydrogen-bonded silanols. Moreover, it was possible to selectively silylate just one of

the three silanols of R7Si7O9(OH)3, particularly if the reaction was carried out in the

presence of an amine base like Et3N. Successive silylations were not as selective and

afforded a mixture of the di- and trisilylated silsesquioxanes.

An increasing number of reactions involving the reaction of R7Si7O9(OH)3

silsesquioxane and its silylated derivatives with metal centres have been reported in the

last years. Several of these metallasilsesquioxanes find applications as model

compounds for metal centres grafted on the surface of silica and silicate-based

heterogeneous catalysts and as homogeneous catalysts themselves.

2.5.5.1 Metallasilsesquioxane catalysts for the epoxidation of alkenes

In 1994 Maschmeyer et al.84 reported the synthesis of (c-C6H11)7Si7O12Ti(η5-C5Ph5)

(Figure 2.11, 18a) from (η5-C5Ph5)2TiCl2 and the trisilanol (c-C6H11)7Si7O9(OH)3 which

led in 1995 to the highly successful grafting procedure of Ti(η5-C5H5)2Cl2 onto silica

MCM-41, yielding the most active (per mole of titanium) heterogeneous epoxidation

catalyst of alkenes with organic hydroperoxides to date.85

Similarly, when Ti(η5-C5H5)Cl3 was reacted with the trisilanol (c-C6H11)7Si7O9(OH)3 the

titanium silsesquioxane (c-C6H11)7Si7O12Ti(η5-C5H5) was produced (Figure 2.11, 18b).86,87

In 1997 Abbenhuis et al.88 tested this complex as catalyst for the epoxidation of alkenes

with tert-butyl hydroperoxide (TBHP) as the oxidant, obtaining good conversions and

selectivities for a number of substrates. The catalyst was reported not to leach the titanium

centre nor the cyclopentadienyl ligand during the epoxidation reaction and has a remarkable

stability towards hydrolysis even in 1M HCl solution. In order to heterogenise this active

homogeneous catalyst, (c-C6H11)7Si7O12Ti(η5-C5H5) was immobilised on the channels of

MCM-41 via a straightforward adsorption.89 The heterogeneous catalyst was tested for

epoxidation showing a good activity, even if lower than that of the homogeneous

(c-C6H11)7Si7O12Ti(η5-C5H5). Leaching of the Ti-silsesquioxane occurred just for Al-containing

MCM-41 samples. The adsorption of (c-C6H11)7Si7O12Ti(η5-C5H5) on all-silica MCM-41

was investigated by spectroscopic techniques, demonstrating that the Ti-silsesquioxane is

unaltered as adsorbed species and suggesting that the catalytically-active Ti-centre points

towards the centre of the MCM-41 pore.90 The interaction between the silsesquioxane

complex and the MCM-41 surface takes place through the cyclohexyl groups.


35

Figure 2.11. Metallasilsesquioxanes (Cpen = c-C5H9, Chex = c-C6H11).

CrOO

OOSi

SiSi

Si

Si

Si

Si

OO

O

OO

OSiMe3

OO

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

VOO

OSi

SiSi

Si

Si

Si

Si

OO

O

OOO

O

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

O

MgOO

Si

SiSi

Si

Si

Si

Si

OO

O

OO

O

OO

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

Ti

Cl

ClCl

MoOO

N

Si

SiSi

Si

Si

Si

Si

OO

O

OO

OSiMe3

OO

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

H

MeMe

PriPri

Si

Si

O

OO

OSi

SiO

OO

Cpen

Cpen

Cpen

OCpen

OSi

Si

SiO

O Cpen

Cpen

Cpen

O

SiOO

Si

SiSi

Si

Si

Si

OO

O

OOO

O

O

Cpen

Cpen

Cpen

Cpen

Cpen

Cpen

O

Cpen

O

Mg

Mg

Mg

Mg

Cl

THF

THF

Cl

OOH

Si

SiSi

Si

Si

Si

Si

OO

O

OO

OSiMe3

OO

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

SiOO

Si

SiSi

Si

Si

Si

OO

O

OO

O

O

Chex

Chex

Chex

Chex

Chex

Chex

O

Chex

O

Ti

Me3SiO

Cp*

Cp* = η5-C5Me5

Chex

O

Si Si

SiSi

Si

Si

Si

Si

O O O

O

OO

O

O

O

O

Chex

Chex

Chex

O

Chex

ChexO

Chex

Chex

Al

Al

n

SiSi

Si Si

O

O

O

O R

R

R

R R

RO

Si Ti

Si Si

O

O

OO

L

OO

O

R

a L = (η5-C5Ph5)

b L = (η5-C5H5)

c L = CH2Ph

d L = NMe2

e L = OSiMe3

f L = OMe

g L = OPri

h L = OBu

R = Cpen, Chex18 19

20 21

22 23

24 25

CrOO

OOSi

SiSi

Si

Si

Si

Si

OO

O

OO

OSiMe3

OO

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

CrOO

OOSi

SiSi

Si

Si

Si

Si

OO

O

OO

OSiMe3

OO

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

VOO

OSi

SiSi

Si

Si

Si

Si

OO

O

OOO

O

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

O

VOO

OSi

SiSi

Si

Si

Si

Si

OO

O

OOO

O

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

O

MgOO

Si

SiSi

Si

Si

Si

Si

OO

O

OO

O

OO

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

Ti

Cl

ClCl

MgOO

Si

SiSi

Si

Si

Si

Si

OO

O

OO

O

OO

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

Ti

Cl

ClCl

MoOO

N

Si

SiSi

Si

Si

Si

Si

OO

O

OO

OSiMe3

OO

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

H

MeMe

PriPri

MoOO

N

Si

SiSi

Si

Si

Si

Si

OO

O

OO

OSiMe3

OO

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

H

MeMe

PriPri

Si

Si

O

OO

OSi

SiO

OO

Cpen

Cpen

Cpen

OCpen

OSi

Si

SiO

O Cpen

Cpen

Cpen

O

SiOO

Si

SiSi

Si

Si

Si

OO

O

OOO

O

O

Cpen

Cpen

Cpen

Cpen

Cpen

Cpen

O

Cpen

O

Mg

Mg

Mg

Mg

Cl

THF

THF

Cl

Si

Si

O

OO

OSi

SiO

OO

Cpen

Cpen

Cpen

OCpen

OSi

Si

SiO

O Cpen

Cpen

Cpen

O

SiOO

Si

SiSi

Si

Si

Si

OO

O

OOO

O

O

Cpen

Cpen

Cpen

Cpen

Cpen

Cpen

O

Cpen

O

Mg

Mg

Mg

Mg

Cl

THF

THF

Cl

OOH

Si

SiSi

Si

Si

Si

Si

OO

O

OO

OSiMe3

OO

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

SiOO

Si

SiSi

Si

Si

Si

OO

O

OO

O

O

Chex

Chex

Chex

Chex

Chex

Chex

O

Chex

O

Ti

Me3SiO

Cp*

Cp* = η5-C5Me5

OOH

Si

SiSi

Si

Si

Si

Si

OO

O

OO

OSiMe3

OO

O

Chex

Chex

Chex

Chex

Chex Chex

OChex

SiOO

Si

SiSi

Si

Si

Si

OO

O

OO

O

O

Chex

Chex

Chex

Chex

Chex

Chex

O

Chex

O

Ti

Me3SiO

Cp*

Cp* = η5-C5Me5

Chex

O

Si Si

SiSi

Si

Si

Si

Si

O O O

O

OO

O

O

O

O

Chex

Chex

Chex

O

Chex

ChexO

Chex

Chex

Al

Al

nChex

O

Si Si

SiSi

Si

Si

Si

Si

O O O

O

OO

O

O

O

O

Chex

Chex

Chex

O

Chex

ChexO

Chex

Chex

Al

Al

Chex

O

Si Si

SiSi

Si

Si

Si

Si

O O O

O

OO

O

O

O

O

Chex

Chex

Chex

O

Chex

ChexO

Chex

Chex

Al

Al

n

SiSi

Si Si

O

O

O

O R

R

R

R R

RO

Si Ti

Si Si

O

O

OO

L

OO

O

R

a L = (η5-C5Ph5)

b L = (η5-C5H5)

c L = CH2Ph

d L = NMe2

e L = OSiMe3

f L = OMe

g L = OPri

h L = OBu

R = Cpen, Chex18

SiSi

Si Si

O

O

O

O R

R

R

R R

RO

Si Ti

Si Si

O

O

OO

L

OO

O

R

a L = (η5-C5Ph5)

b L = (η5-C5H5)

c L = CH2Ph

d L = NMe2

e L = OSiMe3

f L = OMe

g L = OPri

h L = OBu

R = Cpen, Chex18 19

20 21

22 23

24 25

Chapter 2 __________________________________________________________________________________________________________

36

Silsesquioxanes (c-C6H11)7Si7O9(OH)3 and (c-C5H9)7Si7O9(OH)3 were also reacted with

a variety of TiL4 centres (L = CH2Ph, NMe2, OSiMe3, OMe, OPri, OBu) to give

R7Si7O12TiL complexes that were studied as homogeneous epoxidation catalysts and as

model compounds for isolated titanium centres on silicate materials like TS-1 and

MCM-41 (Figure 2.11, 18c-h).74,76,91 All the catalysts were tested for the epoxidation of

1-octene with TBHP: the highest activities were found for L = alkoxy group. The

activities of these catalysts are reported to be comparable to or higher than those of

Ti-MCM-41 (on the basis of turnover frequencies). (c-C5H9)7Si7O12TiOPri is found to

exist as an equilibrium between two forms in solution: a monomer with

tetra-coordinated titanium and a dimer with penta-coordinated titanium.91 In order to

investigate also monodentate and bidentate Ti-silsesquioxanes as models for analogous

species on silicate surfaces, the mono- and disilylated silsesquioxanes

R7Si7O9(OSiMe3)(OH)2 and R7Si7O9(OSiMe3)2OH (R = cyclohexyl, cyclopentyl) were

reacted with Ti-centres.76,92 Studies of the activity of these complexes allowed to

determine that tridentate Ti-silsesquioxanes are more active catalysts than complexes

with lower coordination number of siloxy groups. Moreover, it was shown that

OSi(CH3)3 can weakly coordinate to the Ti-centre, diminishing the access to the active

site and consequently lowering the activity.92 A heterogeneous derivative of a

Ti-complex of (c-C6H11)7Si7O9(OH)3 was prepared by modifying MCM-41 with

(3-glycidyloxypropyl)trimethoxysilane followed by reaction of the oxirane ring with

deprotonated (c-C6H11)7Si7O9(OH)3 and finally by reaction with Ti(OBu)4.93 The

heterogeneous material so-obtained was characterised and tested as catalyst for the

epoxidation of cyclooctene, showing promising results when using H2O2 as oxidant.

2.5.5.2 Metallasilsesquioxane catalysts for the polymerisation of alkenes

The reaction of the disilanol (c-C6H11)7Si7O9(OSiMe3)(OH)2 with CrO3 in the presence

of MgSO4 as dehydrating agent yields the chromium silsesquioxane

(c-C6H11)7Si7O9(OSiMe3)(O2CrO2) (Figure 2.11, 19).94 This complex was found to

catalyse the polymerisation of ethene when activated with Al(CH3)3.

(c-C6H11)7Si7O9(OSiCH3)(O2CrO2) was also proposed as a model compound for the

Phillips catalyst (CrO3 supported on silica or alumina), which is industrially used for the

polymerisation of ethene.


37

Vanadium silsesquioxane (c-C6H11)7Si7O12VO (Figure 2.11, 20) is obtained by the

reaction of (c-C6H11)7Si7O9(OH)3 with (C3H7O)3VO.95 In the presence of Al(CH3)3 or

Al(CH2SiMe3)3 as an activating agent, (c-C6H11)7Si7O12VO catalyses the polymerisation

of ethene and, to a lesser extent, of other alkenes.95,96 A study of the catalytically active

site involving the coordination of the alkylaluminium to the vanadium silsesquioxane

was performed by means of 1H, 13C, 17O, 29Si and 51V NMR spectroscopy.97

The reaction of (c-C6H11)7Si7O9(OH)3 with BuMgEt produces

[(c-C6H11)7Si7O9(OH)O2Mg]n (n = 1,2), which further reacts with TiCl4 to give

[(c-C6H11)7Si7O12MgTiCl3]n (n = 1,2) as a mixture of monomer and dimer (Figure 2.11,

21).75 [(c-C6H11)7Si7O12MgTiCl3]n was studied as a model compound for a

silica-supported catalyst having the same magnesium and titanium content. Remarkably,

the silsesquioxane-based homogeneous catalyst showed a higher activity (per gram of

titanium) than its Ti/Mg/SiO2 heterogeneous counterpart for the polymerisation of

ethene in the presence of Al(CH3)3.

Reaction of (c-C5H9)7Si7O9(OH)3 and its silylated derivatives with Group 4 (Ti, Zr, Hf)

metal centres produced a series of silsesquioxane-based complexes with different

coordination number between the metal and the silsesquioxane structure(s).98-100 These

compounds were used as models for silica-grafted homogeneous alkene-polymerisation

catalysts. Nearly all the prepared silsesquioxane catalysts were active in the

polymerisation of ethene when using methylalumoxane as co-catalyst. Nevertheless, the

metal centre on the silsesquioxane is easily substituted by methylalumoxane and the

so-obtained complex results in the active catalyst.100 When using B(C6F5)3 as milder

cocatalyst, no leaching of the metal centre was registered; still, some of the

metallasilsesquioxanes showed good polymerisation activities.100,101

2.5.5.3 Other metallasilsesquioxanes

The reaction of (c-C5H9)7Si7O9(OH)3 with CH3MgCl in THF gives a complex

containing four Mg atoms and two silsesquioxane ligands (Figure 2.11, 22).102 From

X-ray analysis it was found that the magnesium-chlorine bonds are unusually short,

indicating highly electron deficient magnesium atoms. The complex has been subjected

to transmetalation reactions, showing a low activity.

Aluminosilsesquioxanes as model compounds for acidic sites in zeolites or on silica

surfaces have been obtained by reacting AlMe3 or AlEt3 with (c-C5H9)7Si7O9(OH)3 and

Chapter 2 __________________________________________________________________________________________________________

38

(c-C5H9)7Si7O9(OSiR3)(OH)2, where SiR3 = SiMe3 or SiMe2Ph.103-105 Products have

been characterised by spectroscopic and X-ray techniques.

A series of titanium silsesquioxanes were obtained by the reaction of

(c-C6H11)7Si7O9(OH)3 and its mono- and disilylated derivatives with a number of Ti(III)

and Ti(IV)-centres.106-108 Particularly interesting seems to be the complex obtained by

reacting (c-C6H11)7Si7O9(OSiMe3)(OH)2 with (η5-C5Me5)Ti(η5,η1-C5Me4CH2).107 This

structure contains a Ti-centre coordinated to two silsesquioxane ligands and to a vicinal

Si-OH group (Figure 2.11, 23): this arrangement makes the complex an interesting

model compound for a silica surface Ti-site. However, no catalytic test has been yet

reported for this system.

A number of iron silsesquioxanes has been synthesised by reacting

(c-C5H9)7Si7O9(OH)3 and (c-C5H9)7Si7O9(OSiMe3)(OH)2 with FeCl3P(c-C6H11)3 and

FeCl2(dcpe), where dcpe = bis(dicyclohexylphosphino)ethane.109 Phosphine ligands are

used to prevent oligomerisation of the Fe-centres. These iron silsesquioxanes could be

seen as models for iron zeolites catalysts. Nevertheless, the first attempts to check the

activity of these complexes in the oxidation of benzene were not successful.

The reaction of (c-C6H11)7Si7O9(OSiMe3)(OH)2 with two equivalents of TlOEt affords

the complex (c-C6H11)7Si7O9(OSiMe3)(OTl)2. This compound was then used as

transmetalating agent to obtain (c-C6H11)7Si7O9(OSiMe3)(O2Mo)(CHCMe2Ph)(NAr),

where Ar = 2,6-di-isopropylphenyl) (Figure 2.11, 24).110 The molybdenum

silsesquioxane complex turned out to be a very efficient catalyst for the metathesis of

various alkenes.

Finally, (c-C6H11)7Si7O9(OH)3, (c-C5H9)7Si7O9(OH)3 and related structures have been

reacted with a number of other metal centres,32,86 including Cu,111 Rh,112 W,113 Os,114

Pt,115 Nd and Y116. None of these complexes has yet been tested for catalytic

performance.

2.5.6 R8Si8O11(OH)2 (a8b2)

Two different incompletely condensed silsesquioxane structures correspond to

the general formula R8Si8O11(OH)2 (Figure 2.2, compounds 14 and 15).

Silsesquioxane 14 has been obtained with low yields for R = cyclohexyl as a by-product

of the synthesis of (c-C6H11)7Si7O9(OH)3.1 Since the two silanol groups of this


39

(c-C6H11)8Si8O11(OH)2 silsesquioxane point in opposite directions, the compound can

be used as a bifunctional monodentate ligand for metal species to obtain polymeric

complexes with bridging metal units (Figure 2.11, 25). Reaction of compound 14 with

Al(CH3)3 produced a gel that results in an active heterogeneous catalyst for Diels-Alder

reactions of enones.117 A similar polymeric species is obtained by reacting

compound 14 with TiCl4 or Ti(CH2Ph)4; when tested as catalyst for the epoxidation of

alkenes, the titanium-containing gel undergoes hydrolysis resulting in the reformation of

silsesquioxane 14 and catalytically active Ti-species.88

Silsesquioxane 15 can be obtained by the selective cleavage of a Si-O-Si bond of

R8Si8O12 structures, as described in Paragraph 2.2.2.2,22,25 R8Si8O11(OH)2

silsesquioxane 15 with R = cyclopentyl, cyclohexyl has been used as a ligand for

Al-complexes104 and Ce-complexes.118

2.5.7 Other incompletely condensed structures (a6b2, a7b1, a8b4)

Incompletely condensed silsesquioxanes R6Si6O8(OH)2 and R8Si8O10(OH)4

(Figure 2.2, 10 and 16) have been synthesised by the cleavage of Si-O bonds of,

respectively, R6Si6O923 and R8Si8O12.25 No application of these two structures have been

reported so far. R7Si7O10(OH) silsesquioxanes (Figure 2.2, 12) can be obtained by

dehydration of R7Si7O9(OH)3.1 Few applications of this silsesquioxane, with

R = cyclopentyl or cyclohexyl, have been reported.21,86,92

2.6 Conclusions

The field of oligosilsesquioxane chemistry has expanded significantly over the

last 10 years yielding exciting developments in their synthesis as well as in their

application. It can be expected that these compounds, having left their ‘niche’ existence

(as demonstrated above), will make an increasingly large contribution to new

developments in the fundamental understanding of chemical phenomena (when used as

models for, e.g., zeolites or silica surface chemistry) and to novel applications which are

based on them, either as catalysts or composite material building blocks operating at the

mesostructure level.

Chapter 2 __________________________________________________________________________________________________________

40

References

1 F.J. Feher, D.A. Newman, J.F. Walzer, J. Am. Chem. Soc., 1989, 111, 1741. 2 F.J. Feher, D. Soulivong, A.G. Eklund, Chem. Commun., 1998, 399. 3 R.H. Baney, M. Itoh, A. Sakakibara, T. Suzuki, Chem. Rev., 1995, 95, 1409. 4 M.G. Voronkov, V.I. Lavrent’yev, Top. Curr. Chem., 1982, 102, 199. 5 P.G. Harrison, J. Organomet. Chem., 1997, 542, 141. 6 J.F. Brown, L.H. Vogt, J. Am. Chem. Soc., 1965, 87, 4313. 7 N. Winkhofer, H.W. Roesky, M. Noltemeyer, W.T. Robinson, Angew. Chem. Int. Ed., 1992, 31, 599. 8 T. Takiguchi, J. Am. Chem. Soc., 1959, 81, 2359. 9 H. Ishida, J.L. Koenig, J. Polym. Sci. Polym. Phys. Ed., 1979, 17, 1807. 10 S.S. Al-Juaid, N.H. Buttrus, R.I. Damja, Y. Derouiche, C. Eaborn, P.B. Hitchcock, P.D. Lickiss, J.

Organomet. Chem., 1989, 371, 287. 11 M.M. Sprung, F.O. Guenther, J. Am. Chem. Soc., 1955, 77, 3990. 12 M.M. Sprung, F.O. Guenther, J. Am. Chem. Soc., 1955, 77, 3996. 13 J.F. Brown, J. Am. Chem. Soc., 1965, 87, 4317. 14 T. Kudo, M.S. Gordon, J. Am. Chem. Soc., 1998, 120, 11432. 15 T. Kudo, M.S. Gordon, J. Phys. Chem. A, 2000, 104, 4058. 16 F.J. Feher, T.A. Budzichowski, R.L. Blanski, K.J. Weller, J.W. Ziller, Organometallics, 1991, 10,

2526. 17 L.H. Vogt, J.F. Brown, Inorg. Chem., 1963, 2, 189. 18 C.L. Frye, W.T. Collins, J. Am. Chem. Soc., 1970, 92, 5586. 19 See Chapter 5. 20 P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, Angew. Chem. Int. Ed., 2001, 40,

740. 21 F.J. Feher, D. Soulivong, G.T. Lewis, J. Am. Chem. Soc., 1997, 119, 11323. 22 F.J. Feher, D. Soulivong, F. Nguyen, Chem. Commun., 1998, 1279. 23 F.J. Feher, F. Nguyen, D. Soulivong, J.W. Ziller, Chem. Commun., 1999, 1705. 24 F.J. Feher, R. Terroba, J.W. Ziller, Chem. Commun., 1999, 2153. 25 F.J. Feher, R. Terroba, J.W. Ziller, Chem. Commun., 1999, 2309. 26 D. Herren, H. Bürgy, G. Calzaferri, Helv. Chim. Acta, 1991, 74, 24. 27 C. Marcolli, G. Calzaferri, Appl. Organometal., 1999, 13, 213. 28 F.J. Feher, T.A. Budzichowski, J. Organomet. Chem., 1989, 373, 153. 29 K. Olsson, C. Axen, Arkiv. Kemi., 1964, 22, 237. 30 K. Olsson, C. Gronwall, Arkiv. Kemi., 1961, 17, 529. 31 F.J. Feher, T.A. Budzichowski, J. Organomet. Chem., 1989, 379, 33. 32 F.J. Feher, T.A. Budzichowski, Polyhedron, 1995, 14, 3239. 33 M.P. Besland, C. Guizard, N. Hovnanian, A. Larbot, L. Cot, J. Sanz, I. Sobrados, M. Gregorkiewitz, J.

Am. Chem. Soc., 1991, 113, 1982.


41

34 P. Bussian, F. Sobott, B. Brutschy, W. Scrader, F. Schüth, Angew. Chem. Int. Ed., 2000, 39, 3901. 35 L. Matejka, O. Dukh, J. Brus, W.J. Simonsick jr., B. Meissner, J. Non-Cryst. Solids, 2000, 270, 34. 36 D.P. Fasce, R.J.J. Williams, R. Erra-Balsells, Y. Ishikawa, H. Nonami, Macromolecules, 2001, 34,

3534. 37 R. Bakhtiar, F.J. Feher, Rapid Commun. Mass Spectrom., 1999, 13, 687. 38 W.E. Wallace, C.M. Guttman, J.M. Antonucci, J. Am. Soc. Mass Spectrom., 1999, 10, 224. 39 P. Eisenberg, R. Erra-Balsells, Y. Ishikawa, J.C. Lucas, A.N. Mauri, H. Nonami, C.C. Riccardi, R.J.J.

Williams, Macromolecules, 2000, 33, 1940. 40 J.F. Brown, L.H. Vogt, P.I. Prescott, J. Am. Chem. Soc., 1964, 86, 1120. 41 I.E. Davidova, L.A. Gribov, I.V. Maslov, V. Dufaud, G.P. Niccolai, F. Bayard, J.M. Basset, J. Mol.

Struct., 1998, 443, 67. 42 E. Wiberg, W. Simmler, Z. Anorg. Allg. Chem., 1955, 282, 330. 43 K.A. Adrianov, B.A. Izmaylov, J. Organomet. Chem., 1967, 8, 435. 44 K.A. Adrianov, V.A. Odinets, Izv. Akad. Nauk SSSR Otd. Khim. Nauk, 1959, 460. 45 M.A. Hossain, M.B. Hursthouse, K.M.A. Malik, Acta Cryst., 1979, B35, 2258. 46 G. Koellner, U. Müller, Acta Cryst., 1989, C45, 1106. 47 G. Calzaferri, R. Hoffmann, J. Chem. Soc., Dalton Trans., 1991, 917. 48 C.W. Earley, Inorg. Chem., 1992, 31, 1250. 49 C.W. Earley, J. Phys. Chem., 1994, 98, 8693. 50 K.-H. Xiang, R. Pandey, U.C. Pernisz, C. Freeman, J. Phys. Chem. B, 1998, 102, 8704. 51 P.A. Agaskar, Inorg. Chem., 1991, 30, 2707. 52 A.J. Barry, W.H. Daudt, J.J. Domicone, J.W. Gilkey, J. Am. Chem. Soc., 1955, 77, 4248. 53 M.M. Sprung, F.O. Guenther, J. Am. Chem. Soc., 1955, 77, 6045. 54 M.M. Sprung, F.O. Guenther, J. Polym. Sci., 1958, 28, 17. 55 T.N. Martynova, V.P. Korchkov, Zh. Obshch. Khim., 1982, 52, 1585. 56 V.W. Day, W.G. Klemperer, V.V. Mainz, D.M. Millar, J. Am. Chem. Soc., 1985, 107, 8262. 57 F.J. Feher, K.J. Weller, Organometallics, 1990, 9, 2638. 58 F.J. Feher, K.J. Weller, Inorg. Chem., 1991, 30, 880. 59 J.D. Lichtenhan, Comments Inorg. Chem., 1995, 17, 115. 60 A.R. Bassindale, T.E. Gentle, J. Mater. Chem., 1993, 3, 1319. 61 A.R. Bassindale, T.E. Gentle, J. Organomet. Chem., 1996, 521, 391. 62 F.J. Feher, K.D. Wyndham, Chem. Commun., 1998, 323. 63 P.-A. Jaffrès, R.E. Morris, J. Chem. Soc., Dalton Trans., 1998, 2767. 64 L. Ropartz, R.E. Morris, G.P. Schwarz, D.F. Foster, D.J. Cole-Hamilton, Inorg. Chem. Commun., 2000,

3, 714. 65 L. Ropartz, R.E. Morris, , D.F. Foster, D.J. Cole-Hamilton, Chem. Commun., 2001, 361. 66 F.J. Feher, J.J. Schwab, S.H. Phillips, A. Eklund, E. Martinez, Organometallics, 1995, 14, 4452. 67 B. Hong, T.P.S. Thomas, H.J. Murfee, M.J. Lebrun, Inorg. Chem., 1997, 36, 6146. 68 H.J. Murfee, T.P.S. Thomas, B. Hong, J. Greaves, Inorg. Chem., 2000, 39, 5209.

Chapter 2 __________________________________________________________________________________________________________

42

69 M.G. Voronkov, V.I. Lavrent’yev, V.M. Kovrigin, J. Organomet. Chem., 1981, 220, 285. 70 P.A.. Agaskar, Synth. React. Inorg. Met. Org. Chem., 1990, 20, 483. 71 P.A.. Agaskar, V.W. Day, W.G. Klemperer, J. Am. Chem. Soc., 1987, 109, 5554. 72 H.C.L. Abbenhuis, Chem. Eur. J., 2000, 6, 25. 73 S. Krijnen, R.W.J.M. Hanssen, H.C.L. Abbenhuis, J.H.C. van Hooff, R.A. van Santen, Chem.

Commun., 1999, 501. 74 J.M. Thomas, G. Sankar, M.C. Klunduk, M.P. Attfield, T. Maschmeyer, B.F.G. Johnson, R.B. Bell, J.

Phys. Chem. B, 1999, 103, 8809. 75 J.-C. Liu, Chem. Commun., 1996, 1109. 76 M. Crocker, R.H.M. Herold, A.G. Orpen, Chem. Commun., 1997, 241. 77 H. Ishida, J.L. Koenig, K.C. Gardner, J. Chem. Phys., 1982, 77, 5748. 78 P.D. Lickiss, S.A. Litster, A.D. Redhouse, C.J. Wisener, Chem. Commun., 1991, 173. 79 F.J. Feher, J.J. Schwab, D. Soulivong, J.W. Ziller, Main Group Chem., 1997, 2, 123. 80 T.S. Haddad, B.M. Moore, S.H. Phillips, Polymer Preprints, 2001, 42, 196. 81 M.Unno, K. Takada, H. Matsumoto, Chem. Lett., 1998, 489. 82 T.W. Hambley, T. Maschmeyer, A.F. Masters, Appl. Organomet. Chem., 1992, 6, 253. 83 F.J. Feher, D.A. Newman, J. Am. Chem. Soc., 1990, 112, 1931. 84 L.D. Field, C.M. Lindall, T. Maschmeyer, A.F. Masters, Aust. J. Chem., 1994, 47, 1127. 85 T. Maschmeyer, F. Rey, G. Sankar and J.M. Thomas, Nature, 1995, 378, 159. 86 F.J. Feher, T.A. Budzichowski, K. Rahimian, J.W. Ziller, J. Am. Chem. Soc., 1992, 114, 3859. 87 I.E. Buys, T.W. Hambley, D.J. Houlton, T. Maschmeyer, A.F. Masters, A.K. Smith, J. Mol. Catal.,

1994, 86, 309. 88 H.C.L. Abbenhuis, S. Krijnen, R.A. van Santen, Chem. Commun., 1997, 311. 89 S. Krijnen, H.C.L. Abbenhuis, R.W.J.M. Hanssen, J.H.C. van Hooff, R.A. van Santen, Angew. Chem.

Int. Ed., 1998, 37, 356. 90 S. Krijnen, B.L. Mojet, H.C.L. Abbenhuis, J.H.C. van Hooff, R.A. van Santen, Phys. Chem. Chem.

Phys., 1999, 1, 361. 91 T. Maschmeyer, M.C. Klunduk, C.M. Martin, D.S. Shephard, J.M. Thomas, B.F.G. Johnson, Chem.

Commun., 1997, 1847. 92 M.C. Klunduk, T. Maschmeyer, J.M. Thomas, B.F.G. Johnson, Chem. Eur. J., 1999, 5, 1481. 93 P. Smet, J. Riondato, T. Pauwels, L. Moens, L. Verdonck, Inorg. Chem. Commun., 2000, 3, 557. 94 F.J. Feher, R.L. Blanski, Chem. Commun., 1990, 1614. 95 F.J. Feher, J.F. Walzer, Inorg. Chem., 1991, 30, 1689. 96 F.J. Feher, J.F. Walzer, R.L. Blanski, J. Am. Chem. Soc., 1991, 113, 3618. 97 F.J. Feher, R.L. Blanski, J. Am. Chem. Soc., 1992, 114, 5886. 98 R. Duchateau, H.C.L. Abbenhuis, R.A. van Santen, S.K.-H. Thiele, M.F.H. van Tol, Organometallics,

1998, 17, 5222. 99 R. Duchateau, H.C.L. Abbenhuis, R.A. van Santen, A. Meetsma, S.K.-H. Thiele, M.F.H. van Tol,

Organometallics, 1998, 17, 5663.


43

100 R. Duchateau, U. Cremer, R.J. Harmsen, S.I. Mohamud, H.C.L. Abbenhuis, R.A. van Santen, A.

Meetsma, S.K.-H. Thiele, M.F.H. van Tol, M. Kranenburg, Organometallics, 1999, 18, 5447. 101 R. Duchateau, R.A. van Santen, G.P.A. Yap, Organometallics, 2000, 19, 809. 102 R.W.J.M. Hanssen, A. Meetsma, R.A. van Santen, H.C.L. Abbenhuis, Inorg. Chem., 2001, 40, 4049. 103 R. Duchateau, R.J. Harmsen, H.C.L. Abbenhuis, R.A. van Santen, A. Meetsma, S.K.-H. Thiele, M.

Kranenburg, Chem. Eur. J.,1999, 5, 3130. 104 M.D. Skowronska-Ptasinska, R. Duchateau, R.A. van Santen, G.P.A. Yap, Eur. J. Inorg. Chem., 2001,

133. 105 M.D. Skowronska-Ptasinska, R. Duchateau, R.A. van Santen, G.P.A. Yap, Organometallics, 2001, 20,

3519. 106 F.J. Feher, S.L. Gonzales, J.W. Ziller, Inorg. Chem., 1988, 27, 3440. 107 F.T. Edelmann, S. Giessmann, A. Fischer, Chem. Commun., 2000, 2153. 108 F.T. Edelmann, S. Giessmann, A. Fischer, J. Organomet. Chem., 2001, 620, 80. 109 F. Liu, K.D. John, B.L. Scott, R.T. Baker, K.C. Ott, W. Tumas, Angew. Chem. Int. Ed., 2000, 39,

3127. 110 F.J. Feher, T.L. Tajima, J. Am. Chem. Soc., 1994, 116, 2145. 111 F.T. Edelmann, S. Giessmann, A. Fischer, Inorg. Chem. Commun., 2000, 3, 658. 112 M. Nowotny, T. Maschmeyer, B.F.G. Johnson, P. Lahuerta, J.M. Thomas, J.E. Davies, Angew. Chem.

Int. Ed., 2001, 40, 955. 113 P. Smet, B. Devreese, F. Verpoort, T. Pauwels, I. Svoboda, S. Foro, J. van Beeumen, L. Verdonck,

Inorg. Chem., 1998, 37, 6583. 114 J.-C. Liu, S.R. Wilson, J.R. Shapley, F.J. Feher, Inorg. Chem., 1990, 29, 5138. 115 H.C.L. Abbenhuis, A.D. Burrows, H. Kooijman, M. Lutz, M.T. Palmer, R.A. van Santen, A.L. Spek,

Chem. Commun., 1998, 2627. 116 W.A. Herrmann, R. Anwander, V. Dufaud, W. Scherer, Angew. Chem., 1994, 106, 1338. 117 H.C.L. Abbenhuis, H.W.G. van Herwijnen, R.A. van Santen, Chem. Commun., 1996, 1941. 118 Y.K. Gun’ko, R. Reilly, F.T. Edelmann, H.-G. Schmidt, Angew. Chem. Int. Ed., 2001, 40, 1279.

45

3

A new, efficient route to titanium-silsesquioxane epoxidation catalysts developed by using High-Speed Experimentation

Abstract

High-Speed Experimentation techniques have been applied to the synthesis and testing

of silsesquioxane-based titanium catalysts for the epoxidation of alkenes. In the

optimisation of the hydrolytic condensation of silanes to incompletely condensed

silsesquioxane structures, different solvents and organotrichlorosilanes were employed.

The HSE approach allowed the identification of a new, fast and straightforward method

to synthesise a silsesquioxane precursor for a Ti-catalyst the performance of which is

comparable to the best silsesquioxane-based titanium catalysts reported in literature. At

the same time, some knowledge was gained about the effects of the solvent and of the

organic group R on the silanes on the synthesis of silsesquioxanes.


P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, Angew. Chem. Int. Ed., 2001, 40, 740.

P.P. Pescarmona, J.J.T. Rops, J.C. van der Waal, J.C. Jansen, T. Maschmeyer, J. Mol. Catal. A, 2002, 182-183, 319.


Chapter 3 __________________________________________________________________________________________________________

46

3.1 Introduction

Titanium centres dispersed on, and supported by, various forms of silica are

catalysts with remarkable properties for partial oxidations.1 The character of the active

site varies with the silica support used. Although it has been established that in the best -

on the basis of activity per gram of titanium - heterogeneous catalysts the Ti-centres are

all tetra-coordinated,2-4 it is still unclear whether, for optimum performance, the active

catalytic species requires the specific configuration of four siloxy groups, or if being

partially hydrolysed to, say, one hydroxy and three siloxy ligands is equally good.

Chemical modelling studies by various groups5-8 have shown that Ti-catalysts with

fewer than four siloxy groups should also be active as catalysts. These considerations

indicate that there is still some residual uncertainty regarding the active site, despite the

many experimental and computational studies carried out to resolve it.9-13 Indeed,

several mechanisms, or variations of one underlying mechanism, might operate

depending on the precise Ti-site configuration. Hence, different types of Ti-centres

might exhibit the desired characteristics.

For these heterogeneous catalysts, which are difficult to characterise, hybrid

inorganic-organic compounds have proved to be useful models. Especially

silsesquioxanes14 (RSiO1.5)a(H2O)0.5b have been a focus of attention as model

compounds for silica - as described in Chapter 2. Particularly interesting for

applications in catalysis are the incompletely condensed silsesquioxanes (b ≠ 0), which

contain Si-OH groups and therefore are able to complex catalytically active centres.14-16

The incompletely condensed silsesquioxane a7b3 has found applications as ligand for a

number of metal centres. Particularly, titanium complexes of silsesquioxane a7b3

resulted in very active homogeneous catalysts for the epoxidation of alkenes

(Figure 3.1).5-8 There is considerable interest in these catalysts, but commercial

introduction is severely hindered by the long and expensive preparation method of the

silsesquioxane precursor a7b3.17,18 Therefore, the identification of a new, faster and

efficient way to synthesise silsesquioxane precursors for titanium catalysts active in the

epoxidation of alkenes was set as the first goal of this research.

A new route to Ti-silsesquioxane catalysts discovered using HSE __________________________________________________________________________________________________________

47

Figure 3.1. Complexation of titanium to the incompletely condensed silsesquioxane

R7Si7O12H3 (R = cyclohexyl, cyclopentyl), yielding a catalyst active in the

the epoxidation of alkenes with TBHP, and a proposed mechanism for the

catalytic cycle.12,13,19

OHOH

OH

R

R

R

R

R R

ROO

SiO

Si

O

O

OSiSi

OSiOSi

O

SiTi(OL)4

THF

OL

R

R

R

R

R R

ROO

SiO

Si

O

O

OSiSi

OSiOSi

O

SiO

OO Ti

ButOOH

OL

Ti

But

LO

Ti

O

OH

O

ButL

O

Ti

OH

ButOH

H2C CH C6H13 + ButOOH H2C CH C6H13 + ButOHTi-catalyst O

OHOH

OH

R

R

R

R

R R

ROO

SiO

Si

O

O

OSiSi

OSiOSi

O

SiTi(OL)4

THF

OL

R

R

R

R

R R

ROO

SiO

Si

O

O

OSiSi

OSiOSi

O

SiO

OO TiOH

OH

OH

R

R

R

R

R R

ROO

SiO

Si

O

O

OSiSi

OSiOSi

O

SiTi(OL)4

THF

OL

R

R

R

R

R R

ROO

SiO

Si

O

O

OSiSi

OSiOSi

O

SiO

OO Ti

ButOOH

OL

Ti

But

LO

Ti

O

OH

O

ButL

O

Ti

OH

ButOH

ButOOH

OL

Ti

OL

Ti

But

LO

Ti

O

OH

But

LO

Ti

O

OH

OO

ButL

O

Ti

OH But

LO

Ti

OH

ButOH



Chapter 3 __________________________________________________________________________________________________________

48

3.2 The High-Speed Experimentation approach

On the basis of a primary screening consisting of a survey of the literature about

silsesquioxanes and of chemical knowledge of the subject, some general conclusions

can be drawn:

• Silsesquioxanes are usually synthesised by the hydrolytic condensation of

organosilanes (RSiX3): this is a multiple-step reaction for which an overall

mechanism is not available.

• Many parameters influence the hydrolytic condensation and determine which

silsesquioxane species are formed and in which amounts.

• The hydrolytic condensation of organosilanes usually produces a mixture of

completely and incompletely condensed silsesquioxanes.

• Incompletely condensed silsesquioxanes different from a7b3 may also act as

precursors for titanium complexes catalytically active in the epoxidation of alkenes.

Given all this information, High-Speed Experimentation techniques20 were chosen as

the most suitable method to investigate the system. Since a good knowledge about

silsesquioxanes was already available, no experimental primary screening was

necessary. In the secondary screening, the synthesis of silsesquioxanes was optimised as

a function of the activity of the catalysts obtained after titanium coordination to the

silsesquioxane structures. Therefore, this approach aimed at producing any incompletely

condensed silsesquioxane that would result in active catalysts after titanium

coordination rather than a specific structure (e.g. silsesquioxane a7b3). The epoxidation

of 1-octene with tert-butyl hydroperoxide (TBHP) as the oxidant was chosen as test

reaction for the activity of the catalysts.6

The first step of this secondary screening was to decide which parameter space had to

been screened. It was hypothesised that the solvent and the R-group have the most

relevant role among the parameters influencing the hydrolytic condensation of

organosilanes (see Paragraph 2.2.1). Therefore, the approach was to study a parameter

space defined by the combination of these two parameters. The choice of extensively

studying a small set of parameters rather than investigating a more varied parameter

space also gives the possibility of more clearly identifying trends concerning the effects

of these parameters on the reaction. It is important that each of the parameters is

sufficiently varied within the parameter space in order to minimise the risk of


49

identifying a local maximum or of missing local trends.

3.3 Results and discussion

The parameter space was defined by the combination of 4 solvents and 10

R-groups (or mixtures of them). The set of R-groups consisted of 8 trichlorosilanes

RSiCl3 (R = cyclohexyl, cyclopentyl, phenyl, methyl, ethyl, tert-butyl, n-octyl and allyl)

and 2 equimolar mixtures of n-octyl- and methyltrichlorosilane and of cyclohexyl- and

phenyltrichlorosilane. Since the reaction includes water among the reagents, only

water-miscible solvents were selected: acetone, acetonitrile, methanol and

tetrahydrofuran (THF). The solvents were chosen on a varying scale of polarity to check

the influence of this property on the synthesis of silsesquioxanes. The parameter space

was screened for the activity of the catalysts obtained by coordination of titanium

isopropoxide, Ti(OPri)4, to the silsesquioxane precursors. Since the goal was to identify

a more efficient way to synthesise silsesquioxane structures, the reaction time for the

hydrolytic condensation was set at 18 hours, a much shorter time than that commonly

required for the synthesis of silsesquioxane a7b3.17,18 Concerning the other parameters,

they were chosen as follows: reaction temperature = 50°C, X = Cl, solvent : water = 4 : 1,

trichlorosilane concentration = 0.136M.

The epoxidation activity of the titanium catalysts as a function of the different

solvents and R-groups varied in the synthesis of the silsesquioxanes precursors is

reported in Figure 3.2. The values are normalised to the activity of the complex

obtained by reacting Ti(OPri)4 with the pure cyclopentyl silsequioxane a7b3 in THF.

The results show some general trends:

• Catalysts derived from silsesquioxane structures which were synthesised in

acetonitrile as solvent show the highest catalytic activity, followed, in decreasing

order of activity, by those from acetone, methanol and from tetrahydrofuran

(Figure 3.2). This trend applies for all the R-groups except ethyl and tert-butyl for

which silsesquioxanes synthesised in THF generate more active catalysts than those

synthesised in acetone and methanol. Interestingly, acetonitrile is more effective

than acetone, the solvent commonly reported for the synthesis of incompletely

condensed silsesquioxane a7b3.17,18 The fact that acetonitrile gives the best results

Chapter 3 __________________________________________________________________________________________________________

50

can be explained on the basis of its high polarity (amongst the solvents used,

acetonitrile has the highest dipole moment and dielectric constant). It has been

proposed that in the presence of a polar molecule the activation barrier for the

condensation reactions towards the formation of silsesquioxanes is reduced and,

therefore, the synthesis is sped up.21,22 Moreover, a highly polar solvent might

stabilise incompletely condensed silsesquioxanes by interaction with their silanol

groups, therefore, favouring the synthesis of these silsesquioxanes over that of the

less polar, completely condensed species.

Figure 3.2. Catalytic activity in the epoxidation of 1-octene with TBHP of the

Ti-silsesquioxanes in the screened parameter space.

• With respect to the silanes employed, the highest epoxidation activities are observed

in the order cyclopentyl > cyclohexyl > phenyl > tert-butyl > ethyl > methyl > allyl ∼

THF

methanolacetonitrile

acetone0.0000.1000.200

0.3000.400

0.500

0.600

0.700

0.800

0.900

1.000

relativeactivity

cyclohexyltrichlorosila

ne

cyclopentyltrichlorosila

ne

phenyltrichlorosila

ne

methyltrichlorosila

ne

ethyltrichlorosila

ne

tert-butyltric

hlorosilane

n-octyltrichlorosila

ne

allyltrichlorosila

ne

n-octyl- & methyltrichlorosila

nes (1:1)

cyclohexyl- & phenyltrichlorosila

nes (1:1)THF

methanolacetonitrile

acetone0.0000.1000.200

0.3000.400

0.500

0.600

0.700

0.800

0.900

1.000

relativeactivity


ne


ne

phenyltrichlorosila

ne

methyltrichlorosila

ne

ethyltrichlorosila

ne

tert-butyltric

hlorosilane


ne

allyltrichlorosila

ne


nes (1:1)


nes (1:1)


51

n-octyl ~ 0, for all solvents apart from THF for which tert-butyl- and

ethyltrichlorosilanes yield more active catalysts than phenyltrichlorosilanes.

(Figure 3.2). This trend is in good agreement with the literature, where cyclopentyl-

and cyclohexyltrichlorosilanes are reported to form incompletely condensed

silsesquioxanes in high yields.17,18 In principle, the nature of the organic group might

influence which silsesquioxane structures are produced and in which ratios,14 as well

as the electronic and spatial properties of the actual catalyst. Considering that the

organic groups are linked to the titanium centre through a Si-O unit, electronic

effects should have a negligible influence on the catalytic properties and, since the

organic groups point away from the titanium centre, steric effects are probably also

not relevant. The principal effect of the organic group is, therefore, considered to be

its role in determining which silsesquioxane structures are formed during the

hydrolytic condensation and in which amounts. The trend identified seems to be

related to the size of the organic substituent: bulky groups probably hinder the

formation of completely condensed silsesquioxanes, which are not able to bind to

titanium centres and therefore to generate active catalysts.

From a methodological point of view, the fact that different points in the parameter

space screened show very different activities confirms the hypothesis that the chosen

parameters have a relevant influence on the hydrolytic condensation: this means that a

representative parameter space was studied. An important question that arose from this

experiment was whether the parameter space was narrowed too much using literature

and experience as the starting point. In this respect, the coincidence of finding only

cyclopentyl and cyclohexyl as good candidates while they are also the only ones

reported in the literature for the synthesis of silsesquioxane a7b3,17,18 might be due to

bias for the synthesis condition particularly suited for these silanes. This remark

together with the trend indicating a beneficial effect of polarity suggest that it could be

interesting in further research to broaden the screened parameter space to other polar

solvents (see Chapter 7).

The two 1:1 molar mixtures, of n-octyl- and methyltrichlorosilane and of

cyclohexyl- and phenyltrichlorosilane, were used to investigate possible synergetic

effects between two different organotrichlorosilanes in the formation of incompletely

condensed structures (Figure 3.2). The n-octyl/methyl mixture gave a very low

Chapter 3 __________________________________________________________________________________________________________

52

epoxidation activity. As in the case of the pure n-octyl compound, the hydrolytic

condensation produced a gel, generated by micelle formation (probably due to the

slower hydrolysis rate of long-chain alkylsilanes).23 Water retained in this gel would be

inimical to the Ti-complexation. The cyclohexyl/phenyl mixture showed considerable

activity, suggesting the possibility of synergy between trichlorosilanes with different

organic R-groups (see Chapter 4).

The activities of the catalysts synthesised using High-Speed Experimentation

techniques were compared with those of Ti(OPri)4 and of the

Ti-(cyclopentyl silsesquioxane) catalyst produced by the reaction of pure cyclopentyl

silsesquioxane a7b3 with Ti(OPri)4. The High-Speed Experimentation catalysts with

R = cyclohexyl, cyclopentyl, phenyl, tert-butyl, ethyl, methyl and the 1:1 mixture of

cyclohexyl and phenyl, show an activity higher than Ti(OPri)4, confirming the formation

of Ti-silsesquioxanes as the active species. The Ti-catalyst derived from the

silsesquioxanes synthesised by the hydrolytic condensation of

cyclopentyltrichlorosilane in acetonitrile presents the highest catalytic activity

(Figure 3.2). This activity is 87% of that of the Ti-catalyst obtained using pure

silsesquioxane a7b3 as precursor. The relevance of this result lies in the fact that the

synthesis of these silsesquioxane precursors does not require any purification process

and is much less time-consuming than the synthesis of silsesquioxane a7b3.

3.3.1 Characterisation of the HSE lead

It is important to notice that with this HSE approach it was possible to identify a

lead but that no information was obtained regarding the actual nature of the

silsesquioxane precursor produced by the hydrolytic condensation of

cyclopentyltrichlorosilane in acetonitrile. It was thus interesting to characterise this lead

on a conventional laboratory scale. First, the synthesis and testing of the

Ti-silsesquioxane was repeated on a 125-ml scale (50 times up-scaling) yielding a

catalyst with an equal activity and, therefore, confirming the applicability of HSE

techniques to the synthesis of silsesquioxanes. Then, the silsesquioxane precursor was

characterised prior to reaction with the titanium centre. The reaction of cyclopentyl

trichlorosilane with water in acetonitrile, for 18 hours at 50°C, yields two


53

silsesquioxane fractions: one as a precipitate (A) and the other as solute in the reaction

mixture (B). Fraction B was dried under reduced pressure and redissolved in

tetrahydrofuran. The drying of fraction B induces an increase in the level of

condensation of the silsesquioxane species, as proved by the fact that the gel obtained is

not anymore soluble in acetonitrile. Both fractions were characterised by NMR

spectroscopy and mass spectrometry.

Figure 3.3. 29Si NMR analysis of fraction A, previous to (above) and after (below)

titanium coordination (L = alkoxy group).

-55 -60 -65 ppm

-55 -60 -65 ppm

Ti(OL)4 THF

* #

××××

Si

SiSi

Si

Si

Si

Si

OHO

OOH

O

OO

OH

OO

O

R

R

R

R

R R

RO

#

#

#

*

*

*

××××

*

#

××××

-55 -60 -65 ppm

-55 -60 -65 ppm

Ti(OL)4 THF

-55 -60 -65 ppm

-55 -60 -65 ppm

Ti(OL)4 THFTi(OL)4 THF

* #

××××

Si

SiSi

Si

Si

Si

Si

OHO

OOH

O

OO

OH

OO

O

R

R

R

R

R R

RO

#

#

#

*

*

*

××××

Si

SiSi

Si

Si

Si

Si

OHO

OOH

O

OO

OH

OO

O

R

R

R

R

R R

RO

#

#

#

*

*

*

××××

*

#

××××

Chapter 3 __________________________________________________________________________________________________________

54

Figure 3.4. 29Si NMR analysis of fraction B, previous to (above) and after (below)

titanium coordination (L = alkoxy group).

Fraction A mainly consists of silsesquioxane a7b3,18 as determined by means of 29Si NMR (see Figure 3.3) and MS analysis. Fraction B is a mixture of different

silsesquioxane species, as indicated by 29Si NMR and MS analysis. The region in which

-55 -60 -65 ppm

-55 -60 -65 ppm

Ti(OL)4 THF

-55 -60 -65 ppm

-55 -60 -65 ppm

Ti(OL)4 THFTi(OL)4 THF


55

most of the peaks in the 29Si NMR spectrum of fraction B are found is that of silanol

groups belonging to incompletely condensed silsesquioxanes (see Figure 3.4).24 It is

equally difficult to assign all the species present in the mixture or to separate them.

However, both NMR and MS data suggest that the main species is the silsesquioxane

structure a6b2 shown in Figure 3.5; this structure is in agreement with the three 29Si NMR peaks marked in Figure 3.4 and with the main set of MS peaks,

corresponding to silsesquioxanes with 6 Si atoms. This assignment is supported by the

similarity of the 29Si NMR spectrum with that of cyclohexyl silsesquioxane a6b2, which

can be obtained by cleavage of the completely condensed a6b0 structure.25

Figure 3.5. Schematic representation of cyclopentyl silsesquioxane a6b2

[R = cyclopentyl].

Finally, both fractions were reacted with a titanium alkoxide. The coordination of the

titanium centre to the silsesquioxane structures was confirmed by 29Si NMR

(Figures 3.3 and 3.4). The catalytic activity in the epoxidation of 1-octene of the HSE

lead and of the two fractions in which the latter can be divided (all the three catalysts

are homogeneous) were studied as a function of the reaction time (Figure 3.6). The

activity of fraction A (TOF = 0.97 molepo·molTi-1·min-1) is comparable to that of the

Ti-catalyst obtained with pure silsesquioxane a7b3, which after 4 hours of reaction

gives complete conversion of TBHP towards the epoxide. Fraction B shows a much

lower, but still significant, activity (TOF = 0.16 molepo·molTi-1·min-1): this result confirms

Si

Si

Si

O

O

O

O

R

Si

O

O

Si

Si OH

HO

O

O

R R

R R

R

Si

Si

Si

O

O

O

O

R

Si

O

O

Si

Si OH

HO

O

O

R R

R R

R

Chapter 3 __________________________________________________________________________________________________________

56

the validity of the initial assumption that incompletely condensed silsesquioxanes other

than a7b3 can be precursors for active titanium catalysts. For fraction A, the conversion

reaches a maximum after 30 minutes, while for fraction B the conversion slowly

increases along the 4 hours and does not reach a plateau. The behaviour of the HSE lead

(TOF = 0.43 molepo·molTi-1·min-1) is an average of that of the two fractions: a plateau is

reached after 2 hours of reaction.

Figure 3.6. Activity in the epoxidation of 1-octene with TBHP as a function of the

reaction time.

3.4 Conclusions

In this chapter, it has been shown that High-Speed Experimentation can be a

valuable technique in the study and optimisation of catalysts. A very active

silsesquioxane-based Ti-catalyst for the epoxidation of 1-octene

(cyclopentyltrichlorosilane hydrolysed in acetonitrile) has been prepared by a faster and

cheaper method than conventionally possible. In addition to identifying the optimal

synthesis conditions within the screened parameter space, the substantial number of

experiments performed using High-Speed Experimentation techniques enabled to gain

additional insight in the system under study. Some general behaviours, with respect to

0%

20%

40%

60%

80%

100%

0 30 60 90 120 150 180 210 240

time (minutes)

conv

ersi

on

fraction A

fraction B

HSE lead0%

20%

40%

60%

80%

100%

0 30 60 90 120 150 180 210 240

time (minutes)

conv

ersi

on

fraction A

fraction B

HSE lead

fraction A

fraction B

HSE lead


57

the roles in the synthesis of silsesquioxanes of both the solvent and the R-group of the

organotrichlorosilane have been identified.

The synthesis of the HSE lead was repeated on a conventional lab-scale and the

silsesquioxane precursor was characterised by means of MS and NMR spectroscopy.

3.5 Experimental

Experiments were performed on an automated parallel synthesis workstation26

coupled with a personal computer supplied with software enabling to program the

workstation. Reagents: cyclohexyl-, cyclopentyl-, phenyl-, methyl-, ethyl-, tert-butyl-,

n-octyl-, and allyltrichlorosilanes (RSiCl3); acetone, acetonitrile, methanol and

tetrahydrofuran (THF) as solvents; deionised water. In a typical experiment, 6 times

2-ml aliquots of each of the 4 solvents were dispensed in a rack containing an 6×4 array

of glass tubes, followed by the addition of 340 µmol of each of 6 silanes to the

solvent-containing reaction vessels in such a way that 24 individually different silane

solutions were prepared. Hydrolysis of the silane was started by the addition of 0.5 ml

of water to the reaction vessel and placing the vessel array on an orbital shaker26 at

50°C for 18 hours. After removal of the solvent, and of the excess water and

hydrochloric acid, in a vacuum centrifuge,26 the samples were stored under argon.

Titanium complexation was performed by dissolving the crude silsesquioxane mixture

in 2 ml of THF under argon, followed by the addition of 54 µmol of titanium

isopropoxide (as a solution in isopropanol) to each sample. After 5 hours at 60°C, the

tetrahydrofuran was removed by means of the vacuum centrifuge and the samples were

stored under argon.

The catalytic activity of the materials obtained was determined by the

epoxidation of 1-octene with tert-butyl hydroperoxide (TBHP). The reaction was

performed by adding to each dried sample ~0.5 mmol of TBHP as a ~40%wt solution in

cyclohexane (TBHP : catalyst ratio ≈ 10) and 0.0144 mol of 1-octene, that therefore

acted both as reactant and solvent (1-octene : catalyst ratio = 300). The employed

1-octene contained 2%vol of decane as an internal standard for the following GC

analysis. Samples were taken after reacting for 4 hours at 80°C and analysed on a

UNICAM Pro GC using a CP-Sil-5B column. The epoxidation reaction did not proceed

Chapter 3 __________________________________________________________________________________________________________

58

further at room temperature. The reported activities were obtained by normalising the

1,2-epoxyoctane GC peak area by means of the internal standard. The values for the

activities are the averages of the results from different experiments. The experimental

average error for these values is under 3%, as determined by calibration and averaging

of the results of three identical experimental runs. Since 1-octene was used both as

reactant and solvent, it has not been possible to realise a mass balance. The reaction was

completely selective towards 1,2-epoxyoctane: no other products were detected. The

conversion selectivity of TBHP towards 1,2-epoxyoctane lies between 80 and 94%.

The other catalysts (Ti(OPri)4 and Ti-(cyclopentyl silsesquioxane a7b3) were

tested using the same amount of titanium and the same experimental conditions as

described above.

The 125-ml scale synthesis of the HSE lead was performed by carefully adding

25 ml of deionised water into a round-bottom flask containing a solution of 2.8 ml of

cyclopentyltrichlorosilane (c-C5H9)SiCl3 (97% purity) in 100 ml of acetonitrile

(CH3CN:H2O = 4:1 in volume). The solution was then heated to 50°C while stirring for

18 hours: a white precipitate was formed (fraction A: 1.025 g, 50% yield). The

remaining solution (fraction B) was dried under reduced pressure to afford a dense gel,

which was then dissolved in 10 ml of tetrahydrofuran: an insoluble NH4Cl residue

(produced by the hydrolysis of CH3CN) was removed. Consecutive drying of the

solution afforded a silsesquioxane gel that was characterised by NMR and MS.

Titanium complexation was performed by dissolving each fraction in 50 ml of

tetrahydrofuran, followed by the addition of 0.4 ml of titanium butoxide, Ti(OBu)4, to

each solution. Both samples were then heated to 60°C for 5 hours and finally dried

under reduced pressure. To test the catalytic activity in epoxidation reactions, the two

solid samples were dissolved each in 56.5 ml of 1-octene (with 2%vol of decane as

internal standard for GC analysis), followed by addition of 3 ml of TBHP solution

(~40%wt solution in cyclohexane), stirred and heated to 80°C. Samples were taken at

different times (see Figure 3.6) and analysed with a UNICAM Pro GC using a

CP-Sil-5B column. The activity data were obtained by normalising the 1,2-epoxyoctane

GC peak area by means of the internal standard. For all the three catalysts the reaction

was completely selective towards 1,2-epoxyoctane: no other products were found.


59

29Si NMR spectra were measured on a Varian VXR-400S (79.5 MHz, 1H decoupled,

deuterated THF as solvent, 25°C). Main peaks for fraction A, δ: -57.91, -65.70, -67.29

(3:1:3); fraction B, δ: -55.13, -60.16, -60.33 (1:1:1).

Mass spectrometry analysis was performed using a Micromass Quattro LC-MS with

ESI+ as ionisation technique. The samples were prepared by dissolving ~0.05 g of crude

product in 4 ml of tetrahydrofuran, 1 ml of acetonitrile and 0.1 ml of formic acid 0.1M.

From the MS data it is possible to determine the number of Si atoms (the value of a in

(RSiO1.5)a(H2O)0.5b) of the detected species but not the level of condensation (the value

of b). The reason for this is that in the ionisation process silsesquioxanes can lose water

molecules, as confirmed by the fact that the ratio among the peaks for a given value of a

changes as a function of the cone voltage applied in the MS analysis.

Selected MS data:

Fraction A (cone voltage = 65V). For a = 7, m/z: 875.04 (H+, 100%), 839.00 (H+, 41%),

897.08 (Na+, 16%), 911.09 (H+, 16%), 857.03 (H+, 11%). For a = 8, m/z: 969.03 (H+,

4%), 987.05 (H+, 2%).

Fraction B (cone voltage = 30V). For a = 4, m/z: 543.60 (Na+, 14%), 557.56 (H+,

+ H2O, 13%). For a = 5, m/z: 655.53 (Na+, 30%), 633.61 (H+, 28%). For a = 6, m/z:

763.41 (H+, 100%), 785.39 (Na+, 55%), 745.41 (H+, 38%). For a = 9, m/z: 1157.23

(Na+, 22%), 1175.24 (Na+, 17%), 1135.12 (H+, 18%). For a = 10, m/z: 1287.03 (Na+,

29%), 1305.09 (Na+, 27%), 1265.11 (H+, 25%). For a = 11, m/z: 1417.97 (Na+, 36%),

1396.11 (H+, 26%). For a = 12, m/z: 1547.83 (Na+, 39%), 1529.95 (Na+, 26%).

Chapter 3 __________________________________________________________________________________________________________

60

References

1 B. Notari, Ad. Catal., 1996, 41, 253. 2 G. Bellussi, M.S. Rigutto, Stud. Surf. Sci. Catal., 1994, 85, 177. 3 T. Maschmeyer, F. Rey, G. Sankar, J.M. Thomas, Nature, 1995, 378, 159. 4 M.C. Klunduk, T. Maschmeyer, J.M. Thomas, B.F.G. Johnson, Chem. Eur. J., 1999, 5 (5), 1481. 5 T. Maschmeyer, M.C. Klunduk, C.M. Martin, D.S. Shephard, J.M. Thomas, B.F.G. Johnson, Chem.

Commun., 1997, 1847. 6 M. Crocker, R.H.M. Herold, A.G. Orpen, Chem. Commun., 1997, 2411. 7 H.C.L. Abbenhuis, S. Krijnen, R.A. van Santen, Chem. Commun., 1997, 331. 8 S. Krijnen, H.C.L. Abbenhuis, R.W.J.M. Hanssen, J.H.C. van Hooff, R.A. van Santen, Angew. Chem.

Int. Ed., 1998, 37, 356. 9 L. Marchese, T. Maschmeyer, E. Gianotti, S. Coluccia, J.M. Thomas, J. Phys. Chem. B, 1997, 101,

8836. 10 P.E. Sinclair, G. Sankar, C.R.A. Catlow, J.M. Thomas, T. Maschmeyer, J. Phys. Chem. B, 1997, 101,

4232. 11 D. Tantanak, M.A. Vincent, I.H. Hiller, Chem. Commun., 1998, 1031. 12 P.E. Sinclair, C.R.A. Catlow, J. Phys. Chem. B, 1999, 103, 1084. 13 C.M. Barker, D. Gleeson, N. Kaltsoyannis, C.R.A. Catlow, G. Sankar, J.M. Thomas, Phys. Chem.

Chem. Phys., 2002, 4, 1228. 14 P.P. Pescarmona, T. Maschmeyer, Aust. J. Chem., 2001, 54, 583. 15 T. Maschmeyer, J.M. Thomas, A.F. Masters, NATO ASI Ser., New Trends in Materials Chemistry,

1997, 498, 461. 16 H.C.L. Abbenhuis, Chem. Eur. J., 2000, 6, 25. 17 F.J. Feher, D.A. Newman, J.F. Walzer, J. Am. Chem. Soc., 1989, 111, 1741. 18 F.J. Feher, T.A. Budzichowski, R.L. Blanski, K.J. Weller, J.W. Ziller, Organometallics, 1991, 10,

2526. 19 K.A. Jørgensen, Chem. Rev., 1989, 89, 431. 20 P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, Catal. Lett., 1999, 63, 1, see also

Chapter 1. 21 T. Kudo, M.S. Gordon, J. Am. Chem. Soc., 1998, 120, 11432. 22 T. Kudo, M.S. Gordon, J. Phys. Chem. A, 2000, 104, 4058. 23 P.G. Harrison, J. Organomet. Chem., 1997, 542, 141. 24 See Paragraph 2.3.1. 25 F.J. Feher, F. Nguyen, D. Soulivong, J.W. Ziller, Chem. Commun., 1999, 1705. 26 See Appendix A.

61

4

Fine-tuning of the synthesis of titanium-silsesquioxane epoxidation catalysts by means of High-Speed Experimentation

Abstract

High-Speed Experimentation techniques have been applied to a fine-tuning of the

preparation of titanium silsesquioxane catalysts for the epoxidation of alkenes. The first

part of this optimisation led to an adjustment of the reaction conditions for the synthesis

of the silsesquioxane precursors. Next, the effect of the nature of the titanium centres

bound to the silsesquioxane structures on the catalytic activity was explored: the

coordination of diols to the titanium centres resulted in an improvement of the activity

when using H2O2 as the oxidant.


P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, Angew. Chem. Int. Ed., 2001, 40, 740.

P.P. Pescarmona, T. Maschmeyer, NATO Science Series, 2002, Ser. II Vol. 69, 173.


P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, Chem. Eng. Commun., accepted.

Chapter 4 __________________________________________________________________________________________________________

62

4.1 Introduction

Besides leading to the identification of a new route to synthesise a promising

Ti-catalyst and to an additional knowledge about the synthesis of silsesquioxane

structures, the HSE exploration described in Chapter 3 provided a starting point for

further investigation. One of the directions that can be taken is that of studying other

parameters influencing the catalytic activity of titanium silsesquioxanes. The initial

choice of investigating the effect of the nature of the solvent and of the R-group on the

synthesis of silsesquioxane precursors for Ti-catalysts was based on the assumption that

these two parameters were the most relevant in determining which silsesquioxane

structures were formed and in which amounts. The validity of this hypothesis can be

verified (or falsified)1,2 by studying the effect of the other parameters affecting the

synthesis of silsesquioxanes.3 This study would also provide an adjustment of the

method to prepare silsesquioxane precursors for epoxidation catalysts. After having

completed the optimisation of the synthesis of the silsesquioxane precursors, the next

step of the preparation of the Ti-catalysts can be investigated by varying the nature of

the titanium centre bound to the silsesquioxane structures.


Different HSE approaches can be used to investigate the system. In the case of

the effect of pH variation, the procedure was similar to that reported in Chapter 3. For

all other parameters a different approach was used. The parameter set was reduced from

12 (8 R-groups + 4 solvents) to 4 (3 R-groups + 1 solvent) elements: cyclopentyl-,

cyclohexyl- and phenyltrichlorosilane with acetonitrile as solvent, i.e. the combinations

that led to the highest catalytic activities (see Paragraph 3.3). This reduced parameter

space was then used as a basis to create larger parameter spaces by varying the other

parameters influencing the synthesis. Finally, for the exploration of the nature of the

titanium centres, two silsesquioxane precursors were used to bind to a set of different

titanium compounds. These Ti-silsesquioxanes were then tested for the epoxidation of

1-octene with both TBHP and H2O2.

Fine-tuning of the synthesis of Ti-silsesquioxanes by means of HSE __________________________________________________________________________________________________________

63


4.3.1 Synergetic effect of mixtures of silanes

In Chapter 3 it was shown that Ti-catalysts derived from silsesquioxanes

precursors produced by the hydrolytic condensation of a 1:1 mixture of cyclohexyl- and

phenyltrichlorosilanes displayed higher epoxidation activities than those obtained from

either pure cyclohexyl- or phenyltrichlorosilane. This suggested the existence of a

synergetic effect between trichlorosilanes with different R-groups. In order to further

investigate this possible synergetic effect, a systematic variation of the ternary

composition of cyclopentyl-, cylcohexyl- and phenyltrichlorosilanes mixtures using

acetonitrile as the solvent was explored. The mixtures were screened for the activity as

epoxidation catalysts after the coordination of Ti(OPri)4 (Figure 4.1). The highest

epoxidation activity was found for the pure cyclopentyltrichlorosilane, i.e. the

experiment did not lead to the identification of a more active catalyst. However, the

surface of the activity diagram shows some interesting features, underpinning the

usefulness of the experiment: various mixtures yield catalysts with higher activities than

those from pure cyclohexyl- and phenyltrichlorosilanes. Such behaviour would have

been difficult to identify without a combinatorial approach.

Figure 4.1. Activity in the epoxidation of 1-octene with TBHP of Ti-silsesquioxane

catalysts as a function of the ratio of cyclopentyl-, cyclohexyl- and

phenyltrichlorosilanes in the initial synthesis mixture.

relativeactivity

Phenyl (molar ratio)

Cycloh

exyl

(mola

r rati

o)

Cyclopentyl (molar ratio)

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

00.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.91

0.8

0.9

1

0.5

0.6

0.7

0.2

0.3

0.4

00.1

relativeactivity

Phenyl (molar ratio)

Cycloh

exyl

(mola

r rati

o)

Cyclopentyl (molar ratio)

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

00.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.91

0.8

0.9

1

0.5

0.6

0.7

0.2

0.3

0.4

00.1

Chapter 4 __________________________________________________________________________________________________________

64

4.3.2 Effect of the reaction conditions

With the aim of further optimising the synthesis conditions for silsesquioxane

precursors, the other parameters influencing the hydrolytic condensation were

investigated.3 The activity of the Ti-silsesquioxane catalysts was screened as a function

of the variation of the reaction conditions employed during the hydrolytic condensation.

The reported activities are normalised to the activity of (c-C5H9)7Si7O12TiOCH(CH3)2.4

• Effect of pH variation (Figure 4.2). During the hydrolysis of the trichlorosilanes,

hydrochloric acid is produced, which could influence the level of condensation and

the amounts of silsesquioxanes generated.3 In order to check this potential pH effect,

the synthesis of silsesquioxane precursors was performed using a 0.3M HCl water

solution to hydrolyse the silanes in a similar experiment to that reported in Chapter 3.

Figure 4.2. Catalytic activity in the epoxidation of 1-octene with TBHP of the

Ti-silsesquioxanes obtained by combining different trichlorosilanes and

solvents and using a 0.3M HCl solution as hydrolysing agent.

THF

methanol

acetonitrile

acetone0.0000.1000.2000.3000.4000.5000.6000.700

0.8000.900

1.000


ne


ne

phenyltrichlorosila

ne

methyltrichlorosila

ne


ne

allyltrichlorosila

ne


nes (1:1)


nes (1:1)

relativeactivity

THF

methanol

acetonitrile

acetone0.0000.1000.2000.3000.4000.5000.6000.700

0.8000.900

1.000


ne


ne

phenyltrichlorosila

ne

methyltrichlorosila

ne


ne

allyltrichlorosila

ne


nes (1:1)


nes (1:1)

relativeactivity


65

The trends observed using H2O or the 0.3M HCl water solution are very similar,

though a higher activity is found when using an initially neutral water solution (70%

greater on average, compare Figure 3.2 and Figure 4.2); thus extra hydrochloric acid

is not beneficial.

• Effect of trichlorosilane concentration (Figure 4.3). The activity of the

Ti-silsesquioxane catalysts was screened as a function of the variation of the

trichlorosilane concentration in the hydrolytic condensation over the reduced

parameter space constituted of cyclopentyl-, cyclohexyl- and phenyltrichlorosilane

with acetonitrile as the solvent. In order to deal with a constant amount of

trichlorosilane in all the experiments, the volumes of solvent and water were varied

instead of the amount of trichlorosilane used.


catalysts as a function of the trichlorosilane concentration in the initial

silsesquioxane synthesis mixture.

For most of the trichlorosilane concentration tested, the order of activity is still

cyclopentyl > cyclohexyl > phenyl. At low concentration the order between

cyclohexyl and phenyl reverses but cyclopentyl still results in the best activity. This

1.088

M

0.544

M

0.272

M

0.181

M

0.136

M

0.109

M

0.091

M

0.078

M

phenyltrichlorosilane

cyclohexyltrichlorosilanecyclopentyltrichlorosilane0.000

0.200

0.400

0.600

0.800

1.000

relativeactivity

trichlorosilane concentration

1.088

M

0.544

M

0.272

M

0.181

M

0.136

M

0.109

M

0.091

M

0.078

M


cyclohexyltrichlorosilanecyclopentyltrichlorosilane0.000

0.200

0.400

0.600

0.800

1.000

relativeactivity

trichlorosilane concentration

Chapter 4 __________________________________________________________________________________________________________

66

result confirms the starting hypothesis that the R-group is a more relevant factor than

the concentration of the silane in influencing the hydrolytic condensation. The

maximum of activity corresponds to a different silane concentration for each of the

three R-groups. The maximum for cyclopentyl as R-group - corresponding to a

concentration of silane of 0.181M - displays an activity that is 95% of that of pure

Ti-(cyclopentyl silsesquioxane a7b3): this is a relevant improvement compared to the

optimum obtained previously (87%).5 These results show that the experiment

worked as a fine-tuning of the synthesis of silsesquioxane precursors described in

Chapter 3. From a methodological point of view it is important to notice that in the

wide concentration range applied, no activity low enough to reject any of the three

R-groups - if such trichlorosilane concentration would have been chosen in the first

experiment - was obtained. This significantly strengthens the assumption that the

conditions of the first experiment were sufficiently discriminative towards the

R-groups screened.

• Effect of the H2O concentration (Figure 4.4). The activity of the Ti-silsesquioxane

catalysts was screened as a function of the variation of the amount of H2O used in the

hydrolytic condensation over the reduced parameter space constituted of

cyclopentyl-, cyclohexyl- and phenyltrichlorosilane with acetonitrile as solvent. To

separate the effect of the water concentration from the effect of the trichlorosilane

concentration, the total amount of liquid (solvent + water) in which the synthesis was

carried out was kept constant (2.5 ml). The acetonitrile/water ratio was varied from

32 to 0.25. For any acetonitrile/water ratio, the order of activity is still

cyclopentyl > cyclohexyl > phenyl, apart from an inversion between cyclohexyl and

phenyl at a ratio ≤ 0.5. As for the trichlorosilane concentration experiment, this result

confirms the hypothesis that the nature of the R-group has a directing role in the

formation of silsesquioxanes. Similarly, the position of the maximum of the activity

is different for each R-group: the maximum activity for phenyl is at high water

content while for cyclopentyl and cyclohexyl it is at lower water content. The highest

activity is found for cyclopentyl at an acetonitrile/water ratio of 2: this activity

corresponds to 95% of that of Ti-(cyclopentyl silsesquioxane a7b3), so a relevant

improvement of activity is obtained by tuning this parameter.


67


catalysts as a function of the acetonitrile/water ratio in the initial

silsesquioxane synthesis mixture.

• Effect of the reaction time (Figure 4.5). The hydrolytic condensation of

silsesquioxane precursors in acetonitrile was performed for 4 hours, 18 hours and 2

weeks, the latter time being comparable to the reaction times reported in literature.6,7

After a 4 hours synthesis, the catalysts obtained after titanium complexation show a

comparable activity for all the three R-groups. Increasing the reaction time to

18 hours and to 2 weeks causes two different trends: the activity of the catalysts

obtained from cyclopentyl- and cyclohexyltrichlorosilanes increases with time, while

that of the catalysts generated from phenyltrichlorosilane decreases with time. This

again suggests that the R-group is strongly influencing which silsesquioxanes species

are formed. The activity of the Ti-catalyst derived from the 2-weeks hydrolytic

condensation of cyclopentyltrichlorosilane in acetonitrile displays the same activity

(100%) as the pure Ti-(cyclopentyl silsesquioxane a7b3). This could indicate that

after 2 weeks of reaction with the above conditions, silsesquioxane a7b3 is the only

product. This hypothesis was confirmed by repeating the hydrolytic condensation of

cyclopentyltrichlorosilane in acetonitrile on a 125-ml scale for 2 weeks at 50°C and

by analysing the silsesquioxane precipitate by means of 29Si NMR (see Chapter 5 for

more information).

32 16 8 4 2 1 0.5 0.25

phenyltrichlorosilanecyclohexyltrichlorosilane

cyclopentyltrichlorosilane0.000

0.200

0.400

0.600

0.800

1.000

relativeactivity

acetonitrile / water ratio (vol/vol)

32 16 8 4 2 1 0.5 0.25

phenyltrichlorosilanecyclohexyltrichlorosilane


0.200

0.400

0.600

0.800

1.000

relativeactivity

acetonitrile / water ratio (vol/vol)

Chapter 4 __________________________________________________________________________________________________________

68


catalysts as a function of the silsesquioxane synthesis time.

• Effect of the reaction temperature (Figure 4.6). The syntheses of cyclopentyl,

cyclohexyl and phenyl silsesquioxanes in acetonitrile were performed at 4

temperatures: 25, 50, 75 and 100°C. The silsesquioxanes generated were tested for

epoxidation activity after titanium complexation.


catalysts as a function of the silsesquioxane synthesis temperature.

4 hours18 hours

2 weeks


cyclohexyltrichlorosilane

cyclopentyltrichlorosilane

0.000

0.200

0.400

0.600

0.800

1.000

relativeactivity

silsesquioxane synthesis time

4 hours18 hours

2 weeks




0.000

0.200

0.400

0.600

0.800

1.000

relativeactivity

silsesquioxane synthesis time

25°C50°C

75°C100°C




0.200

0.400

0.600

0.800

1.000

relativeactivity

silsesquioxane synthesis temperature

25°C50°C

75°C100°C




0.200

0.400

0.600

0.800

1.000

relativeactivity

silsesquioxane synthesis temperature


69

Some technical problems were discovered in retrospect, i.e. the leaking of the

reaction vessels at high temperatures. This slightly affected the reliability of the

results for T ≥ 75°C and the data have to be interpreted with this in mind. However,

there seems to be a trend showing that temperature does not have a major effect on

the performance of the silsesquioxanes as precursors for titanium catalysts.

Some general conclusions can be drawn from these experiments. The activity of

the Ti-catalysts based on phenyl silsesquioxanes decreases when increasing the

concentration of the trichlorosilane and with longer reaction times for the hydrolytic

condensation. For cyclohexyl and cyclopentyl the trend concerning the reaction time

seems to be opposite. Further, the dependence on the trichlorosilane and water

concentrations and on the reaction time suggests that the formation of silsesquioxanes is

kinetically controlled. It can be hypothesised that the nature of the R-group influences

the energy of intermediates in the synthesis of silsesquioxanes, determining which

silsesquioxane structure is favoured. If the hypothesis is correct, this means that the

favoured phenyl silsesquioxanes are completely condensed and/or consist of other

structures that are not suitable for complexating titanium. On the other hand, the favoured

cyclohexyl and cyclopentyl silsesquioxane would be silsesquioxane a7b3 and/or another

incompletely condensed structure (see Chapter 5) suitable for complexating titanium in

a way that leads to a very active catalyst. The hypothesis is in agreement with the

characterisation of the HSE lead described in Paragraph 3.3.1 and with the literature,

where the formation of completely condensed phenyl silsesquioxanes8 and of

incompletely condensed cyclohexyl and cyclopentyl silsesquioxane a7b36,7,9 are reported.

4.3.3 Effect of the nature of the titanium centre

After the optimisation of the synthesis of the silsesquioxane precursors, it is

interesting to further study the titanium-silsesquioxane epoxidation catalysts by varying

the nature of the active titanium centre. Different ligands on the titanium could

influence the activity of the catalyst both through steric and electronic effects.10,11 It

would be particularly useful to identify a titanium silsesquioxane that could be used

with aqueous hydrogen peroxide (H2O2) as the oxidant. The advantage of using

hydrogen peroxide rather than TBHP or other organic hydroperoxides, is that hydrogen

Chapter 4 __________________________________________________________________________________________________________

70

peroxide would yield H2O as the by-product with evident environmental and

economical benefits for the process. So far, the use of H2O2 as the oxidant with

homogeneous titanium catalysts has been hindered either by the strong adsorption of

H2O on the titanium centre or by the rapid deactivation of the catalyst due to the

leaching/hydrolysis of the titanium centre and subsequent formation of inactive

oligomeric TiOx species. Few examples of epoxidation of alkenes with H2O2 using

heterogeneous titanium catalysts with a hydrophobic environment have been

reported.12,13

Here, HSE techniques were used for a further secondary screening of the titanium

silsesquioxanes. As silsesquioxane precursors, both the pure cyclopentyl

silsesquioxane a7b3 and the lead mixture obtained from the previous HSE

screenings5,14,15 (cyclopentyltrichlorosilane hydrolysed for 18 hours in acetonitrile) were

used. As oxidants, TBHP in cyclohexane and H2O2 (aq) were used. As titanium

precursors, 5 different titanium alkoxides Ti(OL)4 (L = Me, Et, Pr, Pri, Bu) and

cyclopentadienyltitanium trichloride were chosen. Besides, a series of diols (ethylene

glycol, 2,3-butanediol, 1,2-cyclohexanediol (mixture of cis and trans), pinacol and

benzopinacol) were used as ligands for the Ti-centre, with Ti(OBu)4 as titanium

precursor. The purpose of having a diol coordinating to the Ti-centre (Figure 4.7) is to

reduce its accessibility for polar molecules by increasing the hydrophobicity around the

catalytic centre:13 this may prevent the reaction with water and consequent catalyst

deactivation when using H2O2 as oxidant.

Figure 4.7. Possible interactions between a diol and the Ti-centre on a silsesquioxane.

The resulting 44-element parameter space was screened for activity in the epoxidation

of 1-octene (Figure 4.8). The values are normalised to the activity of the complex

OL

Ti

O

Ti

LO

Ti

OHHO

RIV

RIIIRII

RI

OH OH

HO

RIV

RIII

RIIRI

RIV

RIII

RIIRI

or/and

OL

Ti

O

Ti

LO

Ti

OHHO

RIV

RIIIRII

RI

OH OH

HO

RIV

RIII

RIIRI

RIV

RIII

RIIRI

or/and


71

obtained by reacting Ti(OPri)4 with the pure cyclopentyl silsequioxane a7b3. The

experimental results of the 6 titanium precursors and of the 5 diols are discussed

separately.

Figure 4.8. Screening of the epoxidation activity of Ti-silsesquioxanes with two

different oxidants (TBHP and H2O2) as a function of the nature of the

Ti-centres.

• When varying the 6 titanium precursors, the first general observation is that all the

catalysts are very active with TBHP as the oxidant while they show low activities

when H2O2 is the oxidising agent. This suggests that these catalysts still undergo

deactivation due to the reaction with water, as confirmed by the formation of a white

precipitate (probably TiO2) in the reaction vessels. When TBHP is used as the

benzopinacolpinacol1,2-cyclohexanediol2,3-butanediolethyleneglycol

OHHOOHHO

H3C CH3OH

OH OH

CH3

OH

H3CH3C

H3C OH

Ph

OH

PhPh

Ph

CpdTiC

l 3

Ti(OPri ) 4

Ti(OMe) 4

Ti(OEt) 4

Ti(OPr) 4

Ti(OBu) 4

Ti(OBu) 4

+ ethy

lene g

lycol

Ti(OBu) 4

+ 2,3-b

utane

diol

Ti(OBu) 4

+ 1,2-c

ycloh

exan

ediol

Ti(OBu) 4

+ pinaco

l

Ti(OBu) 4

+ benz

opina

col

HSE silsesquioxane mixture / H2O2

Silsesquioxane a7b3 / H2O2

HSE silsesquioxane mixture / TBHPSilsesquioxane a7b3 / TBHP0.000

0.200

0.400

0.600

0.800

1.000

1.200

relativeactivity


OHHOOHHO

H3C CH3OH

OH OH

CH3

OH

H3CH3C

H3C OH

Ph

OH

PhPh

Ph


OHHOOHHO

H3C CH3OH

OH OH

CH3

OH

H3CH3C

H3C OH

Ph

OH

PhPh

Ph

CpdTiC

l 3

Ti(OPri ) 4

Ti(OMe) 4

Ti(OEt) 4

Ti(OPr) 4

Ti(OBu) 4

Ti(OBu) 4

+ ethy

lene g

lycol

Ti(OBu) 4

+ 2,3-b

utane

diol

Ti(OBu) 4

+ 1,2-c

ycloh

exan

ediol

Ti(OBu) 4

+ pinaco

l

Ti(OBu) 4

+ benz

opina

col




0.200

0.400

0.600

0.800

1.000

1.200

relativeactivity

CpdTiC

l 3

Ti(OPri ) 4

Ti(OMe) 4

Ti(OEt) 4

Ti(OPr) 4

Ti(OBu) 4

Ti(OBu) 4

+ ethy

lene g

lycol

Ti(OBu) 4

+ 2,3-b

utane

diol

Ti(OBu) 4

+ 1,2-c

ycloh

exan

ediol

Ti(OBu) 4

+ pinaco

l

Ti(OBu) 4

+ benz

opina

col




0.200

0.400

0.600

0.800

1.000

1.200

relativeactivity

Chapter 4 __________________________________________________________________________________________________________

72

oxidant, the pure silsesquioxane a7b3 turns out being a slightly better precursor than

the HSE mixture for any Ti-centre. On average, the ratio between the activities of the

HSE mixture and the pure a7b3 catalysts is 0.87. Considering the various Ti-centres,

hardly any variation in activity was observed among the titanium alkoxides with

L = Me, Et, Pr, Bu: most of the differences fall within the same range, given that the

experimental error of these HSE data is ± 3%. The activity with Ti(OPri)4 and

cyclopentadienyltitanium trichloride are slightly, but noticeably, lower. This trend

can be explained by considering that the isopropoxy and cyclopentadienyl groups are

bulkier than the linear alkoxides. Therefore, the accessibility to the titanium site is

reduced, causing a decrease in catalytic activity.16 On the basis of literature data,16 a

trend in activity among the linear alkoxides was expected (the shorter the alkyl chain,

the higher the activity). The experimental data reported here were not conclusive in

confirming this trend, due to the level of inaccuracy of the HSE techniques. From a

methodological point of view this indicates a limit of HSE techniques, which may

not be able to discern between activities within a very narrow range of values.

• For what concerns the addition of a diol to the Ti-centre, the expected effect should

be a lower catalytic activity when using TBHP as oxidant due to the lower

accessibility of the site (similarly to what is described above for isopropoxy and

cyclopentadienyl groups). On the other hand, a higher activity with H2O2 may occur

if the diol ligand were protecting the titanium site from reaction with water. The data

reported in Figure 4.8 show the expected decrease in activity with TBHP when

ethylene glycol, 2,3-butanediol and 1,2-cyclohexanediol are used, while the effect is

less relevant with the two bulkiest diols, pinacol and benzopinacol. This indicates

that the coordination of pinacol and benzopinacol to the titanium centres on the

silsesquioxanes is probably sterically hindered, particularly for the Ti-centre on the

pure silsesquioxane a7b3. The same effect to a minor extent can explain the

intermediate activity obtained with 1,2-cyclohexanediol. The fact that the presence of

a diol causes a much higher decrease in activity with the HSE mixture of

silsesquioxanes than with pure silsesquioxane a7b3 points to the different nature of

the Ti-complexes present in the HSE mixture, some of which probably become

completely inaccessible for catalysis in the presence of one or more diol ligands.

When H2O2 is used as oxidant, a positive effect of the diols can be seen mainly with

ethylene glycol and 2,3-butanediol, in agreement with the result for TBHP, where for


73

these two diols the biggest decrease in activity is observed. Unfortunately, the

activity is still rather low, with the best result being 8% of conversion of H2O2 into

1,2-epoxyoctane in the case of 2,3-butanediol with pure silsesquioxane a7b3. This

indicates that the effect of the diols, although detectable – thus, the concept works, is

not sufficient to produce a good and robust catalyst for the epoxidation of 1-octene

with hydrogen peroxide.

4.4 Conclusions

The effect of different variables on the preparation of Ti-silsesquioxane catalysts

was explored by means of HSE techniques. Fine-tuning the parameters influencing the

synthesis of the silsesquioxane precursors led to slight improvements in the activity of

the Ti-catalysts and to some knowledge in regards to how these parameters affect the

hydrolytic condensation of trichlorosilanes. The initial hypothesis that, among these

parameters, the nature of the solvent and of the R-group were dominant in influencing

the formation of silsesquioxanes was confirmed. The optimisation of the features of the

catalytic titanium centre anchored to the silsesquioxane structures evidenced the

importance of the accessibility of the titanium site in determining the activity of the

catalyst.

4.5 Experimental

All the experiments described in this chapter were performed by means of the

HSE automated parallel synthesis workstation described in Appendix A. The

experimental procedure followed for the fine-tuning of the parameters influencing the

synthesis of silsesquioxane precursors is analogous to that of the first HSE study

described in Chapter 3. For the HSE study of the Ti-centres, titanium complexation was

performed by dissolving either 4.7·10-5 mol of the pure cyclopentyl silsesquioxane a7b3

or a corresponding amount of the HSE-optimised cyclopentyl silsesquioxane mixture in

2 ml of THF, followed by the addition of 54 µmol of one of the titanium precursors as a

solution in an organic solvent (the corresponding alcohol for the titanium alkoxides,

Chapter 4 __________________________________________________________________________________________________________

74

dichloromethane for cyclopentadienyltitanium trichloride). Samples were stirred in an

orbital shaker for 5 hours at 60°C. In the case of the experiment with diols, THF

solutions, each containing 0.1 mmol of a diol, were added to the appropriate vials.

Samples were then stirred in the orbital shaker for 10 hours at 50°C. Finally, THF was

removed by means of a vacuum centrifuge and the samples were stored under argon.

The activity of the Ti-silsesquioxanes in the epoxidation of 1-octene was determined by

adding to each dried sample 2.26 ml of 1-octene (with 2%vol of decane as internal

standard), that acted both as reactant and as solvent, and 121 µl of TBHP solution

(~45%wt solution in cyclohexane) or 70 µl of H2O2 (35%wt solution in H2O). H2O2 and

1-octene are not miscible: to obtain a monophasic system, the reactions were

alternatively performed with 1-propanol as solvent (1.13 ml of 1-propanol, 1.13 ml of

1-octene and 70 µl of aqueous H2O2). However, the activities obtained in these

conditions were lower than with the biphasic system, probably because of faster catalyst

deactivation caused by reaction with water. Samples were taken after having been

stirred in the orbital shaker for 4 hours at 80°C (TBHP) or for 6 hours at 60°C (H2O2)

and analysed on the UNICAM Pro GC. The reported activities were obtained by

normalising the 1,2-epoxyoctane GC peak area by means of the internal standard. As a

reference, a blank reaction with 1-octene and H2O2 and without catalyst was performed,

giving no epoxidation activity.


75

References

1 K.R. Popper, The Logic of Scientific Discovery, Hutchinson of London, 1959. 2 K.R. Popper, Conjectures and Refutations. The Growth of Scientific Knowledge, Routledge and Kegan

Paul, 1969, 3rd edition. 3 P.P. Pescarmona, T. Maschmeyer, Aust. J. Chem., 2001, 54, 583, see also Paragraph 2.2.1. 4 See Paragraph 3.3. 5 P. P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, Angew. Chem. Int. Ed., 2001, 40,

740, see also Chapter 3. 6 F. J. Feher, D. A. Newman, J. F. Walzer, J. Am. Chem. Soc., 1989, 111, 1741. 7 F. J. Feher, T. A. Budzichowski, R. L. Blanski, K. J. Weller, J. W. Ziller, Organometallics, 1991, 10,

2526. 8 J.F. Brown, L.H. Vogt, P.I. Prescott J. Am. Chem. Soc., 1964, 86, 1120. 9 J.F. Brown, L.H. Vogt, J. Am. Chem. Soc., 1965, 87, 4313. 10 M.C. Klunduk, T. Maschmeyer, J.M. Thomas, B.F.G. Johnson, Chem. Eur. J., 1999, 5, 1481. 11 J.M. Thomas, G. Sankar, M.C. Klunduk, M.P. Attfield, T. Maschmeyer, B.F.G. Johnson, R.G. Bell, J.

Phys. Chem. B, 1999, 103, 8809. 12 F. Figueras, H. Kochkar, S. Caldarelli, Micropor. Mesopor. Mat., 2000, 39, 249. 13 J.M. Fraile, J.I. García, J.A. Mayoral, E. Vispe, J. Catal., 2000, 189, 40. 14 P.P. Pescarmona, T. Maschmeyer, NATO Science Series, 2002, Ser. II Vol. 69, 173. 15 P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, Chem. Eng. Commun., accepted

for publication. 16 T. Maschmeyer, M.C. Klunduk, C.M. Martin, D.S. Shephard, J.M. Thomas, B.F.G. Johnson, Chem.

Commun., 1997, 1847.

77

5

Fast and high-yield syntheses of cyclopentyl and cyclohexyl silsesquioxanes using acetonitrile as reactive solvent

Abstract

New and efficient methods for the synthesis of cyclopentyl and cyclohexyl

silsesquioxanes are described. In the case of cyclopentyl silsesquioxanes, structure a7b3

was obtained selectively with an isolated yield of 64% within 18 hours. The synthesis

was monitored by means of mass spectrometry and in-situ infrared spectroscopy in

combination with chemometric analysis methods. This study allowed to identify the

silsesquioxane species present in the reaction mixture, to explain the high selectivity of

the reaction towards silsesquioxane a7b3 and to propose a mechanism for its formation.

In the case of cyclohexyl silsesquioxanes, the synthesis gave high yields but was not

selective towards a particular structure type. The main products were: silsesquioxanes

a6b0, a6b2, a7b1 and a7b3. The crucial role of acetonitrile as a reactive solvent in the

syntheses is discussed.

____________________ The contents of this chapter have been submitted to:

P.P. Pescarmona, J.C. van der Waal, T. Maschmeyer, Eur. J. Inorg. Chem.

P.P. Pescarmona, M.E. Raimondi, J. Tetteh, B. McKay, T. Maschmeyer, J. Phys. Chem. B, accepted.

Chapter 5 __________________________________________________________________________________________________________

78

5.1 Introduction

Chapter 3 and 4 described a study of the synthesis of silsesquioxanes1 as

precursors for Ti-catalysts active in the epoxidation of alkenes, realised by means of

High-Speed Experimentation (HSE) techniques.2,3 The best catalysts were obtained with

cyclopentyl and cyclohexyl silsesquioxanes synthesised by the hydrolytic condensation

of the corresponding trichlorosilanes with acetonitrile as solvent. Here, a follow-up

study of these two HSE leads by means of other laboratory techniques is reported, with

the aim to fully characterise the silsesquioxane intermediates and products and to

understand the role of the solvent.

The hydrolytic condensation of trichlorosilanes in acetone to produce incompletely

condensed silsesquioxanes was first reported by Brown and Vogt4 (for R = cyclohexyl)

and more recently it was optimised and studied in detail by Feher et al. (for

R = cyclopentyl and cyclohexyl).5,6 In the case of R = cyclopentyl, silsesquioxane a7b3

(Figure 5.1) was the only product and was obtained with a yield of 29% after 3 days of

reaction.6 When R = cyclohexyl, the hydrolytic condensation proceeded more slowly

and produced a mixture of three silsesquioxanes: a6b0, a7b3 and a8b2.5 Silsesquioxane

a7b3 was the main product, but with a low yield (8% after 12 weeks of reaction).

Figure 5.1. Silsesquioxanes a7b3, R7Si7O9(OH)3 [R = cyclopentyl, cyclohexyl].

OHOH

OH

R

R

OO

SiO

Si

O

O

OSiSi

OSiOSi

O

Si

a7b3

RR

R

R

R

OHOH

OH

R

R

OO

SiO

Si

O

O

OSiSi

OSiOSi

O

Si

a7b3

RR

R

R

R

Synthesis of cyclopentyl and cyclohexyl silsesquioxanes __________________________________________________________________________________________________________

79

More information about the hydrolytic condensation of organosilanes can be found in

Paragraph 2.2.1. The reaction scheme is reproduced below since it will be useful to refer

to it throughout this chapter:

RSiCl3 + 3 H2O RSi(OH)3 + 3 HCl [1]

a RSi(OH)3 (RSiO1.5)a(H2O)0.5b + (1.5a – 0.5b) H2O [2]


5.2.1 Synthesis of cyclopentyl silsesquioxane a7b3

The hydrolytic condensation of cyclopentyltrichlorosilane in acetonitrile in the

presence of an excess of water led to the formation of a white precipitate. After 18 hours

of reaction at 50°C, a crude precipitate was obtained in ~50 % yield.I It mainly

contained silsesquioxane a7b3 [(c-C5H9)7Si7O9(OH)3], as determined on the basis of 29Si NMR and MS analyses.7 Small amounts of impurities, predominantly the cubic

completely condensed silsesquioxane a8b0, (c-C5H9)8Si8O12, were present in the

precipitate. If the reaction time was extended from 18 hours to 2 weeks,3 the yield of

precipitated silsesquioxanes was nearly quantitative with the main fraction still being

silsesquioxane a7b3. The synthesis was further optimised according to the fine-tuning of

the reaction conditions described in Paragraph 4.3.2 and on the basis of the observation of

Feher et al., that the yield of the synthesis of cyclopentyl silsesquioxane a7b3 in acetone

can be significantly improved by refluxing the reaction mixture.5 Therefore, the synthesis

method was modified by increasing the reaction temperature from 50°C to reflux

conditions. This led to an increase of the yield to 81% after 18 hours of reaction, with the

high selectivity towards silsesquioxane a7b3 being preserved. Silsesquioxane a7b3 was

then purified by pyridine extraction5,6 and the pure compound (as determined by

I Relative to silsesquioxane a7b3 (the yield is calculated by dividing the grams of product by the grams of

silsesquioxane a7b3 that would be obtained if the reaction quantitatively gave silsesquioxane a7b3).

Chapter 5 __________________________________________________________________________________________________________

80

29Si NMR, cf. Figure 5.2 and Figure 3.3) could be isolated in an overall yield of 64% - a

significant improvement compared to the previously reported yield of 29% in 3 days.6

Figure 5.2. 29Si NMR spectrum of cyclopentyl silsesquioxane a7b3.

This improved synthesis procedure differs from that reported by Feher et al.6 regarding

a number of parameters: the lower initial trichlorosilane concentration, the higher

solvent-to-water ratio and above all the nature of the solvent (acetonitrile), which

proved to be a dominant parameter in influencing the synthesis of silsesquioxanes.1-3

After the hydrolytic condensation had proceeded for 18 hours under reflux, the reaction

mixture separated into two layers: a water phase (bottom) and an acetonitrile phase

(top). This liquid-liquid phase separation is related to the hydrolysis of acetonitrile, a

side reaction being favoured by the presence of an acid (cf. reaction [3]),8,9 in this case

the hydrochloric acid produced during the hydrolysis of (c-C5H9)SiCl3 (cf. reaction [1]):

CH3CN + 2H2O + HCl → CH3COOH + NH4Cl [3]

The ammonium chloride obtained is much more soluble in water than in acetonitrile,

causing phase separation (0.41 g of ammonium chloride are needed to induce phase

separation of a mixture of 12.5 ml of water and 37.5 ml of acetonitrile). The preferential

solvation of ammonium chloride by water, as compared to that of silsesquioxanes or

-57.83

-65.71

-67.24

ppm-68 -70-66 -72-64-62-60-54 -58-52 -56

-57.83

-65.71

-67.24

ppm-68 -70-66 -72-64-62-60-54 -58-52 -56


81

acetonitrile, drives reaction [2] towards the formation of more condensed species. This

effect is very pronounced towards the end of the experiment, when actual phase

separation is observed. At the same time, the hydrolysis of acetonitrile in presence of an

acid plays an important role due to the fact that the formation of ammonium chloride

(partially) neutralises the reaction mixture. A detrimental effect of the acidity of the

solution was already observed in the synthesis of silsesquioxane as Ti-catalyst

precursors.2 Since under acidic conditions silsesquioxanes undergo fast

hydrolysis/condensation reactions,1 the species formed might be partially decomposed

or further condensed before they would start to precipitate from the solution. The partial

neutralisation of the reaction mixture could slow down the hydrolysis/condensation

reactions, favouring the precipitation of the species that are only poorly soluble in this

medium, like the incompletely condensed silsesquioxane a7b3.II

A further contributing factor regarding the improved yield when using acetonitrile as the

solvent is its higher polarity compared to acetone, the solvent used in literature for the

synthesis of silsesquioxane a7b3.6 A highly polar solvent can better stabilise

incompletely condensed species by interaction with their silanol groups, therefore,

favouring the synthesis of these silsesquioxanes over completely condensed structures.

Moreover, it has been proposed that in the presence of a polar molecule the activation

barrier for the condensation reactions towards the formation of silsesquioxanes is

reduced and, therefore, the synthesis is sped up.10,11

In this context, it is also relevant to note that acetonitrile has a higher boiling point than

acetone (82°C against 56°C). Since the reaction is performed under reflux, the synthesis

is accelerated.

High-Speed Experimentation was an efficient means to determine the optimum

conditions for the high selectivity synthesis of silsesquioxane a7b3,2 but could not

directly be used to elucidate the mechanism of formation. In order to shed light on this

mechanism, to identify the species formed during the process and to try to explain the II While the partial neutralisation of the HCl formed during the hydrolysis of trichlorosilanes might have a

role in accelerating the formation of silsesquioxane precipitates, the complete neutralisation is

detrimental. A decrease of the yield (from 79% to 44%, as crude precipitate) was observed by performing

the synthesis of cyclohexyl silsesquioxanes in the presence of triethylamine (C2H5)3N. This indicates that

without HCl the hydrolysis/condensation reactions are significantly slowed down from the beginning of

the synthesis.

Chapter 5 __________________________________________________________________________________________________________

82

high selectivity towards structure a7b3 of the optimised synthetic method described

above (64% yield in 18 hours), the synthesis of cyclopentyl silsesquioxane a7b3 was

monitored by means of Electrospray Ionisation Mass Spectrometry (ESI MS)12-14 and

in-situ Attenuated Total Reflection Fourier Transform Infrared (ATR FTIR)

spectroscopy.15,16 Spectroscopic data from the latter were analysed using chemometric

methods to identify the pure spectra and relative concentration profiles.

5.2.1.1 Monitoring by means of mass spectrometry

The mechanistic study by means of mass spectrometry was performed by analysing

samples of the reaction mixture at regular intervals throughout the experiment. The MS

spectra recorded between t = 0 and t = 1440 min. show peaks corresponding to

cyclopentyl silsesquioxane structures with 1 ≤ a ≤ 13.

Figure 5.3. ESI MS plot after 90 minutes of reaction and list of possible cyclopentyl

silsesquioxanes with 1 ≤ a ≤ 8. Analytical parameters for MS measurements

were set as follows: flow rate (of the syringe pump) = 40 µl/min,

RF lens = 0.31V, capillary = 3.20 kV, cone = 30 V, extractor = 4 V, source

block temperature = 80°C, desolvation temperature = 300°C, nebuliser gas

flow = 85 l/hour, desolvation gas flow = 450 l/hour.

a = 4

a = 5

a = 6

a = 8a = 7

a = 9a = 10 a = 11 a = 12

a = 13

a b m/z a b m/z a b m/z a b m/z a b m/z

1 3 148.06 4 0 484.16 6 0 726.24 7 1 856.29 8 0 968.32 4 2 502.17 6 2 744.25 7 3 874.30 8 2 986.33

2 0 242.08 4 4 520.19 6 4 762.27 7 5 892.32 8 4 1004.352 2 260.10 4 6 538.21 6 6 780.29 7 7 910.33 8 6 1022.372 4 278.11 6 8 798.30 7 9 928.35 8 8 1040.38 5 3 632.22 8 10 1058.40

3 3 390.14 5 5 650.24 3 5 408.16

5 7 668.25

a = 4

a = 5

a = 6

a = 8a = 7

a = 9a = 10 a = 11 a = 12

a = 13

a = 4

a = 5

a = 6

a = 8a = 7

a = 9a = 10 a = 11 a = 12

a = 13

a b m/z a b m/z a b m/z a b m/z a b m/z

1 3 148.06 4 0 484.16 6 0 726.24 7 1 856.29 8 0 968.32 4 2 502.17 6 2 744.25 7 3 874.30 8 2 986.33

2 0 242.08 4 4 520.19 6 4 762.27 7 5 892.32 8 4 1004.352 2 260.10 4 6 538.21 6 6 780.29 7 7 910.33 8 6 1022.372 4 278.11 6 8 798.30 7 9 928.35 8 8 1040.38 5 3 632.22 8 10 1058.40

3 3 390.14 5 5 650.24 3 5 408.16

5 7 668.25


83

Peaks due to a = 7 species (m/z = 839.3, 857.3, 875.3, 893.3) are present in all the

recorded spectra (see Figure 5.3 for an example). The relative abundance of these peaks

compared to that of the peaks of other silsesquioxane species is rather low at any

reaction time, indicating a low concentration in solution. Knowing that the precipitate

produced by the synthesis is mainly silsesquioxane a7b3, it can be inferred that this

compound has a low solubility in the reaction mixture and tends to precipitate as it gets

formed. Its lower solubility with respect to other cyclopentyl silsesquioxanes is

probably due to the tendency of silsesquioxane a7b3 to form dimers,5,17 which are

insoluble in polar solvents such as acetonitrile. This is in agreement with the observed

formation of a white precipitate after 1 hour of reaction.

The concentration of the a = 7 species in solution can, therefore, be considered

approximately constant during the course of the reaction. This assumption allowed to

normalise the intensity of the peaks of the other silsesquioxanes relative to the a = 7

peaks, making it possible to compare the spectra measured at the various reaction

times.III The relative concentrations of the principal silsesquioxane species with

increasing a values are plotted against reaction time in Figure 5.4.

This plot shows how various silsesquioxane species formed and disappeared over time.

At t = 0, silsesquioxane species with 1 ≤ a ≤ 6 were present in considerable amounts.

No cyclopentyltrichlorosilane was detected, indicating that the hydrolysis step [1] was

complete before the first sample was injected in the mass spectrometer.17 The presence

of various species at t = 0 shows that the process of condensation of silsesquioxane

a1b3, (c-C5H9)Si(OH)3, (step [2]) leading to the formation of larger silsesquioxane

structures is also very fast. At t = 0, species with a = 4 were the most abundant.

Silsesquioxanes with 1 ≤ a ≤ 4 can be seen as the building blocks for larger

silsesquioxane structures (see Figure 5.5 for a proposed synthesis scheme). They are

likely to be reactive species that tend to condense further, as confirmed by the fact that

their concentrations diminished quickly as the reaction proceeded. The smaller the

structures, the faster they disappeared to form larger ones. At t = 30 minutes,

silsesquioxane a1b3 had already disappeared from the reaction mixture. Silsesquioxanes

III A normalisation of the MS data is necessary, because the mass spectrometer produces plots in which

the intensities of the peaks are not absolute but normalised to the most intense peak, the value of which is

set at 100%.

Chapter 5 __________________________________________________________________________________________________________

84

with a = 2 disappeared after 1 hour of reaction and those with a = 3 reached a very low

concentration after 1 hour of reaction and could not be detected anymore after 6 hours.

The concentration of silsesquioxanes with a = 4 decreased less rapidly. These species

were still present in small amounts after 10 hours and were only fully consumed close to

the end of the reaction, after 24 hours.

Figure 5.4. Relative intensity of the ESI MS peaks of the various silsesquioxane species

present in solution, as a function of the reaction time. All the intensities are

normalised to the intensity of the peaks of a = 7 species.

0.00

1.00

2.00

3.00

0 60 120 180 240 300 360 420 480 540 600 660 720 780 840 900 960 1020 1080 1140 1200 1260 1320 1380 1440

time (minutes)

rela

tive

inte

nsity

a = 8

a = 9

a = 10

a = 11

a = 12

a = 13

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

0 60 120 180 240 300 360 420 480 540 600 660 720 780 840 900 960 1020 1080 1140 1200 1260 1320 1380 1440

time (minutes)

rela

tive

inte

nsity

a = 1a = 2a = 3a = 4a = 5a = 6a = 8

0 60 120 180 270 360 480 600 1440

0 60 120 180 270 360 480 600 1440

3.00

2.00

1.00

0.000.00

1.00

2.00

3.00

0 60 120 180 240 300 360 420 480 540 600 660 720 780 840 900 960 1020 1080 1140 1200 1260 1320 1380 1440

time (minutes)

rela

tive

inte

nsity

a = 8

a = 9

a = 10

a = 11

a = 12

a = 13

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

0 60 120 180 240 300 360 420 480 540 600 660 720 780 840 900 960 1020 1080 1140 1200 1260 1320 1380 1440

time (minutes)

rela

tive

inte

nsity

a = 1a = 2a = 3a = 4a = 5a = 6a = 8

0 60 120 180 270 360 480 600 14400 60 120 180 270 360 480 600 1440

0 60 120 180 270 360 480 600 14400 60 120 180 270 360 480 600 1440

3.00

2.00

1.00

0.00

3.00

2.00

1.00

0.00


85

For a = 1, the only possible structure is the trisilanol a1b3.1 For a = 2 and a = 3,

more structures are possible but a2b4 and a3b5 are the only two that do not present a large

geometrical strain. For a = 4, two structures are the most likely: a4b4 and a4b6, the first

being a ring and the second a linear structure. The ring structure is part of the structure of

the final product a7b3, and so it can be assumed that a4b418 is the precursor species. These

structures are schematised in Figure 5.5, together with their most likely paths of reaction.

For most of the reaction, silsesquioxanes with a = 5 and a = 6 were the two

major species in solution. Their concentrations increased in the beginning of the

reaction to reach a maximum after 1 hour and then gradually decreased. This happened

more rapidly for species with a = 5 that were only present in very small amounts at the

end of the reaction, than for silsesquioxanes with a = 6, which were still present in

relevant amounts after 24 hours of reaction. The two most likely structures with a = 5

and a = 6 are silsesquioxanes a5b5 and a6b4, which are represented in Figure 5.5,

together with their possible pathways of formation.

For silsesquioxanes with a = 8, two groups of peaks with different relative

concentration profiles exist. The different behaviour as a function of the reaction time

suggests that the two groups of peaks belong to two different species with a = 8. The

first group presented a main peak at m/z = 987.15, which corresponds to an a8b2

silsesquioxane. This species has a low concentration during the entire reaction, with

only a slightly positive slope towards the end (see Figure 5.4, top plot). The second

group of peaks presented a main peak at m/z = 1041.20, its intensity reached a

maximum at t = 30 minutes and then decreased rapidly to disappear after 3 hours of

reaction (see Figure 5.4, bottom plot). The intensity profile for these species is very

similar to those of species with 9 ≤ a ≤ 13, indicating a maximum species concentration

at t = 30 minutes for a = 9, 10 and at t = 60 minutes for 11 ≤ a ≤ 13. Given the short

lifetime of the species with 8 ≤ a ≤ 13 and the fact that they were present in the early

stages of the reaction, it can be suggested that these species are instable aggregates of

two smaller silsesquioxanes.IV For example, species with a = 8 are probably dimers of

a = 4 species, those with a = 9 are constituted by a = 4 and a = 5 species, and so on.

IV A similar effect can be seen measuring a mass spectrum of pure cyclopentyl silsesquioxane a7b3 with

similar experimental conditions but at two different cone voltages: with cone = 30V two main peaks are

present, at m/z = 875.33 and at m/z = 1749.89; with cone = 65V, the peak at m/z = 1749.89, due to the

a7b3 dimer, disappears.

Chapter 5 __________________________________________________________________________________________________________

86

Figure 5.5. Proposed mechanism for the synthesis of cyclopentyl silsesquioxane a7b3.

The silsesquioxane structures are represented in a schematic way. Each

circle symbolises a siloxane unit [(c-C5H9)SiO3]: silicon atoms are

represented by the circles and the oxygen atoms by the lines; non-bridging

lines represent the -OH groups. The cyclopentyl groups are not shown.

On the basis of all the information collected from this MS study, it is possible to

propose a mechanism for the formation of silsesquioxane a7b3 (Figure 5.5).

Silsesquioxanes with 1 ≤ a ≤ 4 are formed in the early stages of the synthesis and react

very quickly to form more condensed species. It is assumed that the very reactive a1b3

will not be formed again by hydrolysis reactions of more condensed species: this means

that the reactions in which a1b3 takes part are effectively irreversible and that the

compound will only be available for the reactions in the early stages of the synthesis.

Silsesquioxanes with 2 ≤ a ≤ 4 are then going to react with each other to form the more

condensed silsesquioxanes with 4 ≤ a ≤ 8. These reactions are more easily reversible,

-H2O

precipitate precipitate

a1b3 a2b4a3b5

a4b4a6b4

a5b5

a7b3

a8b2

a8b0

-H2O

precipitate precipitate

a1b3 a2b4a3b5

a4b4a6b4

a5b5

a7b3

a8b2

a8b0


87

meaning that the structures that are formed by condensation can be hydrolysed back to

the original compounds or to other less condensed species.

As mentioned above, silsesquioxane a7b3 is less soluble than other

silsesquioxanes present in the reaction mixture and precipitates as a white solid. This

will influence the equilibrium composition and consequently drive the reactions that

involve its formation towards the product; therefore, these reactions are considered

irreversible. Silsesquioxanes with a = 6 exhibit a maximum of concentration in solution

after 1 hour of reaction; they are then rapidly consumed up until 3 hours of reaction,

after which the reaction proceeds more slowly. Since no a = 6 silsesquioxane is present

in the precipitate, it is inferred that the compound gets slowly hydrolysed to smaller

species which then recombine to give the product a7b3.6,19 The absence of an a = 6

silsesquioxane among the products is, therefore, not due to the instability of the

compound but rather to its higher solubility in the reaction mixture. The completely

condensed silsesquioxane a8b0, obtained by condensation of a8b2, is only present in

traces in the precipitate. This means that, although silsesquioxane a8b0 might be

expected to be the most thermodynamically favoured structure, its formation is

unfavourable compared to that of silsesquioxane a7b3, suggesting that the reaction is

kinetically controlled.

An alternative way to represent the results of the MS analysis is by plotting the

fraction of the total silsesquioxane structures for each value of a (with 1 ≤ a ≤ 8), as a

function of the reaction time (Figure 5.6). This graph gives a measure of the distribution

of the various species in solution at various stages of the reaction. This plot does not,

however, provide any information about the total concentration of silsesquioxane

species in solution (the value of which is always set at 100%).III(page 83) This graph nicely

shows how the smaller silsesquioxane species constitute the main fraction in the early

stages of the synthesis and then tend to gradually be consumed in favour of bigger

structures as the reaction proceeds. Towards the end of the synthesis, silsesquioxanes

with a = 6 are the major fraction in solution, though the relative fractions of species

with a = 7 and a = 8 are increasing. This is in agreement with the hypothesis of the

concentrations of silsesquioxanes a = 7 and a = 8 being constant in solution and that of

silsesquioxanes a = 6 slowly decreasing upon hydrolysis and consecutive condensation

towards silsesquioxane a7b3.

Chapter 5 __________________________________________________________________________________________________________

88

Figure 5.6. Fraction of the total silsesquioxane structures in solution for 1 ≤ a ≤ 8, as a

function of the reaction time (based on ESI MS data).

5.2.1.2 Monitoring by means of infrared spectroscopy

Attenuated Total Reflection (ATR) FTIR spectroscopy allows in-situ monitoring of

liquid-phase reactions by collecting the infrared spectra for solution species directly in

contact with the infrared probe.20 In the general case of a reaction in which various

compounds are present, each ATR FTIR spectrum will consist of the overlapping

spectra of all the pure components present in solution at a specific reaction time. During

the reaction, the concentration of reagents and products will change, thus influencing

the spectrum profile. This may allow appropriate chemometric methods to be used to

deconvolute the pure component spectra and relative concentration profiles as a

function of time. If the IR spectra of some components are known in advance (e.g. the

solvent, reagents and selected products), it is possible to use them as references to

improve the fidelity of the deconvolution.

In the ATR FTIR study of the synthesis of cyclopentyl silsesquioxane a7b3,

in-situ ATR FTIR spectra of the reaction mixture were collected every 2 minutes during

the reaction time. The spectra obtained were plotted as a function of the reaction time

(Figure 5.7). The pure component spectra and relative concentration profiles were

0%

10%

20%

30%

40%

50%

60%

0 60 120 180 240 300 360 420 480 540 600 660 720 780 840 900 960 1020 1080 1140 1200 1260 1320 1380 1440

time (minutes)

perc

enta

gea = 1a = 2a = 3a = 4a = 5a = 6a = 7a = 8


89

subsequently recovered using a Multivariate Curve Resolution (MCR)21 technique

based on a modified target factor analysis algorithm.22

Figure 5.7. In-situ ATR FTIR spectra of the silsesquioxane reaction mixture, as a

function of the reaction time for a window of 670-1000 cm-1. Analytical

parameters were as follows: IR analysis (670-4000 cm-1) was performed for

10 hours with an acquisition of a spectrum every 2 minutes, 16 scans per

spectrum and a resolution of 4 cm-1.

Principal Component Analysis (PCA)23 was first used to determine the number

of independently varying chemical species present and to provide initial estimates of the

spectral shapes resulting from these species and of their concentration profiles.

Reference ATR FTIR spectra for a number of components (the solvent acetonitrile and

water, the reagent cyclopentyltrichlorosilane and the product a7b3) were measured to

assist in the deconvolution of the data. Frequency windows were selected that allowed

the best discrimination between the reference compounds (725-775 cm-1 for acetonitrile,

850-900 cm-1 for water and the silsesquioxane). Finally, the MCR technique was

Chapter 5 __________________________________________________________________________________________________________

90

applied to the data in the selected frequency windows to find the component spectra and

relative concentration profiles that best fit the observed spectra.

Figure 5.8. Comparison of the measured reference spectra (dotted line) with the pure

component spectra (solid line) obtained by chemometric deconvolution of

the complex ATR FTIR spectra reported in Figure 5.7.

From Figure 5.8 it can be seen that the pure component spectra obtained by the

MCR technique closely match the reference spectra for acetonitrile (with a correlation

coefficient R = 0.987), water (R = 0.993) and silsesquioxane (R = 0.953). A poor match

(R = 0.214) was obtained for the cyclopentyltrichlorosilane reference spectra, indicating

that cyclopentyltrichlorosilane cannot be detected at any stage during the reaction. This

supports the MS results that also suggest that trichlorosilane is immediately hydrolysed

once water is added to the solution.

acetonitrile

silsesquioxanespecies

water


acetonitrile

silsesquioxanespecies

water



91

Relative concentration profiles for the identified species are given in Figure 5.9.

The concentrations of acetonitrile and water were reasonably constant during the

reaction, reflecting the fact that both acetonitrile and water were present in large

quantities, and that any change in concentration is effectively negligible.

Figure 5.9. Relative concentration profiles of the three species identified by

chemometric deconvolution of the complex ATR FTIR spectra reported in

Figure 5.7.

The relative concentration profile of the silsesquioxane component, which has a slightly

less accurate fit to the silsesquioxane a7b3 reference spectrum (R = 0.953), reaches a

maximum after 2 hours and becomes approximately constant after 4 hours of reaction.

From the MS study (see Paragraph 5.2.1.1) it is inferred that silsesquioxane a7b3 has a

very low and approximately constant concentration in solution throughout the entire

reaction, since the compound is only sparingly soluble, and precipitates at higher

concentrations. This suggests that the silsesquioxane monitored by ATR FTIR is not

represented by structure a7b3. Other incompletely condensed cyclopentyl

silsesquioxanes are expected to have infrared spectra very similar to that of a7b3 in the

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0 100 200 300 400 500time (minutes)

norm

aliz

ed in

tens

ity

silsesquioxane species

acetonitrile

waternorm

alis

ed a

ctiv

ity

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0 100 200 300 400 500time (minutes)

norm

aliz

ed in

tens

ity

silsesquioxane species

acetonitrile

waternorm

alis

ed a

ctiv

ity

Chapter 5 __________________________________________________________________________________________________________

92

studied spectral regions.12,17,V Therefore, it is proposed that the species identified by

MCR is an a = 5 or an a = 6 structure, which presents an MS concentration profile

similar to that obtained by ATR FTIR, or a mixture of more silsesquioxane structures.

The hypothesis of an a = 6 structure is supported by the fact that the ATR FTIR

concentration profile for the silsesquioxane species as a function of time is rather

similar to the MS concentration profile obtained for the a = 6 silsesquioxane (compare

Figures 5.4 and 5.9). The slightly different position of the maximum in the two plots is

considered to be due to the longer time required in the ATR FTIR experiment to reach

the reflux temperature.VI

The combined MS and ATR FTIR results point to the fact that the concentration

of the silsesquioxane species in solution becomes almost constant after 4 hours of

reaction. This might imply that a reaction time shorter than 18 hours could be sufficient

to obtain silsesquioxane a7b3 in high yield. Therefore, the synthesis was repeated with

the same conditions but with a reaction time of 6 hours: silsesquioxane a7b3 was

isolated with a yield of 54%. After collection of the precipitate, the reaction mixture was

allowed to react for 12 more hours yielding more precipitate and the total yield of 64%

was achieved. Although the yield after 6 hours is lower than that obtained after 18 hours

of reaction, the experiment confirmed that most of the yield of the desired a7b3

structure is generated within the first hours of reaction. This finding would certainly be

of practical importance for a potential up-scaling of silsesquioxane a7b3 production.

V This similarity was confirmed by measuring and comparing the transmission IR spectra (KBr pellet) of

cyclopentyl silsesquioxane a7b3 with that of the mixture of silsesquioxanes (mainly containing a = 5 and

a = 6 species) obtained by removing the solvent from the reaction mixture after 18 hours of reaction at

50°C (instead of reflux conditions). The two spectra present corresponding sets of peaks (within 6 cm-1)

in the region 670-1000 cm-1.

VI This explanation was confirmed by repeating part of the MS study of the synthesis of

silsesquioxane a7b3 at 50°C (instead of reflux temperature). In these conditions the hydrolytic

condensation is slower and, though the same silsesquioxane species were present in solution, species with

1 ≤ a ≤ 4 disappeared more slowly and those with a = 5 and a = 6 reached the maximum of the

concentration profile after a longer reaction time.


93

5.2.2 Synthesis of cyclohexyl silsesquioxanes

The hydrolytic condensation of cyclohexyltrichlorosilane in acetonitrile, in the

presence of an excess of water, produced a white precipitate in 42 % yieldI(page 79) after 18

hours of reaction at 50°C. The crude precipitate contained a mixture of silsesquioxane

species, as determined by 29Si NMR. The formation of the precipitate could be

accelerated by performing the synthesis under reflux, in analogy to the synthesis of

cyclopentyl silsesquioxanes (see Paragraph 5.2.1). This caused an increase of the yield to

79% after 18 hours of reaction, with the precipitate still being a complex mixture of

silsesquioxanes (see Figure 5.10 for the 29Si NMR spectrum and the MS plot).

Figure 5.10. 29Si NMR spectrum (top) and APCI MS plot (bottom) of the cyclohexyl

silsesquioxane mixture obtained by the hydrolytic condensation of

cyclohexyltrichlorosilane in acetonitrile.

ppm-69-68-67-66-65-64-63-62-61-60-59-58-57-56

-56.18

-56.72

-57.16

-58.66-58.87

-59.79 -62.32 -62.57 -66.76

-67.93

-68.45-69.33

811.36

829.43937.12

955.18

973.19

1081.07 1099.07

ppm-69-68-67-66-65-64-63-62-61-60-59-58-57-56

-56.18

-56.72

-57.16

-58.66-58.87

-59.79 -62.32 -62.57 -66.76

-67.93

-68.45-69.33

ppm-69-68-67-66-65-64-63-62-61-60-59-58-57-56

-56.18

-56.72

-57.16

-58.66-58.87

-59.79 -62.32 -62.57 -66.76

-67.93

-68.45-69.33

811.36

829.43937.12

955.18

973.19

1081.07 1099.07

811.36

829.43937.12

955.18

973.19

1081.07 1099.07

Chapter 5 __________________________________________________________________________________________________________

94

MS analysis indicated that the silsesquioxanes present in the mixture were mainly

constituted of 6,7 or 8 siloxane units (6 ≤ a ≤ 8). The most intense peak in the 29Si NMR

spectrum (δ -56.72, 40% of the crude precipitate)VII belongs to the completely

condensed silsesquioxane (c-C6H11)6Si6O9 (a6b0, see Figure 5.11), which can be easily

separated from the incompletely condensed species by extraction with pyridine, in

which just the latter silsesquioxanes are soluble.5

Figure 5.11. Silsesquioxanes a6b0, a6b2 and a7b1 (α and β isomers) [R = cyclohexyl].

The main species present in the remaining mixture of incompletely condensed

silsesquioxanes were assigned by combining 29Si NMR and MS data:

(c-C6H11)6Si6O8(OH)2 (a6b2, 11% of the crude precipitate,VII see Figure 5.11),

VII Percentages based on the integrals of the peaks present in the 29Si NMR spectrum of the crude product.

O

OHOO

SiO

Si

OO

OSiSi

OSiOSi

O

Si

R

RR

R R

RR

O

OH

OO

SiO

Si

OO

OSiSi

OSiOSi

O

Si

R

RR

R R

R

Ra7b1

O

OO

SiO

Si

OO

SiSiO O

Si

O

Si

R

R

RR

R

R

a6b0

OH

RO

OSi

O

Si

Si

O

SiSi

O

O

Si OO

HO

R

R

R

R

R

a6b2

(α) (β)

O

OHOO

SiO

Si

OO

OSiSi

OSiOSi

O

Si

R

RR

R R

RR

O

OHOO

SiO

Si

OO

OSiSi

OSiOSi

O

Si

R

RR

R R

RR

O

OH

OO

SiO

Si

OO

OSiSi

OSiOSi

O

Si

R

RR

R R

R

R

O

OH

OO

SiO

Si

OO

OSiSi

OSiOSi

O

Si

R

RR

R R

R

Ra7b1

O

OO

SiO

Si

OO

SiSiO O

Si

O

Si

R

R

RR

R

R

a6b0

O

OO

SiO

Si

OO

SiSiO O

Si

O

Si

R

R

RR

R

R

a6b0

OH

RO

OSi

O

Si

Si

O

SiSi

O

O

Si OO

HO

R

R

R

R

R

a6b2

OH

RO

OSi

O

Si

Si

O

SiSi

O

O

Si OO

HO

R

R

R

R

R

OH

RO

OSi

O

Si

Si

O

SiSi

O

O

Si OO

HO

R

R

R

R

R

a6b2

(α) (β)


95

(c-C6H11)7Si7O10OH (a7b1, 29%, see Figure 5.11) and (c-C6H11)7Si7O9(OH)3 (a7b3,

7%, see Figure 5.1). Silsesquioxane a7b3 can be separated by precipitation from a THF

solution of the incompletely condensed silsesquioxane mixture by the slow addition of

acetonitrile, in which it has a very low solubility. Silsesquioxane a6b2 is, on the other

hand, the compound with the highest solubility in acetonitrile among those present in

the mixture. Therefore, it can be separated by slowly adding more acetonitrile to the

THF/acetonitrile solution, thus inducing the precipitation of all the other silsesquioxane

structures. The solid obtained after removing silsesquioxanes a6b2 and a7b3 still

contained a mixture of incompletely condensed silsesquioxanes that have very similar

solubilities in many organic solvents and, therefore, resisted all attempts of further

purification. The main component of this mixture was assigned to an a7b1

silsesquioxane structure (two isomers are possible for this silsesquioxane, see

Figure 5.11).

The assignments were confirmed by silylation of the silanol groups of the

incompletely silsesquioxanes by means of reaction with trimethylchlorosilane

(CH3)3SiCl.24 The mixture of incompletely condensed structures obtained after

separation of silsesquioxane a6b0 was reacted with an excess of (CH3)3SiCl. The MS

analysis of the products of the silylation reaction clearly showed the presence of four

main species: disilylated a6b2, monosilylated a7b1, disilylated a7b3 and disilylated

a8b2 (Figure 5.12).

Figure 5.12. APCI MS analysis of the incompletely condensed cyclohexyl silsesquioxanes,

after silylation with (CH3)3SiCl.

1027.25

1117.32 1243.33

973.37

1027.25

1117.32 1243.33

973.37

Chapter 5 __________________________________________________________________________________________________________

96

The monosilylated a7b1 was the species present in the highest concentration. On the

basis of these experiments, the assignment of a7b1 as the main incompletely condensed

species appears unequivocal. However, the 29Si NMR peaks belonging to this

silsesquioxane a7b1, both silylated or not, did not correspond to those of the analogous

silsesquioxane reported in the literature.5,VIII The peaks corresponding to the silylated

a7b1 reported by Feher were present in the 29Si NMR spectrum just as the second main

species. This suggests that the two isomers of silsesquioxane a7b1 (Figure 5.11) were

formed with the synthetic method reported here, with the one known from the literature

being the less abundant (α:β ≈ 1:2). The two isomers differ only in the spatial

orientation of the silanol group. They are both possible products of the hydrolytic

condensation of cyclohexyltrichlorosilane, while only the α-isomer of silsesquioxane

a7b1 is obtained by the dehydration of silsesquioxane a7b3 reported by Feher.5

After 18 hours of hydrolytic condensation under reflux and removal of the

precipitate, the reaction was left to continue at room temperature for 20 weeks: more

white precipitate was formed (additional 10% yield).I(page 79) The crude product was a

mixture of three silsesquioxane species, with some impurities, as determined by 29Si NMR: a6b0 (21%),VII(page 94) a6b2 (27%) and a7b3 (51%). A similar increase of the

ratio of silsesquioxane a7b3 among the products as the reaction time becomes longer

had already been reported by Feher et al..5

The synthesis procedure described here is the first reported method to obtain

silsesquioxanes a6b2 and a7b1 (both isomers) via the hydrolytic condensation of

cyclohexyltrichlorosilane. Cyclohexyl silsesquioxane a6b2 had previously been VIII Peaks corresponding exactly to the 29Si{1H} NMR spectrum of the α-isomer of a7b1 reported by

Feher5 (assumed to have been measured in C6D6 although reported to be in CDCl3) were found as the

second main species in the 29Si{1H} NMR spectrum of the mixture of incompletely condensed

silsesquioxanes measured in C6D6 [δ -55.12, -57.05, -57.15, -66.06, -66.40 (1:1:2:1:2)]. In the 29Si{1H} NMR spectra measured in THF or CDCl3, only the two most intense peaks due to the α-isomer

of a7b1 could be assigned [in THF, δ -59.47, -68.45; in CDCl3, δ -57.76, -67.12], whilst the three less

intense peaks could not be unambiguously assigned among the remaining peaks of the silsesquioxane

mixture.


97

synthesised via the acid-mediated cleavage and rearrangement of silsesquioxane a6b0.25

The α-isomer of cyclohexyl silsesquioxane a7b1 was formerly obtained by dehydration

of silsesquioxane a7b3.5 The synthesis also afforded cyclohexyl silsesquioxane a7b3 in

significant yields, although its separation from the rest of the mixture is not

straightforward. This new method presents a number of differences from that reported

by Feher et al. for the hydrolytic condensation of cyclohexyltrichlorosilane,5 namely:

the solvent acetonitrile, the lower initial trichlorosilane concentration and the higher

solvent-to-water ratio. Additionally, performing the reaction under reflux caused a

noticeable increase in precipitate yield in this synthesis, while the rate of silsesquioxane

formation could not be improved for the reaction reported in literature. Comparing the

two synthetic methods, it is evident that the formation of silsesquioxanes is much faster

with the method reported here, albeit less selective. Such an increase of the rate of

formation of silsesquioxanes suggests that the conditions for the precipitation of

silsesquioxanes are met more readily and for a larger number of different structures.

This fact points to the relevance of using acetonitrile as a solvent, similarly to what is

observed for the synthesis of cyclopentyl silsesquioxanes (see Paragraph 5.2.1).

Analogously, after 18 hours of reaction under reflux, the hydrolysis of acetonitrile and

the consequent formation of NH4Cl caused a liquid-liquid phase separation of the

reaction mixture. In order to further study the effect of the presence of NH4Cl, two

additional experiments were performed:

1) Some NH4Cl (0.073M) was added to the starting synthesis mixture, leading, after

18 hours of reaction under reflux, to a silsesquioxane precipitate with a yield of 82%.I(page 79) 29Si NMR analysis showed that the ratio of a6b0 increased compared to the other

(incompletely condensed) silsesquioxanes. Thus, the addition of NH4Cl increased the

yield but also drove the reaction towards completely condensed species, confirming the

influence of the preferential solvation of NH4Cl by water on the silsesquioxane

synthesis.

2) The hydrolytic condensation of cyclohexyltrichlorosilane was performed in acetone

(the solvent used by Brown4 and Feher5) in presence of NH4Cl (0.15M) as additive:

after 18 hours of reaction under reflux no precipitate was formed (as in the case with no

NH4Cl). This indicates that the preferential solvation of NH4Cl by water alone is not

sufficient to accelerate the formation of silsesquioxanes.

Chapter 5 __________________________________________________________________________________________________________

98

The lower selectivity of the synthesis of cyclohexyl silsesquioxanes compared to that of

cyclopentyl silsesquioxanes seems to be connected to the nature of the organic group.1

The larger steric hindrance of cyclohexyl might account for the formation of the

completely condensed silsesquioxane a6b0. The Si-O framework of this molecule is

more strained than in the case of silsesquioxane a7b3, therefore its formation might be

expected to be unfavourable. On the other hand, in this structure the organic groups on

the silicon atoms point away from each other: this may favour the formation of this

molecule for bulkier organic groups like cyclohexyl. Another effect of the organic

group is to influence the solubility of the silsesquioxanes.1 Cyclohexyl

silsesquioxane a7b3 has been reported to be more soluble in organic solvents than

cyclopentyl silsesquioxane a7b3.5,6 The low solubility of the cyclopentyl structure can

account for its selective formation as precipitate, in agreement with what discussed in

Paragraph 5.2.1.1.

5.3 Conclusions

New methods for the synthesis of cyclopentyl and cyclohexyl silsesquioxanes

have been developed. In both cases, acetonitrile plays a strong role in accelerating the

formation of a silsesquioxane precipitate. The influence of the hydrolysis of acetonitrile

on the synthesis has been ascertained.

In the case of cyclopentyl silsesquioxanes, the synthesis is very selective towards the

formation of structure a7b3, which can be obtained in 64% yield after 18 hours of

reaction. This method provides a fast and high-yield way to prepare silsesquioxane

a7b3, therefore making this compound more readily available for its many applications.

By coupling MS and ATR FTIR analytical techniques, it was possible to obtain detailed

information on the time-resolved mechanism of silsesquioxane formation, with the ATR

FTIR results being consistent with the mechanism based on the interpretation of the MS

data. The study has shown that the formation of the desired silsesquioxane a7b3 follows

a complex pathway, via initial formation of short-lived smaller oligomers. The final,

high selectivity for the a7b3 species is ascribed to its lower solubility, which impedes

re-dissolution and equilibrium reactions yielding other silsesquioxanes.


99

In the case of the synthesis of cyclohexyl silsesquioxanes, four silsesquioxane structures

are produced in significant yields: a6b0, a6b2, a7b1 (both isomers) and a7b3. This

method offers a facile way to synthesise the non-easily accessible silsesquioxane

structures a6b2 and a7b1, though their purification proved to be difficult. The lower

selectivity of this reaction seems to be connected to the higher solubility of cyclohexyl

silsesquioxanes as compared to that of cyclopentyl silsesquioxanes.

5.4 Experimental

29Si NMR characterisation of the silsesquioxanes was performed on a Varian

VXR-400S (79.5 MHz, 1H decoupled, 25°C) and on a Varian Inova-300 (59.6 MHz, 1H

decoupled, 25°C). The position of the silsesquioxanes peaks in the 29Si NMR spectra is

influenced by the solvent in which the silsesquioxanes are dissolved. Spectra were

recorded in THF and CDCl3, with TMS as a reference.

Mass spectrometry characterisation of the silsesquioxanes was performed on a

Micromass Quattro LC-MS with ESI+ (electrospray ionisation) and APCI+

(atmospheric pressure chemical ionisation) as ionisation techniques. Mass spectrometry

analysis allows identifying the number of siloxane units constituting the silsesquioxane

structures (i.e. the value of a) but not their level of condensation (i.e. the value of b).

The reason for this is that during the ionisation process silsesquioxanes can lose water

molecules, as confirmed by the fact that the ratio of peak intensities for a given value of

a is influenced by the cone voltage applied during the MS analysis.

In-situ infrared analysis was performed by means of an ASI ReactIR™1000 Reaction

Analysis System, based on Attenuated Total Reflection (ATR) Fourier Transform

Infrared (FTIR) spectroscopy. The spectrometer was equipped with a zinc selenide

(ZnSe) crystal cell, a diamond probe and a mercury cadmium telluride (MCT) detector.

During the analysis, the IR diamond probe had to be positioned just below the surface

of the reaction mixture in order to avoid coating of the probe by the silsesquioxane

precipitate formed during the synthesis. The reaction was carried out under reflux

conditions, which led to bubble formation at the surface of the solution. This affected

Chapter 5 __________________________________________________________________________________________________________

100

the contact between the probe surface and the reaction mixture, causing some noise in

the recorded spectra. Noisy or spiky data may also be caused by impact on the ATR

probe of the precipitate formed during the reaction.

5.4.1 Synthesis of cyclopentyl silsesquioxane a7b3

Typically, 37.5 ml of deionised water were carefully added into a 500-ml

round-bottom flask containing a solution of 5.6 ml of cyclopentyltrichlorosilane

((c-C5H9)SiCl3, 97% purity, 3.3·10-2 moles) in 150 ml of acetonitrile

(CH3CN:H2O = 4:1 in volume). This homogeneous solution was then heated to 50°C or

under reflux while vigorously stirring for 18 hours: a white precipitate was formed. The

crude precipitate was dissolved in 20 times its weight of pyridine and stirred overnight.

Next, the solution was filtered to remove the insoluble completely condensed species

(traces) and poured into an equal volume of ice-cold aqueous HCl (37%): a white

precipitate formed. The solid was filtered, carefully washed with water and dried

overnight at 45°C: NMR-pure cyclopentyl silsesquioxane a7b3 was obtained.

Selected spectroscopic data. 29Si{1H} NMR (THF as solvent), δ -57.83, -65.71, -67.24 (3:1:3).

MS (ESI+, cone voltage = 60 V, THF 78%, CH3CN 20%, HCO2H 0.1M 2%), m/z:

875.04 ([a7b3 + H]+, 100%), 839.00 ([a7b3 - 2H2O + H]+, 41%), 897.08 ([a7b3 + Na]+,

16%), 911.09 ([a7b3 + 2H2O + H]+, 16%), 857.03 ([a7b3 - H2O + H]+, 11%).

For the mechanistic study by mass spectrometry and in-situ ATR FT-IR spectroscopy,

the synthesis of cyclopentyl silsesquioxane a7b3 was performed on a smaller-volume

scale: 9.4 ml of deionised water were carefully added to a 100-ml round-bottomed flask

containing a solution of 1.4 ml of cyclopentyltrichlorosilane ((c-C5H9)SiCl3, 97%

purity) in 37.5 ml of acetonitrile. The homogeneous solution was vigorously stirred and

heated under reflux conditions in an oil bath set at 90°C. A white precipitate was

already visible after 1 hour. At the end of the reaction, the white precipitate (mainly

composed of silsesquioxane a7b3) was present on the walls of the flask and a

liquid-liquid phase separation of the reaction solution was observed.

For the mechanistic study by means mass spectrometry, ESI+ was used as ionisation

technique. At regular intervals throughout the experiment 200-µl samples were taken


101

from the reaction supernatant liquid, ensuring no aspiration of solids into the syringe,

and directly injected into the mass spectrometer by means of a syringe pump. The first

sample (t = 0) was taken straight after adding water to the acetonitrile solution of

cyclopentyltrichlorosilane, and subsequent samples were taken after 30, 60, 90, 120,

180, 270, 360, 480, 600, 1440 minutes. The presence of water and acids (hydrochloric

and acetic) in the reaction mixture yielded favourable ionisation conditions for MS.

A list of possible cyclopentyl silsesquioxane structures and their masses, based on the

equations that define the a and b values in the formula (RSiO1.5)a(H2O)0.5b (see

Paragraph 2.1), is shown in Figure 5.3, together with a MS plot of the silsesquioxane

synthesis mixture.

Main MS peaks for the different values of a: For a = 1, m/z: 257.55 (H+, a1b3 + 6H2O),

275.61 (H+, a1b3 + 7H2O), 293.62 (H+, a1b3 + 8H2O), 316.67 (H+, 275.61 + CH3CN).

For a = 2, m/z: 261.59 (H+), 279.59 (H+), 302.65 (H+, 261.59 + CH3CN), 320.65 (H+,

279.59 + CH3CN). For a = 3, m/z: 373.64 (H+, 391.64 - H2O), 391.64 (H+), 414.64 (H+,

373.64 + CH3CN), 432.64 (H+, 391.64 + CH3CN). For a = 4, m/z: 485.57 (H+), 503.57

(H+), 521.57 (H+), 544.56 (H+, 503.57 + CH3CN). For a = 5, m/z:. For a = 5, m/z:

597.49 (H+, 633.49 – 2H2O), 615.49 (H+, 633.49 – H2O), 633.49 (H+), 638.48 (H+,

597.49 + CH3CN), 651.43 (H+). For a = 6, m/z: 727.42 (H+), 745.36 (H+), 763.36 (H+),

781.36 (H+). For a = 7, m/z: 839.28 (H+, 857.28 – H2O), 857.28 (H+), 875.28 (H+),

893.28 (H+), 911.22 (H+). For a = 8, m/z: 969.20 (H+), 987.20 (H+), 1005.19 (H+),

1023.20 (H+), 1041.20 (H+). For a = 9, m/z: 1099.06 (H+), 1117.12 (H+), 1135.19 (H+),

1153.12 (H+), 1171.13 (H+). For a = 10, m/z: 1211.05 (H+), 1229.00 (H+), 1247.05 (H+),

1265.05 (H+), 1282.99 (H+), 1301.00 (H+). For a = 11, m/z: 1340.91 (H+), 1358.91 (H+),

1376.97 (H+), 1394.91 (H+), 1412.91 (H+). For a = 12, m/z: 1470.90 (H+), 1488.90 (H+),

1506.84 (H+), 1524.84 (H+), 1542.84 (H+). For a = 13, m/z:1600.82 (H+), 1618.76 (H+),

1636.76 (H+), 1654.76 (H+).

5.4.2 Synthesis of cyclohexyl silsesquioxanes

Typically, 125 ml of deionised water were carefully added into a 1-l

round-bottom flask containing a solution of 15 ml of cyclohexyltrichlorosilane

((c-C6H11)SiCl3, 97% purity, 8.2·10-2 moles) in 500 ml of acetonitrile (CH3CN:H2O = 4:1

in volume). This homogeneous solution was then heated to 50°C or under reflux while

Chapter 5 __________________________________________________________________________________________________________

102

vigorously stirring for 18 hours.IX The crude precipitate formed was dissolved in 20

times its weight of pyridine and stirred overnight. The solution was then filtered to

remove the insoluble completely condensed species (a6b0) and poured into an equal

volume of ice-cold aqueous HCl (37%) to precipitate the incompletely condensed

silsesquioxanes. The solid mixture of silsesquioxanes was filtered, carefully washed

with water and dried overnight at 45°C. Silsesquioxane a7b3 was separated by

dissolving the mixture of incompletely condensed species in 5 times its weight of

tetrahydrofuran (THF) and then by slowly layering an equal weight of acetonitrile over

the THF solution. Upon these conditions, a white precipitate mainly constituted of

silsesquioxane a7b3 was obtained. The remaining solution was partially dried under

reduced pressure and more acetonitrile was slowly added until a precipitate deposited.

The solid was separated by filtration and the solution dried under reduced pressure to

afford silsesquioxane a6b2.

The silylation of the mixture of incompletely condensed silsesquioxanes was performed

by adding 200 µl of trimethylchlorosilane (CH3)3SiCl to a solution of 0.3 g of the

silsesquioxanes in 30 ml of toluene and 1.5 ml of triethylamine (C2H5)3N. The turbid

solution was stirred overnight at room temperature. Next, the solvent and the

triethylamine were removed under reduced pressure to afford a white solid. The solid

was extracted with pentane, the solution was filtered to remove the insoluble

(C2H5)3N·HCl and dried under reduced pressure to give a solid containing the silylated

silsesquioxanes.

Selected spectroscopic data. 29Si{1H} NMR: a6b0 (in THF), δ -56.72; a6b2 (in THF) δ -57.16, -62.32, -62.57 (2:2:2),

(in CDCl3) δ -55.20, -61.72, -61.92 (2:2:2); a7b1 (in THF) δ -56.18, -58.66, -58.87,

-66.76, -67.93 (1:2:1:1:2), (in CDCl3) δ -56.23, -57.95, -58.58, -66.82, -67.66 (1:1:2:1:2);

silylated a7b1 isomers (in CDCl3): (α) δ 9.06 ((CH3)3Si), -56.32, -59.02, -66.67, -68.74,

-68.83 (1:2:1:1:2, (c-C6H11)Si) and (β) δ 8.20 ((CH3)3Si), -56.28, -58.70, -66.89, -67.35,

-68.14 (1:2:1:1:2, (c-C6H11)Si); a7b3 (in THF) δ -59.79, -67.86, -69.33 (3:1:3).

IX The synthesis under reflux was performed both on a 100-ml scale and on a 1000-ml scale, yielding to

the formation of the same silsesquioxane mixture.


103

MS: crude precipitate (APCI+, cone voltage = 15 V, CH2Cl2, see Figure 5.10), for a = 6,

m/z: 811.36 ([a6b0 + H]+, 100%), 829.43 ([a6b2 + H]+, 20%); for a = 7, m/z: 937.12

([a7b1 – H2O + H]+, 14%), 955.18 ([a7b1 + H]+, 38%), 973.19 ([a7b3 + H]+, 23%); for

a = 8, m/z: 1081.07 ([a8b0 + H]+, 8%), 1099.07 ([a8b2 + H]+, 7%); silylated

incompletely condensed silsesquioxanes (APCI+, cone voltage = 15 V,

toluene/acetonitrile (2:1), few drops of CH3CO2H; see Figure 5.12), m/z: 973.37

([disilylated a6b2 + H]+, 40%), 1027.25 ([monosilylated a7b1 + H]+, 100%), 1117.32

([disilylated a7b3 + H]+, 22%), 1243.33 ([disilylated a8b2 + H]+, 15%).

Chapter 5 __________________________________________________________________________________________________________

104

References 1 P.P. Pescarmona, T. Maschmeyer, Aust. J. Chem., 2001, 54, 583, see also Chapter 2. 2 P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, Angew. Chem. Int. Ed., 2001, 40,

740, see also Chapter 3. 3 P.P. Pescarmona, T. Maschmeyer, NATO Science Series, 2002, Ser. II Vol. 69, 173, see also Chapter 4. 4 J.F. Brown, L.H. Vogt, J. Am. Chem. Soc., 1965, 87, 4313. 5 F. J. Feher, D. A. Newman, J. F. Walzer, J. Am. Chem. Soc., 1989, 111, 1741. 6 F. J. Feher, T. A. Budzichowski, R. L. Blanski, K. J. Weller, J. W. Ziller, Organometallics, 1991, 10,

2526. 7 See Paragraph 3.3.1. 8 V.K. Krieble, C.I. Noll, J. Am. Chem. Soc., 1939, 61, 560. 9 A.J. Belsky, T.B. Brill, J. Phys. Chem. A, 1999, 103, 3006. 10 T. Kudo, M.S. Gordon, J. Am. Chem. Soc., 1998, 120, 11432. 11 T. Kudo, M.S. Gordon, J. Phys. Chem. A, 2000, 104, 4058. 12 M.G. Voronkov, V.I. Lavrent’yev, Top. Curr. Chem., 1982, 102, 199. 13 P. Bussian, F. Sobott, B. Brutschy, W. Schrader, F. Schüth, Angew. Chem. Int. Ed., 2000, 39, 3901. 14 R. Bakhtiar, F.J. Feher, Rapid Commun. Mass Spectrom., 1999, 13, 687. 15 N.J. Harrick, J. Phys. Chem., 1960, 64, 1110. 16 J. Fahrenfort, Spectrochim. Acta, 1961, 17, 698. 17 J.F. Brown, L.H. Vogt, J. Am. Chem. Soc., 1965, 87, 4313. 18 T.S. Haddad, B.M. Moore, S.H. Phillips, Polymer Preprints, 2001, 42, 196. 19 F.J. Feher, R. Terroba, J.W. Ziller, Chem. Commun., 1999, 2153. 20 N.J. Harrick, Internal Reflection Spectroscopy, John Wiley & Sons, New York, 1967. 21 R. Tauler , B. Kowalski, S. Fleming, Anal. Chem., 1993, 65, 2040. 22 E. Metcalfe, J. Tetteh, Polym. Mater. Sci. Eng., 2000, 83, 88. 23 J.E. Jackson, A users guide to principal components, John Wiley & Sons, New York, 1991. 24 F.J. Feher, D.A. Newman, J. Am. Chem. Soc., 1990, 112, 1931. 25 F. J. Feher, F. Nguyen, D. Soulivong, J. W. Ziller, Chem. Commun., 1999, 1705.

105

6

Osmium silsesquioxane as model compound and homogeneous catalyst for the dihydroxylation of alkenes

Abstract

A homogeneous dihydroxylation catalyst was synthesised by complexation of OsO4

with a silsesquioxane ligand containing a tetrasubstituted olefin moiety. The advantage

of this procedure consists in avoiding the presence of the highly toxic and volatile OsO4

in solution during the dihydroxylation reaction. The Os-silsesquioxane complex was

used as a homogeneous catalyst and as a model of the catalytic site of the heterogeneous

analogue. The homogeneous catalyst gave 84% conversion with 99% selectivity for the

dihydroxylation of cyclopentene and 99% conversion with 99% selectivity for the

dihydroxylation of cyclohexene.


P.P. Pescarmona, A. F. Masters, J.C. van der Waal, T. Maschmeyer, Angew. Chem. Int. Ed.

Chapter 6 __________________________________________________________________________________________________________

106

6.1 Introduction

Osmium tetroxide (OsO4) is generally considered the best catalyst for the

cis-dihydroxylation of double bonds (Figure 6.1).1,2 The disadvantage of using this

compound as homogeneous catalyst lies in its high volatility and toxicity.3 Therefore,

many attempts have been carried out to immobilise osmium tetroxide on different

supports.4 Recently, Jacobs et al. reported an efficient and robust heterogeneous catalyst

obtained by binding OsO4 to a tetrasubstituted olefin covalently linked to a silica

support.5,6 The tetrasubstituted diolate ester which was obtained at one side of the

Os-centre provided a stable connection between the osmium centre and the silica

support, since it does not undergo hydrolysis under the reaction conditions applied. The

remaining osmium coordination site is available for the catalytic reaction. The catalytic

cycle is then reduced to the right part of the scheme in Figure 6.1.

Figure 6.1. Proposed catalytic cycles for the dihydroxylation of alkenes with OsO4 as

the catalyst.

This silica-supported Os-catalyst proved to be active in the dihydroxylation of a number

of alkenes with N-methylmorpholine-N-oxide (NMO) as the oxidant5,7 (NMO may be

R

R

R

R

OHOH

R ROH2

oxidant

oxidant OH2

OH

OH

R

R

OsO

OO

O

R

R

VI

OsO

OO

O

R

R

O

R

R

VI

OsO

OO

O

R

R

OR

R O

VIIIR

OsO

OO

O

RO

VIII

OsO

OO

O

VIII

R

R

R

R

OHOH

R ROH2

oxidant

oxidant OH2

OH

OH

R

R

OsO

OO

O

R

R

OsO

OO

O

R

R

VI

OsO

OO

O

R

R

O

R

ROs

O

OO

O

R

R

O

R

R

VI

OsO

OO

O

R

R

OR

R O

OsO

OO

O

R

R

OR

R O

VIIIR

OsO

OO

O

RO

VIIIR

OsO

OO

O

RO

VIII

OsO

OO

O OsO

OO

O

VIII

Os-silsesquioxane as dihydroxylation catalyst __________________________________________________________________________________________________________

107

regenerated by oxidation with H2O2, see Figure 6.2).6 Careful heterogeneity tests

confirmed the absence of OsO4 leaching. Given the relevance of these results, the

preparation, characterisation and catalytic test of a silsesquioxane8-based homogeneous

analogue of this heterogeneous silica-supported catalyst becomes interesting in order to

establish the surface chemistry unequivocally and to broaden the scope for

immobilisation (e.g. silsesquioxane complexes of titanium have been successfully

immobilised on MCM-41, yielding highly active, non-leaching heterogeneous

catalysts).9-11

Figure 6.2. NMO regeneration by oxidation of N-methylmorpholine (NMM) with H2O2,

using vanadyl acetylacetonate as peroxide activator.


The synthesis of the silsesquioxane ligand for the Os-catalyst was performed

analogously to that of the functionalised silica support described by Severeyns et al..5

Cyclopentyl silsesquioxane a7b3 (1)12 was used as the precursor for the synthesis of

silsesquioxane 3, which contains a tetrasubstituted olefin moiety to which OsO4 can be

anchored (Figure 6.3). In the first step, silsesquioxane 1 was reacted with

3-aminopropyltrimethoxysilane to obtain the amino-functionalised silsesquioxane 2 as

the only product (as assessed by means of ESI MS). In the second step, the amino group

of silsesquioxane 2 was reacted with 3,4-dimethylcyclohex-3-enylcarbonyl chloride, C,

to produce silsesquioxane 3. The synthesis was performed under a N2 flow and in the

presence of a stoichiometric amount of triethylamine to remove HCl formed during the

reaction. The formation of silsesquioxane 3 was confirmed by means of ESI MS and 13C NMR analysis.

The acid chloride, C, was prepared by means of a three-steps synthesis (Figure 6.4):

1) Diels-Alder reaction of 2,3-dimethyl-1,3-butadiene with ethyl acrylate to give ethyl

NMO

NMM

Voxidised

Vreduced

H2O2

H2O

NMO

NMM

Voxidised

Vreduced

H2O2

H2O

Chapter 6 __________________________________________________________________________________________________________

108

3,4-dimethylcyclohex-3-enyl carboxylate, A.13 2) Hydrolysis of the ester, A, to produce

3,4-dimethylcyclohex-3-enyl carboxylic acid, B. 3) Conversion of the carboxylic

acid, B, with SOCl2 to obtain the desired acid chloride, C. The reaction was carried out

under reflux and by flushing with N2 to remove the HCl formed, which would otherwise

attack the double bond in the cyclohexene ring. All the steps of the synthesis of

compound C were monitored by 13C NMR analysis.

Figure 6.3. Procedure for the preparation of the Os-silsesquioxane complex

[RI = cyclopentyl].

Figure 6.4. Procedure for the synthesis of 3,4-dimethylcyclohex-3-enylcarbonyl

chloride.

+

O

O

O

O

O

HO

O

ClSOCl2

N2

H2O

A B C

12

3

45

6

7

8

910

11

1 3

45

6

7

8

9 1 3

45

6

7

8

92 2

+

O

O

O

O

O

HO

O

ClSOCl2

N2

H2O

A B C

12

3

45

6

7

8

910

11

1 3

45

6

7

8

9 1 3

45

6

7

8

92 2

(CH3O)3Si(CH2)3NH2

K2OsO2(OH)4, NMO

Et3N,

N2

O

OO

O

RI

OO

SiO Si

OO

OSiSi

OSiOSi

O

SiNHSi

RI

RI RI

RIRI

RI

O

OOs

O

O

4

OH OH

OH

RI

OO

SiO Si

OO

OSiSi

OSiOSi

O

Si

RI

RI RI

RIRI

RI

1

OO

O

RI

OO

SiO Si

OO

OSiSi

OSiOSi

O

Si

Si

RI

RI RI

RIRI

RI

NH2

2

O

OO

O

RI

OO

SiO Si

OO

OSiSi

OSiOSi

O

SiNHSi

RI

RI RI

RIRI

RI

3

12

3

45

6

7

8

9

10

11

12

O

Cl C

(CH3O)3Si(CH2)3NH2

K2OsO2(OH)4, NMO

Et3N,

N2

O

OO

O

RI

OO

SiO Si

OO

OSiSi

OSiOSi

O

SiNHSi

RI

RI RI

RIRI

RI

O

OOs

O

O

4

O

OO

O

RI

OO

SiO Si

OO

OSiSi

OSiOSi

O

SiNHSi

RI

RI RI

RIRI

RI

O

OOs

O

O

4

OH OH

OH

RI

OO

SiO Si

OO

OSiSi

OSiOSi

O

Si

RI

RI RI

RIRI

RI

1

OH OH

OH

RI

OO

SiO Si

OO

OSiSi

OSiOSi

O

Si

RI

RI RI

RIRI

RI

1

OO

O

RI

OO

SiO Si

OO

OSiSi

OSiOSi

O

Si

Si

RI

RI RI

RIRI

RI

NH2

2

OO

O

RI

OO

SiO Si

OO

OSiSi

OSiOSi

O

Si

Si

RI

RI RI

RIRI

RI

NH2

2

O

OO

O

RI

OO

SiO Si

OO

OSiSi

OSiOSi

O

SiNHSi

RI

RI RI

RIRI

RI

3

12

3

45

6

7

8

9

10

11

12

O

OO

O

RI

OO

SiO Si

OO

OSiSi

OSiOSi

O

SiNHSi

RI

RI RI

RIRI

RI

3

12

3

45

6

7

8

9

10

11

12

O

Cl C

O

Cl C


109

The next step of the synthesis of the catalytic complex is the addition of OsO4 to the

double bond of the functionalised silsesquioxane 3 (in 1:1 molar ratio). To avoid the

risks connected with the handling of OsO4, potassium osmate K2OsO2(OH)4 was used

as osmium source.14 A solution of potassium osmate was oxidised to OsO4 by means of

N-methylmorpholine-N-oxide (NMO). In the case of the synthesis of the Os-silsesquioxane

catalyst, care must be taken to avoid the reaction of the Os-centre with two silsesquioxane

ligands 3, since this would lead to the formation of a stable, inactive bis-chelate complex.

The reaction of compound 3 with OsO4 prepared by oxidation of potassium osmate with

NMO using stoichiometric amounts of the reagents, did not lead to the desired

Os-silsesquioxane 4. It was found that the oxidation of potassium osmate with NMO did

not proceed if a stoichiometric amount of the oxidant was used. It was then inferred that

an excess of NMO should be used to oxidise the potassium osmate (20:1).

In order to prevent the formation of the bis-chelate while using an excess of NMO,

compound 3 was added to the reaction mixture together with an excess of cyclohexene.

The double bond of the cyclohexene would then compete with that of compound 3 for

binding to the osmium centre. Being less sterically hindered and being in large excess

(19:1) compared to compound 3, cyclohexene will be favoured in reacting with the

osmium centre and will tend to occupy both catalytic sites. Cyclohexene is not a

tetrasubstituted alkene, therefore its bond with the osmium centre is labile and in the

presence of H2O and of NMO it undergoes oxidative hydrolysis, yielding

1,2-cyclohexanediol and restoring the catalytic site on the osmium. This process proceeds

according to the catalytic cycle for the dihydroxylation of alkenes (Figure 6.1). As the

concentration of cyclohexene in solution decreases, it becomes more likely that

compound 3 binds to the osmium centre to produce complex 4 and the

silesquioxane-Os-cyclohexane bisdiolate complex 5 (Figure 6.5). This latter complex

presents an osmium centre bound to one side to the tetrasubstituted olefin of compound 3

and on the other side to cyclohexene. The cyclohexane diolate can be hydrolysed

providing the catalytic site for dihydroxylation reactions, while the tetrasubstituted diolate

is stable under the employed reaction conditions and, therefore, prevents the formation of

free OsO4 during the catalytic process. The synthesis was monitored by means of ESI

MS: after 3 days of reaction, compound 5 is the main product, with small amounts of

compound 4 and of unreacted compound 3 (Figure 6.6). A mixture of H2O, ButOH and

CH2Cl2 was used as solvent during the synthesis: H2O was needed to dissolve the

Chapter 6 __________________________________________________________________________________________________________

110

potassium osmate and as reactant for the dihydroxylation of cyclohexene, CH2Cl2 was

used to dissolve the silsesquioxane ligand 3 and ButOH was used to obtain a monophasic

solution. At the end of the reaction, dichloromethane was removed by evaporation under

reduced pressure and H2O was added, thereby causing the precipitation of the

silsesquioxane-based compounds that were then collected as a grey-brown powder.

Figure 6.5. Os-silsesquioxane complexes 5 and 6 [RI = cyclopentyl].

Figure 6.6. ESI MS analysis of the synthesis of Os-silsesquioxane complexes 4 (m/z: 1418.47)

and 5 (m/z: 1470.60). The isotope pattern of osmium is clearly recognisable.

OO

OOO

SiO Si

OO

OSiSi

O SiOSi

O

Si

O

NO

OOs

O

O

OSi

5RI

RI

RI

RI

RI

RI

RI

OO

OOO

SiO Si

OO

OSiSi

O SiOSi

O

Si

O

NO

OOs

O

O

OSi

6RI

RI

RI

RI

RI

RI

RI

OO

OOO

SiO Si

OO

OSiSi

O SiOSi

O

Si

O

NO

OOs

O

O

OSi

5RI

RI

RI

RI

RI

RI

RIO

O

OOO

SiO Si

OO

OSiSi

O SiOSi

O

Si

O

NO

OOs

O

O

OSi

5RI

RI

RI

RI

RI

RI

RI

OO

OOO

SiO Si

OO

OSiSi

O SiOSi

O

Si

O

NO

OOs

O

O

OSi

6RI

RI

RI

RI

RI

RI

RIO

O

OOO

SiO Si

OO

OSiSi

O SiOSi

O

Si

O

NO

OOs

O

O

OSi

6RI

RI

RI

RI

RI

RI

RI


111

A catalytic test of complex 5 was performed under the same reaction conditions used for

testing the heterogeneous silica-supported analogue of the Os-silsesquioxane complex.5

Two different substrates were used: cyclopentene and cyclohexene. The results of the

catalytic test for the dihydroxylation of these two substrates are reported in Table 6.1.

Conversions and selectivities were determined by GC and GC-MS analysis. The

catalyst was active and very selective in the dihydroxylation of both substrates: after 3

hours reaction a TONcyclopentene of 130 and a TONcyclohexene of 414 were observed. The

conversion was higher for cyclohexene than for cyclopentene, probably because the

complex between the osmium centre and cyclopentene ring is geometrically strained

and its formation is, therefore, less favourable than for the complex with cyclohexene.

The homogeneous Os-silsesquioxane showed similar trends but higher activity than its

silica-supported analogue.5 For the dihydroxylation of cyclopentene, a conversion of

84% was achieved after 24 hours of reaction, while with the silica-supported catalyst a

conversion of 83% was obtained after 48 hours. The difference in performance is even

more evident for cyclohexene, for which total conversion is achieved after 48 hours

with the heterogeneous catalyst and after just 3 hours with the Os-silsesquioxane

catalyst.

Table 6.1. Catalytic test of the Os-silsesquioxane in the dihydroxylation of cyclopentene

and cyclohexene.

After the dihydroxylation of cyclopentene proceeded for 24 hours, the catalyst was

analysed by means of ESI MS: complex 6 (Figure 6.5) was the only Os-complex present

substrate time conversion selectivitycyclopentene 3 h 36% 99%

6 h 50% 99%24 h 84% 99%

cyclohexene 3 h 99% 99%

OH

OH

OH

OH

NMO, H2O

Os-silsesquioxaneOs-silsesquioxane

NMO, H2O

substrate time conversion selectivitycyclopentene 3 h 36% 99%

6 h 50% 99%24 h 84% 99%

cyclohexene 3 h 99% 99%

OH

OH

OH

OH

NMO, H2O


NMO, H2O OH

OH

OH

OH

NMO, H2O


NMO, H2O

Chapter 6 __________________________________________________________________________________________________________

112

in solution. This finding is in agreement with the proposed mechanism for the

dihydroxylation reaction: the cyclohexane diolate of complex 5 was hydrolysed

restoring the catalytic site, which was then active in the dihydroxylation of

cyclopentene. The silsesquioxane ligand, bound to the osmium centre via a

tetrasubstituted diolate, did not undergo hydrolysis and, therefore, prevented formation

of the undesired OsO4 in solution.

6.3 Conclusions

An Os-silsesquioxane complex was synthesised, characterised and tested for

activity in dihydroxylation reactions. The complex provided a model for a

silica-supported Os-catalyst. Characterisation of the homogeneous Os-silsesquioxane

confirmed the proposed nature of the heterogeneous catalytic site, where the osmium

centre is on one face bound in a stable fashion to the silsesquioxane ligand and on the

other face is available for catalysing the dihydroxylation of olefins. The

dihydroxylations of cyclopentene and cyclohexene were used as test reactions. In both

cases, the homogeneous catalyst displayed higher turnover frequencies than the

heterogeneous one, while keeping the same selectivity (99%).

6.4 Experimental

13C NMR spectra were measured at 25°C in CDCl3 as solvent on a Varian

Inova-300 (75.5 MHz, 1H decoupled) and on a Bruker DPX-300 (75.5 MHz, 1H

decoupled).

Mass spectrometry analysis was performed on a Micromass Quattro LC-MS with ESI as

ionisation technique and on a ThermoQuest Finnigan LCQ Deca equipped with ESI

source and controlled by Xcalibur software. The desolvation temperature was set to

300°C, the cone voltage to 20 V. The samples had a concentration of ~0.01 mg of

silsesquioxane-based compound diluted in 1 ml of methanol (or of a

methanol/dichloromethane 1:1 mixture) with addition of a drop of acetic acid if

necessary.


113

GC analysis was performed on a Hewlett-Packard 5890 equipped with a split/splitless

capillary injector, FID detector and a Phenomenex Zebron ZB-5 column (30 m). In the

temperature program, the temperature was set at 30°C for 1 minute, and then it was

raised to 250°C with a rate of 10°C/min and hold at that temperature for 5 minutes.

GC-MS analysis was performed on a Finnigan Polaris Q ion trap mass spectrometer

with a Trace GC. The GC was equipped with a HP1-MS column (25 m). The GC

method started with a temperature of 40°C for 2 minutes, then the temperature was

raised to 200°C at a rate of 10°C/min and hold for 1 minute. The ionisation technique

for MS analysis was EI (70 eV).

Synthesis of compound 2. 2.0·10-3 moles of cyclopentyl silsesquioxane a7b3 1 were

dissolved in 70 ml of tetrahydrofuran (THF) and 2.2·10-3 moles of

3-aminopropyltrimethoxysilane, (CH3O)3Si(CH2)3NH2, were dissolved in 50 ml of

THF. The two solutions were mixed and allowed to react for two days at room

temperature while stirring. The solvent was then removed under reduced pressure and

the residue was washed with acetonitrile to remove any 3-aminopropyltrimethoxysilane

that might still be present. The residue, in the form of a white gel, was redissolved in

dichloromethane and dried again under reduced pressure yielding compound 2 as a

white solid (1.484g, corresponding to a yield of 78%). MS data, m/z: 958.22

(compound 2 + H+, 100%), other peaks belonging to impurities with intensities < 3%.

Synthesis of compound B. 0.15 moles of 3,4-dimethylcyclohex-3-enyl ethyl

carboxylate A were added to 250 ml of 10%wt NaOH solution in H2O. ~0.30g of

tetrabutylammonium bromide were added to the biphasic system to favour mixing

between the ester A and water. The reaction mixture was stirred overnight under reflux

to give a monophasic solution. The ethanol formed in the hydrolysis of compound A

was removed from the solution under reduced pressure. H2SO4 1N was added to the

solution until an acidic pH was reached: the carboxylic acid B precipitated as a white

solid that was then dried overnight under reduced pressure. Total conversion of

compound A to compound B was achieved. 13C{1H} NMR data: for compound A,

δ 14.32 (11), 18.87 (7), 19.01 (8), 25.96 (6), 31.12 (5), 33.84 (2), 40.33 (1), 60.20 (10),

124.02 (4), 125.29 (3), 176.00 (9); for compound B, δ 19.05 (7), 19.17 (8), 25.80 (6),

31.05 (5), 33.64 (2), 40.14 (1), 123.94 (4), 125.59 (3), 182.04 (9).

Chapter 6 __________________________________________________________________________________________________________

114

Synthesis of compound C. 4.0·10-2 moles of thionyl chloride (SOCl2) were added to

1.6·10-2 moles of the carboxylic acid B and the solution obtained was stirred for 3 hours

and 30 minutes under reflux, while the surface of the reaction mixture was flushed with

N2. Gaseous HCl leaving the reaction mixture through the reflux cooler was redirected

into an aqueous solution of NaOH. The reaction mixture was dried under reduced

pressure to remove the thionyl chloride still present. The remaining liquid contained the

acid chloride C (~70%) together with some carboxylic acid B (~30%), as determined on

the basis 13C NMR analysis. The sample was stored under N2. 13C{1H} NMR data for

compound C, δ 19.04 (7), 19.33 (8), 25.38 (6), 30.74 (5), 33.15 (2), 41.31 (1), 123.51

(4), 125.58 (3), 171.81 (9).

Synthesis of compound 3. 1.6·10-3 moles of compound 2 were dissolved in a solution of

1.6·10-3 moles of triethylamine in 30 ml of dichloromethane. Subsequently, the solution

was added to the mixture of acid chloride C and carboxylic acid B and stirred overnight

at room temperature under N2 flow. The solution was then concentrated to a volume of

5 ml under reduced pressure and 60 ml of acetonitrile were added: a white solid

precipitated. The solid was washed with 0.8%wt NaOH(aq) to remove B and C that might

still be present. The yield in compound 3 was ~68%. MS data, m/z: 1094.15

(compound 3 + H+, 13%), 1116.11 (compound 3 + Na+, 100%), 1132.00 (compound 3 +

K+, 8%). 13C{1H} NMR data for compound 3, δ 9.36 (12), 18.84 (7), 18.98 (8), 23.26

(11) 26.56 (6), 31.14 (5), 34.54 (2), 41.47 (1), 42.33 (10), 124.09 (4), 125.48 (3), 175.76

(9); δ 22.23 (ipso carbons of the cyclopentyl groups on the silsesquioxane), 27.03, 27.30

(multiplets; remaining carbons of the cyclopentyl groups on the silsesquioxane).

Synthesis of complexes 4 and 5. 4.0·10-5 moles of potassium osmate, K2OsO2(OH)4,

were dissolved in 2 ml of H2O to give a dark-red solution. 8.0·10-4 moles of

N-methylmorpholine-N-oxide (NMO) were dissolved in 0.5 ml of H2O. The two

aqueous solutions were mixed together with 4 ml of tert-butanol and the obtained pink

solution was stirred overnight at room temperature. Upon addition of 0.1 ml of a 0.42M

aqueous solution of acetic acid the solution turned colourless. Next, 0.044 g (~4·10-5

moles) of compound 3 and 7.6·10-4 moles of cyclohexene, previously dissolved in 5 ml

of tert-butanol and 3 ml of dichloromethane, were added to the Os-containing solution:

the solution immediately turned dark brown, indicating the formation of the Os-diolate.

After stirring the reaction mixture for 3 days at room temperature, a sample was taken

and analysed by MS. Then, dichloromethane was removed under reduced pressure from


115

the homogeneous dark brown solution. Next, 15 ml of H2O were added to the slightly

turbid solution and a precipitate was formed. The solution was filtered on a folded paper

filter and the grey-brown residue was collected (0.028 g, corresponding to ~1.7·10-5

moles and, therefore, to a yield of ~42% in Os-silsesquioxane complexes). MS data,

m/z: 1470.60 (complex 5 + Na+, 100%), 1418.47 (complex 4 + 3Na+, 18%), 1116.67

(compound 3 + Na+, 21%), 1094.67 (compound 3 + H+, 8%).

Catalytic test. 0.014 g of the grey-brown solid (~8.4·10-6 moles of Os-silsesquioxane

complexes) were dissolved in 2.5 ml of dichloromethane and 5 ml of tert-butanol. To

this solution, 4.0·10-3 moles of alkene (cyclopentene or cyclohexene), 2.0·10-3 moles of

n-nonane (internal standard for GC analysis), 4.0·10-3 moles of NMO and 500 µl of H2O

were added. The yellow-brown solution was stirred at room temperature for 24 hours.

Samples for GC and GC-MS analysis were taken after 3, 6 and 24 hours of reaction.

MS analysis of the catalyst after the dihydroxylation of cyclopentene, m/z: 1456.53

(complex 6 + Na+).

Safety precautions: all handling of osmium compounds was performed in a fume hood,

while wearing a class P2 (particulate) respirator. The reactions in which OsO4 was

formed were carried out in sealed flasks or, if any gas was leaving the synthesis flask, it

was led through a bottle containing oil, rich in unsaturated fatty acids, in order to trap

any OsO4 that might have been developed from the reaction mixture.

Chapter 6 __________________________________________________________________________________________________________

116

References

1 M. Schroeder, Chem. Rev., 1980, 80, 187. 2 H.C. Kolb, M.S. VanNieuwenhze, K.B. Sharpless, Chem. Rev., 1994, 94, 2483. 3 R.J. Lewis, Sax’s Dangerous Properties of Industrial Materials, Wiley-Interscience, 10th edition, 2000. 4 A. Severeyns, D.E. De Vos, P.A. Jacobs, Top. Catal., 2002, 19, 125. 5 A. Severeyns, D.E. De Vos, L. Fiermans, F. Verpoort, P.J. Grobet, P.A. Jacobs, Angew. Chem. Int. Ed.,

2001, 40, 586. 6 A. Severeyns, D.E. De Vos, P.A. Jacobs, Green Chem., 2002, 4, 380. 7 V. VanRheenen, R.C. Kelly, D.Y. Cha, Tetrahedron Lett., 1976, 23, 1973. 8 P.P. Pescarmona, T. Maschmeyer, Aust. J. Chem., 2001, 54, 583, see also Chapter 2. 9 S. Krijnen, H.C.L. Abbenhuis, R.W.J.M. Hanssen, J.H.C. van Hooff, R.A. van Santen, Angew. Chem.

Int. Ed., 1998, 37, 356. 10 S. Krijnen, B.L. Mojet, H.C.L. Abbenhuis, J.H.C. van Hooff, R.A. van Santen, Phys. Chem. Chem.

Phys., 1999, 1, 361. 11 P. Smet, J. Riondato, T. Pauwels, L. Moens, L. Verdonck, Inorg. Chem. Commun., 2000, 3, 557. 12 See Chapter 5. 13 J. Monnin, Helv. Chim. Acta, 1958, 225, 2112. 14 W.D. Lloyd, B.J. Navarette, M.F. Shaw, Synthesis, 1972, 610.

117

7

New tert-butyl and phenyl silsesquioxane precursors for epoxidation titanium catalysts discovered by means of High-Speed Experimentation

Abstract

A set of new titanium-silsesquioxane epoxidation catalysts was discovered by exploring

the hydrolytic condensation of a series of trichlorosilanes in highly-polar solvents by

means of High-Speed Experimentation techniques. The three most promising

silsesquioxane leads were resynthesised in a larger-volume scale and characterised. The

lead obtained from the hydrolytic condensation of ButSiCl3 in water consisted of a

single silsesquioxane structure: But2Si2O(OH)4. This is the first reported example of the

use of this silsesquioxane as precursor for active titanium catalysts. The Ti-complexes

prepared with But2Si2O(OH)4 were supported on silica to produce an active

heterogeneous epoxidation catalyst. The lead generated by the hydrolytic condensation

of ButSiCl3 in DMSO consisted of a set of incompletely condensed silsesquioxane

structures. The lead obtained by the hydrolytic condensation of PhSiCl3 in water was

constituted of polymeric silsesquioxanes.


P.P. Pescarmona, J.C. van der Waal, T. Maschmeyer, Chem. Eur. J.

Chapter 7 __________________________________________________________________________________________________________

118

7.1 Introduction

Chapter 3 and 4 reported the successful application of High-Speed

Experimentation (HSE) techniques to the optimisation of silsesquioxane precursors for

titanium catalysts active in the epoxidation of alkenes.1,2 This work resulted in a new

and efficient way to synthesise silsesquioxane precursors and generated knowledge

about the effect of the various parameters involved in the synthesis. Particularly, one of

the observed trends indicated a favourable effect induced by high-polarity solvents. This

result stimulated further investigation: here, the application of High-Speed

Experimentation techniques to the optimisation of the synthesis of silsesquioxanes in

very polar solvents is presented. The experimental approach is similar to that used in

previous studies1,2 and it is thoroughly described in Paragraph 3.2.


7.2.1 The High-Speed Experimentation screening

The synthesis of silsesquioxane precursors for titanium catalysts via the

hydrolytic condensation of 6 different trichlorosilanes (R = cyclohexyl, cyclopentyl,

phenyl, methyl, ethyl and tert-butyl) in 3 highly-polar solvents (dimethyl sulfoxide

(DMSO), water and formamide) was performed by means of High-Speed

Experimentation techniques. Previous work showed that the R-group on the

organosilane and the solvent in which the hydrolytic condensation is carried out are the

most relevant parameters influencing the synthesis of silsesquioxanes.3 The parameter

space studied was defined by the full-factorial combination of the 6 trichlorosilanes and

the 3 solvents. This parameter space was screened as a function of the activity in the

epoxidation of 1-octene with tert-butyl hydroperoxide (TBHP)4 displayed by the

catalysts which were obtained after coordination of Ti(OBu)4 to the silsesquioxane

structures. In Figure 7.1, the relative activities of the titanium silsesquioxanes are shown

together with those of the titanium silsesquioxanes obtained from silsesquioxanes

synthesised in acetonitrile, which gave the best results among the solvents tested in

previous work.1,2 The values are normalised to the activity of the most active titanium

New Ti-silsesquioxane catalysts discovered using HSE __________________________________________________________________________________________________________

119

silsesquioxane found so far, i.e. that obtained by reacting Ti(OBu)4 with cyclopentyl

silsesquioxane a7b3 [(c-C5H9)7Si7O12TiOC4H9].4,5

Figure 7.1. Screening of the epoxidation activity of the titanium silsesquioxanes as a

function of the solvent and of the trichlorosilanes used in the synthesis of

the silsesquioxane precursors.

The highest catalytic activities were found for the titanium complexes obtained from

tert-butyl silsesquioxanes synthesised in DMSO and water, with respectively 84% and

74% of the activity of the reference catalyst (c-C5H9)7Si7O12TiOC4H9.I These two

catalysts exhibited almost the same activity as the previous best HSE catalyst (87%)

obtained from cyclopentyl silsesquioxanes synthesised in acetonitrile1,2,6 and are the

first reported examples of tert-butyl silsesquioxanes as precursors for very active I With the experimental conditions employed, (c-C5H9)7Si7O12TiOC4H9 gives complete and selective

conversion of TBHP towards the epoxide: therefore, the relative activities of the reported catalysts

correspond to their TBHP conversions towards 1,2-epoxyoctane.




methyltrichlorosilane

ethyltrichlorosilane

tert-butyltrichlorosilane

formamide

H2O

DMSO

acetonitrile

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

relativeactivity




methyltrichlorosilane

ethyltrichlorosilane

tert-butyltrichlorosilane

formamide

H2O

DMSO

acetonitrile

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

relativeactivity

Chapter 7 __________________________________________________________________________________________________________

120

titanium catalysts. Relevant catalytic activities were also obtained with cyclohexyl

silsesquioxanes synthesised in DMSO (67%) and with phenyl silsesquioxanes

synthesised in H2O (61%).

Besides the identification of these leads, the High-Speed Experimentation screening

provided some more information about the system under study. No regular trend can be

identified, i.e. for each solvent the order of activity as a function of the R-groups is

different. This can be explained on the basis of the different nature of the solvents

employed. Water acts both as solvent and reagent, and the fact that silsesquioxanes with

only a very low level of condensation are soluble in water7 causes the fast precipitation

of condensed structures from solution. Formamide has a high boiling point (220°C),

requiring high temperatures to remove the solvent from the silsesquioxane products. At

these high temperatures, the silsesquioxanes tend to condense further to produce

completely condensed structures, which are not suitable for binding metal centres.8 This

can explain the weak activity observed for the catalysts obtained from silsesquioxanes

synthesised in formamide.

Finally, all the methyl and ethyl silsesquioxanes are poor precursor for titanium-based

epoxidation catalysts, in agreement with what found in previous work.1,2

Although High-Speed Experimentation techniques proved to be a powerful tool

to identify leads and gain knowledge about the system under study, they did not provide

any information about the actual silsesquioxane structures that were formed.6 To gain

such information, selected HSE leads can be synthesised on a conventional laboratory

scale and characterised by appropriate analytical techniques (e.g. spectroscopy and

chromatography). Here, the phenyl silsesquioxanes synthesised in water and the

tert-butyl silsesquioxanes synthesised in DMSO and water (i.e. the silsesquioxane

precursors that gave more active catalysts than those synthesised in acetonitrile,1,2 see

Figure 7.1) were selected for further investigation.

7.2.2 Phenyl silsesquioxanes synthesised in H2O

The hydrolytic condensation of phenyltrichlorosilane in H2O was performed in a

25-fold up-scaling of the HSE synthesis. The white precipitate obtained from the reaction

was characterised by NMR Spectroscopy. The liquid-phase 13C NMR spectrum presents

two sets of partially overlapping peaks due to the phenyl rings, indicating that two


121

different types of phenyl siloxane units, PhSiO1.5(H2O)0.5x, are present in the silsesquioxane

product. This observation is confirmed by the solid-state 29Si NMR spectrum (Figure 7.2),

which presents two broad bands of comparable intensity, one in the region of Si atoms

exclusively connected to other Si atoms through oxygen bridges (PhSiO1.5, δ -79.4 ppm),

the other in the region of Si atoms connected to an -OH group (PhSiO(OH), δ -69.8 ppm).8,9

The liquid-phase 29Si NMR spectrum does not show any signal due to silsesquioxane

species: this is probably caused by the viscosity of the sample, suggesting that the

product is constituted of polymeric species. On the basis of these data it is proposed that

the product consists of ladder-type structures10 in which around half of the connections

between the ‘steps’ of the ladder are not fully condensed and, therefore, present silanol

groups to which titanium centres can be coordinated (Figure 7.2).

Figure 7.2. Solid-state 29Si NMR spectrum and proposed structure of the phenyl

silsesquioxanes obtained by the hydrolytic condensation of PhSiCl3 in H2O.

The complex (I) obtained by reaction of titanium butoxide with the phenyl

silsesquioxanes was tested for activity in the epoxidation of 1-octene with TBHP. The

amount of titanium butoxide was chosen to yield a ratio of one titanium centre per two

silanol groups, under the assumption that half of the silicon atoms of the silsesquioxanes

-69.8

-79.4

PhPh

Ph

Ph

Ph

Si

SiO

OH

Si Si Si

Si

Si

Si

Si Si

OO

O

O O

O

O

O

OOHO

Ph

PhPh

Ph

Ph

OH

HO

-69.8

-79.4

-69.8

-79.4

PhPh

Ph

Ph

Ph

Si

SiO

OH

Si Si Si

Si

Si

Si

Si Si

OO

O

O O

O

O

O

OOHO

Ph

PhPh

Ph

Ph

OH

HO

PhPh

Ph

Ph

Ph

Si

SiO

OH

Si Si Si

Si

Si

Si

Si Si

OO

O

O O

O

O

O

OOHO

Ph

PhPh

Ph

Ph

OH

HO

Chapter 7 __________________________________________________________________________________________________________

122

contain -OH groups. This titanium to silsesquioxane ratio is almost double than the one

employed in the HSE experiment. The catalyst (I) showed a similar activity (per mole

of titanium) to the corresponding HSE-catalyst (61% TBHP conversion and >99%

selectivity towards 1,2-epoxyoctane after 4 hours of reaction, see Table 7.1 on

page 130). Differently from other reported titanium catalysts containing phenyl

silsesquioxanes,11 these titanium silsesquioxanes are not soluble in the epoxidation

reaction mixture and remain a distinct powder. Finally, the catalyst was tested for the

epoxidation of 1-octene with aqueous H2O2 as the oxidant, resulting in negligible

activity accompanied by catalyst deactivation.

7.2.3 tert-butyl silsesquioxanes synthesised in H2O

The hydrolytic condensation of tert-butyltrichlorosilane ButSiCl3 in water was

performed in a 25-fold up-scaling of the HSE synthesis. ButSiCl3 is a solid at room

temperature: since the HSE workstation employed can only handle liquids, the

trichlorosilane was dissolved in a minimum of acetonitrile. The same procedure was

used for the up-scaled synthesis. After 18 hours of reaction at 50°C, the reaction

mixture contained a white precipitate, which was isolated by filtration (fraction A).

Upon drying of the filtrate, a white solid was obtained (fraction B). The two fractions

were exsiccated in an oven at 100°C. Liquid-phase NMR characterisation showed that both

fractions contained the same single silsesquioxane structure constituted of equivalent

ButSiO1.5(H2O)0.5x units (Figure 7.3). The position of the 29Si NMR peak (δ -49.55 ppm)

indicates that this silsesquioxane contains one or more silanol groups (x ≥ 1).8 The

structure was assigned to silsesquioxane But2Si2O(OH)4 (a2b4, Figure 7.4)8 on the basis

of these NMR data and of single-crystal X-ray diffraction analysis, which provided the

same cell parameters of the tert-butyl silsesquioxane a2b4 reported in literature.12 The

synthetic procedure reported here allows complete and selective conversion of ButSiCl3

into silsesquioxane But2Si2O(OH)4. This method is more straightforward and results in a

higher isolated yield (90%) than the literature method (65%) in which the compound is

synthesised by addition of ButSiCl3 in diethyl ether to a mixture of KOH, water and

methanol.12 The selectivity of the synthesis is ascribed to the bulkiness of the tert-butyl

group, which can hinder further reaction of But2Si2O(OH)4, and to the use of water as

the solvent, which disfavours condensation.


123

Figure 7.3. 1H, 13C and 29Si NMR spectra of tert-butyl silsesquioxane But2Si2O(OH)4,

obtained by the hydrolytic condensation of tert-butyltrichlorosilane in H2O. In

the 1H NMR spectrum (top) two peaks in a 9:2 ratio are present: the signal at

δ 1.46 ppm is due to the methyl resonances of the tert-butyl groups and the

signal at δ 8.26 ppm is due to the four silanol groups. The three additional

peaks in the region 7-9 ppm are due to pyridine (present in small amount in

the deuterated solvent). In the 13C NMR spectrum (middle) two peaks in 3:1

ratio are present: the signal at δ 27.36 ppm is due to the three equivalent

methyls of the tert-butyl groups, while the signal at δ 18.25 ppm is due to the

ipso carbon of the tert-butyl group. The peaks in the region 120-150 ppm are

due to pyridine (the solvent). The peak shown in the 29Si NMR spectrum

(bottom) is due to the two equivalent silicon atoms present in structure a2b4.

0

1 0 0 0

2 0 0 0

3 0 0 0

4 0 0 0

5 0 0 0

6 0 0 0

29Si NMR

13C NMR

1H NMR

ppm20406080100120140

ppm-70-50 -80-60-40-20-10 -30

ppm2468 79 510 3 1

8.26

1.46

27.36

18.25

-49.55

0

1 0 0 0

2 0 0 0

3 0 0 0

4 0 0 0

5 0 0 0

6 0 0 0

29Si NMR

13C NMR

1H NMR

ppm20406080100120140

ppm-70-50 -80-60-40-20-10 -30

ppm2468 79 510 3 1

8.26

1.46

27.36

18.25

-49.55

Chapter 7 __________________________________________________________________________________________________________

124

Figure 7.4. tert-butyl silsesquioxane a2b4, But2Si2O(OH)4.

The fact that silsesquioxane a2b4 was obtained as the only product both as a precipitate

and by drying the remaining reaction mixture prompted an investigation of whether a

shorter reaction time would lead to the same product. Therefore, the hydrolytic

condensation was performed for 20 minutes at room temperature. Next, the solvent was

removed by filtration to afford a white solid that was then exsiccated in an oven at

100°C. 13C and 29Si NMR analysis showed that, besides silsesquioxane a2b4, the

product contained a second species assigned to silsesquioxane ButSi(OH)3 (a1b3)13 on

the basis of the position of the 29Si NMR peak (δ -40.06 ppm).8 The molecular ratio

between the a1b3 and a2b4 structures was ~2:1. This experiment indicates that the

formation of tert-butyl silsesquioxane a2b4 takes place by the slow condensation of

tert-butyl silsesquioxane a1b3. The silsesquioxane a2b4 does not react further and it

starts to precipitate as its concentration in solution increases. Further investigation

showed that 7 hours of hydrolytic condensation at 50°C are necessary to obtain

silsesquioxane a2b4 as the only product.

The nature of the solvent in which the ButSiCl3 is dissolved before being added

to the water does not seem to influence the reaction: the same product was obtained if

acetone was used instead of acetonitrile. On the other hand, the amount of this

co-solvent has a relevant role: when ButSiCl3 was dissolved in a 5-fold smaller volume

of acetone, after 18 hours at 50°C the hydrolytic condensation produced large amounts

of precipitate containing viscous polymeric silsesquioxanes together with structure a2b4

(as determined by NMR analysis). After removing the precipitate by filtration, the

filtrate afforded silsesquioxane a2b4 upon drying. The lower selectivity and the

a2b4

OH

But

OHSiSi O

HO

HO

But

a2b4

OH

But

OHSiSi O

HO

HO

ButOH

But

OHSiSi O

HO

HO

But


125

increased amount of precipitate indicate that when less co-solvent is present in solution,

the formation and the precipitation of silsesquioxanes takes place more rapidly,

probably as a consequence of the reduced solubility of the silsesquioxanes in the

reaction mixture. This behaviour is similar to that reported above for the hydrolytic

condensation of phenyltrichlorosilane in H2O, where no co-solvent was used.

These observations confirm the complexity of the mechanism of formation of

silsesquioxanes: small changes in the reaction conditions result in significantly different

amounts and types of structures.

Besides the intrinsic value of the identification of a new, selective and high yield

method to synthesise silsesquioxane But2Si2O(OH)4, this experiment proved that

tert-butyl silsesquioxane a2b4 is a suitable precursor for titanium catalysts. Thus,

silsesquioxane structures different from the known precursor silsesquioxane a7b3

[R7Si7O9(OH)3]4,5,14 can effectively coordinate titanium centres to yield almost equally

active epoxidation catalysts. In order to investigate how the titanium centre coordinates

to tert-butyl silsesquioxane a2b4 and which is the optimum number of titanium centres

that the structure can accommodate, catalysts with different titanium to silsesquioxane

ratios were prepared, characterised and tested. The HSE lead had a titanium to

silsesquioxane ratio of ~ 1:3, the ratio having been tuned for more condensed

silsesquioxanes (e.g. structure a7b3) which present a low average number of -OH

groups per silicon atom. Silsesquioxane a2b4 has four -OH groups that may react with

titanium alkoxide complexes. Coordination of a titanium centre to two -OH groups on

the same silicon would produce a geometrically strained complex and is, therefore,

unlikely. In principle, each titanium could coordinate to two -OH groups on two

different silicons to form a polymeric chain with alternate titanium centres and

silsesquioxanes (Figure 7.5). This chain presents a 1:1 titanium to silsesquioxane molar

ratio. Less ordered structures with a titanium to silsesquioxane molar ratio > 1 are also

possible. Therefore, complexes were prepared by reacting Ti(OBu)4 with tert-butyl

silsesquioxane a2b4 in 1:1 (catalyst II) and 2:1 (catalyst III) molar ratios. The two

catalysts were characterised and tested for epoxidation activity.

Chapter 7 __________________________________________________________________________________________________________

126

Figure 7.5. Ideal structure of a polymeric chain obtained from the reaction of Ti(OBu)4

with But2Si2O(OH)4 in a 1:1 molar ratio.

The liquid-phase 29Si NMR spectrum of catalyst II shows a single peak at -49.55 ppm

due to the unreacted tert-butyl silsesquioxane a2b4, while the spectrum of catalyst III

presents a very weak and broad signal around -56 ppm. These data indicate that a

titanium to silsesquioxane ratio > 1 is needed to react all the silsesquioxane a2b4 and

suggest that both catalysts consist of polymeric titanium complexes (which are difficult

to detect in liquid-phase 29Si NMR). The solid-state 29Si NMR spectrum of catalyst II

(Figure 7.6, top) shows the peak due to tert-butyl silsesquioxane a2b4 (already visible

in the liquid-phase spectrum) and two broad signals centred at -56 ppm and -66 ppm

(integral ratio ~ 1:2:1). The complexation of titanium causes the silicon peak position

to move towards higher shifts,6,15 suggesting that the signals around -56 ppm originate

from siloxy groups to which one titanium centre is coordinated, while the signals at

-66 ppm derive from siloxy groups to which two titanium centres are coordinated. The

solid-state 29Si NMR spectrum of catalyst III (Figure 7.6, bottom) presents two broad

and partially overlapping signals. These two peaks are in the same position as those

found for catalyst II, but exhibit a different ratio (4:1 as compared to 2:1), consistent

with catalyst III containing proportionally more siloxy groups to which only one

titanium centre is coordinated.

But

Si

Si

But

O Ti

O

O

O

O

m

But

Si

Si

But

O Ti

O

O

O

O

m


127

Figure 7.6. Solid-state 29Si NMR spectra of catalyst II (top) and of catalyst III (bottom).

The liquid-phase 13C NMR spectra of both catalysts confirm that the titanium

complexation took place and generated different types of complexes: a set of

overlapping signals is present in the region of the methyl resonances of the tert-butyl

group on the silsesquioxanes (27.3 to 27.5 ppm) and another in the region of the ipso

carbon of the tert-butyl group (18.1 to 18.3 ppm), see Figure 7.7 (top). Besides these

peaks, the spectra present four peaks of equal intensity due to n-butanol.

The solid-state 13C NMR spectra of both catalysts contain two broad signals in the

expected 3:1 ratio, due to the methyl groups and to the ipso carbon of the tert-butyl

group on the silsesquioxanes respectively, consistent with what was observed in the

liquid-phase 13C NMR (Figure 7.7). Besides these signals, the spectra present four broad

peaks assigned to a butoxy group coordinated to a titanium.16,17 The extreme broadness

of the peak centred at 79 ppm - due to the carbon in the α position to the oxygen of the

butoxy group - can be attributed to the presence of different types of titanium centres

-40 -50 -60 -70 -80 ppm

-40 -50 -60 -70 -80 ppm

-56

-56

-66

-66

-50

catalyst II

catalyst III

-40 -50 -60 -70 -80 ppm

-40 -50 -60 -70 -80 ppm

-56

-56

-66

-66

-50

catalyst II

catalyst III

-40 -50 -60 -70 -80 ppm

-40 -50 -60 -70 -80 ppm

-56

-56

-66

-66

-50

catalyst II

catalyst III

Chapter 7 __________________________________________________________________________________________________________

128

and/or to fast relaxation caused by poor mobility of the α-carbon. Different degrees of

complexation of titanium butoxide to the silsesquioxane a2b4 may have occurred: the

butoxide groups can be partially or completely substituted by coordination to the

silsesquioxanes. In addition, the butoxide signals could also be due to n-butanol that

was released during the complexation of titanium butoxide to the silsesquioxane

structures and that may still be coordinated to the titanium centres. The difference in

peak position of the α-carbon in the solid-state and liquid-phase 13C NMR indicates that,

in solution, the butoxy groups originally coordinated to the titanium are exchanged with

pyridine (the solvent), which is known to coordinate to titanium centres.18

Figure 7.7. Liquid-phase (top) and solid-state (bottom) 13C NMR spectra of catalyst III.

ppm102030405060708090

61.72 35.74

27.4

19.54

18.2

14.17

ppm102030405060708090

79.5

35.8

27.4

19.9

18.0

14.6

ppm102030405060708090

61.72 35.74

27.4

19.54

18.2

14.17

ppm102030405060708090

79.5

35.8

27.4

19.9

18.0

14.6


129

Gel permeation chromatography of catalyst II showed two bands. The first, with

a number average molecular weight of 273 g/mol, corresponds to the unreacted

tert-butyl silsesquioxane a2b4 (MW = 254.53 g/mol), also detected by NMR

spectroscopy. The polydispersity of this band is 1.24, in agreement with the presence of

a single species. The second band has a polydispersity of 2.14, indicating the presence

of different structures with a number average molecular weight of 1615 g/mol. This

suggests that the titanium silsesquioxane complexes are on average constituted of 4-5

a2b4 structures and 5-6 titanium centres.

Catalysts II and III were tested for epoxidation activity with TBHP as the

oxidant (Figure 7.8 and Table 7.1), using for both the same titanium to substrate ratio.

The two catalysts displayed higher activity (per mole of titanium) than the HSE lead

(74% TBHP conversion towards 1,2-epoxyoctane after 4 hours of reaction): catalyst II

gave 80% conversion after 3 hours and reached a plateau at 93% conversion

(TOF = 0.44 molepo·molTi-1·min-1); catalyst III gave 90% conversion after 3 hours and

reached a plateau at 97% conversion (TOF = 0.54 molepo·molTi-1·min-1). Both catalysts

exhibited 97% selectivity towards 1,2-epoxyoctane.

Figure 7.8. Activity in the epoxidation of 1-octene with TBHP, for Ti-a2b4 complexes with

1:1 and 2:1 titanium to silsesquioxane ratios. The concentration of titanium and

the titanium to TBHP ratio were the same in the two catalytic tests.

60 120 180 240 300 360 420

time (minutes)

conv

ersi

on

Ti:silsesquioxane = 1:1 (catalyst II)

Ti:silsesquioxane = 2:1 (catalyst III)

00%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

60 120 180 240 300 360 420

time (minutes)

conv

ersi

on





00%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Chapter 7 __________________________________________________________________________________________________________

130

Table 7.1. Conversions and selectivities in the epoxidation of 1-octene with TBHP for

the catalysts reported in this chapter and for (c-C5H9)7Si7O12TiOC4H9 (see

also Figure 3.6).II The same titanium:substrate ratio was used in all the

catalytic tests. (a supported on silica; b supported on silylated silica). c It is

interesting to note that the selectivity towards the epoxide is slightly lower in

the case of the heterogeneous catalysts even though virtually no formation of

diol as side-product can be detected. This is consistent with a separate

oxidation pathway involving radicals of homolytically disassociated TBHP - a

path promoted by the presence of a high surface area solid.

Catalyst III is homogenously dissolved in the epoxidation mixture, while catalyst II

does not dissolve completely: the insoluble fraction is probably constituted of the

unreacted silsesquioxane a2b4, which is not soluble in 1-octene. The difference in

activity between catalyst II and III probably originated from the different nature of the

titanium centres (e.g. accessibility), as determined by solid-state 29Si NMR analysis.

The catalytic sites of these titanium silsesquioxanes strongly interact with alcohols: if

the epoxidation test of catalyst II was performed in a 1:1 mixture of 1-octene and

1-propanol, very low activity was found. This deactivation is reversible: the expected

II The experimental conditions of the catalytic tests reported in this thesis were chosen in order to adapt to

the HSE equipment, implying a rather high catalyst concentration. The same conditions were also used

for the experiments performed on a conventional laboratory scale. These conditions are rather different

from those generally reported in literature for the epoxidation of alkenes with titanium catalysts4,5,14,15,19-24

and make it difficult to directly compare the results. Nevertheless, all the catalytic activities reported in

this thesis are compared to that of (c-C5H9)7Si7O12TiOL complexes, which proved to be extremely active

homogeneous catalysts for the epoxidation of alkenes.4,5

catalyst silsesquioxane precursor Ti:silsesq. TOF (min -1 ) conversion (3h) selectivity c

I PhSiCl3 in H2O ~1:1 61% (4h) 99%

II ButSiCl3 in H2O 1:1 0.44 80% 97%

III ButSiCl3 in H2O 2:1 0.54 90% 97%IVa ButSiCl3 in H2O 1:1 94% 92%Vb ButSiCl3 in H2O 1:1 94% 92%

VI ButSiCl3 in DMSO ~1:1 0.22 84% 98%A (c- C5H9)7Si7O9(OH)3 1:1 0.97 >99% 99%


131

epoxidation activity was recovered when reusing the catalyst without 1-propanol as

co-solvent. The same activity loss was observed when methanol was used as co-solvent

instead of 1-propanol. This reversible coordination of alcohols to the titanium is in

agreement with the results of the 13C NMR analysis of the titanium-silsesquioxane a2b4

complexes. The catalyst deactivation is ascribed to the interaction between the hydroxyl

group and the titanium centre,5 which would reduce the accessibility to the catalytic site.

The deactivation caused by coordination of alcohols to the titanium centre is not

encountered when using (c-C5H9)7Si7O12TiOC4H9 as catalyst, hence pointing to a

different nature of the catalytic centre.

The main drawback of homogeneous catalysts lies in the difficulty to recover

them from the reaction mixture. In order to overcome this downside, the

titanium-silsesquioxane a2b4 complex was supported on high surface area silica and the

materials obtained were tested as heterogeneous catalysts for the epoxidation of

1-octene.14,19-24 Titanium silsesquioxane a2b4 was supported by adsorption on two types

of silica: one untreated and the other dehydroxylated by derivatisation with (CH3)2SiCl2

(catalysts IV and V, respectively). The larger size of the titanium-silsesquioxane

complexes compared to that of the other molecules present in the epoxidation mixture

should favour the adsorption of the complexes as well as prevent their leaching from the

silica supports (entropic driving force). After adsorption of the complexes on the silica

supports, possible leaching species were removed by prolonged Soxhlet extraction in

tetrahydrofuran. The epoxidation test gave a 94% TBHP conversion and 92% selectivity

towards 1,2-epoxyoctane after 3 hours of reaction for both catalysts. Leaching of the

titanium silsesquioxanes from the silica support was checked by filtering the solid from

the epoxidation mixture and by measuring the catalytic activity of the possible soluble

titanium silsesquioxanes: the soluble fractions presented negligible epoxidation activity

(equal to that of a blank sample), hence proving that no leaching of active species

occurred. The heterogeneous catalyst supported on untreated silica (IV) was reused for

3 cycles and yielded comparable epoxidation activity (± 4% of the activity obtained in

the first catalytic test).

The titanium-silsesquioxane a2b4 complexes were supported on both untreated and

silylated silica in order to investigate the nature of the adsorption. The similar activity

and selectivity obtained with the two supports suggests that the adsorption of the

Chapter 7 __________________________________________________________________________________________________________

132

complexes takes place by physical interaction with the silica surface rather than through

chemical anchoring on the -OH groups (totally or almost absent on the silica derivatised

with (CH3)2SiCl2).

Remarkably, the heterogeneous catalysts displayed an epoxidation activity (per mole

of titanium) similar to that of the homogeneous titanium-silsesquioxane a2b4

complexes, although with a lower selectivity towards 1,2-epoxyoctane (see Table 7.1

on page 130).

Neither the homogeneous titanium-silsesquioxane a2b4 complexes nor the

silica-supported heterogeneous catalysts showed good activity in the epoxidation of

1-octene with aqueous H2O2 as the oxidant.

7.2.4 tert-butyl silsesquioxanes synthesised in DMSO

The hydrolytic condensation of tert-butyltrichlorosilane in DMSO was

performed in a 25-fold up-scaling of the HSE synthesis. Similarly to what has been

described for the synthesis of tert-butyl silsesquioxane a2b4, the

tert-butyltrichlorosilane was first dissolved in acetonitrile and then added to the DMSO.

After 18 hours of reaction at 50°C, the clear solution obtained was distilled under

reduced pressure to give a yellow gel that still contained DMSO.III Adding water to the

gel caused the precipitation of the silsesquioxane product that could then be separated

from the DMSO/H2O solution. 13C and 29Si NMR analysis showed that the sample

consisted of a number of oligomeric silsesquioxane species. The 29Si NMR spectrum

(Figure 7.9) presents two sets of peaks, one in the region of silicons connected to three

other silicons via oxygen bridges (-55 ppm < δ < -61 ppm), and the other of silicons

connected to one or more -OH groups (-45 ppm < δ < -51 ppm).8 The latter set of peaks

has a much higher intensity than the former, indicating that most of the silsesquioxanes

are incompletely condensed structures with a low level of condensation.

III Removing DMSO from the silsesquioxanes is necessary to prevent competition between the oxidation

of DMSO to dimethyl sulfone and the epoxidation of 1-octene to 1,2-epoxyoctane. Complete removal of

DMSO was attained in the HSE experiment by drying the samples in a vacuum centrifuge that, therefore,

proved to be more effective than a distillation under vacuum.


133

Figure 7.9. Liquid-phase 29Si NMR spectrum of the tert-butyl silsesquioxanes obtained

by the hydrolytic condensation of tert-butyltrichlorosilane in DMSO.

Characterisation by means of mass spectrometry showed that the main species is an

a = 4 structure (i.e. constituted of four ButSiO1.5(H2O)0.5x units) followed, in decreasing

concentration order, by a = 5, a = 3, a = 6 and a = 7 silsesquioxanes. Due to possible

intramolecular condensation accompanied by loss of water molecules during the

ionisation process, the level of condensation of the silsesquioxanes (i.e. the b value)

cannot be determined by means of MS analysis of the crude product. More information

about the number of silanol groups present in the silsesquioxane structures can be

generated by their silylation with trimethylchlorosilane (CH3)3SiCl.25,26 MS analysis of

the products of the silylation reaction showed a main peak corresponding to a disilylated

a4b4 structure; other species in relevant concentrations were mono- and disilylated

a3b3, mono- and disilylated a4b2, monosilylated a4b4, mono- and disilylated a5b3,

di- and trisilylated a5b5. These MS data indicated the presence of a3b3, a4b4 and a5b5

structures while did not allow establishing whether structures a4b2 and a5b3 were

actually present or were formed in the mass spectrometer. By joining the results form

NMR and MS analyses it is proposed that the major components of the product were

silsesquioxanes a3b3 [But3Si3O3(OH)3], a4b4 [But

4Si4O4(OH)4] and a5b5

[But5Si5O5(OH)5], for which the most likely structures are reported in Figure 7.10.

-45.37

-46.50

-47.96

-48.27

-49.06

-48.94

-49.37

-56.56 -57.79-58.32

ppm-60-58-56-54-52-50-48-46

-47.46-45.37

-46.50

-47.96

-48.27

-49.06

-48.94

-49.37

-56.56 -57.79-58.32

ppm-60-58-56-54-52-50-48-46

-47.46

Chapter 7 __________________________________________________________________________________________________________

134

Figure 7.10. tert-butyl silsesquioxanes a3b3, a4b4 and a5b5.

In order to prepare and test the titanium-silsesquioxane catalyst (VI), the product of the

hydrolytic condensation of tert-butyltrichlorosilane in DMSO was reacted with Ti(OBu)4.

The amount of titanium precursor was chosen in order to have a 1:1 ratio between

titanium and silsesquioxanes, by approximating the composition of the silsesquioxane

mixture to that of the major species, silsesquioxane a4b4. Such ratio is almost double than

the one employed in the HSE experiment. The titanium silsesquioxane was tested for

catalytic activity in the epoxidation of 1-octene with TBHP (Figure 7.11 and Table 7.1 on

page 130). The homogeneous catalyst (VI) gave a lower turnover than titanium

silsesquioxane a2b4 (TOF = 0.22 molepo·molTi-1·min-1, cf. Figure 7.8 and 7.11). However, it

presented TBHP conversion towards 1,2-epoxyoctane comparable (per mole of titanium) to

that of the corresponding HSE lead; after 7 hours of reaction almost complete TBHP

conversion was achieved, with 98% selectivity towards 1,2-epoxyoctane. An epoxidation

test with aqueous H2O2 as the oxidant gave negligible activity accompanied by catalyst

deactivation.

a3b3 a4b4

a5b5

or

SiO

HO

SiSiO

O

But But

OH

OH

But

SiO

O

OSiSi O

Si

But

OHHO

But

But

OHHO

But

SiO

O

O

SiSi

Si

Si

But

OHHO

But

But

OHHO

But

OO

But

OHSiO

O

OSiSi O

Si

But

OHHO

But

But

OHO

But

Si

But

OHHO

a3b3 a4b4

a5b5

or

SiO

HO

SiSiO

O

But But

OH

OH

But

SiO

HO

SiSiO

O

But But

OH

OH

But

SiO

O

OSiSi O

Si

But

OHHO

But

But

OHHO

ButSi

OO

OSiSi O

Si

But

OHHO

But

But

OHHO

But

SiO

O

O

SiSi

Si

Si

But

OHHO

But

But

OHHO

But

OO

But

OH

SiO

O

O

SiSi

Si

Si

But

OHHO

But

But

OHHO

But

OO

But

OHSiO

O

OSiSi O

Si

But

OHHO

But

But

OHO

But

Si

But

OHHO

SiO

O

OSiSi O

Si

But

OHHO

But

But

OHO

But

Si

But

OHHO


135

Figure 7.11. Activity in the epoxidation of 1-octene with TBHP, for the catalyst obtained

by insertion of titanium butoxide in the silsesquioxane structures synthesised

by the hydrolytic condensation of ButSiCl3 in DMSO as the solvent.

7.3 Conclusions

The effect of highly polar solvents on the hydrolytic condensation of a set of

trichlorosilanes to produce silsesquioxane precursors for titanium catalysts active in the

epoxidation of alkenes was studied by means of High-Speed Experimentation

techniques. The HSE screening allowed the identification of a number of leads. The

three most promising ones were studied in detail. It was determined that these three

leads consisted of three very different types of silsesquioxane structures. This result

showed that silsesquioxanes with high structural diversity can be suitable precursors for

active titanium catalysts. The most interesting lead is the tert-butyl silsesquioxane a2b4

synthesised via the hydrolytic condensation of tert-butyltrichlorosilane in water. The

catalysts obtained by complexation of titanium butoxide to the silsesquioxane a2b4

displayed very good activity and selectivity in the epoxidation of 1-octene with TBHP.

Characterisation showed that these catalysts are constituted of various titanium centres

and silsesquioxanes units linked to each other in different arrangements. Titanium

silsesquioxane a2b4 complexes can be supported on silica to produce an active,

non-leaching, recyclable, heterogeneous, epoxidation catalyst.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 60 120 180 240 300 360 420

time (minutes)

conv

ersi

on

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 60 120 180 240 300 360 420

time (minutes)

conv

ersi

on

Chapter 7 __________________________________________________________________________________________________________

136

7.4 Experimental

7.4.1 The High-Speed Experimentation screening

The preparation of the HSE samples was performed by means of a Hamilton

Dual-Arm liquid-handling robotic workstation27 coupled with a personal computer

supplied with software enabling to program the workstation. Reagents for the hydrolytic

condensation: cyclohexyl-, cyclopentyl-, phenyl-, methyl-, ethyl- and

tert-butyltrichlorosilanes; the latter was dissolved beforehand in acetonitrile (340 µmol

in 0.5 ml of CH3CN); acetonitrile, dimethyl sulfoxide (DMSO), deionised water and

formamide as solvents; deionised water. The silsesquioxane precursors were prepared

by dispensing 2-ml aliquots of each of the solvents in racks containing 6x4 arrays of

glass tube reactors (3 ml working volume), followed by the addition of 340 µmol of

each of the trichlorosilanes and by the addition of 0.5 ml of deionised water to each

reaction tube. Subsequently, the racks were placed in a steel heater block mounted on an

orbital shaker27 and heated to 50°C for 18 hours. After evaporation of the liquids in a

vacuum centrifuge27 overnight, the silsesquioxane samples were dissolved in

tetrahydrofuran (THF), 54 µmol of titanium butoxide, [Ti(OBu)4, 98% purity] were

added to each sample and the rack was stirred for 6 hours at 60°C. After titanium

complexation, THF was removed by vacuum centrifugation. Each sample was then

tested for epoxidation activity by adding 2.26 ml of 1-octene (with 2%vol of decane as

internal standard) and 121 µl of TBHP (~45%wt solution in cyclohexane). After having

been stirred in the orbital shaker for 4 hours at 80°C, samples were analysed on an

automated Unicam Pro GC27 under isothermal conditions. The activities were obtained

by normalising the 1,2-epoxyoctane GC peak area by means of the internal standard.

The reported results are the average of multiple runs.

7.4.2 Phenyl silsesquioxanes synthesised in H2O

1.35 ml of phenyltrichlorosilane [PhSiCl3, 97% purity] were carefully added to

62.5 ml of deionised water. As soon as the trichlorosilane was added, a white solid

started to precipitate. After having been stirred for 18 hours at 50°C, the solution was

filtered. The filtrate was evaporated under reduced pressure to afford negligible amounts


137

of solid. The precipitate was dried overnight in an oven at 100°C to afford 1.005g of

white solid (92% yield on the hypothesis of a silsesquioxane structure containing 50%

of PhSiO1.5 units and 50% of PhSiO2H units). The solid is soluble in acetone, THF and

pyridine; dissolution is slow.

The titanium complexation and the catalytic test were performed analogously to what is

described for the HSE study. The molar ratio between titanium and TBHP and the

concentration of TBHP in the epoxidation mixture were the same in all the catalytic

tests reported in this chapter.

NMR characterisation of the silsesquioxane product was performed on a Varian

VXR-400S (1H decoupled, 25°C). Selected data: 13C{1H} NMR (liquid phase, deuterated

acetone as the solvent), two sets of partially overlapping peaks due to the phenyl groups:

δ 128.77 ppm, 128.85 ppm (carbons meta to the ipso carbons), multiplet around

δ 131.53 ppm (ipso carbons and carbons para to it), δ 134.93 ppm, 135.07 ppm (carbons

ortho to the ipso carbons); 29Si{1H} NMR (solid state), δ -69.8 ppm, -79.4 ppm.

7.4.3 tert-butyl silsesquioxanes synthesised in H2O

1.63 g of tert-butyltrichlorosilane [ButSiCl3, >95% purity] were dissolved in

12.5 ml of acetonitrile. This solution was added to 50 ml of deionised water. After 18

hours of reaction at 50°C upon stirring, a white, fine precipitate was formed. The

precipitate was collected by filtration (fraction A) and the filtrate solution was dried

under reduced pressure to yield a white, fine powder (fraction B). To remove water that

might still be present, the two fractions were dried for 24 hours in an oven at 100°C:

after drying, 0.335 g of fraction A and 1.502 g of fraction B were present. Both

fractions were constituted of tert-butyl silsesquioxane a2b4, (t-C4H9)2Si2O(OH)4

[MW = 254.43 g/mol]. The yield in the isolated silsesquioxane a2b4 is 90%, with >99%

selectivity. The compound melts between 192 and 204°C and is soluble in pyridine,

DMSO and, to a lesser extent, in THF. In order to obtain suitable crystals for

single-crystal X-ray diffraction analysis, 0.1 g of the compound were dissolved in 10 ml

of THF at 50°C and recrystallised by carefully adding 1-ml aliquots of acetonitrile (5 ml

in total) to the THF solution.

Chapter 7 __________________________________________________________________________________________________________

138

The synthesis of catalyst II and III was performed by titanium complexation to the

silsesquioxane structure with similar experimental conditions to those described for the

HSE study.

CAB-O-SIL fumed silica EH-5 [surface area: 380 m2/g, ~4 hydroxyl groups/nm2] was

used as support for the heterogenisation of the titanium-silsesquioxane a2b4 complexes.

Dehydroxylation of the silica was performed by adding 1 ml of (CH3)2SiCl2 to a

suspension of 1 g of silica in 30 ml of diethyl ether and by stirring the suspension for

1 hour at room temperature. The volatiles were removed under reduced pressure. The

titanium-silsesquioxane a2b4 complexes (catalyst II) were supported on the silica (both

untreated and dehydroxylated) by adding a 150-ml THF solution of the complexes

(containing 5·10-4 mol of titanium) to a suspension of 1 g of silica in 100 ml of THF and

by stirring the suspension for 3 hours at 60°C. After isolation of the solid by filtration,

non-adsorbed species were removed by Soxhlet extraction with THF for 7 hours.

Finally, the silica-supported catalysts (IV and V) were dried overnight at 120°C under

reduced pressure.

The catalytic tests were performed analogously to what is described for the HSE study.

The same ratio between titanium and TBHP and the same concentration of TBHP in the

reaction mixture were employed in all the catalytic tests reported in this chapter.

For single-crystal X-ray diffraction of But2Si2O(OH)4, the crystal was mounted along an

arbitrary axis on an Enraf-Nonius CAD-4 diffactometer. Lattice constants were determined

from 25 reflections with 10° < θ < 17º (standard deviations are within parentheses):

a = 6.226(6), b = 6.223(6), c = 9.956(4); α = 85.06(5), β = 80.72(5), γ = 70.51(9).

NMR spectra were measured on a Varian VXR-400S (1H decoupled, 25°C) and on a

Varian Inova-300 (1H decoupled, 25°C).

Selected data for silsesquioxane a2b4, liquid-phase NMR (deuterated pyridine as the

solvent): 1H NMR, δ 1.46 ppm, (But; relative integral = 9), 8.26 ppm (Si-OH; relative

integral = 2); 13C{1H} NMR, δ 18.25 ppm (ipso carbon of But; relative integral = 1)

δ 27.36 ppm (CH3 groups of But; relative integral = 3); 29Si{1H} NMR, δ -49.55 ppm

(ButSi(OH)2O0.5).

Selected data for the silsesquioxane attributed to the a1b3 structure, liquid-phase NMR

(deuterated pyridine as the solvent): 13C{1H} NMR, δ 18.32 ppm (ipso carbon of But;


139

relative integral = 1) δ 27.51 ppm (CH3 groups of But; relative integral = 3); 29Si{1H} NMR, δ -40.06 ppm (ButSi(OH)3).

Selected data for catalyst II: solid-state 29Si{1H} NMR, δ -50 ppm (relative integral

~ 1), δ -56 ppm, (broad signal; relative integral ~ 2), δ -66 ppm (broad signal; relative

integral ~ 1).

Selected data for catalyst III: liquid-phase 13C{1H} NMR (deuterated pyridine as the

solvent), δ 14.17 ppm (δ), 19.54 ppm (γ), 35.74 ppm (β), 61.72 ppm (α) (1:1:1:1; δCH3

γCH2βCH2

αCH2OH), δ 18.2 ppm, 27.4 ppm (~1:3; multiplets; respectively: ipso

carbon and CH3 groups of the But-group on the silsesquioxane species); solid-state 13C{1H} NMR, δ 14.6 ppm (δ), 19.9 ppm (γ), 35.8 ppm (β), 79.5 ppm (α) (~1:1:1:1;

broad peaks, α very broad; δCH3γCH2

βCH2αCH2OTi), δ 18.0 ppm, 27.4 ppm (~1:3;

broad peaks; respectively: ipso carbon and CH3 groups of the But-group on the

silsesquioxane species); solid-state 29Si{1H} NMR, δ -56 ppm, (broad band; relative

integral ~ 4), δ -66 ppm (broad signal; relative integral ~ 1).

Gel permeation chromatography of catalyst II was performed on a Waters 410

differential refractometer gel permeation chromatographer equipped with a Waters

Styragel HT 6E column calibrated on polystyrene standards. THF was used as the

eluent.

7.4.5 tert-butyl silsesquioxanes synthesised in DMSO

1.63 g of tert-butyltrichlorosilane [ButSiCl3, >95% purity] were dissolved in

12.5 ml of acetonitrile. The solution was added to 37.5 ml of DMSO, followed by

addition of 12.5 ml of deionised water. The transparent, colourless solution was stirred

for 18 hours at 50°C and then distilled under reduced pressure to remove volatile

species and to isolate the silsesquioxanes. The yellow gel obtained still contained

DMSO, which was removed by adding 50 ml of H2O: the silsesquioxanes precipitated

as a sticky, soft, solid that glued to the glass walls of the reaction flask, from which the

water/DMSO solution was easily removed. The silsesquioxane mixture is soluble in

THF and pyridine.

The silylation of the silsesquioxane mixture was performed by adding 172 µl of

trimethylchlorosilane (CH3)3SiCl to a solution containing 0.04 g of silsesquioxanes in

Chapter 7 __________________________________________________________________________________________________________

140

30 ml of toluene and 1 ml of triethylamine (C2H5)3N. The turbid solution was stirred

overnight at room temperature. Next, the toluene and the triethylamine were removed

under reduced pressure and a white solid was obtained. The solid was extracted with

pentane; the solution was filtered to remove the insoluble (C2H5)3N·HCl and dried under

reduced pressure to yield a solid containing the silylated silsesquioxanes.

The titanium complexation and the catalytic test were performed similarly to what is

described for the HSE study. The molar ratio between titanium and TBHP and the

concentration of TBHP in the reaction mixture were the same in all the catalytic tests

reported in this chapter. The NMR spectrum of the silsesquioxane mixture was collected on a Varian Inova-300

(1H decoupled, 25°C). Selected data: 29Si{1H} NMR (liquid phase, deuterated pyridine

as the solvent), δ -45.37 ppm (relative integral ~ 1), -46.50 ppm (~3), -47.46 ppm (~1),

-47.96 ppm (~2), -48.27 ppm (~4.5), -48.94 ppm (~6.5), -49.06 ppm (~8.5), -49.37 ppm

(~11), -56.56 ppm (~1), -57.79 ppm (~1), -58.32 ppm (~1).

Mass spectrometry analysis was performed on a Micromass Quattro LC-MS with

APCI+ as ionisation technique. For the analysis, ~0.03 g of the sample were dissolved

in 5 ml of CH2Cl2/THF (4:1) and few drops of CH3CO2H were added.

Silsesquioxane product (cone voltage = 45 V): for a = 3, m/z: 337.50 ([a3b1 + H]+,

20%), 355.54 ([a3b3 + H]+, 17%), 409.50 ([a3b5 + 2H2O + H]+, 22%); for a = 4, m/z:

437.47 ([a4b0 + H]+, 100%), 455.52 ([a4b2 + H]+, 99%), 509.54 ([a4b6 + H2O + H]+,

32%), 527.39 ([a4b6 + 2H2O + H]+, 28%); for a = 5, m/z: 537.45 ([a5b1 – H2O + H]+,

40%), 555.56 ([a5b1 + H]+, 25%), 573.48 ([a5b3 + H]+, 52%), 627.43 ([a5b7 + H2O +

H]+, 14%); for a = 6, m/z: 655.35 ([a6b0 + H]+, 15%), 673.39 ([a6b2 + H]+, 33%),

691.25 ([a6b4 + H]+, 11%); for a = 7, m/z: 773.37 ([a7b1 + H]+, 14%), 791.17 ([a7b3 +

H]+, 7%), 809.34 ([a7b5 + H]+, 9%); for a = 8, m/z: 873.34 ([a8b0 + H]+, 4%), 891.20

([a8b2 + H]+, 6%), 909.06 ([a8b4 + H]+, 6%), 927.17 ([a8b6 + H]+, 5%); for a = 9, m/z:

991.30 ([a9b1 + H]+, 7%), 1027.15 ([a9b5 + H]+, 9%), 1045.32 ([a9b7 + H]+, 6%); for

a = 10, m/z: 1145.23 ([a10b6 + H]+, 11%).

Silylated silsesquioxanes (cone voltage = 15 V): for a = 3, m/z: 427.66 ([silylated a3b3

+ H]+, 37%), 499.61 ([disilylated a3b3 + H]+, 22%); for a = 4, m/z: 455.64 ([a4b2 +

H]+, 16%), 527.53 ([silylated a4b2 + H]+, 51%), 545.53 ([silylated a4b4 + H]+, 38%),


141

599.53 ([disilylated a4b2 + H]+, 22%), 617.60 ([disilylated a4b4 + H]+, 100%), for

a = 5, m/z: 645.52 ([silylated a5b3 + H]+, 37%), 717.46 ([disilylated a5b3 + H]+, 17%),

735.40 ([disilylated a5b5 + H]+, 19%), 807.41 ([trisilylated a5b5 + H]+, 11%); for a = 6,

m/z: 763.45 ([silylated a6b4 + H]+, 7%), 835.33 ([disilylated a6b4 + H]+, 7%); for a = 7,

m/z: 953.32 ([disilylated a7b5 + H]+, 7%).

Chapter 7 __________________________________________________________________________________________________________

142

References

1 P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, Angew. Chem. Int. Ed., 2001, 40,

740, see also Chapter 3. 2 P.P. Pescarmona, J.J.T. Rops, J.C. van der Waal, J.C. Jansen, T. Maschmeyer, J. Mol. Cat. A, 2002,

182-183, 319, see also Chapter 3. 3 P.P. Pescarmona, T. Maschmeyer, NATO Science Series, 2002, Ser. II Vol. 69, 173, see also Chapter 4. 4 M. Crocker, R.H.M. Herold, A.G. Orpen, Chem. Commun., 1997, 2411. 5 T. Maschmeyer, M.C. Klunduk, C.M. Martin, D.S. Shephard, J.M. Thomas, B.F.G. Johnson, Chem.

Commun.,1997, 1847. 6 P.P. Pescarmona, J.C. van der Waal, T. Maschmeyer, Catal. Today, 2003, 81, 347, see also Chapter 3. 7 J.F. Brown, L.H. Vogt, J. Am. Chem. Soc., 1965, 87, 4313. 8 P.P. Pescarmona, T. Maschmeyer, Aust. J. Chem., 2001, 54, 583, see also Chapter 2. 9 F.J. Feher, J.J. Schwab, D. Soulivong, J.W. Ziller, Main Group Chem., 1997, 2, 123. 10 J.F. Brown, L.H. Vogt, P.I. Prescott, J. Am. Chem. Soc., 1964, 86, 1120. 11 G. Blanco-Brieva, J.M. Campos-Martín, M.P. de Frutos, J.L.G. Fierro, Chem. Commun., 2001, 2228. 12 P.D. Lickiss, S.A. Litster, A.D. Redhouse, C.J. Wisener, Chem. Commun., 1991, 173. 13 N. Winkhofer, H.W. Roesky, M. Noltemeyer, W.T. Robinson, Angew. Chem. Int. Ed., 1992, 31, 599. 14 S. Krijnen, H.C.L. Abbenhuis, R.W.J.M. Hanssen, J.H.C. van Hooff, R.A. van Santen, Angew. Chem.

Int. Ed., 1998, 37, 356. 15 H.C.L. Abbenhuis, S. Krijnen, R.A. van Santen, Chem. Commun., 1997, 331. 16 E. Albizzati, L. Abis, E. Pettenati, E. Giannetti, Inorg. Chim. Acta, 1986, 120, 197. 17 S.I. Chervina, E.G. Maksimenko, R.S. Barshtein, N.V. Shabanova, A.K. Bulai, Y.I. Kotov, I.Y. Slonim,

Kinet. Catal., 1987, 28, 947. 18 M. Altaf Hossain, M.B. Hursthouse, M.A. Mazid, A.C. Sullivan, Chem. Commun., 1988, 1305. 19 T. Maschmeyer, F. Rey, G. Sankar, J.M. Thomas, Nature, 1995, 378, 159. 20 S. Thorimbert, S. Klein, W.F. Maier, Tetrahedron, 1995, 51, 3787. 21 R. Hutter, T. Mallat, A. Baiker, J. Catal., 1995, 153, 177. 22 A. Corma, U. Díaz, V. Fornés, J.L. Jordá, M. Domine, F. Rey, Chem. Commun., 1999, 779. 23 Z. Shan, E. Gianotti, J.C. Jansen, J.A. Peters, L. Marchese, T. Maschmeyer, Chem. Eur. J., 2001, 7,

1437. 24 J. Jarupatrakorn, T.D. Tilley, J. Am. Chem. Soc., 2002, 124, 8380. 25 F.J. Feher, D.A. Newman, J. Am. Chem. Soc., 1990, 112, 1931. 26 P.P. Pescarmona, J.C. van der Waal, T. Maschmeyer, submitted to Eur. J. Inorg. Chem., see also

Chapter 5. 27 See Appendix A.

143

8

Study of the synthesis of zeolite beta using High-Speed Experimentation

Abstract

Zeolites are a well-known family of crystalline microporous materials with broad

applications as heterogeneous catalysts. The application of HSE techniques to the study

of the synthesis of aluminium-rich zeolite beta (Si/Al ratio from 2.5 to 5) allowed the

investigation of the effect of the Si/Al and TEA+/Al ratios on the formation of such

zeolite. A Si/Al = 5 was found as the lowest ratio to produce pure zeolite beta. Lower

Si/Al ratios led to mixtures of zeolites (mainly beta and NaP1).

____________________

The contents of this chapter have been published in:

P.P. Pescarmona, J.J.T. Rops, J.C. van der Waal, J.C. Jansen, T. Maschmeyer, J. Mol. Catal. A, 2002, 182-183, 319.

P.P. Pescarmona, T. Maschmeyer, NATO Science Series, 2002, Ser. II Vol. 69, 173.

Chapter 8 __________________________________________________________________________________________________________

144

8.1 Introduction

In 1756, Crønstedt reported the discovery of the first natural zeolite.1,2 Upon

heating, the zeolite released occluded water: this property gave the material its general

name, zeolite, after the Greek zeo (to boil) and lithos (stone). Since then, many zeolite

structures - both natural and synthetic - have been discovered and studied.

Zeolites3 are crystalline aluminosilicates characterised by a framework consisting of Si

and Al atoms connected by oxygen bridges. Each Si4+ and Al3+ atom is tetrahedrally

surrounded by four O2- atoms (Figure 8.1). The presence of Al3+ results in a negative

charge into the framework, which is compensated by the presence of cations (e.g. H+,

Na+, Ca2+). In the case these counterions are protons, the zeolite exhibits Brønsted

acidity. Since Al-O-Al bonds are not encountered in zeolites (Löwenstein rule), the

Si/Al molar ratio is ≥ 1. By varying the Si/Al ratio of a zeolite, the amount and strength

of acid sites can be changed: the number of acid sites increases with increasing amount

of Al atoms while their strength decreases. This decrease in acidity is caused by the

lower electronegativity of AlO4-tetrahedral units compared to SiO4-tetrahedral units.

Figure 8.1. Schematic representation of a zeolite in the H-form.

The SiO4- and AlO4-tetrahedral units can be connected in many different ways, thus

generating a variety of possible zeolite structures. The arrangement of the tetrahedral

units generates frameworks characterised by three-dimensional networks of cages and

channels. The pore openings are in the 0.3-1.0 nm range: a comparable size to that of

small molecules. For these characteristics, zeolites are referred to as crystalline

microporous materials with uniform pore dimensions.

The properties described above account for the many applications of zeolites, and

OSi

OO

OAl

O O

OSi

OSi

O OO O

OAl

OSi

O

O OO O

H H

OSi

OO

OAl

O O

OSi

OSi

O OO O

OAl

OSi

O

O OO O

H H

Synthesis of zeolite beta by means of HSE __________________________________________________________________________________________________________

145

particularly: size and shape selective heterogeneous catalysts, ion-exchangers and

molecular sieves.3

8.1.1 Zeolite beta

Zeolite beta (BEA) is a so-called large pore zeolite characterised by three

families of mutually perpendicular channels with 12-membered ring apertures

(Figure 8.2),4,5 which make it suitable as heterogeneous catalyst for numerous organic

reactions6 (e.g. incorporation of titanium into zeolite beta leads to a versatile oxidation

and Lewis acid catalyst).7,8 Synthetic zeolite beta can be prepared with a wide range of

Si/Al ratios (from 5 to ∞).9,10 The control of the Si/Al ratio is important since it allows

tuning the number and strength of acid sites as well as the hydrophobicity of the zeolite.

Here, the optimisation of the synthesis of aluminium-rich zeolite beta in order to reach

the lowest possible Si/Al ratio is described. The synthesis of an aluminium-rich zeolite

beta is particularly interesting also because Tschernichite, the natural form of beta, has

Si/Al = 3.11

Figure 8.2. Three-dimensional structure of zeolite beta viewed along [100].

0.76 nm0.76 nm

Chapter 8 __________________________________________________________________________________________________________

146

The preparation of zeolites involves three main constituents: the source of the

framework elements (Si and Al), the mobiliser or mineraliser (OH- or F-), which

provides the appropriate pH for the reaction mixture, and the template (typically an

organic or inorganic cation), which directs the synthesis towards the formation of a

specific zeolite structure. Zeolite beta is commonly synthesised by hydrothermal

crystallisation from alkaline media, using tetraethylammonium (C2H5)4N+ (TEA+) as the

template.12,13


Similarly to what has been described for the formation of silsesquioxanes,14 the

synthesis of zeolites consists of a multiple-step reaction for which an overall mechanism

is not available. Various parameters determine which structures are formed and in

which amounts:12,15-17

- the nature of the silica and alumina sources and the ratio between the two

- the nature and the concentration of the template

- the presence, type and concentration of alkali cations

- the water content

- the pH

- the temperature

- the reaction time.

Given the complexity of the system, a High-Speed Experimentation18 approach

allowing the fast screening of broad parameter spaces is particularly suitable to study

and optimise the synthesis zeolite beta. Taking into account the good level of

knowledge available about the synthesis of zeolite beta, it was decided that no primary

screening was necessary. The HSE workstation19 employed in the project can only

handle low-viscosity liquids: this influenced the choice of the silica and alumina sources

used in the experiments. As silica and alumina sources, the cheap and easy to handle

Ludox-HS 40 colloidal silica and an aqueous solution of sodium aluminate were chosen.

Hydrothermal synthesis using tetraethylammonium ions (TEA+) as templates in an

alkaline environment was chosen among the methods commonly used for the

preparation of aluminium-rich zeolite beta for its adaptability to the HSE workstation.


147

The method applied is similar to that described by Borade and Clearfield,20,21 with a

different silica source and a higher water content in order to allow handling by the HSE

workstation. Given these restrictions, the Si/Al ratio and the TEA+/Al ratio in the

starting reaction mixture were selected as the most relevant parameters determining the

formation of aluminium-rich zeolite beta. The screening was performed by means of an

X-ray powder diffractometer supplied with an automated sampler. The intensity of the

strongest peak in the diffractogram for each zeolite species present was taken as a

measure of crystallinity (Figure 8.3).

Figure 8.3. X-ray diffractogram of zeolite beta.


The parameter space chosen was defined by the combination of 6 Si/Al ratios and

4 TEA+/Al ratios and screened as a function of the presence of zeolite beta. The range in

which the two parameters were varied had been chosen in order to investigate the

possibility of lowering the Si/Al in zeolite beta and the amount of the expensive TEA+

employed with respect to the synthesis method proposed by Borade and Clearfield.20,21

5 15 25 35 45

angle 2θ (degrees)

inte

nsity

Chapter 8 __________________________________________________________________________________________________________

148

Figure 8.4. Presence of zeolite beta (top) and of zeolite NaP1 (bottom) in the screened

parameter space.

54.5

43.5

32.5

1.25

1

0.75

0.5

0

500

1000

1500

2000

2500

3000

Si/Al ratio

TEA+/Al ratio

intensity (a.u.)

Zeolite beta

5 4.5 4 3.5 32.5

1.25

1

0.75

0.5

0

500

1000

1500

2000

2500

3000

Si/Al ratio

TEA+/Al ratio

Zeolite NaP1

intensity (a.u.)

54.5

43.5

32.5

1.25

1

0.75

0.5

0

500

1000

1500

2000

2500

3000

Si/Al ratio

TEA+/Al ratio

intensity (a.u.)

Zeolite beta

5 4.5 4 3.5 32.5

1.25

1

0.75

0.5

0

500

1000

1500

2000

2500

3000

Si/Al ratio

TEA+/Al ratio

Zeolite NaP1

intensity (a.u.)


149

The products of the 24 syntheses mainly consist of a mixture of two zeolites: zeolite

beta and zeolite NaP1 (Figure 8.4). NaP1 is a gismondine-type of zeolite (GIS), with a

Si/Al ratio of 1.5 and Na+ ions as template ions. Hence, NaP1 is found as a product at

low values of Si/Al ratio in the starting reaction mixture, while beta is formed at higher

values of Si/Al ratio. Pure zeolite beta was obtained at a Si/Al ratio of 5 and

TEA+/Al ratio of 1.25 in the starting reaction mixture, while pure NaP1 was obtained at

a Si/Al ratio of 2.5 for the entire screened TEA+/Al ratio. In between, mixtures of both

zeolites were observed. Some samples also contained amorphous silica and small

amounts of crystalline impurities (consistent with analcime (ANA)). A higher value of

TEA+/Al ratio causes a positive trend towards the formation of zeolite beta, as

competition takes place between the two templates (Na+ for NaP1 and TEA+ for beta).

This experiment suggests that a Si/Al ratio of 5 and a TEA+/Al ratio of 1.25 in the

original reaction mixture are the limiting values to obtain pure zeolite beta under

hydrothermal synthesis conditions and using TEA+ as the template. The corresponding

sample was further characterised to determine the actual Si/Al in the zeolite framework.

From ICP OES elemental analysis, a Si/Al ratio of 4.7 and an Al/Na ratio of 0.8 were

determined. 27Al NMR was used to check if all the aluminium species in the sample

were part of the zeolite framework. The 27Al NMR spectrum shows a high-intensity

peak at 59 ppm corresponding to tetrahedral aluminium species and a low-intensity

peak at 4 ppm corresponding to octahedral aluminium species (Figure 8.5). Tetrahedral

aluminium species are assigned to aluminium present in the zeolite structure, while

octahedral species are likely to be due to non-zeolitic aluminium or extra-framework

species. The ratios between the integrals of the two peaks is 92/8. This implies that a

maximum of 8% of the total aluminium is not part of the zeolite framework.

Considering just the tetrahedral aluminium as part of the zeolite framework, the

Si/Al ratio has then to be corrected to a value of 5.1. This value is slightly higher than

the Si/Al ratio of 4.5 reported by Borade and Clearfield20,21 and is obtained using a

higher TEA+/Al ratio (1.25 against 0.8). The highest TEA+/Al ratio needed can be a

consequence of the higher water content used in this method. A relevant advantage of

the synthesis method described here is the use of Ludox HS-40 as silica source instead

of fumed silica, since the former is cheaper, easier and safer to handle and forms a less

dense synthesis gel.

Chapter 8 __________________________________________________________________________________________________________

150

Figure 8.5. 27Al NMR of the pure zeolite beta obtained with Si/Al = 5 and

TEA+/Al = 1.25 in the starting reaction mixture.

8.3.1 Up-scaling of the HSE lead

Up-scaling of the synthesis of zeolite beta with a Si/Al ratio of 5 and a TEA+/Al

ratio of 1.25 in the original reaction mixture to a 50-ml autoclave led to a similar result

though small amounts of NaP1 were present as by-products.

To further study the effects of differences in scale on the crystallisation of zeolites, a

synthesis gel for zeolite beta was crystallised in a 50-ml autoclave and in 3-ml

autoclaves in a HSE 24-array. Zeolite beta with a Si/Al ratio ~ 11 was synthesised using

the method proposed by Pérez-Pariente et al.,12 which has been verified by the

International Zeolite Association (IZA). Analysing the synthesis gel after only 20 hours

of crystallisation results in amorphous material for the 50-ml autoclaves, while in the

3-ml autoclaves the formation of zeolite beta already started, as can be seen in

Figure 8.6. After 40 hours of crystallisation in the 50-ml autoclaves, pure crystalline

zeolite beta is obtained using this method.

Differences in results using HSE 3-ml autoclaves and conventional larger-volume

autoclaves are probably due to the higher heating rate of the small autoclaves, which is

mainly caused by their lower thermal conductance (due to the thinner walls of the


151

autoclaves and of the inserts), and to the higher surface/volume ratio of the small

autoclaves, which favours heterogeneous nucleation and, therefore, speeds up the

crystallisation process.

Figure 8.6. Crystallisation of zeolite beta as a function of the synthesis time and of the

reactor scale.

8.4 Conclusions

In this chapter it has been shown that High-Speed Experimentation techniques

can be successfully applied to study the synthesis of zeolites. This application, together

with the results of the HSE of silsesquioxanes reported in the previous chapters, shows

the versatility of HSE techniques as a tool for research for both heterogeneous and

homogeneous catalysts. From this HSE study, it has been possible to identify a facile

and less costly method to produce pure zeolite beta and pure zeolite NaP1. Moreover,

the effect of the Si/Al and TEA+/Al ratios on the formation of these two zeolites was

studied.

Time

Scale

Synthesis of zeolite beta in a 50-ml autoclave(20 hours of crystallisation)

0

500

1000

1500

5 15 25 35 45

angle 2θ (degrees)

inte

nsity

(a.u

.)

0

500

1000

1500

2000

5 15 25 35 45

inte

nsity

(a.u

.)


angle 2θ (degrees)

010002000300040005000

5 15 25 35 45


angle 2θ (degrees)

inte

nsity

(a.u

.)

Time

Scale


0

500

1000

1500

5 15 25 35 45

angle 2θ (degrees)

inte

nsity

(a.u

.)


0

500

1000

1500

5 15 25 35 45

angle 2θ (degrees)

inte

nsity

(a.u

.)

0

500

1000

1500

2000

5 15 25 35 45

inte

nsity

(a.u

.)


angle 2θ (degrees)

0

500

1000

1500

2000

5 15 25 35 45

inte

nsity

(a.u

.)


angle 2θ (degrees)

010002000300040005000

5 15 25 35 45


angle 2θ (degrees)

inte

nsity

(a.u

.)

010002000300040005000

5 15 25 35 45


angle 2θ (degrees)

inte

nsity

(a.u

.)

Chapter 8 __________________________________________________________________________________________________________

152

8.5 Experimental

Experiments were performed on an automated parallel synthesis workstation19

coupled with a personal computer supplied with software enabling to program the

workstation. Reactants: Ludox-HS 40 colloidal silica (40%wt suspension of SiO2 in

water) as silica source; sodium aluminate (40-45%wt Na2O, 50-56%wt Al2O3) as alumina

source; tetraethylammonium hydroxide (TEAOH) (35% solution in water) as template

source; deionised water. Appropriate amounts of reactants for the zeolite synthesis were

dispensed by the D-arm of the workstation in a 6×4 matrix of small 3-ml teflon-lined

stainless steel autoclaves. An equal amount of alumina source was dispensed in each

autoclave in order to set the pH of all the samples at ~ 14. In a typical experiment, after

dispensing the alumina, the template and the silica sources, the autoclaves were filled up

to 75% of their volume with water. Next, the obtained aluminosilicate gels were stirred

in an ultrasonic bath (T ≈ 60°C) for 1 hour. The autoclaves were then closed and heated

statically at 170 °C. After 48 hours, heating was stopped and the autoclaves were cooled

down to room temperature. Finally, the autoclaves with the formed solids were washed

twice with deionised water, centrifuged at 1500 r.p.m. for 1 hour and dried in a vacuum

centrifuge.19 After the experiment, the teflon inserts were cleaned overnight with NaOH

(4M) at reaction temperature or with HF (10%) at room temperature.

Powder diffraction data were collected on a Philips PW1840 diffractometer generating

Cu Kα radiation. Data were obtained from 5° to 50° 2θ with a step of 0.01° and a

counting time of 0.5 s/step. The receiving slit had a width of 0.3 mm. The tube voltage

was 40 kV and the tube current was 50 mA. In order to reduce the analysis time, an

automatic sampler was installed, thus allowing X-ray analysis of 24 samples in

15 hours. Elemental analyses were performed on a Perkin-Elmer 3000 DV ICP OES.

After dissolution of the zeolites in a 1% HF (40%)/1.3% H2SO4 solution, the

concentrations of silicon, aluminium and sodium were measured with induced coupled

plasma optical emission spectroscopy (ICP OES). The samples were analysed twice as

independent duplicates to get an indication of the precision of the analysis. The

accuracy in the analysis of Si and Al is estimated at about ± 5%, for Na at ±10%.

High-resolution solid-state 27Al MAS NMR spectra were recorded on a Varian

VXR-400S spectrometer.


153

All the HSE results reported in this chapter are averages of the values obtained in

different sets of experiments. The deviation of the values of each set of experiments

from the average is larger than what was found for the HSE study of the synthesis of

silsesquioxanes.22 This means that while the identified trends are reliable, the absolute

values suffer of a level of uncertainity. For example, considering the results reported in

Figure 8.4, for Si/Al ratio = 2.5, the yield in NaP1 is not constant for all the 4 TEA+/Al

ratios: this is more likely to be due to small experimental fluctuations rather than being

related to the different TEA+/Al ratio. The experimental error of these HSE results is

bound to be connected to technical problems that occurred during the synthesis of

zeolites, particularly those caused by the leaking of the autoclaves.

Chapter 8 __________________________________________________________________________________________________________

154

References

1 A.F. Crønstedt, Akad. Handl. Stockholm, 1756, 18, 120. 2 R.W. Tschernich, Zeolites of the world, Geoscience, 1992. 3 H. van Bekkum, E.M. Flanigen, P.A. Jacobs, J.C. Jansen (editors), Introduction to Zeolite Science and

Practice, Elsevier, 2001, 2nd edition. 4 M.M.J. Treacy, J. M. Newsam, Nature, 1988, 332, 249. 5 J.M. Newsam, M.M.J. Treacy, W. T. Koetsier, C. B. De Gruyter, Proc. R. Soc. Lond. A, 1988, 420, 375. 6 P.B. Venuto, Microporous Mater., 1994, 2, 297. 7 J.C van der Waal, M.S. Rigutto, H. van Bekkum, Appl. Catal. A-Gen., 1998, 167, 331. 8 J.C van der Waal, P.J. Kunkeler, K. Tan, H. van Bekkum, J. Catal., 1998, 173, 74. 9 R.L. Wadlinger, G.T. Kerr, E.J Rosinski, US Pat. Appl. 3.308.069, 1967. 10 J.C van der Waal, M.S. Rigutto, H. van Bekkum, J. Chem. Soc., Chem. Commun., 1994, 1241. 11 J.V. Smith, J.J. Pluth, R.C. Boggs, D.G. Howard, J. Chem. Soc., Chem. Commun., 1991, 363. 12 J. Pérez-Pariente, J.A. Martens, P.A. Jacobs, Appl. Catal., 1987, 31, 35. 13 F. Vaudry, F. Di Renzo, P. Espiau, F. Fajula, P. Schulz, Zeolites, 1997, 19, 253. 14 P.P. Pescarmona, T. Maschmeyer, Aust. J. Chem., 2001, 54, 583, see also Chapter 2. 15 M.A. Camblor, J. Pérez-Pariente, Zeolites, 1991, 11, 202. 16 M.A. Camblor, A. Misfud, J. Pérez-Pariente, Zeolites, 1991, 11, 792. 17 M.J. Eapen, K.S.N. Reddy, V.P. Shiralkar, Zeolites, 1994, 14, 294. 18 P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, Catal. Lett., 1999, 63, 1, see also

Chapter 1. 19 See Appendix A. 20 R.B. Borade, A. Clearfield, Microporous Mater., 1996, 5, 289. 21 R.B. Borade, A. Clearfield, Chem. Commun., 1996, 625. 22 See Paragraph 3.5.

The HSE equipment __________________________________________________________________________________________________________

155

Appendix A: High-Speed Experimentation equipment

A.1 The HSE automated workstation

A Hamilton Dual Arm liquid-handling robotic workstation was used for the HSE

synthesis of samples described and discussed in this thesis (Figure A.1). The robot is

suited for dispensing and transferring chemicals in the liquid phase by means of two

types of devices: the single needle arm and the D-arm.

Figure A.1. The Hamilton robot.

1) Two kind of operations can be performed using the single stainless steel needle arm:

A. Transfer set amounts of liquids from a rack of vials to another.

B. Dispense desired amounts of liquids to a rack of vials from an array of 24 bottles

containing different liquids, connected to the needle through 3 teflon valves each

with 8 exits (Figure A.2).

Appendix A __________________________________________________________________________________________________________

156

Both operations are carried out by means of two syringes (5 ml and 1 ml volume) that

can aspirate and dispense set amounts of liquids accurately and precisely, with an error

of ± 3 µl. The syringes use a perfluoro compound as reservoir liquid to prevent the

syringes from contamination with the transferred or dispensed liquids (a modification

that was part of this dissertation).1 Perfluroalkanes are used since they are immiscible

with the majority of other liquids and inert. The set-up is also equipped with an in-house

designed junction,1 which prevents the formation of bubbles at the interface between the

reservoir liquid and the transferred or dispensed liquid. Bottles, valves, syringes,

junction and needle are connected to each other by teflon lines with an internal diameter

of ~1.6 mm (1/16 inch) (Figure A.2).

Figure A.2. Schematic representation of the single needle arm part of the robot.

2) The D-arm is used to transfer liquids even more precisely than with the single needle

arm. This is achieved by means of a positive air displacement device similar to that used

in the disposable-tip pipettes employed in many biochemical laboratories. The arm

picks up a new disposable plastic tip for each transfer operation, therefore excluding

reservoirsyringe 1

bottle13

bottle 14

bottle 12

bottle 9

bottle 10

bottle 11

bottle 16

bottle15

valve 4

needle

bottle 21

bottle22

bottle20

bottle 17

bottle18

bottle 19

bottle 24

bottle 23

bottle 5

bottle 6

bottle 4

bottle 1

bottle 2

bottle 3

bottle 8

bottle 7

vials racks

waste

reservoirliquid

sample

valve2

valve 3

valve 1

syringe 2

junction

The HSE equipment __________________________________________________________________________________________________________

157

contamination risks. Another advantage is that the D-arm can dispense more viscous

liquids than the single needle arm. On the other hand, differently from the single needle

arm, it cannot dispense liquids through septa and is therefore not adequate for

experiments that require a closed environment (e.g. air-sensitive compounds that require

an inert atmosphere).

Various kinds of racks have been used in this thesis, with variable number of vessels.

For silsesquioxane experiments, the most common reaction racks consisted of 6×4

arrays of 24 glass tubes, each with a working volume of 3.5 ml. The tubes can be

closed with a set of teflon septa, therefore allowing reactions in a closed environment

(Figure A.3). For the synthesis of zeolites, the racks consisted of a 6×4 matrix of 24

small 3-ml teflon-lined stainless steel autoclaves, each supplied with a safety valve to

avoid pressure to exceed the value of 30 bar (Figure A.3).

All the operations performed by the robot are controlled by a computer supplied with a

Hamilton software. An in-house written module for this software allows programming

of the robot by writing simple files in a Microsoft Excel format.

Figure A.3. 6×4 glass-tube rack (left) and 6×4 autoclave rack (right).

A.2 The heating blocks

Steel heating blocks able to arrange the 6×4 type of racks were used to heat the

samples for a selected time to a desired temperature, with a maximum allowed value of

Appendix A __________________________________________________________________________________________________________

158

500°C. The blocks are placed on an orbital shaker that can provide a gentle stirring of

the liquid samples contained in the racks, with a maximum allowed rotation speed of

525 rpm.

A.3 The vacuum centrifuge

A centrifuge connected to a vacuum pump was used to remove solvents from

samples. This vacuum centrifuge can arrange up to 4 6×4 racks. The chamber of the

centrifuge is also connected to a gas-dispensing line, allowing setting the samples under

argon (inert) atmosphere.

Note: all the equipment described in this appendix was kindly provided by Avantium

Technologies, Amsterdam-Delft.

References

1 P. J. van den Brink, P.P. Pescarmona, J.C. van der Waal, Patent EPA99310598, WO0148443,

AU2389701, 1999.

Glossary __________________________________________________________________________________________________________

159

Appendix B: Glossary

B.1 Concepts

Parameter space: the space defined by the combination of each element of a chosen set

of experimental parameters with each element of each of the other chosen sets of

experimental parameters (see Chapter 1). The dimension of the parameter space is

given by the number of sets of experimental parameters; the number of elements

contained in a parameter space is given by the product of the number of elements

presents in each of the sets of experimental parameters. As an example, in the

optimisation of the synthesis of the silsesquioxane precursors reported in

Chapter 3, two parameters were investigated: the solvent and the R-group. The

set of solvents consisted of 4 elements, while the set of R-groups consisted of 10

elements. Therefore, the parameter space had dimension 2 and was constituted of

40 elements.

Library: a collection of compounds related to each other by a common feature (e.g. a

library of catalysts, a library of products). Combinatorial Chemistry and High-Speed

Experimentation techniques allow fast preparation and screening of libraries of products

(see Chapter 1).

Primary screening: a first, general investigation of the system under study (see

Paragraph 1.3).

Secondary screening: a successive, more focussed investigation of the system under

study (see Paragraph 1.3).

Appendix B __________________________________________________________________________________________________________

160

B.2 Abbreviations

APCI: Atmospheric Pressure Chemical Ionisation. ATR FTIR: Attenuated Total Reflection Fourier Transform Infrared. Bu: butyl. But: tert-butyl. CC: Combinatorial Chemistry. DMSO: dimethyl sulfoxide. ESI: Electrospray Ionisation. Et: ethyl. GC: Gas Chromatography. HSE: High-Speed Experimentation. IR: Infrared. MCR: Multivariate Curve Resolution. Me: methyl. MS: Mass Spectrometry. NMO: N-methylmorpholine-N-oxide. NMR: Nuclear Magnetic Resonance. PCA: Principal Component Analysis. Ph: phenyl. Pr: propyl. Pri: isopropyl. TBHP: tert-butyl hydroperoxide. TEA+/TEAOH: tetraethylammonium (ion)/tetraethylammonium hydroxide. THF: tetrahydrofuran.

Summary __________________________________________________________________________________________________________

Paolo P. Pescarmona: “An Exploration of Silsesquioxanes and Zeolites using

High-Speed Experimentation”

Summary

This thesis describes the most interesting results of the Ph.D. research of the

author. As stated in the title, the subject of this research is the exploration of

silsesquioxanes (synthesis and catalysis) and zeolites (synthesis) using High-Speed

Experimentation (HSE).

Combinatorial Chemistry and High-Speed Experimentation techniques are recently

developed methods with an increasing number of applications to many fields of

research: originally introduced in the search of pharmaceutical active compounds, these

techniques have now found useful applications in other fields such as catalysis and

materials science. One of the basic ideas behind Combinatorial Chemistry and

High-Speed Experimentation is to enable an evolutionary process in order to identify

suitable compounds or conditions for a chosen purpose. This means that with these

techniques many parameters are combined to produce large numbers of samples that are

then tested for the desired property. Automated workstations are normally employed to

prepare and screen these samples. In this dissertation, the attention is focussed on the

application of High-Speed Experimentation techniques to catalysis. (Chapter 1).

The study and optimisation of the synthesis of silsesquioxanes is one of the main

subjects of this thesis. Silsesquioxanes are a family of compounds of the general

formula (RSiO1.5)a(H2O)0.5b. They are constituted of tetrahedral units in which a silicon

is connected to three oxygens and to an organic group R (or to a hydrogen). Therefore,

silsesquioxanes are characterised by a Si-O framework similar to that of silica or

zeolites. The presence of the organic groups makes these compounds soluble in a

number of organic solvents. Here, particular attention is given to incompletely

condensed silsesquioxanes, which contain free silanol groups on which metal centres

can be anchored. Incompletely condensed silsesquioxanes have found applications as

model compounds for silica surfaces and as ligands for homogeneous catalysts.

(Chapter 2).

161

Summary __________________________________________________________________________________________________________

Complexes of titanium with silsesquioxane a7b3 have been used as models for

heterogeneous silica-based catalysts and as homogenous catalysts for the epoxidation of

alkenes. Here, the synthesis of silsesquioxane precursors for Ti-catalysts was studied by

means of High-Speed Experimentation with the aim to either improve the otherwise

long and expensive preparation method of silsesquioxane a7b3 or to generate other

silsesquioxane structures suitable for forming active epoxidation Ti-catalysts. For this

purpose, the hydrolytic condensation of trichlorosilanes to produce silsesquioxanes was

optimised as a function of the epoxidation activity of the catalysts obtained after

reaction of the silsesquioxane precursors with a titanium centre. In a first stage, the two

parameters that were assumed to have the most relevant influence on determining the

selectivity and the yield of the synthesis of silsesquioxanes were studied: the R-group of

the organotrichlorosilanes RSiCl3 and the solvent in which the hydrolytic condensation

occurs. The High-Speed Experimentation approach allowed the rapid screening of a

large number of trichlorosilanes and solvents and, consequently, the identification of

some trends concerning these two experimental parameters. The best silsesquioxane

precursor for an epoxidation catalyst was obtained by the hydrolytic condensation of

cyclopentyltrichlorosilane in acetonitrile. Characterisation by means of NMR and MS

showed that the lead contained a mixture of silsesquioxanes, with silsesquioxane a7b3

as the main component. (Chapter 3).

In a second stage, the best synthesis methods identified were further optimised by

varying the other parameters influencing the hydrolytic condensation. This fine-tuning,

also performed by means of High-Speed Experimentation, allowed adjusting the

conditions for the synthesis of the silsesquioxane precursors and to gain knowledge

about the effect of the various parameters. Next, the effect of the nature of the titanium

centre bound to the silsesquioxane was investigated using High-Speed Experimentation.

(Chapter 4).

The two best Ti-catalysts identified by means of HSE were obtained with cyclopentyl

and cyclohexyl silsesquioxanes synthesised in acetonitrile. The synthesis of the two

precursors was further improved and the silsesquioxane products were fully

characterised. In the case of cyclopentyl silsesquioxanes, structure a7b3 was selectively

obtained in 64% yield after 18 hours of reaction – a significant improvement compared

to the formerly known methods. The synthesis was monitored by MS and in-situ IR in

order to identify the silsesquioxanes present in the reaction mixture, to explain the high

162

Summary __________________________________________________________________________________________________________

selectivity of the reaction towards structure a7b3 and to propose a mechanism for its

formation. In the case of cyclohexyl silsesquioxanes, the synthesis had a high yield but

was less selective than in the case of cyclopentyl. The main products were:

silsesquioxanes a6b0, a6b2, a7b1 and a7b3. The role of acetonitrile as a reactive

solvent in the syntheses was ascertained. (Chapter 5).

Cyclopentyl silsesquioxane a7b3, synthesised with the new, improved method, was

used to prepare an Os-complex in which the poisonous OsO4 is rendered stable. This

Os-silsesquioxane complex was used both as model compound for a silica-supported,

heterogeneous Os-catalyst and as homogeneous catalyst for the dihydroxylation of

alkenes. The Os-silsesquioxane catalyst displayed higher turnover frequencies than the

heterogeneous one, while keeping the same selectivity (99%). (Chapter 6).

The first screening of the synthesis of silsesquioxane precursors for epoxidation

Ti-catalysts using High-Speed Experimentation showed a positive effect of the polarity

of the solvent (Chapter 3). This trend encouraged the investigation of the effect of other

highly polar solvents on the synthesis of the silsesquioxane precursors. The results of

this HSE study did not show clear trends but allowed the identification of a number of

very active epoxidation Ti-catalysts, the most promising of which were fully

characterised: each lead consisted of a different type of silsesquioxane structures. The

lead synthesised by the hydrolytic condensation of ButSiCl3 in water was exclusively

constituted of silsesquioxane a2b4. This was the first reported example of the use of this

silsesquioxane structure to prepare active Ti-catalysts. These complexes were supported

on silica to produce an active, non-leaching, heterogeneous, epoxidation catalyst. The

lead obtained by the hydrolytic condensation of ButSiCl3 in DMSO consisted of a

number of incompletely condensed structures (mainly with 3 < a < 5). The lead

produced by the hydrolytic condensation of PhSiCl3 in water consisted of polymeric

silsesquioxanes. (Chapter 7).

After the successful application of High-Speed Experimentation to the synthesis of

silsesquioxanes, these techniques were used to study the synthesis of zeolite beta.

Zeolites are well-known microporous crystalline materials with many applications as

heterogeneous catalysts, ion-exchangers and molecular sieves. Here, the synthesis of

aluminium-rich zeolite beta was investigated by performing the synthesis with different

Si/Al and TEA+/Al ratios (where TEA+ is the template). It was found that Si/Al = 5 and

TEA+/Al = 1.25 in the starting reaction mixture are the lowest ratios to obtain pure

163

Summary __________________________________________________________________________________________________________

zeolite beta with the employed hydrothermal method. Mixtures of zeolites (mainly beta

and NaP1) were observed at lower Si/Al and TEA+/Al ratios. (Chapter 8).

Altogether, it was shown that High-Speed Experimentation techniques, by coupling

discovery and understanding, are a powerful research tool for applied and fundamental

research in the field of catalysis.

164

Samenvatting __________________________________________________________________________________________________________

Paolo P. Pescarmona: “Een Exploratie van Silsesquioxanen en Zeolieten door middel

van High-Speed Experimentation”.

Samenvatting

In dit proefschrift worden de meest interessante resultaten van het

promotieonderzoek van de auteur weergegeven. Zoals al blijkt uit de titel, is het

onderwerp van het promotieonderzoek de exploratie van silsesquioxaanchemie en

zeolietsynthese met behulp van High-Speed Experimentation (HSE) technieken.

Combinatorial Chemistry en High-Speed Experimentation technieken zijn recent

ontwikkelde methoden met toenemend aantal toepassingen binnen verschillende

weteschappenlijke disciplines. Oorspronklijk zijn ze ontwikkeld om farmaceutisch

actieve verbindingen te ontdekken, maar ze worden tegenwoordig ook toegepast in

andere vakgebieden, met name in de katalyse en de materiaalkunde. Het idee achter

Combinatorial Chemistry en High-Speed Experimentation is het vinden van de voor een

bepaald doel optimale verbinding of condities, door middel van een evolutie proces.

Met deze technieken, worden door systematische combinatie van veel parameters grote

aantallen monsters geproduceerd, die dan voor de gewenste eigenschap getest kunnen

worden. Geautomatiseerde werkstations worden gewoonlijk gebruikt om deze grote

aantallen monsters te bereiden en testen. Binnen dit promotie onderzoek is de aandacht

gericht op de toepassing van High-Speed Experimentation in de katalyse. (Hoofdstuk 1).

De bestudering en optimalisatie van de synthese van silsesquioxanen is een van de

belangrijkste onderwerpen van dit proefschrift. Silsesquioxanen zijn een familie van

organosilicium verbindingen met de algemene structuurformule (RSiO1.5)a(H2O)0.5b. Ze

bestaan uit tetraëders waarin een siliciumatoom gebonden is aan drie zuurstofatomen en

een organische groep (of een waterstofatoom). Ze bezitten daardoor een met silica en

zeolieten vergelijkbare Si-O-structuur. Dankzij de organische groepen zijn

silsesquioxanen in een aantal organische oplosmiddelen oplosbaar. In het

promotieonderzoek is de aandacht gericht op incompleet gecondenseerde

silsesquioxanen: structuren die nog silanol groepen bevatten, waardoor ze aan

metaalcentra verankerd kunnen worden. Dankzij deze eigenschappen zijn incompleet

gecondenseerde silsesquioxanen uitermate geschikt als modelverbindingen voor

silica-oppervlakken en als liganden voor homogene katalysatoren. (Hoofdstuk 2).

165

Samenvatting __________________________________________________________________________________________________________

Complexen van titanium met silsesquioxaan a7b3 zijn als model voor heterogene

silica-gedragen titaniumkatalysatoren en homogene titaniumkatalysatoren voor de

epoxidatie van alkenen bekend. In dit promotieonderzoek is vooral de synthese van

silsesquioxaan precursors voor titaniumkatalysatoren door middel van High-Speed

Experimentation bestudeerd. Het doel was de lange en dure synthesemethode voor

silsesquioxaan a7b3 te verbeteren of andere geschikte structuren te vinden om actieve

epoxidatie Ti-katalysatoren te genereren. Hiertoe werd de hydrolytische condensatie van

trichloorsilanen die silsesquioxanen produceert als functie van de epoxidatie activiteit

van de katalysatoren, verkregen door de reactie van het ruwe silsesquioxaanmengsel

met een titaniumcentrum, geoptimaliseerd. In de eerste fase zijn de twee parameters met

naar verwachting de grootste invloed op de selectiviteit en opbrengst van de

silsesquioxaansynthese bestudeerd: de organische R-groep van het organotrichloorsilaan

RSiCl3 en het oplosmiddel waarin de hydrolytische condensatie plaatsvindt. De

High-Speed Experimentation aanpak maakte een snelle test mogelijk van een groot

aantal trichloorsilanen en oplosmiddelen en, derhalve, de identificatie van belangrijke

trends over de twee bestudeerd experimenteel parameters. De beste

silsesquioxaanprecursor voor een epoxidatiekatalysator is door de hydrolytische

condensatie van cyclopentyltrichloorsilaan in acetonitril verkregen. Karakterisering

door NMR en MS toonde aan dat de beste lead uit een mengsel van silsesquioxanen

bestaat, waarvan structuur a7b3 de belangrijkste component is. (Hoofdstuk 3).

In een volgende fase werden de beste synthesemethoden verder geoptimaliseerd door

het variëren van de overige parameters die een invloed op de hydrolytische condensatie

hebben. High-Speed Experimentation leidde tot een verfijning van de synthesecondities

en tot nieuwe kennis over de effecten van de verschillende parameters. Tevens is het

effect van het aan het silsesquioxaan gebonden titaniumcentrum onderzocht.

(Hoofdstuk 4).

De twee beste katalysatoren, welke door middel van HSE geïdentificeerd waren, waren

de in acetonitril gesynthetiseerde cyclopentyl- en cyclohexylsilsesquioxanen. De

synthese van de twee silsesquioxanen werd verder verbeterd en de silsesquioxaan

producten werden volledig gekarakteriseerd. Voor cyclopentylsilsesquioxanen werd

structuur a7b3 selectief verkregen in 64% opbrengst na 18 uur reactie; een relevante

verbetering ten opzichte van de vroeger bekende methoden. Om de hoge selectiviteit

naar structuur a7b3 te verklaren en een synthetisch mechanisme voor te kunnen stellen,

166

Samenvatting __________________________________________________________________________________________________________

werd de synthese met MS en in-situ IR gevolgd om de verschillende silsesquioxaan

species in de reactiemengsel te identificeren. Ook voor cyclohexyltrichloorsilaan heeft

de synthese een hoge opbrengst maar de selectiviteit is lager dan die voor

cyclopentyltrichloorsilaan. De belangrijkste producten zijn silsesquioxanen a6b0, a6b2,

a7b1 en a7b3. Daarnaast is de rol van acetonitril als reagerend oplosmiddel

opgehelderd. (Hoofdstuk 5).

De cyclopentylsilsesquioxaan a7b3 werd daarna gebruikt om een Os-complex te

bereiden waarin OsO4 is geimmobiliseerd. Het Os-silsesquioxaan-complex is als model

voor een silica-gedragen heterogene katalysator en als homogene katalysator voor de

dihydroxylatie van alkenen gebruikt. De Os-silsesquioxaan katalysator toont hogere

turnover frequenties dan de heterogenene katalysator bij gelijke, hoge selectiviteit

(99%). (Hoofdstuk 6).

Een positief effect van de polariteit van het oplosmiddel was een belangrijke conclusie

uit eerder onderzoek van de synthese van silsesquioxaan precursors voor epoxidatie

Ti-katalysatoren door middel van High-Speed Experimentation (Hoofdstuk 3).

Aangemoedigd door de gevonden trend, richtte het onderzoek zich op het effect van

andere zeer polaire oplosmiddelen op de synthese van silsesquioxanen. De resultaten

van deze HSE-bestudering toonden geen duidelijke trends maar leverden wel een aantal

zeer actieve epoxidatie-katalysatoren op. De meest interessante katalysatoren zijn

gekarakteriseerd: elke lead bestaat uit verschillende soorten silsesquioxaanstructuren.

De lead verkregen uit de hydrolytische condensatie van ButSiCl3 in water, bestaat

uitsluitend uit silsesquioxaan a2b4. Dit is het eerste bekende voorbeeld van het gebruik

van deze silsesquioxaanstructuur om een actieve Ti-katalysator te bereiden. Het

complex werd op silica gezet om zo een actieve, niet-uitlogende heterogene katalysator

voor epoxidatie van alkenen te krijgen. De lead verkregen uit de hydrolytische

condensatie van ButSiCl3 in DMSO bestaat uit een aantal incompleet gecondenseerde

structuren (vooral met 3 < a < 5). De lead verkregen uit de hydrolytische condensatie

van PhSiCl3 in water als oplosmiddel bestaat uit een polymeer silsesquioxaan.

(Hoofdstuk 7).

Na de positieve resultaten van de toepassing van High-Speed Experimentation in de

synthese van silsesquioxanen werden de technieken gebruikt om de synthese van

zeolieten te bestuderen. Zeolieten zijn microporeuze materialen met toepassingen als

heterogene katalysatoren, als ionen-wisselaars en als moleculaire zeven. In dit

167

Samenvatting __________________________________________________________________________________________________________

onderzoek werd de synthese van aluminium-rijk zeoliet beta met verschillende Si/Al en

TEA+/Al verhoudingen bestudeerd met het TEA+ kation als het templaat. De laagste

gevonden verhoudingen in het reactiemengsel die met de gebruikte hydrothermale

methode nog tot zuiver puur zeoliet beta leidden waren Si/Al = 5 en TEA+/Al = 1.25.

Bij lagere Si/Al en TEA+/Al verhoudingen werden mengsels van zeolieten (vooral beta

en NaP1) gevonden. (Hoofdstuk 8).

Een belangrijke conclusie van het promotieonderzoek is dat High-Speed

Experimentation een krachtig hulpmiddel blijkt te zijn voor toegepast en fundamenteel

katalyse-onderzoek.

168

Acknowledgements __________________________________________________________________________________________________________

Acknowledgements

Here is the end of this thesis, and with it comes the time to take a look back at

the four and a half nice years of my Ph.D.. Through this experience I met a lot of

persons who helped me with my work and, more important, who created a lively,

multicultural environment that made the days spent together fine and pleasant. I want to

express my gratitude to this people.

First of all I would like to thank my two supervisors. Thomas, for having given

me the opportunity to come to Delft for my Ph.D., for your advices, for your friendly

attitude, for always having a good word to encourage and motivate me and for the

excellent balance between independence and guidance that you gave me in my research

during these years. Jan Kees, for your advices, for the nice and useful scientific

discussions (‘is it only one peak?’), for sharing with me your knowledge, for your

positive mood and overall for the good time we had together in the lab. Thomas and

Jan Kees, it was a great pleasure to work with the two of you and I hope that we will

have the chance to collaborate again in the future.

Next, I am grateful to all the staff of TOCK/BOC and more in general of

DelftChemTech (Bibliotheek, Technische dienst, Facilitaire dienst, IT-groep) for the

kind help provided. Adrie Knol, Anton van Estrik, Ben Norder, Ernst Wurtz, Lars

Könemann, Mieke Jacobs, Mieke van der Kooij, Niek van de Pers: hartelijk dank! A

special acknowledgement goes to Anton Sinnema and Kristina Djanashvili for the

measurement and help in the interpretation of numerous NMR spectra and for the nice

talking in between. Moreover, I would like to thank Greet de Loos, Henk van

Koningsveld, Herman van Bekkum, Joop Peters, Koos Jansen, Leen Maat and Ulf

Hanefeld for having always been available for fruitful discussions.

In these years spent in the lab I met many Ph.D.’s, post-docs and students: with

some of them I shared many good moments, with others the contact was short but

nevertheless pleasant. I know it is not going to be easy to keep in contact with all of

you; anyhow, I will be happy whenever I meet you again. Thanks to all of you: Frank,

mijn Nederlandse spreektaal leraar. Inma, por tu cariño. Lars, for your sincerity.

Luca G., perché nonostante ci sia di mezzo il Po, si è buoni compagni. Martijn, for

seldom agreeing but still liking to discuss. Paulo, por tu amistad. Rute, for all the strong

feelings and the many good and few bad times we shared. Silvia P., for your sensitivity.

169

Acknowledgements __________________________________________________________________________________________________________

Hans, it was a very positive experience to have you as my student and to know you as

the kind person you are. The ‘Combi’ group: Leon, Silvia G., SP and JK; the good days

spent together at the congress in Ischia (and surroundings) will always be a special

memory. Abdel, Allard, Andrea, Angel, Aniel, Anne B., Anne P., Annemieke, Arné,

Bruno, Carla, Carlos G., Carlos P., César, Chrétien, Cindy, Dean, Delia, Elena, Eva,

Ewald, Fatma, Filip, Gema, Göran, Guillaume, Heidi, Hilda, Jan, José Miguel, Leszek,

Lingqiu, Luca F., Luka, Luuk, Maikel, Manuela, Margreth, Marina, Mathias, Maxim,

Menno, Mette, Michel P., Michel V., Michiel H., Michiel v.V., Mike, Mireia, Moira,

Naseem, Nazely, Nina, Paloma, Pedro, Petra, Raffaella, Sameh, Sandrine, Scoob,

Stefan, Stephan, Tanya, Teresa, Ton, Weiwen Ma, Xavier, Yuxin Li, Zbigniew,

Zhiping: it was nice to meet all of you!

I would like to thank all the people from Avantium with whom I have been in

touch and particularly: Maria, for our rare but nice meetings, Arie, Bas, Ben, Chris,

François, John, Lisa, René, Roger.

During my short and intense stay at the University of Sydney I had the pleasure

to meet and work with a number of persons: Tony, thanks for the good period I had

there and for coming all the way from Down Under for my Ph.D. defence, Jim Beattie,

Keith Fisher, Kelvin Picker, Ian Luck, Dave, John and Ricky.

Finally, I am grateful to all the members of the Examination Committee for

accepting to take part to my Ph.D. defence.

There are many other persons who do not have any direct relationship with my

Ph.D. but to whom I am happy to be close (although often not geographically): my

parents and family, friends from Italy, friends met in the Netherlands and many others I

met around the world.

170

Publications and oral presentations __________________________________________________________________________________________________________

Publications and oral presentations

Publications

1. G. Pacchioni, P. Pescarmona, “Structure and stability of oxygen vacancies on

sub-surface, terraces, and low-coordinated surface sites of MgO: an ab initio study”,

Surf. Sci., 1998, 412/413, 657.

2. P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, “Combinatorial

Chemistry, High-Speed Screening and catalysis”, Catal. Lett., 1999, 63, 1.

3. P. Pescarmona, M. Chiesa, M.C. Paganini, E. Giamello, “OH-D2 (OD-H2) isotopic

exchange at the surface of CaO monitored by electron paramagnetic resonance”,

Magn. Reson. Chem., 2000, 38, 833.

4. P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, “A new,

efficient route to titanium-silsesquioxane epoxidation catalysts developed by using

High-Speed Experimentation techniques”, Angew. Chem. Int. Ed., 2001, 40, 740.

5. P.P. Pescarmona, T. Maschmeyer, “Oligomeric silsesquioxanes: synthesis,

characterization and selected applications”, Aust. J. Chem., 2001, 54, 583.

6. P.P. Pescarmona, J.T.T. Rops, J.C. van der Waal, J.C. Jansen, T. Maschmeyer,

“High-Speed Experimentation techniques applied to the study of zeolites and

silsesquioxanes”, J. Mol. Cat. A, 2002, 182-183, 319.

7. P.P. Pescarmona, T. Maschmeyer, “Serial and parallel ways to enhance and

accelerate catalyst testing”, NATO Science Series, 2002, Ser. II Vol. 69, 173.

8. P.P. Pescarmona, J.C. van der Waal, T. Maschmeyer, “Application of Combinatorial

Chemistry and High-Speed Experimentation concepts to the study of

titanium-silsesquioxane catalysts”, Catal. Today, 2003, 81, 347.

9. P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, “High-Speed

Experimentation techniques applied to the synthesis of titanium-silsesquioxane

epoxidation catalysts”, Chem. Eng. Comm., accepted.

10. P.P. Pescarmona, M.E. Raimondi, J. Tetteh, B. McKay, T. Maschmeyer, “A

mechanistic study of silsesquioxane synthesis by mass spectrometry and in-situ

ATR FT-IR spectroscopy”, J. Phys. Chem. B, accepted.

171


11. P.P. Pescarmona, J.C. van der Waal, T. Maschmeyer, “Fast and high-yield

silsesquioxane syntheses using acetonitrile as reactive solvent”, Eur. J. Inorg.

Chem., submitted.

12. P.P. Pescarmona, A. F. Masters, J.C. van der Waal, T. Maschmeyer, “Osmium

silsesquioxane as model compound and homogeneous catalyst for the

dihydroxylation of alkenes”, Angew. Chem. Int. Ed., submitted.

13. P.P. Pescarmona, J.C. van der Waal, T. Maschmeyer, “New silsesquioxane-based

homogeneous and heterogeneous epoxidation catalysts developed using

High-Speed Experimentation”, Chem. Eur. J., submitted.

Patents

1. P. J. van den Brink, P.P. Pescarmona, J.C. van der Waal, “Liquid dispensing device”,

Patent EPA99310598, WO0148443, AU2389701, 1999.

Oral presentations and posters

1. P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, “Development

of a new, efficient route to titanium-silsesquioxane epoxidation catalysts using

High-Speed Experimentation techniques”, Frontiers in Homogeneous Catalysis,

Joint meeting of the NIOK and the Katalyseverbund NRW, Eindhoven University

of Technology, Netherlands, October 9, 2000: poster.


of titanium-silsesquioxane epoxidation catalysts using High-Speed Experimentation

techniques”, 2000 Annual Meeting of the American Institute of Chemical

Engineers (AIChE), Los Angeles, United States of America, November 12-17,

2000: oral presentation.


of new titanium-silsesquioxane epoxidation catalysts using High-Speed

Experimentation techniques”, NCCC2, Noordwijkerhout, Netherlands, March 5-7,

2001: oral presentation.

172


4. P.P. Pescarmona, J.T.T. Rops, J.C. van der Waal, J.C. Jansen, T. Maschmeyer,

“Synthesis of zeolites and silsesquioxanes using High-Speed Experimentation

techniques”, 10th International Symposium on Relations between Homogeneous and

Heterogeneous Catalysis, Lyon, France, July 2-6, 2001: oral presentation.

5. P.P. Pescarmona, J.C. van der Waal, T. Maschmeyer, “Study of the synthesis of

incompletely condensed silsesquioxanes by means of High-Speed Experimentation

techniques”, NCCC3, Noordwijkerhout, Netherlands, March 4-6, 2002: poster.

6. P.P. Pescarmona, “Study of the synthesis of silsesquioxanes by means of

High-Speed Experimentation techniques”, University of Torino, Italy, May 29, 2002:

oral presentation.

7. P.P. Pescarmona, J.C. van der Waal, T. Maschmeyer, “Study of the synthesis of

incompletely condensed silsesquioxanes by means of High-Speed Experimentation

techniques”, EuroCombiCat, European Workshop on Combinatorial Catalysis, Ischia,

Italy, June 2-5, 2002: poster.

8. P.P. Pescarmona, “High-Speed Experimentation techniques applied to the synthesis

of Ti-silsesquioxanes as epoxidation catalysts”, University of Sydney, Australia,

October 25, 2002: oral presentation.

9. P.P. Pescarmona, J.K. Beattie, A. F. Masters, J.C. van der Waal, T. Maschmeyer,

“An osmium silsesquioxane as a catalyst for the dihydroxylation of alkenes”,

Conference of the Inorganic Chemistry Division of the Royal Australian Chemical

Institute, Melbourne, Australia, February 2-6, 2003: poster.

10. P.P. Pescarmona, J.C. van der Waal, M.E. Raimondi, J. Tetteh, T. Maschmeyer,

“Improved synthesis of silsesquioxane a7b3 [(c-C5H9)7Si7O9(OH)3]”, NCC4

(Netherlands Catalysis & Chemistry Conference) Noordwijkerhout, Netherlands,

March 10-12, 2003: oral presentation and poster.

11. P.P. Pescarmona, J.K. Beattie, A. F. Masters, J.C. van der Waal, T. Maschmeyer,

“An osmium silsesquioxane as a catalyst for the dihydroxylation of alkenes”, 6th

International Symposium on Catalysis Applied to Fine Chemicals (CAFC 6), Delft,

Netherlands, April 6-10, 2003: poster.

173

Curriculum vitae __________________________________________________________________________________________________________

Curriculum vitae Paolo Pescarmona was born on the 2nd of November 1973 in Torino, Italy. He

obtained his secondary education diploma in 1992 at the Liceo Scientifico Statale

Galileo Ferraris, Torino, Italy, with a grade of 60/60. In October 1992 he started his

university studies in Chemistry at the Università di Torino, Italy. He received his Laurea

in Chimica (Master of Science in Chemistry, 5 years) on the 4th of December 1997 with

a grade of 110/110 cum laude et mentione. His final project, “Defective centres at the

surface of alkaline-earth oxides: spectroscopic studies (EPR) and quantum mechanical

calculations”, was carried out under the supervision of Prof. Elio Giamello (Università

di Torino) and of Prof. Gianfranco Pacchioni (Università di Milano). From December

1997 till October 1998 he fulfilled his Civil Service at the Foreigners Office of the City

of Torino, Italy. In February 1999 he started his Ph.D. research at the Laboratory of

Applied Organic Chemistry and Catalysis of the Technische Universiteit Delft,

Netherlands, under the supervision of Prof. dr. Thomas Maschmeyer and Dr. ir. Jan

Kees van der Waal on the “Exploration of silsesquioxanes and zeolites using

High-Speed Experimentation”. The results of this research are described in this thesis.

174

Propositions

belonging to the thesis:

“An Exploration of Silsesquioxanes and Zeolites using High-Speed Experimentation”

by Paolo P. Pescarmona

1. The term ‘polarity’, although being one of the most commonly used in chemistry, is still rather undefined.

2. The capability of pronouncing the word ‘silsesquioxanes’ is an effective criterion

to recognise people that worked with these compounds. 3. Although colour blindness may generate some difficulties for the experimental

chemist, it can favour a critical attitude towards observation. 4. A multidisciplinary approach to scientific research can be compared to a

multiethnic society: once a common language is found and the reciprocal differences are understood and accepted, the blend will produce great benefits.

5. (Scientific and technological) progress should be driven by its social value rather

than by its economic potential. 6. One of the risks of economic endeavours is to confuse the well-being of companies

with that of people. 7. Nice would be a world where all the persons are equal, but not uniform; alas,

uniformity is growing whilst equality is decreasing. 8. Ties, depilation and hijab (headscarf) have a common feature: making use of them

should be a free individual choice and not an imposition. 9. One of the most reliable methods to recognise whether somebody is Dutch, is to

ask her/him to write an ‘8’.

10. A rather underestimated skill often acquired in University and Ph.D. studies is the interpretation of the handwriting of professors.

Stellingen

behorende bij het proefschrift:

“An Exploration of Silsesquioxanes and Zeolites using High-Speed Experimentation”

van Paolo P. Pescarmona

1. De term ‘polariteit’, hoewel een van de meest gebruikte in de scheikunde, is nog niet goed gedefinieerd.

2. De manier waarop het woord ‘silsesquioxanen’ wordt uitgesproken, is een goed

criterium om mensen die met deze verbindingen werken te herkennen. 3. Hoewel kleurenblindheid bepaalde problemen voor de experimentele chemicus

genereert, kan het een kritische houding voor het waarnemen begunstigen. 4. Een multi-disciplinaire aanpak voor wetenschappelijk onderzoek is met een

multi-etnische maatschappij vergelijkbaar: als een gemeenschappelijke taal gevonden is en wederzijdse verschillen begrepen en geaccepteerd zijn, zal het mengsel grote voordelen opleveren.

5. (Wetenschappelijke en technologische) vooruitgang behoort geleid te worden door

zijn sociale waarde eerder dan door het economische potentieel. 6. Een van de risico’s van de moderne economie is het welzijn van bedrijven met die

van mensen te verwarren. 7. Mooi zou een wereld zijn waarin alle mensen gelijk waren maar niet uniform;

helaas groeit tegenwoordig de uniformiteit terwijl de gelijkheid vermindert. 8. Dassen, ontharing en hijab (hoofddoek) hebben een gemeenschappelijke

eigenschap: het gebruik ervan behoort een vrije keuze te zijn en niet een verplichting.

9. Een van de meest betrouwbare manieren om een Nederlander te herkennen is

haar/hem te vragen een ‘8’ te schrijven.

10. Een doorgaans onderschatte kennis die vaak gedurende een universitaire of Ph.D.-studie wordt opgedaan is de interpretatie van het handschrift van professoren.

an exploration of silsesquioxanes and zeolites using high-speed experimentation

Documents