C-H Activation through Photocatalytic Sulfoxidation of … · C-H Activation through Photocatalytic...

C-H Activation through Photocatalytic Sulfoxidation

of Alkanes

Der Naturwissenschaftlichen Fakultät der Friedrich-Alexander-Universität Erlangen-Nürnberg

zur Erlangung des Doktorgrades

vorgelegt von Francesco Parrino

aus Palermo (Italien)

Als Dissertation genehmigt von der Naturwissenschaftlichen Fakultät der Universität Erlangen-Nürnberg

Tag der mündlichen Prüfung: 27. 03. 2009

Vorsitzender der Promotionskommission: Prof. Dr. Eberhard Bänsch

Erstberichterstatter: Prof. Dr. Horst Kisch

Zweitberichterstatter: Prof. Dr. Ulrich Zenneck

1

Acknowledgements

I wish to thank Prof. Dr. Horst Kisch for the supervision of this work and many fruitful discussions. I am particularly grateful for his advice, his trust in me, and his generous support of my work.

Parts of this work would not have been possible without the help of several people. I thank Christina Wronna for elemental analyses, Cornelia Damm for time resolved photovoltage measurements, Siegfried Smolny for surface area measurements, Helga Hildebrand for XPS measurements, and Martin Bachmueller for MS analysis. Manfred Weller, Peter Igel and their colleagues from the “Werkstatt” are acknowledged for the assistance with technical problems. I am also indebted to Dr. Matthias Moll for his manifold help, Uwe Reißer for his help with electronic equipment, Ronny Wiefel for glass work, and to Dr. Jörg Sutter for computer assistance.

Many thanks to all my colleagues for contributing to the very good atmosphere in the group – Marc, Ayyappan, Sakthi, Joachim, Sina, Christian, Darek, Przemek and Radim – and also to all my friends outside the institute.

I am very grateful towards my parents for their lifelong love and encouragement, and I dedicate this work to them.

3

“This work presents very novel results that may have a significant impact not only to

the field of semiconductor photocatalysis but also in terms of organic synthesis. If scale-

up of this process proves viable there could be very important industrial implications.”

Unknown referee

Trova il tempo di pensare, di pregare, di ridere

E’ la fonte del potere più grande sulla terra,

E’ la musica dell’anima.

Trova il tempo per giocare, per amare ed essere amato

Trova il tempo di dare

E’ il segreto dell’eterna giovinezza, il privilegio dato da Dio.

La giornata è troppo corta per essere egoisti.

Trova il tempo di leggere, di essere amico, di lavorare

E’ la fonte della saggezza, la strada della felicità,

Il prezzo del successo

Trova il tempo di fare la carità

È la chiave del paradiso.

Madre Teresa di Calcutta

5

Die vorliegende Arbeit entstand in der Zeit von April 2006 bis Februar 2009 am

Institut für Anorganische Chemie der Universität Erlangen-Nürnberg unter Anleitung

von Herrn Prof. Dr. Horst Kisch.

7

Contents

Acknowledgements.............................................................................................. 1

Contents................................................................................................................ 7

1. Introduction .................................................................................................. 9

2. Literature Review....................................................................................... 15

2.1 Heterogeneous Photocatalysis ........................................................................15 2.1.1 Fundamental Properties of Semiconductors ........................................................15 2.1.2 Properties of Titanium Dioxide ...........................................................................17 2.1.3 Photocatalysis Mechanism ..................................................................................20 2.1.4 Visible Light Activity of Modified Titania .........................................................22

2.2 Activation of C-H Bonds by Metal Complexes..............................................24 2.2.1 Activation via σ-Organyl Derivatives .................................................................24 2.2.2 Activation through Electron and Hydrogen Transfer ..........................................31 2.2.3 Indirect Activation via Reactive Oxygen Species ...............................................32

2.3 Alkane Activation at Titania...........................................................................35

3. Visible-light Photocatalysis by a Titania-Rhodium (III) Complex........ 39

3.1 Introduction.....................................................................................................39 3.2 Synthesis and Characterization.......................................................................41

3.2.1 Desorption Experiments ......................................................................................41 3.2.2 Diffuse Reflectance Spectra ................................................................................42

3.2.2.1 Principles of DRIFTS................................................................................................42 3.2.2.2 Diffuse Reflectance Measurements...........................................................................45

3.2.3 Determination of Band Edge Positions ...............................................................50 3.2.3.1 Semiconductor-Electrolyte Interface.........................................................................50 3.2.3.2 Concept of Quasi Fermi Level ..................................................................................53 3.2.3.3 Photovoltage Measurements .....................................................................................53

3.2.4 Photocatalytic Activity ........................................................................................57 3.2.5 Understanding the Mechanism ............................................................................60

3.3 Experimental...................................................................................................63 3.3.1 Instruments ..........................................................................................................63 3.3.2 Determination of Absorptivity of [RhCl6]3- ........................................................63 3.3.3 Preparation of Photocatalysts ..............................................................................64 3.3.4 Preparation of 4.0%RhCl3/TH and Charge Transfer Absorption ........................64 3.3.5 Determination of Cl/Rh Ratio .............................................................................65 3.3.6 Measurement of Quasi-Fermi Potentials .............................................................65 3.3.7 Desorption Experiments ......................................................................................65 3.3.8 Photodegradation Procedure and Product Analysis ............................................66

3.4 Conclusions.....................................................................................................68

4. C-H Activation through Catalytic Photosulfoxidation of Alkanes........ 69

4.1 Introduction.....................................................................................................69 4.1.1 Industrial Importance of Photosulfoxidation.......................................................70 4.1.2 Industrial Processes .............................................................................................70

8

4.1.2.1 Mechanism of Industrial Sulfoxidation ....................................................................70 4.1.2.2 Light-Water-Process..................................................................................................73 4.1.2.3 Acetic Anhydride Process .........................................................................................75 4.1.2.4 Other Processes .........................................................................................................76

4.1.3 State of Knowledge before this work ..................................................................78 4.2 Results and Discussion: Reaction in Liquid Alkanes.....................................80

4.2.1 Product Characterization .....................................................................................81 4.2.1.1 Principle of IPC and Measurements .........................................................................81 4.2.1.2 IR Spectra and Amount of Sulphate .........................................................................84 4.2.1.3 Elemental Analysis ...................................................................................................87

4.2.2 Dependence on Photocatalyst Concentration ......................................................88 4.2.3 Deactivation and Regeneration of the Photocatalyst ...........................................89 4.2.4 Surface Modifications of the Catalyst .................................................................91 4.2.5 Interaction between SO2 and TiO2 ......................................................................95 4.2.6 XPS Results .........................................................................................................97

4.2.6.1 XPS Principles ..........................................................................................................97 4.2.6.2 XPS Spectra ..............................................................................................................99

4.2.7 PEMF Results ....................................................................................................101 4.2.7.1 PEMF Basics...........................................................................................................101 4.2.7.2 PEMF Measurements and Discussion ....................................................................103

4.2.8 Reaction Mechanism .........................................................................................110 4.2.8.1 Evidences of Radical Chain Reaction.....................................................................112 4.2.8.2 Evidence for Formation of a TiO2-SO2 CT Complex .............................................117

4.2.9 Oxidation Products ............................................................................................123 4.2.10 Preparative Synthesis of Heptanesulfonic Acid Sodium Salt............................124

4.3 Results and Discussion: Reaction in Acetic Acid ........................................124 4.3.1 System Description and Product Characterization ............................................125 4.3.2 Acetic Acid Adsorption at TiO2 ........................................................................127 4.3.3 Influence of Water .............................................................................................129

4.4 Experimental part..........................................................................................130 4.4.1 Materials ............................................................................................................130 4.4.2 Standard Photosulfoxidation .............................................................................130 4.4.3 Instruments and Methods ..................................................................................133 4.4.4 Surface Modification of Titania ........................................................................139

5. Summary ................................................................................................... 141

6. Zusammenfassung .................................................................................... 146

7. References ................................................................................................. 151

1. Introduction

9

1. Introduction

Alkanes are major constituents of natural gas and petroleum, the feedstocks of

chemical industry, but there are very few processes for converting them directly to

more valuable products such alcohols, ketones, acids, peroxides. The reason of this is

reflected in the other name “paraffins” (from the Latin “parum affinis”) that means

“not enough affinity”. This chemical inertness arises from the strong and localized C-

C and C-H bonds (the value of the C-H dissociation energy for methane is 104

kcal/mol), so that the molecules have no empty orbitals of low energy or filled orbitals

of high energy that could readily participate in a chemical reaction.1

Alkanes may be called the “noble gases of organic chemistry”; however this

comparison is not fully accurate. In fact, whereas noble gases do not react easily with

any usual compound, there is at least one well-known substance which activates

paraffins very well: this substance is the oxygen. Alkanes undergo complete oxidation

in air in the presence of a catalyst and at high temperature to produce water and carbon

dioxide.2 The currently prevalent use of alkanes in combustion applications exploits

their energy content, but not their potential as precursors for more important

chemicals. Most of them, especially oxygenates, are produced from olefins, in turn

obtained from alkanes by fairly inefficient and energy intensive processes like

cracking and thermal dehydrogenation.1 Furthermore, some active reagents, such as

atoms, free radical and carbenes can react with saturated hydrocarbons at room

temperature. But these reactive species are usually demanding to make and offer little

control over product selectivity.

Milder and better-controlled direct conversion of alkanes into e.g. olefins and

alcohols may thus offer large chemical, energetic, and economic benefits. Exploitation

of natural gas resources is hampered by the high cost of both gas transportation and

conversion hydrocarbon into more readily transportable liquid. The available

conversion methods are indirect, involving production of synthesis gas followed by

conversion to the desired product. Developing efficient strategies for the direct

1. Introduction

10

conversion of alkanes to the final products could thus allow us to exploit hydrocarbons

more efficiently. 1

For these reasons C-H activation is certainly a problem of global importance, a

“holy grail” of chemistry as noted by Bard, Whitesides, Zare and McLafferty in

19953,4 and this can be documented by the steady growth of interest in this field in

recent years ( Fig. 1.1).

1985 1990 1995 2000 20050

50

100

150

200

Num

ber o

f pub

blic

atio

ns/Y

ear

Year

Figure 1.1: Annual number of papers published in which “C-H activation” are mentioned in the last 23 years (till July 2008). Literature search was done using SciFinder® Scholar™ searching tool.

In 1969 the first activation of C-H bonds in alkanes was discovered. It was found

that platinum(II) salts catalyze the H-D exchange between methane and D2O at

100°C.5 In the 1970s it was shown that alkanes are oxidized by platinum(IV)6,

palladium(II)7, ruthenium(IV)8, and cobalt(III)9 compounds and that complexes of

iridium(III)10 and titanium(II)11 catalyze the H-D exchange. The field took off during

the 1980s, when there was a dramatic increase in the number of metal salts and

complexes that were found to initiate C–H activation by oxidative addition. But the

drawback was that most of these transformations required equal amounts, in moles, of

the hydrocarbon and the metal, and both partners were consumed during the reaction.12

The next decade was marked by an explosion of interest in the use of catalytic

reactions for bringing about oxidative addition for C–H activation. In these catalytic

1. Introduction

11

processes, the oxidative addition product is a transient intermediate that immediately

undergoes subsequent transformation. The catalytic metal is released, so that it can

attack another molecule of hydrocarbon. An excellent example is the recent discovery

that rhodium catalysts directly convert the C–H bonds at the ends of alkane chains into

carbon–boron bonds;13,14 the products of such reactions are very useful for synthetic

organic chemistry. Another example is the discovery that methane can be converted

into methanol derivatives with unusually high yields using platinum complexes in

strong acid solution.12

As discussed above, most research is aimed at the use of metal complexes, and this

approach derives from several important considerations: first, the central metal can be

varied in its reactivity by ligand design. Second, many metals do insert into

unactivated C-H bonds,4,14-16 and they often do it in a predictable fashion.17 Third,

there is a good reason to mimic enzymatic C-H activation with chemical methods.

In fact, nature utilizes a variety of enzymes which oxidize alkanes efficiently and

selectively. Cytochrome P450 enzymes typically catalyse the conversion of C-H bonds

to C-O bonds in organic compounds. In humans these enzymes are involved in making

cholesterol and other lipids. They also metabolize drugs, converting them to highly

oxidized compounds that can be excreted from the body. The active sites of these

enzymes contain iron. The proposed mechanism is shown in Eqs. 1.1 - 4.

[Fe]=O ↔ [Fe]-O· (1.1)

[Fe]-O· + R-H → [Fe]-OH + R· (1.2)

[Fe]-OH + R· → [Fe] + R-OH (1.3)

[Fe] + ½O2 → [Fe]=O (1.4)

It involves formation of a highly reactive iron-oxygen double bond [Fe]=O

which activates a C-H bond yielding a carbon radical and a complex bearing an iron-

hydroxyl group. After hydroxyl group transfer to the carbon radical, the remaining iron

complex is oxidized back to its original form by molecular oxygen closing the cycle.

1. Introduction

12

However, the details of this process are still a source of controversy. Another example

of enzymatic C-H activation is the transformation of methane to methanol by methane

monooxygenase. This iron containing enzyme was discovered in a class of bacteria

that lives at the interface of aerobic and anaerobic environments. The direct use of

such biological organisms for industrial alkane conversion is in principle possible but

the scale-up step seems to be problematic.12

Another approach to tackle this problem is the metal-free alternative. Many

reactions claimed to be metal-catalyzed produce similar results even in the absence of

a metal. Many transition metal C-H activations are oxygen and water-sensitive and

sometimes involve very expensive and poisonous metals. Furthermore, a costly water

treatment is required often. On the other hand high-temperature radical or even

surface-catalyzed reactions are easy to carry out on a large scale. In general these

reactions exhibit poor selectivity, except some highly selective alkane

functionalizations, induced by nitroxyl radicals and radical reactions conducted under

phase-transfer conditions.18

Figure 1.2: Alkane C-H bond activation paths with uncharged (E) and charged (E+) electrophiles as well as with radicals (E· ) and electron-acceptors (-e-). (Adapted from ref.18)

R-H

E

E+

-e-

E·

δ+· R--H--E δ-·

-EH·

R·

δ+· R--H--E δ+· -EH

TS1: molecule-induced homolytic

δ+· R--H--E δ+·-EH

R·

R---H + ·

- H+ - H·

R· R+

solution gas-phase

TS4: oxidative

TS3: radical TS2: electrophilic

R+

1. Introduction

13

Mechanistic studies of alkanes reacting with electron deficient species reveal

common mechanistic features arising from the formation of radical-cationic transition

states or intermediates.19 The recently suggested H-coupled electron transfer

mechanistic model19 agrees well with the experimental features of alkane

halogenations with electrophiles and may be extended to a wide range of alkane

transformations: from uncharged metaloxo species20 and, possibly, dioxiranes21

(molecule induced homolytic path, TS1, Fig. 1.2) to charged electrophiles like

nitronium salts (´”electrophilic” path, TS2, Fig. 1.2). The “oxidative” path fit into the

same mechanistic regime. The radical path consists of a hydrogen atom abstraction

(TS3, Fig. 1.2). The activation energy of this reaction is usually low and, when a

branched alkane is used as the substrate, the reactivity follows the “normal” selectivity

of the C-H bonds ( tertiary > secondary > primary ) whereas transition metal centers

preferentially activate terminal C-H bonds.2 An example of metal-free alkane

activation is the photosulfoxidation of alkanes applied in industry (see Chapter 4).

Although many promising systems have been developed and our comprehension

on this topic has progressed considerably, profitable, practical applications have not

yet been performed. It remains challenging to conciliate chemical, with economic and

engineering requirements.

The topic of the present dissertation fits in with this huge and complex landscape.

It is a study of a novel C-H functionalization of alkanes through TiO2 photocatalyzed

sulfoxidation accomplished with visible light irradiation to obtain sulfonic acids. A

rare example of an industrially applied process of direct functionalization of saturated

hydrocarbons is the photosulfoxidation of liquid alkanes by sulfur dioxide and oxygen

(Eq. 1.5). The reaction requires the presence of UV light which is generated by large

medium pressure mercury vapor lamps.

R-H + SO2 + ½ O2 + hν → RSO3H (1.5)

1. Introduction

14

In the case of linear C16-20 chain alkanes the resulting alkanesulfonic acids are used

as biodegradable surfactants.

The aim of the present work was to find out if a similar reaction could be

performed with visible light irradiation and without use of toxic mercury containing

lamps.

At the beginning we focused our research on a photocatalyst known to be active in

visible light photooxidations of 4-chlorophenol (4-CP), an ubiquitous pollutant. This

novel RhCl3-modified titania was synthesized, characterized and its role in 4-CP

photodegradation a me was investigated. Chapter 3 presents our results on RhCl3/TiO2.

Further investigations allowed us to observe that commercial titania P25 (75%

anatase / 25% rutile) was even more active in the visible photocatalytic sulfoxidation

although, as known, TiO2 absorbs only in UV range. Chapter 4 of this dissertation is

devoted to this reaction.

2. Literature Review

15


2.1 Heterogeneous Photocatalysis

Heterogeneous photocatalysis is a field of chemistry focused on liquid or gas

catalytic reactions occurring in presence of irradiated solids (normally

semiconductors).

Semiconductors can be excited by light with higher energy than the band gap and

an energy-rich electron-hole pair is formed. This mechanism can either be used

directly to generate electricity in photovoltaic solar cells or to drive a chemical

reaction in which the oxidation and reduction of substrates occur simultaneously. In

general, photocatalytic reactions are aimed to organic synthesis, degradation of

pollutants or to some special reactions like fixation of nitrogen or splitting of water.

In this section a brief overview on the fundamental properties of semiconductors

will be given and the basics of semiconductor photocatalysis will be discussed.

2.1.1 Fundamental Properties of Semiconductors

A very useful way to examine the properties of a semiconductor is the energy band

model.22-26 A solid can be viewed as “very large molecule” in which the number of

molecular orbital is so high that we can neglect the energy difference between them

until they can be considered a continuum forming energy band. The key electronic and

optical properties of a semiconductor are determined by two bands: the lowest

unoccupied energy band, called conduction band, and the highest occupied energy

band, called conduction band. The band gap is defined as the energy difference

between the lower edge of the conduction band and the higher edge of the valence

band.

Metals are good electronic conductors because their conduction and valence band

overlap allowing electron to occupy empty states in the conduction band and to move

freely in the lattice. Materials with Ebg ≥ 4 eV are defined as insulators and it is hardly

possible to promote an electron to the conduction band. In semiconductors the band

gap is small enough (typically 1.0 – 4.0 eV) to allow for increase in conductivity


16

through thermal- or photo-excitation. An electron can be excited to the conduction

band leaving mobile holes (positively charged vacancies) behind in the valence band.

The energy level at the top of the valence band determines the oxidizing ability of

photoholes, while the energy level at the bottom of the conduction band is actually the

reduction potential of the photoelectrons.

Figure 2.1: Formation of energy bands in the Si crystal. 3s and 3p orbitals of a single Si atom (a) become mixed to form 4 hybridized sp3 orbitals (ψhyb) (b). (c) The hybridized ψhyb orbitals on two neighboring Si atoms can overlap to form a bonding (full) orbital (ψB) and an antibonding (empty) orbital (ψA). (d) A cluster of Si atoms. With increasing the number of atoms the overlapping bonding and antibonding orbitals become more numerous and more closely spaced in energy, which, finally, leads to continuous bands of energy levels (e) in a Si crystal – the valence band (full) and the conduction band (empty) separated by a bandgap (Eg); (taken from Refs. 27 and 24).

The energy at which the probability of a level being occupied by an electron is 0.5

is referred to as Fermi level EF. From a thermodynamical point of view the Fermi level

is the electrochemical potential of the electron in the solid. In intrinsic semiconductors

the Fermi level is approximatively in the middle of the band gap and, when the band

gap is small enough, some electrons can be thermally excited from the valence band to

the conduction band at ambient temperature. In “extrinsic” semiconductors the

conductivity can be achieved by doping. When donor energy levels are present just


17

under the conduction band edge, the electrons can be easily thermally excited into the

conduction band, the electrons are the majority charge carrier and the material is a n-

type semiconductor. Accordingly, in an n-type semiconductor the Fermi level is right

below the conduction band. Similarly if acceptor energy levels are present above the

valence band edge, the electron can be thermally excited into these states leaving thus

positive holes in the valence band behind. In this case the holes are the majority

carriers, the semiconductor is a p-type and its Fermi level lies just above the valence

band edge.

When a semiconductor is irradiated by light with energy greater than the band gap

energy, an electron-hole pair is generated. Obviously, when the equilibrium state is

perturbed by light excitation, the semiconductor tends to return to the equilibrium.

Normally, this happens through recombination whereby the energy excess can be

dissipated as heat (radiationless process) or emitted as a photon (radiative emission).

In competition to this process the charges can be trapped at reactive surface states

capable of exchanging electrons with appropriate substrates.

2.1.2 Properties of Titanium Dioxide

Due to oxygen vacancies, TiO2 is an n-type semiconductor. These vacancies are

formed according to the Eq. 2.1

O0x → V0

hh + 2e- + ½ O2 (2.1)

where the Kröger-Vink defect notation is used to explain that inside TiO2 a positive

(2+) charged oxide ion vacancy (V0hh) is formed upon release of two electrons and

molecular oxygen. For example this reaction can be induced by heating (in an oxygen-

poor environment). Formation of Ti2+ centers is necessary for charge neutralization.28

Fig. 2.2 shows the electronic structure of TiO2. The conduction band edge states have

predominantly the Ti 3d character, while the valence-band edge states have the O 2p

character.


18

Figure 2.2: Molecular orbital bonding scheme for anatase TiO2 (taken from Ref. 29): (a) atomic levels, (b) crystal- field split levels, and (c) final levels. The thin-solid and dashed lines represent large and small contributions, respectively. Note that the Fermi level (in the scheme drawn in the middle of the bandgap) will be, in reality, shifted to the vicinity of the conduction band edge due to oxygen vacancies.

Titanium dioxide is mainly found in three modifications: rutile (tetragonal),

anatase (tetragonal) and brookite (orthorhombic). Rutile is the stablest phase but

anatase is the polymorph most widely used for photocatalytic applications. The

reported band gap of anatase (3.2 eV) and rutile (3.06 eV) are very similar but the

former modification in general has a better photocatalytic activity.30 The reasons are

yet a source of debate. It has been suggested that the slightly higher anatase Fermi

level, the higher capacity to adsorb oxygen, the better electrons mobility, and the

higher degree of hydroxylation might explain this fact.31-34 That is why in the first part

of this work anatases modification with rhodium salts was performed. However, the

photocatalytic sulfoxidation of heptane turned out to give the best yield by using


19

Degussa P25, a commercial mixed phase of rutile (25%) and anatase (75%). The

enhanced activity of P25 is a general trend. As proposed by Thurnauer et al., the

electron generated on rutile is transferred to anatase. This special separation largely

prevents recombination and therefore increases the photocatalytic activity.35

The TiO2 surface is schematically depicted in Fig. 2.3. In reality both the titanium

and oxygen atoms are not coordinatively saturated giving rise to the so-called surface

defects. Water molecules through dissociative adsorption can fill up these sites so that

the surface presents a high concentration of OH groups (about 5 OH groups per nm2).

Some of them are bidentately bound (Ti(OH)Ti) and have acid character (pKS = 2.9)

while the monodentately bonded OH groups (≡Ti-OH) have basic character (pKS =

12.7).

For these reasons the TiO2 surface can be protonated or deprotonated depending on

the pH value of the aqueous suspension.

Figure 2.3: Simplified scheme of the protonation and deprotonation of hydroxylated TiO2 surface leading to positive and negative net charge at the surface, respectively.(Taken from ref.36) The isoelectric point (pHIEP) of TiO2 is typically about 5.8–7.5.37-41 The pKA values of monodentate and bidentate OH groups are reported to be 12.7 and 2.9, respectively.40

The pH value at which the net surface charge is zero is called pHIEP (isoelectric point).

If pH > pHIEP there are not positive charges on the surface, if pH < pHiep there are not

negative charges on the surface. Such a pH dependence on the surface charge


20

influences also the energetic position of the valence and conduction bands (see

Paragraph 3.2.1).

2.1.3 Photocatalysis Mechanism

The key steps of a photocatalytic reaction at a small semiconductor particle are

illustrated in Fig. 2.4. The charges generated upon light absorption (process 1) can

either undergo primary recombination, emitting light or heat (process 2), or can be

trapped at reactive surface sites (process 3). In the case of TiO2, electrons are reported

to be trapped as TiIII centers and holes as surface-bound hydroxyl radicals

≡TiIVOH•+.42 Trapping of holes proceeds in 10-100 ns, while this process is faster

for electrons and requires some hundreds of picoseconds.42 The trapped charges can

either recombine (secondary recombination, process 4), or undergo an interfacial

electron transfer process (IFET), whereby the electron reduces an electron acceptor

species A to a primary reduction product A– • (process 5), and the hole oxidizes an

electron donor species to D+• (process 6). In order to avoid back electron transfer

(process 9) A– • and D+• must then undergo a rapid conversion to the final products Ared

and Dox (processes 7 and 8).

Figure 2.4: Schematic representation of the key processes of a photocatalytic reaction at a semiconductor particle. For details see the text. (Taken from ref. 36)


21

Hence, for example, in a typical photocatalytic oxidation of organic water

pollutants on TiO2 the reacting holes are scavenged either directly by the pollutant or

by adsorbed hydroxyl ions to produced hydroxyl radicals which can then oxidize the

pollutant due to their high oxidizing power. Simultaneously, the photogenerated

electrons reduce molecular oxygen to a superoxide radical which can then undergo

further reactions to produce hydroxyl radicals via following reactions:33,42-45

O2 + eCB– → O2

•– (2.2)

O2•– + H+ → HO2

• (2.3)

HO2• + HO2

• → H2O2 + O2 (2.4)

O2•– + HO2

• → O2 + HO2– (2.5)

HO2– + H+ → H2O2 (2.6)

H2O2 + O2•– → •OH + OH– (2.7)

H2O2 + eCB– → •OH + OH– + O2 (2.8)

The thus produced hydroxyl radicals can again contribute to the mineralization of

the pollutant.

The general equation of degradation of 4-CP on TiO2 can be written as follows.

2p-ClC6H4OH + 13O2 → 12CO2 +2H+ +2Cl- + 4H2O (2.9)

in the case of total mineralization.


22

2.1.4 Visible Light Activity of Modified Titania

Ideally, a semiconductor photocatalyst should be chemically and biologically inert,

photocatalitically stable, easy to produce and to use, cheap and without risks for

environment or humans. Titanium dioxide is close to being an ideal photocatalyst, the

only problem is that it does not absorb visible light and only 4% of the solar spectrum

can be utilized. A shift of the absorption edge of TiO2 to larger wavelengths is

expected to increase the solar light conversion efficiency.

Visible light photosensitization consists of photoinduced electron transfer. It can

be achieved by four main approaches: 1) bulk doping, 2) formation of coupled

semiconductors, 3) metal-semiconductor composites, and 4) surface modification.

Bulk doping29,46-48 in a very few cases induces a band gap narrowing. The most

important example is the Cr doped anatase produced by ion implantation.49 Coupled

semiconductors50-54 are formed when TiO2 interacts with particles of semiconductors

of smaller band gap and different band edge potentials like CdS (Eg = 2.5 eV). Visible

light irradiation leads to charge separation only in the smaller band gap semiconductor

and subsequently an electron can be injected into the CB of TiO2 whereas the hole

remains in the valence band of the other semiconductor avoiding charge

recombination.

The most commonly used technique of TiO2 photosensitization is surface

modification. It involves formation of covalent or ionic bonds between a

semiconductor surface and chromophoric molecules (sensitization mechanism) or

formation of chromophoric surface species upon interaction with chromogenic

molecules (charge transfer complex mechanism CTC) .55

In the first case diverse organic dyes (erythrosin B, porphyrins56,57,

phtalocyanines,58,59 thiacarbocyanine dyes60) and metal complexes (Fe(II),55 Pt(IV)

complexes61) adsorbed on the TiO2 surface, absorb a photon of visible light (the

adsorbate itself is colored) and subsequently inject an electron into the conduction

band of the semiconductor. This can result in destruction of the dye and potentially of

other solutes. The above described mechanism, known also as photoinduced electron

transfer, is summarized in Eqs. 2.9 and 2.10.


23

TiO2---S + hν → TiO2---S* (2.9)

TiO2---S* → TiO2(e-cb)---S+· (2.10)

The CTC mechanism can allow visible light activation of a system in which neither

the catalyst nor the adsorbate (generally enediols, chlorophenols62) absorbs visible

light by itself. The adsorbate is colorless but becomes colored when adsorbed to TiO2.

This mechanism is also known as optical electron transfer and could be summarized

as in Eqs. 2.11 and 2.12.

TiO2 + S ↔ [TiO2-S] (2.11)

[TiO2-S] + hνCT → TiO2(e-cb)-S+· (2.12)

The mechanism presented in the Chapter 3 of this work, devoted to the synthesis and

characterization of the novel RhCl3-TiO2 catalyst, can be seen as a sensitization

mechanism whereas the catalytic photosulfoxidation of alkanes (Chapter 4) is initiated

through a CTC mechanism.


24

2.2 Activation of C-H Bonds by Metal Complexes

The knowledge of the most important theories and experimental evidences about

the role of the metal center in C-H activation is obviously a necessary prerequisite for

understanding the mechanism of the reaction presented in this dissertation. Therefore,

in this part a very brief outline of the most significant reactions will be given. C-H

activation is conveniently classified into three groups based on the mutual interaction

between the metal center and the hydrocarbon. The reactions frequently occur in

solutions at room temperature although sometimes heating is required and certain

reactions are stimulated by irradiation. Either light or heat is essential for the

abstraction of several ligands from the initial complex to form a coordinatively

unsaturated species capable of oxidatively adding the C-H compound.

2.2.1 Activation via σ-Organyl Derivatives

In this group are collected the reactions in which there is a direct contact

between the C-H bond and the metal ion. An organometallic derivative, i.e. a

compound containing a M-C σ-bond63, is formed in which a metal centre interact with

the electron pair forming the C-H σ bond. Fig. 2.5 shows two possible intermediate

structures. Structure (a) is the most probable one because of direct analogies with well

known agostic species64.

a b

Figure 2.5: Two possible structures of intermediates: a) the simplest structure analogous to that observed in agostic complexes b) structure with two coordinated C-H bonds, suggested by theoretical calculations.

LnM CHR

H

H

CH2R

H LnM


25

However, theoretical calculations suggest a range of other possible structure as

well.65 No stable σ-complexes have yet been isolated but many evidences suggest that

they should exist. The M-C bond should be cleaved in order to make the process

catalytic. Once an alkane complex has formed, the coordinated C-H bond is cleaved to

yield the product.

The cleavage of the C-H bond by direct participation of a transition metal ion

proceeds via an oxidative addition mechanism or an electrophilic substitution

mechanism. The former is typical for electron-rich, low-valent complexes of the

transition metals at the right side of the periodic table (Re, Fe, Ru, Os, Rh, Ir, Pt).

Metals in high oxidation states take part in electrophilic substitution reactions.

A general oxidative addition reaction is illustrated in Fig. 2.6.

Figure 2.6: General scheme of oxidative addition reactions

A metal atom (M) inserts itself between the atoms of the C–H bond. The metals

oxidation state is two units higher in the organyl hydride complex than it was in the

initial metal compound. This step is formally analogous to the interconversion between

dihydride and dihydrogen complexes, which is often extremely facile.2

An estimation of the heat of oxidative addition via the following Eq. 2.13

LnM + RH → R-LnM-H (2.13)

shows that this reaction is usually endothermic with ΔH ≈ +10 Kcal/mol. The above

consideration refers to complexes of the first series of transition metals, as the alkyl-

metal bonds may be stronger in the case of heavy metals.66

[LnMx] + RH → LnMx+2

H

R


26

The oxidative addition of alkanes to form alkyl hydride complexes was definitively

demonstrated by Bergman in 1982 in studies using iridium complexes67(Fig. 2.7).

Figure 2.7: Oxidative addition of cyclohexane with Cp*Ir(H)2PMe3.

The complex Cp*Ir(H)2PMe3 (Cp* = pentamethylcyclopentadienyl) was irradiated

in a cyclohexane solution to produce the complex Cp* PMe3Ir (H)(C6H11) which was

then converted into the more stable derivative Cp* PMe3Ir (Br)(C6H11), by treatment

with CHBr3 at -60°C.

It is interesting to note that thermal and photochemical activation sometimes afford

the formation of different products. Heating a solution of RhHBPz*(CO)(C2H4) (Pz* =

3,5 dimethylpirazole) in benzene entails the elimination of the ethylene π ligand and

formation of a phenylrhodium hydride complex.68 However irradiation of the same

solution causes the hydride ligand to add to the ethylene molecule, rather than to metal

atom, which results in the appearance of an σ-ethyl group in the complex.69

A considerable proportion of processes initially proceed by an oxidative addition

mechanism. The first stage forms an alkyl hydride complex which undergoes further

transformations. The resulting reaction may be a H-D exchange, dehydrogenation, or

the introduction of a functional group into a C-H compound.2 Transformation such as

dehydrogenation (Eq. 2.14) and carbonylation (Eq. 2.15)

RCH2CH3 → RCH2=CH2 + H2 (2.14)

CH4 + CO → CH3CHO (2.15)

Ir (PMe3)

IrH

PMe3 H

H


27

should be readily accomplished at centers that activate C-H bonds, but both of these

reactions are thermodynamically uphill anywhere near room temperature. The partial

oxidation of alkanes to alcohols, aldehydes, acids and other oxygenates, is in contrast

thermodynamically favored. However, many of the known C-H activating centers are

highly sensitive to oxygen and other oxidants.1

Various transition metal complexes readily abstract hydrogen atoms from alkanes to

produce π-olefin complexes. If the π-complex is relatively unstable, the π-ligand

dissociates making the reaction catalytic with respect to metal complex. An example

of thermally induced reaction of this type is shown in Fig. 2.8.70

Figure 2.8: Dehydrogenation of cyclooctane. L: PPh(CH2CH2PPh2)2.

Continuous removal of the hydrogen evolved (through reflux or by adding a

molecular hydrogen acceptor such as norbonene) displaces the equilibrium toward the

olefin.

Complexes of rhodium and iridium are known to be the most effective

photocatalysts for alkane dehydrogenation. The full light irradiation of a solution of

RhCl(CO)(PMe3)2 in cyclohexane at room temperature by a mercury high pressure

lamp, induce the formation of cyclohexene and molecular hydrogen.71 Analogously,

the dehydrogenation of n-hexane and alkylcyclohexane affords a mixture of hexenes

and alkylcyclohexenes respectively.72 After photolytic expulsion of a carbonyl ligand

from the complex RhCl(CO)(PMe3)2, oxidative addition to RhCl(PMe3)2 is rapidly

established, followed by elimination of hydrogen from the β-position to yield the

olefin.73 UV light was used in almost all cases but complexes a74 and b75 in Fig. 2.9

catalyze the dehydrogenation of alkanes under irradiation with λ > 375 nm and λ > 450

nm, respectively.

LReH5 or LWH6

Hydrogen acceptor


28

(a) (b)

Figure 2.9: Two complexes which catalyze the dehydrogenation of alkanes under visible light irradiation.

Complex RhCl(CO)(PMe3)2 turned out to be an efficient catalyst for photochemically

introducing a CO group into alkanes.76 The following mechanism was proposed for

this reaction (RH: n-pentane, n-decane).

RhCl(CO)(PMe3)2 + hν → RhCl(CO)(PMe3)2* (2.16)

RhCl(CO)(PMe3)2* + RH → RRhH(Cl)(CO)(PMe3)2 (2.17)

RRhH(Cl)(CO)(PMe3)2 + CO → RCHO + RhCl(CO)(PMe3)2(2.18)

It proceeds through photolytic activation of the metal complex (Eq. 2.16), oxidative

addition of the alkane (Eq. 2.17) followed by reductive elimination of RCHO (Eq.

2.18). At the end, a CO molecule restores catalytically the starting complex.

Other functional groups can be also inserted into C-H bonds through the same

mechanism. Rhodium complexes catalyzed halogenation of adamantane.77 Pt(II) salts

and Pt(II) + Pt(IV) systems have been used to catalyze the oxidation of C-H

compounds with various strong oxidants. For example the reaction of methane with

chlorine in water at 125°C in the presence of platinum chlorides affords methyl

chloride which is partially hydrolyzed to methanol in situ.78

RhCl(PMe3)

Ir Ir S

Ph2P

Ph2P

PPh2

PPh2

CO OC


29

Complexes of metals in high oxidation state normally activate alkanes through

electrophilic attack usually in a strongly polar medium such as water or an anhydrous

strong acid.79

This type of reaction is illustrated in Eq. 2.19 without showing the presumed

organometallic intermediate [LnMx+2(R)(X)].

LnMx+2X2 + R-H → [LnMx] + R-X + HX (2.19)

The earliest example is the platinum chemistry discovered by Shilov and co-

workers. Eqs. 2.20 - 22 illustrates the mechanism of oxidation of CH4 to the

corresponding alcohol at 120 °C.80,81

PtIICl2(H2O)2 + CH4 → PtIICl(CH3)(H2O)2 + HCl (2.20)

PtIICl(CH3)(H2O)2 + PtIVCl62- → PtIVCl3(CH3)(H2O)2 + PtIICl4

2- (2.21)

PtIVCl3(CH3)(H2O)2 + H2O → PtIICl2(H2O)2 + CH3OH + HCl (2.22)

Although the overall conversion is catalytic in Pt(II), it requires stoichiometric

amounts of Pt(IV) and is thus impractical. Moreover the catalytic species are not very

stable in solution and eventually precipitate as metallic platinum. Furthermore,

selectivities in this system are generally not quite as high as in the case of oxidative

addition reactions.

From a practical perspective, perhaps the most impressive accomplishment to

date is the oxidation of methane to methyl bisulphate by sulphuric acid, catalysed by a

Pt(II) complex.81 The reaction appears to be an example of electrophilic

functionalization, in addition, the organometallic complex used in this system is quite

stable and there is no platinum metal formation. A second improvement is the fact that

product can be obtained with up to 90% selectivity. This feature is in part due to the

protective power of the bisulphate group: to achieve the observed yield, the relative


30

reactivity of methane versus methyl bisulphate must be of the order of 100:1.

However, the high selectivity achieved comes at a price: the product is of little direct

use and would need to be separately converted to a more useful compound, such as

methanol.

Another type of reaction which occurs between a saturated C-H bond and a

highvalent metal complex is the metathesis, which occurs according to Eq. 2.23:

M-R + R′-H → M-R′ + R-H (2.23)

For example, the exchange between a methyl complex of lutetium or yttrium

(Cp*2MCH3) and labeled methane apparently proceeds as metathesis via the transition

state showed in Fig. 2.10 a, rather than via an oxidative addition involving

intermediate (Fig. 2.10 b.)82

(a) (b)

Figure 2.10: Intermediates for metathesis (a) and oxidative addition mechanism (b).

M

Cp*

Cp*

CH3

13CH3

H M

Cp*

Cp*

CH3

13CH3

H


31

2.2.2 Activation through Electron and Hydrogen Transfer

In this second group we include reactions in which a metal complex cleaves a C-H

bond but no σ-C-M bond is formed. The function of the metal complex usually

consists of abstracting an electron or a hydrogen atom from the hydrocarbon RH. The

radical ions RH+· or radicals R· formed thereof interact with other species, such as

molecular oxygen which is present in the solution or with one of the ligands of the

metal complex. For example, in the hydroxylation of an alkane by an oxo complex of a

high-valent metal, an alkyl radical is generated and subsequently reacts with a

hydroxyl ligand according to Eq. 2.24:

RH + O=[Mn+] → R· + HO-M(n-1)+ → ROH + M(n-2)+ (2.24)

The remaining metal complex can be then oxidized back to its original form. In

this reaction the metal-oxo complex is a strong oxidant of the type CrO42- or MnO4

- or

an oxoferryl species.83 The addition of strong acids accelerates the reaction giving rise

to protonated species such as O=Cr(OH)3+ and HCrO3

+. Furthermore, the oxidation of

alkanes by oxoderivatives of Cr(VI) is greatly accelerated by irradiation.84,85 Acetic

acid and acetonitrile have been used as solvents and the products were alcohol and

carbonyl compounds. Also complexes containing ruthenium(IV) oxidize in the same

way alkanes both in the dark and under irradiation.86,87

In many cases it can not be decided if the hydrogen transfer step is preceeded by

electron transfer activation step. The final result would be also a metal hydride and an

alkyl radical as showed in Eq. 2.25.

[Mn+] + RH → [RH]+· + [M(n-1)+] → [Mn+]-H + R· (2.25)

The mechanism of aerobic alkane photooxidation catalyzed by metal oxo

complexes includes the formation of a photoexcited species which is capable of


32

abstracting a hydrogen atom from an alkane. The alkyl radical thus formed rapidly

adds a molecule of oxygen. The resulting species eventually forms an alkyl

hydroperoxide which decomposes to produce a ketone and an alcohol.

MVI=O + hν → MV-O· (2.26)

MV-O· + RH → MV-OH + R· (2.27)

R· + O2 → ROO· (2.28)

ROO· + MV-OH → ROOH + MVI=O (2.29)

2MV-OH + ½O2 → 2MVI=O + H2O (2.30)

2.2.3 Indirect Activation via Reactive Oxygen Species

Whereas the reactions included in the first and second group require direct contact

between a molecule of the C-H compound and the metal complex (albeit via the

ligand), complexes belonging to the third type initially activate some other reactant

(e.g. O2 or H2O2) to form a reactive species which then attacks the hydrocarbon

molecule. The reactive species is usually a radical, such as a hydroxyl radical, which

attacks the hydrocarbon independent of any participation of the metal complex.

Many industrial processes are based on these reactions and this is a field of

growing interest due to the mechanistic similarities with enzymatic reactions.

The industrial oxidation of alkanes proceeds through heating them under oxygen

at rather high temperature (usually above 100°C). These reactions are always

termodynamically favored due their high exothermicity. The main problem is to

prevent various parallel and consecutive reactions which make these processes

unselective. Being radical chain processes, any additive which can react with free

radicals and form stable adducts, will inhibit deep oxidation.88 Ions of transition metals

are used as catalyst in these reactions.89 The role of the metal ion Mn+ is to produce

free radical reacting with a molecule AB according to the Eqs. 2.31 and 2.32.


33

Mn+ + A-B → M(n+1)+ + A· + B- (2.31)

M(n+1)+ + A-B → Mn+ + A+ + B· (2.32)

The metal ion can alternately increase or decrease its oxidation state; thus, it plays

the role of a catalyst. Furthermore, it can also reacts with hydroperoxides formed in the

course of the oxidation to produce new free radicals as in Eq. 2.33 - 35 with Co (II/III)

as catalyst.90,91

ROOH + CoIII → ROO· + H+ + CoII (2.33)

CoII + ROO· → ROO- + CoIII (2.34)

ROOH + CoII → RO· + OH- + CoIII (2.35)

Thermal reactions operating at milder conditions have been developed recently.

The weakly solvated acetonitrile complex [Co(NCMe)4](PF6)2, catalyzes the air

oxidation of cyclohexane and adamantane at 75 °C to the corresponding ketones and

alcohols.92 Halogenated metalloporphyrins are catalysts for the selective air oxidation

of light alkanes to the corresponding ketones.93 Worth to mention are the Gif systems94

(from Gif-sur-Yvette) for selective oxidation and oxidative functionalization of

alkanes in mild conditions by molecular oxygen in the presence of a reducing agent, an

iron complex, a carboxylic acid and pyridine. Despite of numerous works devoted to

Gif systems, their mechanism is not clear and they are of no technical relevance.

In 1989 a few groups simultaneously described the aerobic photo-oxygenation of

alkanes in solution containing catalytic amounts of metal oxo complexes to alcohols

and ketones. These complexes were heteropolymetalates95 and polyoxotungstate96 in

various solvents. Other oxo compounds which also photocatalyze alkane oxygenation

include K2Cr2O7,97 CrO3,98 (nBu4N)Cr4O13,99 and UO2Cl2.100 Also molybdenum and

vanadium complexes have been reported to catalyze alkane photooxidation.101


34

Also iron(III) chloride has been found to be a photocatalyst for alkane oxidation

with atmospheric oxygen.102 The first step of this process seems to be the

photoexcitation of the iron chloride followed by homolysis of the Fe-Cl bond. The

chlorine radical then attacks the alkane. The resulting Fe(II) can be back oxidized by

molecular oxygen or by an alkylperoxo radical. Final products are the corresponding

alcohols and ketones. Other transition metal chlorides such as CuCl2, AuCl4-, PtCl6

2-,

CrCl3 also catalyze this reaction. The ketone:alcohol ratio can be varied through

choosing different solvents. Furthermore, in the presence of small amount of

hydroquinone, the formation rate of ketone sharply decreases while the formation rate

of alcohol does not. Thus, free radicals may participate in the ketone formation while

alcohol formation does not involve them.

In summary, although a large group of alkane activation reactions have been

reported in the literature, no technical application could be developed from these basic

studies.


35

2.3 Alkane Activation at Titania

Heterogeneous catalysis in selective organic synthesis is not frequently employed,

although nowadays the demand for replacement of traditional oxidation methods with

cleaner ones is increasing. TiO2 sensitized organic photosynthetic reactions include

oxidation and oxidative cleavage, reduction, isomerization, substitution and

polymerization.103,104 Alkanes activation through UV-light irradiation of TiO2

suspensions is well documented in literature. Highly oxidizing OH radicals (the

oxidizing potential of this radical is 2.8 V, being exceeded only by fluorine) formed

through reductive and/or oxidative paths are able to oxidize alkanes to alkyl radicals

which then afford the end products through a radical chain.

VB holes can also react directly with organic compounds before they are trapped.

A thermodynamic estimation for the concerted process in aqueous solution reveals that

RH → R· + H+ + e- (2.36)

the reaction (Eq. 2.36) is endergonic by at least 1.85 eV. This value can be obtained by

using E0 (H+/H) = -2.42 V (H20)105 and a bond dissociation energy of 3.22 eV (for n-

heptane) and converting these values to the ΔG(H2O) values by substracting 0.1 eV

for the solvent contribution. Assuming a potential of 2.5 V for the photogenerated hole

in the valence band of TiO2 make this oxidation thermodynamically possible.

In this sense almost every photosynthetic reaction with TiO2 can be included in the

third class of the classification made in Paragraph 2.1. In fact, TiO2 may be seen as an

non conventional Ti(IV) complex which forms active species which in turn activate a

substrate. Generally, photocatalytic oxidation of alkanes at TiO2 affords products

depending on the reaction medium. The oxidation of neat liquid n-heptane and 2,2-

dimethylbutane at Pt/TiO2 under UV light irradiation (the presence of deposited

platinum is, in fact, not necessary), is reported to lead to three ketones and one ketone,

respectively, as is expected if the oxidation takes place only on secondary C-atoms. No

cleavage products were found in contrast to the same reaction performed with the


36

corresponding gases. Furthermore, different selectivities can be obtained by operating

in the gas phase or in neat-liquid phase, depending on the organic substrate. In fact, the

liquid alkane can act as a solvent of the products of primary oxidation and prevent

them from further oxidation.104

The oxidation of cycloalkanes leads to ketones as major products. The highest

reactivity is achieved for cyclohexane. Cyclohexanol and cyclohexanone, key products

in the synthesis of adipic acid and caprolactam, are obtained conventionally by

catalytical oxidation of cyclohexane with molecular oxygen at elevated temperatures

and pressures in a series of liquid-phase reactors. The single step conversion is kept

low, usually under 10% to minimize deep oxidation and formation of CO2. Using

photocatalysis with TiO2 under UV light irradiation, the oxidation of cyclohexane can

be obtained in the liquid phase at room temperature and pressure.40,106,107 Utilizing a

proper solvent (that minimizes the adsorption strength of the desired products on TiO2,

does not compete with cyclohexane and oxygen for adsorption sites, and does not form

radicals on the illuminated TiO2 surface) leads to an increase of the reaction rate and

the selectivity to cyclohexanol and cyclohexanone and a higher ratio

cyclohexanol/cyclohexanone over the use of neat cyclohexane (in which 85% of the

product was cyclohexanone, 2% was cyclohexanol and 12% CO2). The highest

product formation rate mentioned in literature is observed for dichloromethane as

solvent, which preferentially adsorbs on the TiO2 surface forming a reactive radical

which then abstracts a hydrogen atom from cyclohexane.

It is worth to note that the concentration of the monooxygenation products

increases rapidly in the partial pressure range from 10 to 100 Torr and that above 200

Torr the dependence on pO2 becomes negligible. After formation of alkyl radicals,

cyclohexanol is likely obtained by reaction of species formed via valence band

oxidation process as shown in Eqs. 2.37 - 41.

R· + (OH·)ads → (ROH)ads (2.37)

R· + O2 → ROO· (2.38)

ROO· + RH → R· + ROOH (2.39)


37

ROOH + e- (TiIII) → RO· + TiIV-OH (2.40)

RO· + RH → ROH + R· (2.41)

Cyclohexanone is mainly formed by reaction of intermediate radicals with

activated oxygen species as illustrated in Eqs. 2.42 - 45.

(R2CHOH)ads + 2OH· → R2C=O + 2H2O (2.42)

R2CHOO· + e- (TiIII) → R2C=O + TiIV-OH (2.43)

R2CH· + O2- → R2C=O + OH- (2.44)

R2CH· + HO2· → RO + H2O (2.45)

An interesting study of functionalization was performed with adamantane.108 This

C-H activation has been obtained with TiO2 under UV light irradiation through either

oxygen incorporation109 or C-C bond forming reactions.108 Both oxygen and

inorganic/organic oxidants have been used as electron scavengers. Whereas in MeCN

under air 1- and 2-adamantanol and adamantanone are produced with limited

degradation and preference for functionalization at the 1-position, the oxidation is less

selective in CH2Cl2. In N2-flushed CH3CN solutions with Ag+ as electron acceptor,

products from trapping of both 1-adamantyl radical (adamantyl methyl ketone) and

cation (N-adamantylacetamide) are obtained. Irradiation in a mixed solvent of CH3CN

(2.0 g) and C3H7CN (1.96 g) containing adamantane (40 mg) and TiO2 powder as the

photocatalyst at λ < 340 nm affords 30 μmol of 1-adamatanol, 5 μmol of 2-

adamantanone and minor amounts of other isomers.

In summary, it is recalled that all these photoactivation reactions require the use of

expensive UV light. In artificial systems it is produced by gas discharge lamps

containing toxic mercury, whereas in the natural system (solar light) it is present only

in a very small amount (4%).


38

The starting point of this dissertation was to find a TiO2-based catalyst which

enables alkane activation under visible light irradiation. The novel rhodium modified

titania, described in the next chapter, induces a fast degradation of the model pollutant

4-chlorophenol and it is quite active in the visible catalytic photosulfoxidation of

alkanes described in Chapter 4.

3. Visible-Light Photocatalysis by a Titania-Rhodium (III) Complex

39

3. Visible-light Photocatalysis by a Titania-Rhodium (III) Complex∗

3.1 Introduction

In the Kisch group was recently found that surface-modification of titania by

platinum(IV) chloride affords photocatalysts active in the mineralization of 4-

chlorophenol (4-CP) and many other pollutants with visible light (λ ≥ 455 nm).61,110

These novel materials are easily obtained through stirring a suspension of anatase

powder (TH) in hexachloroplatinate solution and subsequent thermal treatment. From

desorption experiments it was concluded that chemisorption took place affording an

oxygen bound surface complex of the proposed composition [Ti]OPtCl4Ln-, (L =

H2O, OH-, n = 1, 2,)61 abbreviated as Pt(IV)/TH in the following. Thus, the

semiconductor may be considered as an unconventional ligand in a transition metal

coordination complex. The quasi-Fermi level of electrons (nEF*) of this hybrid

semiconductor was changed considerably as a function of surface loading. At pH 7 the

value of −0.54 V as found for the anatase hydrate powder TH was shifted to −0.49,

−0.45, and −0.28 V (vs. NHE) when the surface was covered by 1.0, 2.0, and 4.0wt%

of platinum, respectively. This resembles the corresponding anodic shift observed

upon adsorption of fluoride ions.111 One of the most active photocatalysts in 4-CP

mineralization was 4.0%Pt(IV)/TH, a high surface area (260 m2g-1) material

containing 4.0wt% of platinum. Both upon visible and ultraviolet excitation this novel

titania complex is a superior photocatalyst as compared to previously known titania

materials. It even catalyzes the mineralization of cyanuric acid, which is usually the

final product in atrazine degradation by titania and other advanced oxidation

processes. Fig. 3.1 summarizes the proposed mechanism for the primary reaction

steps.112,113

∗ Part of this work has been published.


40

O2/O2–

CB

VB

TiO2

LnPtIII···Cl0

LnPtIV - Cl

hν

ArOH

BET - H+

ArO·

O2/O2–

CB

VB

TiO2

LnPtIII···Cl0

LnPtIV - Cl

hν

ArOH

BET - H+

ArO·

Figure 3.1: Mechanistic scheme of titania sensitization by Pt(IV) chloride complexes (according to ref.61 ).

According to this the excited platinum surface complex undergoes first homolytic

PtIV-Cl bond cleavage affording a PtIII intermediate and a surface bound chlorine

atom.48,114-117 In the reductive reaction path the platinum(III) species injects an electron

into the titania conduction band from where it is subsequently transferred to oxygen.

Since the titania semiconductor ligand is covalently attached to the chloroplatinate

chromophore, a strong electronic coupling is expected rendering this step fast enough

to efficiently compete with the undesired back electron transfer (Fig. 3.1, process

BET). In the oxidative reaction pathway 4-chlorophenol (ArOH) transfers an electron

to the chlorine atom. As summary of both pathways the PtIV-Cl fragment is reformed.

To find out whether analogous surface modification is feasible also with

chlorides of other d6 metals, we report here on the preparation and photocatalytic

properties of a rhodium (III) chloride modified titania.


41

3.2 Synthesis and Characterization

The novel hybrid photocatalysts x%RhCl3/TiO2 containing 0.5, 1.0, 2.0, and

5.0wt% of rhodium were prepared by stirring a suspension of TiO2 in a corresponding

amount of aqueous rhodium(III) chloride and subsequent heating at 200 0C. The best

commercial titania material for this modification turned out to be the titania hydrate

TH. Modified Hombikat and P25 exhibited less activity in the standard degradation of

4-CP. Therefore in the following we will consider only TH modification.

Maximum loading was observed at 5.0wt% of rhodium since the use of higher

metal chloride concentrations afforded powders from which excess rhodium is

completely removed during washing (see experimental part). In order to understand

the role of the halogen ligand, also 2.0%RhBr3/TH and 4.0%RhBr3/TH were

synthesized.

3.2.1 Desorption Experiments

RhCl3/TH and RhBr3/TH have a pink and a dark yellow color, respectively, and

are surprisingly stable to desorption of the rhodium component as compared with the

previously reported platinum modified TH.61,117 In aqueous suspension upon stirring

either in the dark or under irradiation with visible light, no dissolved rhodium complex

was detectable by UV-vis absorption spectroscopy.

Although it is known that F- ions form very stable Ti-F bonds, both

4.0%RhCl3/TH and 4.0%RhBr3/TH did not undergo desorption of the rhodium

surface-complex even after stirring for five days in the dark in 0.5 M KF. Thus, by

analogy with Pt(IV)/TH one can conclude that Rh(III) is covalently bound to titania

through a bridging oxygen ligand. This fluorinated sample, within experimental error,

exhibited the same photoactivity as the unfluorinated samples in the standard

degradation of 4-CP.

Whereas in 0.1 mol dm-3 HCl the previously reported 4.0%Pt(IV)/TH upon UV

irradiation61 for 24 h suffers almost complete desorption to [PtCl6]2-, only 40% of

[RhCl6]3− were detectable in the case of 4.0%RhCl3/TH. This difference may reflect

the fact that the metal-oxygen bond is about 40 kJ mol-1 stronger in the case of


42

rhodium.118 However, in the presence of 6 mol dm-3 HCl complete desorption of the

rhodium complex is observed. In strongly alkaline suspension the chloride ligands are

completely displaced, as also observed for platinum(IV) chloride modified TH.61 Since

from the amount of chloride produced in this experiment one can conclude that three

chloride ligands are present in the surface rhodium complex, a composition of

[TiO2]-O-RhCl3(H2O)2− is suggested (Fig. 3.2). An analogue structure is proposed

for the RhBr3/TH catalysts.

Figure 3.2: Proposed structure of the rhodium(III) surface complex; L = H2O.

3.2.2 Diffuse Reflectance Spectra

3.2.2.1 Principles of DRIFTS

One of the fundamental electronic properties of a semiconductor is size and

location of bandgap. The excitation of an electron from the valence band to the

conduction band is indicated by a sudden increase in absorptivity at the wavelength

corresponding to the energy difference between the two bands. In case of a

semiconductor powder conventional transmission spectroscopy is rarely applicable due

to problems in preparing transparent plates. For such samples diffuse reflectance

spectroscopy is the method of choice. 119,120

A diffuse reflectance spectrum is obtained by measuring the ratio of light scattered

from the sample and from a non-absorbing material like BaSO4, as a function of the


43

wavelength. Assuming wavelength independent scattering, the absorption coefficient

of the powder can be considered proportional to the Kubelka-Munk function F(R∞)

that can be obtained from diffuse reflectance data as

F(R∞) = ∞

∞−RR

2)1( 2

= sα (3.1)

where R∞ is diffuse reflectance from an infinitely thick sample layer relative to the

reflectance of a standard (e.g. BaSO4). α and s are the absorption- and scattering-

coefficients, respectively. Eq. 3.1 is valid only under well defined conditions:

• Monochromatic irradiation

• Infinitely thick sample (normally about 5 mm)

• Low sample concentration

• Uniform distribution

• Absence of fluorescence.

The absorption coefficient varies with wavelength and its magnitude depends on

whether the semiconductor is a direct or indirect semiconductor.

In direct semiconductors the minimum of the conduction band has the same

momentum k of the maximum of the conduction band and, since for any collision

energy and momentum must be conserved, the direct transition requires only the

absorption of a photon. In contrast for an indirect semiconductor the absorption of a

photon leads to a change in momentum. Therefore the absorption of a photon with its

negligible momentum is not enough to cause this change and a third particle with

significant momentum, a phonon (quantisized lattice vibration) must be emitted or

absorbed additionally. The indirect transition is a two-step process and it is less

probable than the single step direct transition, which typically leads to higher

absorption coefficients for direct semiconductors as compared to indirect ones.


44

The dependence of the absorption coefficient α on the photon energy hν can be

described as26

α ∝ ν

−ν

h)Eh( 2

n

g (3.2)

where n is a constant depending on the nature of the optical transition:

n = 1/2 for allowed direct transitions

n = 3/2 for forbidden direct transitions

n = 2 for allowed indirect transitions

n = 3 for forbidden indirect transitions.

When the scattering coefficient s is assumed to be independent of the wavelength and

proportional to the absorption coefficient, can be written as in Eq. 3.3

F(R∞) ∝ α (3.3)

And combining with the Eq. 3.2 leads to Eq. 3.4

(F(R∞) hν)2/n ∝ (hν − Eg) (3.4)

In case of amorphous semiconductors an energy dependence of k was found to be

as follows121

(F(R∞) (hν)2)1/2 ∝ (hν − Eg) (3.5)

A plot of (F(R∞) hν)1/2 versus the incident photon energy hν and a subsequent linear

extrapolation defines the band gap energy for an indirect, crystalline semiconductor.


45

Analogously a plot of (F(R∞) hν)2 and (F(R∞) (hν)2)1/2 versus hν affords the band gap

of a direct crystalline and amorphous semiconductor, respectively.

3.2.2.2 Diffuse Reflectance Measurements

Comparison of the diffuse reflectance spectra of TH and 4.0%Rh(III)/TH clearly

indicates novel absorption at 400 - 500 nm and 500 – 700 nm (Fig. 3.3, curves a, c).

400 500 600 700 800

0,0

0,1

0,2

0,3

0,4

F(R

∞)

Wavelength / nm

a

b

c

Figure 3.3: Diffuse reflectance spectra of TH, 2.0%RhCl3/TH and 2.0%RhBr3/TH. The Kubelka-Munk function, F(R∞), is used as the equivalent of absorbance; a: TH, b: 2.0%RhCl3/TH, c: 2.0%RhBr3/TH.

The shoulder at about 500 nm compares well with the lowest d,d transition of

[RhCl6]3− observed in hydrochloric acid at 518 nm 122 (Fig. 3.4).


46

400 500 600 7000,000

0,005

0,010

0,015

0,020

0,025

0,030

Abso

rban

ce /

a.u.

λ / nm

Figure 3.4: Absorption spectrum of [RhCl6]3- in HCl.

At wavelengths shorter than about 550 nm a strong absorption increase suggests that it

does not originate exclusively from the second d,d-transition occurring in [RhCl6]3− at

410 nm with about the same intensity as the 510 nm band. It rather may originate from

a rhodium-to-titanium charge transfer transition (MMCT) as also reported for other

titania-metal-complex systems like [Fe(CN)6]3−/TiO2.123 This is corroborated by the

fact that the silica analogue 2.0%RhCl3/SiO2 does not exhibit a strong absorption

increase at λ ≤ 550 nm, most likely because, different from titania, silica does not have

a low lying conduction band (Fig. 3.5, curve b).


47

300 400 500 6000,0

0,1

0,2

0,3

F(R

∞)

λ / nm

c

a

d

b

380nm

Figure 3.5: Diffuse reflectance spectra of 2.0%RhCl3/TH (a), 2.0%RhCl3/SiO2 (b), TH (c). Spectrum d = a – (b + c).

In the corresponding difference spectrum an unsymmetrical absorption band is

observed at a maximum of 380 nm. In the case of 2.0%RhBr3/TH a similar comparison

with 2.0%RhBr3/SiO2 afforded the MMCT maximum at 390 nm (Fig. 3.6).


48

300 400 500 600 700 8000,0

0,1

0,2

F(R

∞)

λ / nm

a

390 nm

b

c

d

Figure 3.6: Diffuse reflectance spectra of 2.0%RhBr3/TH (a), 2.0%RhBr3/SiO2 (b), TH (c). Spectrum d = a – (b + c).

Assuming that all samples are indirect crystalline semiconductors, as is anatase,

the bandgap energy can be obtained by extrapolation of the linear part of a plot of

[F(R∞)hν]1/2 vs. the energy of exciting light as showed in Fig 3.7.121


49

2 3 4

0

2

4

6

8

(F(R

∞)E

)1/2

E / ev

b a

Figure 3.7: Transformed diffuse reflectance spectra of TH and 5.0%RhCl3/TH. The bandgap energy was obtained by extrapolation of the linear part; a: TH, b: 5.0%RhCl3/TH.

From this the bandgap of TH, 0.5, 1.0, 2.0, and 5.0%RhCl3/TH and of 2.0%RhBr3/TH,

can be calculated as 3.29, 3.26, 3.25, 3.22, 3.21 and 3.1 eV, respectively as

summarized in Tab. 3.1.

Photocatalyst Ebg / eV

TH

0.5% RhCl3/TH

1.0% RhCl3/TH

2.0% RhCl3/TH

5.0% RhCl3/TH

2.0% RhBr3/TH

3.29

3.26

3.25

3.22

3.21

3.1

Table 3.1: Bandgap energies of TH, 0.5%, 1.0%, 2.0%, 5.0%RhCl3/TH and 2.0%RhBr3/TH. Reproducibility was better than ± 0.05 eV.


50

3.2.3 Determination of Band Edge Positions

3.2.3.1 Semiconductor-Electrolyte Interface

The Gerischer model25,124-126 provides a useful description of the semiconductor-

electrolyte interface. The electrochemical potential of electrons Eredox in an electrolyte

containing a redox system is given by the Nernst equation:

red

oxredoxredox a

anFRTEE ln0 += (3.6)

Where E0redox is the standard reduction potential, R is the universal gas constant, T

is the absolute temperature, F is the Faraday constant, n is the number of electrons

transferred and aox and ared are the activities of the oxidized and reduced species

respectively. In analogy to the Frank-Condon principle the Gerischer model assumes

that the electron transfer is much faster than the reorientation of the solvation shell.

The reduced form of a redox species Redsolv,red is surrounded by its corresponding

solvation cage. Removing an electron affords the oxidized species Oxsolv,red with the

same solvation shell. During the following relaxation the low-energy equilibrium state

Oxsolv,ox is reached through releasing the reorganization energy λ, which normally

assumes values between 0.5 and 2 eV depending on the strength of interaction between

the solvent and the molecule.127 Analogously, the reduction process of Oxsolv,ox affords

Redsolv,red through releasing the reorganization energy from the intermediate Redsolv,ox.

The fluctuations of the solvent shell demand that the energetic level of a redox system

involved in a charge transfer process can not be described by one discrete value E0

red/ox but rather by a harmonic oscillation behavior; therefore, the electronic empty and

occupied states of a redox couple are represented by a Gaussian type of distribution

Wox(E) and Wred(E), respectively. The density of states Dox(E) and Dred(E) are defined

as the areas under Wox(E) and Wred(E), respectively, and are proportional to the

concentrations of the oxidized and reduced species cox and cred.


51

)()( EWcED oxoxox = (3.7)

)()( EWcED redredred = (3.8)

Figure 3.8: Electron energies of a redox system and the corresponding distribution functions D: (a) cox = cred; (b) cox << cred (note that EF,redox is shifted according to the Nernst equation). E0

ox is actually an electron affinity A and E0red corresponds to ionization

energy I. (Adapted from Ref. 124)

A semiconductor particle can be considered as a microelectrode and described in

the same way. After the contact between the electrolyte and the semiconductor surface,

the equilibrium between the Fermi level of the semiconductor and the redox potential

of the solution must be established after the two phases get into contact.23,124 Since the

number of available states in the solution (for concentrated solutions) is much higher

than those in the semiconductor, the Fermi level will adjust to the redox potential of

the electrolyte. In the case of an n-type semiconductor this process happens through

electron transfer from the solid to the solution, which will leave a positively charged

layer behind (space charge layer). This region is defined depletion layer W because it

is depleted of its majority charge carriers.


52

Figure 3.9: Schematic energy model of the n-type semiconductor/electrolyte interface before (a) and after (b) the establishment of equilibrium. (Taken from ref.36)

On another hand, in the solution is formed an excess of negative charges (because

negative ions and dipoles will accumulate at the interface) to form the so called

Helmholtz layer. An electrical field arises and finally stops further electron transfer so

that the equilibrium is established. This energy barrier is reflected in an upward

bending of the band edges for an n-type semiconductor (downward for a p-type

semiconductor) and the height of the barrier corresponds to the potential drop in the

space charge layer Us showed in Fig. 3.9.

It must be noted that when the Fermi level of the semiconductor is equal to the redox

potential of the electrolyte, band bending does not occurs. But also in the case of band

bending a situation of flat band potential can be achieved applying an external

potential to the semiconductor. In this case only the band edges in the bulk, beyond the

depletion layer can be shifted, the band position at the interface remaining unaffected.

Therefore an applied potential can vary the magnitude and the direction of band

bending. For an n-type semiconductor the situation of flat band can be achieved

applying a negative potential. The potential where no bending corresponds to the flat

band potential Efb and therefore to the Fermi level in the absence of an applied


53

potential. The flat band potential is particularly important because it allows to extimate

the position of the conduction band in a n-type semiconductor and of the valence band

in a p-type semiconductor. Assuming that both semiconductors are heavily doped, the

distance of the flat band potential to the band edges should be so small (0.05 - 0.1 eV)

that it can be neglected for the purpose of this investigation.

3.2.3.2 Concept of Quasi Fermi Level

As already mentioned the Fermi level represents the free energy of the electron and

holes under equilibrium conditions, i.e. in the dark. In an n-type semiconductor in the

dark there is a higher concentration of electrons respect to the holes. Therefore one can

neglect them and the Fermi level lies close to the conduction band. Nevertheless under

radiative conditions, the concentration of holes increases dramatically whereas the

concentration of electron does not change considerably. Therefore the free energy of

electrons and holes can not be expressed by only one Fermi level, thus it is splitted

into two “quasi-Fermi levels” one for holes pEF* and one for electrons nEF*.126 The

former lies near the conduction band, the latter near the valence band. If the particle is

big enough to form a space charge layer, the photogenerated exciton is separated due

to the associated electric field: the electron move to the bulk and the hole to the

interface.

Several methods have been used in order to measure the flat band potential. In the

case of semiconductor powders it can be measured with a method first developed by

Bard et al. based on photocurrent measurements128-130 and then modified by Roy at al.

who performed photovoltage measurements.131

3.2.3.3 Photovoltage Measurements

The quasi Fermi level of the novel rhodium(III) modified titania catalysts was

determined by the method of Roy et al.131 In short, the pH dependence of the potential

of a platinum electrode immersed in an irradiated semiconductor suspension is

recorded in the presence of an electron acceptor with pH independent reduction

potential. A schematic view of the experimental set up for these measurements is

showed in Fig. 3.10.


54

Figure 3.10: (a) Schematic view of the experimental set-up for the determination of quasi-Fermi level of electrons using the method of Roy. A 60 ml solution of KNO3 (0.1 M) with a small amount of MV2+ or DP2+ is bubbled through with nitrogen in order to avoid the reoxidation of MV+• by dissolved oxygen. (b) Determination of the value of pH0 from a typical potential vs. pH curve. (Taken from ref. 36)

In the case of RhCl3/TH the electron acceptor used was the MV2+ (methyl

viologen; 1,1’-dimethyl-4,4’-bipyridinium dichloride; EMV2+/+· = -0.45 V vs NHE)

whereas in the case of RhBr3/TH the electron acceptor used was DP2+ (4,5-dihydro-

3a,5a-diaza-pyrene dibromide; EDP2+/+· = -0.27 V vs NHE)

Figure 3.11: Structures of MV2+ (a) and DP2+ (b).

The basis of the method is represented in the Fig. 3.12.


55

Figure 3.12: Schematic view of the principle of the determination of *EFn (Taken from ref. 36). Upon increasing the pH of the solution, the band edges of a semiconductor shift to more negative potentials. At pH = pH0 the value of *EFn matches the reduction potential of a pH-independent redox species in the electrolyte (e.g., MV2+)

Increasing the pH of the solution shifts to more negative potentials the band edges

of TiO2 enabling the electron in the conduction band of the semiconductor to reduce

the electron acceptor in the electrolyte. The inflection point (pH0) of the potential-pH

curve determines the pH value at which nEF* coincides with the redox potential of the

electron acceptor. The presence of oxygen in this procedure must be avoided because

O2 will be preferentially reduced by the conduction band electrons instead of MV2+.

Therefore, photovoltage measurements have to be conducted under inert gas

atmosphere and the solutions have to be degassed prior to experiments.

The dependence of the flat band potential on the pH can be written as

Efb(pH) = Efb(pH = 0) − k • pH (3.11)

Where k is a constant factor of 59 mV.132

At pH = pH0 with methyl viologen as electron acceptor


56

E0MV2+/+· = Efb(pH0) = Efb(pH = 0) − k · pH0 (3.12)

Thus, subtracting Eq. 3.11 from Eq. 3.12 the value of nEF* at pH = 7 can be obtained

according to Eq. 3.13

Efb(pH) = E0MV2+/+· + k (pH0 − pH) (3.13)

The experimental results obtained with our catalysts are shown in Fig. 3.13.

4 6 8 10

-0.2

0.0

0.2

0.4

0.6

edcba

Pho

tovo

ltage

/ V

pH

Figure 3.13: Photovoltage recorded for 20 mg of TH (a) and 20 mg of 0.5%(b), 1.0%(c), 2.0%(d), and 5.0%RhCl3/TH (e) suspended in 100 cm3 of 0.1 mol dm-3 KNO3 in the presence of 15 mg of methylviologen dichloride and irradiated with UV light (λ ≥ 320 nm); The position of the inflection point pH0 is marked with a dotted line.


57

The quasi-Fermi level of electrons is shifted gradually more anodic upon increasing

the rhodium loading. Thus, the value of −0.55 V (vs. NHE, at pH = 7) as observed for

unloaded TH is shifted to −0.53, −0.48, −0.46, and −0.34 V upon loading with 0.5, 1.0,

2.0, and 5.0% of rhodium, respectively (Fig. 3.13 and Tab. 3.2). The quasi Fermi level

in the case of 2.0%RhBr3/TH was -0.32 V.

Photocatalyst nEF*(pH=7, NHE) / V

TH

0.5% RhCl3/TH

1.0% RhCl3/TH

2.0% RhCl3/TH

5.0% RhCl3/TH

2.0% RhBr3/TH

-0.55

-0.53

-0.48

-0.46

-0,34

-0.32

Tab. 3.2: Quasi-Fermi potentials of TH, 0.5%, 1.0%, 2.0%, 5.0%RhCl3/TH and 2.0%RhBr3/TH. Reproducibility was better than ± 0.02 V.

3.2.4 Photocatalytic Activity

To investigate the photocatalytic activity, the disappearance and mineralization of

4-CP, an ubiquitous pollutant in water, was performed in the presence of air.

Surprisingly, the activity of 5.0%RhCl3/TH was very high and after 60 min of visible

light irradiation (λ ≥ 455 nm) 95% of 4-CP had disappeared whereas 75% of 4-CP

were completely mineralized (Fig. 3.14).


58

0 10 20 30 40 50 600.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

d

c

b

a

TOC

/ TO

C0

4-C

P, c

/ c 0

time / min

Figure 3.14: 4-CP disappearance and mineralization upon visible light irradiation (λ ≥ 455 nm); c0 = 2.5×10-4 mol dm-3; photocatalyst dosage: 0.5 g dm-3; a: TH, b: P25, c and d: 5.0%RhCl3/TH.

2.0%RhBr3/TH exhibited a photoactivity comparable to that of 2.0%RhCl3/TH.

The unmodified powders TH, Hombikat and P25 were inactive under these

experimental conditions.

The photocatalytic activity increases with increasing rhodium loading, exhibiting

the highest value for 5.0%RhCl3/TH (Fig. 3.15). This resembles our recent findings on

the surface-loading of TH with [PtCl6]2-.61,110


59

0 10 20 30 40 50 600.0

0.2

0.4

0.6

0.8

1.0

Ad

c

b

a

4-C

P, c

/ c 0

time / min

0 10 20 30 40 50 600.0

0.2

0.4

0.6

0.8

1.0

B d

c

b

a

4-C

P, T

OC

/ TO

C0

time / min

Figure 3.15: 4-CP disappearance (A) and mineralization (B) as function of rhodium content; λ ≥ 455 nm; c0 = 2.5×10-4 mol dm-3; photocatalyst dosage: 0.5 g dm-3; a: 0.5%RhCl3/TH, b: 1.0%RhCl3/TH, c: 2.0%RhCl3/TH, d: 5.0%RhCl3/TH.


60

3.2.5 Understanding the Mechanism

To obtain experimental evidence for a mutual formation of OH radicals under

visible light irradiation (λ ≥ 400 nm), the photodegradation of benzoic acid in the

presence of 4.0%RhCl3/TH and oxygen (Fig. 3.16) was investigated by monitoring the

production of salicylic acid.110,133,134

Figure 3.16: Oxidation of benzoic acid to salicylic acid through OH radicals.

Surprisingly, no salicylic acid was detectable in solution. A likely reason for this

could be a fast photodegradation of small amounts of initially produced salicylic acid.

To test this hypothesis, photodegradation of salicylic acid was carried out under

identical experimental conditions. The results show that salicylic acid is efficiently

adsorbed onto 4.0%RhCl3/TH (ca. 63% after 12 h of dark adsorption from a 1.0 x 10-4

mol dm-3 solution) and that its photodegradation is very fast. About 96% of salicylic

acid had disappeared after 10 min of irradiation. These results suggest that salicylic

acid formed from benzoic acid largely remains adsorbed and is efficiently decomposed

before being desorbed into solution.

To test if the photocatalytic activity of RhCl3/TH is also initiated by a

homolytic M-Cl bond cleavage, as proposed previously for platinum(IV) chloride

modified TH,61,117 the photodegradation of phenol under visible light irradiation (λ ≥

455 nm) was carried out. Formation of chlorophenol would evidence the presence of

intermediate chlorine atoms. However, no significant amount of chlorophenol was

detectable. This differs from Pt(IV)/TH, in which case chlorophenol formation was

observable.61

OH·


61

Furthermore cyanuric acid, a molecule which is mineralized in the presence of

platinum(IV) modified TH117, is not decomposed by 4.0%RhCl3/TH.

Figure 3.17: Tautomeric structures of cyanuric acid.

These significant differences indicate that in the case of rhodium(III) modification

visible light induced cleavage of the metal-halogen bond is not a major primary

photoprocess. More likely seems a mechanism as proposed for UV light induced

oxidation reactions in the presence of Rh(III) doped nanosized titania colloids.135

[TiO2]O-Rh3+ + hν → [TiO2]O-Rh4+ + e-CB (3.14)

[TiO2]O-Rh4+ + 4-CP → [TiO2]O-Rh3+ + 4-CP+• (3.15)

O2 + e-CB → O2

-• (3.16)

O2-• + H+ → HO2

• (3.17)

HO2• + HO2

• → H2O2 + O2 (3.18)

H2O2 + O2-• → OH• + OH- + O2 (3.19)

H2O2 + e-CB → OH• + OH- (3.20)

Visible light excitation within the MMCT band of RhCl3/TH (Fig. 3.5) affords as

primary products an electron in the titania conduction band and a Rh(IV) center (Eq.

3.14).


62

The energetic position of the latter can be estimated by adding the energy of the

visible absorption onset (1.77 eV) to the quasi-Fermi level as depicted in Fig. 3.18

(Assuming that light absorption originates rather from transitions between rhodium

and conduction band energy levels than within localized rhodium energy states). The

resulting potential of 1.43 V is positive enough to oxidize water or more likely 4-

chlorophenol to the radical cation (Eq. 3.15), which finally breaks down to CO2, HCl,

and H2O, as well known from the UV photodegradation in the presence of unmodified

TiO2.112,113 The electron generated in the conduction band reduces oxygen to

superoxide (Eq. 3.16) produces an OH radical through the reaction sequence according

to Eq. 3.17 - 20136-138 which in turn induces oxidation of 4-CP.

The described mechanism is schematically depicted in Fig. 3.18.

Figure 3.18: Mechanistic scheme of titania sensitization by rhodium(III) complexes. Depicted values apply for 5.0%RhCl3/TH at pH = 7.

4-CP, 1.18 V

1.43 V

hν

VB

CB

[Ti]-O-RhIV

OH/OH–; 2.4 V

O2/O2–; – 0.16 V

– 0.34 V

1.87 V

BET


63

3.3 Experimental

3.3.1 Instruments

Diffuse reflectance spectra of the solids were recorded on a Shimadzu UV-2401PC

UV-Vis recording spectrophotometer. Samples were spread onto BaSO4 plates, the

background reflectance of BaSO4 was measured before. Reflectance was converted by

the instrument software to F(R∞) values according to the Kubelka-Munk theory. The

bandgap was obtained from a plot of F(R∞ )1/2 vs energy of exciting light assuming that

TH and 5.0%Rh(III)/TH are indirect crystalline semiconductors.

TOC measurements were made on a Shimadzu Total Carbon Analyzer TOC-

500/5050 with NDIR optical system detector.

Chloride was measured by ion chromatography (Dionex-120, Ion Pac AS 14

column, conductivity detector, NaHCO3/NaCO3 = 0.001/0.0035 M as eluating agent).

UV-vis spectra were recorded on a Shimadzu UV-3101 PC UV-Vis-NIR Scanning

Spectrophotometer, Quarz cuvette with d = 1 cm.

Specific surface measurements were carried out on a Gemini 2370 according to the

Brunauer – Emmet – Teller theory.

3.3.2 Determination of Absorptivity of [RhCl6]3-

22.5 mg of pure RhCl3×3H2O were dissolved in 100 cm3 of 6 M HCl, and then

refluxed for 4 h. Under these conditions, almost all of rhodium is present as [RhCl6]3-.

The resulting solution was diluted with 6 M HCl to a volume of 250 cm3. Thereafter

the absorbance of [RhCl6]3- was measured by UV-vis spectroscopy at the maximum of

the LMCT band at 252 nm. The absorptivity of [RhCl6]3- at 252 nm at 20 0C was

determined as 1.97 × 104 mol-1 dm3 cm-1. In the same spectrum one can observe the

two weak bands at 410 and 518 nm corresponding to 1A1g → 1T1g and 1A1g → 2T2g

transitions, respectively.

A similar procedure was performed to obtain absorption spectrum of [RhBr6]3- .


64

3.3.3 Preparation of Photocatalysts

To a suspension 1.0 g of titania (TH, Titanhydrat-0, Kerr-McGee) in 10 ml of H2O

were added appropriate amounts of RhCl3×3H2O or RhBr3x3H2O followed by

sonication for 10 min. A pH value of 3-4 was measured for these suspensions. After

stirring for 24 h in the dark, water was removed in vacuo and the residue was dried

under vacuum at room temperature for 3 h. Careful washing with water removed

physisorbed rhodium chloride as indicated by UV-vis absorption spectroscopy. The

resulting powder was heated in air for 2 h at 200 0C, washed four times with 50 ml

portions of doubly distilled water affording acidic solutions of pH 1-2. Drying as

described above and a subsequent second heating for 2 h at 200 0C, gave pink

RhCl3/TH and a dark yellow RhBr3/TH having a specific surface area of 230 m2/g as

obtained from BET measurements. The amount of rhodium present in the RhCl3/TH

powder was determined as follows. 80 mg of RhCl3/TH were suspended in 30 cm3 of 6

M HCl, and then refluxed for 4 h. The resulting suspension was diluted with 6 M HCl

to a volume of 50 cm3. After filtration of the photocatalyst with a Millipore membrane

filter (0.22 μm, Merck), the filtrate was analyzed as described above.

3.3.4 Preparation of 4.0%RhCl3/TH and Charge Transfer Absorption

Attempts to modify SiO2 (Silica gel 60, Merck) through the procedure described

above for TiO2 failed. In fact, during washing the surface complex was almost

removed, affording very weak colored powders. Therefore, modification of SiO2 was

accomplished by grinding an appropriate quantity of RhCl3x3H2O with SiO2 to obtain

4.0%RhCl3/SiO2. The diffuse reflectance spectrum of 4.0%RhCl3/SiO2 was multiplied

by the factor 2.3 in order to obtain the same Kubelka-Munk function as measured for

4.0%RhCl3/TH at λ = 518 nm.


65

3.3.5 Determination of Cl/Rh Ratio

60 mg of 5.0%RhCl3/TH (0.029 mmol Rh) were suspended in 40 ml of

concentrated NaOH and refluxed for 24 h. After filtration 5 ml of the solution were

neutralized with 0.1 ml of concentrated sulfuric acid. Quantitative determination by

ion chromatography afforded 0.081 mmol/40 ml of chloride from which a Cl/Rh ratio

of 2.79 is obtained.

3.3.6 Measurement of Quasi-Fermi Potentials

Quasi-Fermi levels of electrons were measured according to the literature [14]. 20

mg of TH (a) and 20 mg of 0.5%(b), 1.0%(c), 2.0%(d), and 5.0%RhCl3/TH(e) were

suspended in 100 cm3 of 0.1 M KNO3 in the presence of 15 mg of methylviologen

dichloride. Irradiation was performed with UV light (λ ≥ 320 nm, the light source was

the same as used in the photodegradation). Suspensions were stirred and bubbled with

N2 prior to and during the measurement. The pH was adjusted with HNO3 and NaOH

solutions and monitored with a pH - meter. A large surface platinum flag (5 cm2) and

Ag/AgCl were working and reference electrodes, respectively. Stable photovoltages

were recorded about 2 min after adjusting the pH value. In the case of 2.0%RhBr3/TH

the electron acceptor used was DP2+. The measured pH0 values were converted to the

Fermi potential at pH 7 by the equations EF(pH = 7) = -0.44 + 0.059 (pH0 -7) and

EF(pH = 7) = -0.27 + 0.059 (pH0 -7) when MV2+ and DP2+ were used, respectively.

3.3.7 Desorption Experiments

A suspension of 5.0%RhCl3/TH (30 mg) in 0.1 M HCl (15 cm3) was irradiated

with UV light (λ ≥ 320 nm) for 24 h as described below. After filtration of the

photocatalyst with a Millipore membrane filter (0.22 μm, Merck), the filtrate (10 cm3)

was added to 12 M HCl (10 cm3) and then refluxed and analyzed as described above.


66

3.3.8 Photodegradation Procedure and Product Analysis

The photocatalytic degradation of 4-CP was carried out in a jacketed cylindrical 15

cm3 quartz cuvette attached to an optical train.

Figure 3.19: Schematic front and side view of the quartz cuvette

Irradiation was performed with an Osram XBO 150 W xenon arc lamp (Io (400-

520 nm) = 2 × 10−6 Einstein s−1 cm−2) installed in a light condensing lamp housing

(PTI, A1010S) on an optical train.

Figure 3.20: Emission spectrum of the XBO 150 W xenon arc lamp.

350 400 450 500 550 6000,0

0,5

1,0

1,5

2,0

pow

er /

a.u.

λ / nm


67

A water cooled cylindrical 15 cm3 quartz cuvette was mounted at a distance of 30 cm

from the lamp. Appropriate cut-off filters were placed in front of the cuvette. The

suspension was stirred magnetically. In the standard experiment, 15 cm3 of 0.5 g l-1

powder suspension containing 2.5 × 10-4 mol l-1 of 4-CP was sonicated for 15 min and

then transferred to the cuvette. During an illumination run ca. 1.2 cm3 of the reaction

solution was sampled at given time intervals. The samples were filtered through a

Millipore membrane filter (0.22 µm) and then analyzed by UV-vis spectroscopy and

TOC analysis. In Fig. 3.20 the absorption spectrum of 4-CP shows two maxima at 280

nm and 221 nm which correspond to a n-π* and π-π* transition, respectively.

Figure 3.20: UV-Vis spectrum of an aqueous solution of 4-CP (1.25 × 10−4 M)

The same procedure was applied in the photocatalytic degradation of cyanuric

acid followed by TOC measurements. The starting suspension was adjusted to pH = 9

with NaOH solutions and irradiated with visible light (λ ≥ 455 nm).

200 300 400 500 6000.0

0.2

0.4

0.6

0.8

1.0π - π*

n - π*

Abs

. / a

.u.

λ / nm


68

3.4 Conclusions

Titania hybrid photocatalysts containing 0.5, 1.0, 2.0, and 5.0wt% of rhodium(III)

were prepared by chemisorption of RhCl3×3H2O onto anatase hydrate powder (TH).

Analytical data suggest that a titania-trichlororhodate complex is produced containing

a [TiO2]-O-Rh bond.

Similar results are found in the case of modification by RhBr3x3H2O. Diffuse

reflectance spectra exhibit an absorption shoulder throughout the visible region down

to 700 nm. Photoelectrochemical measurements indicate that the quasi-Fermi level of

electrons is gradually shifted to more anodic potentials with increasing rhodium

loading reaching a value of -0.34 V at pH 7 (vs. NHE) in the case of 5.0%RhCl3/TH.

This is more anodic by 210 mV as compared to unmodified TH. Upon visible light

irradiation this photocatalyst induces a fast mineralization of 4-chlorophenol but not of

cyanuric acid. Since the latter is mineralized in the presence of 4.0%H2PtCl4/TH, the

rhodium modified titania photocatalyses 4-CP oxidation by a different mechanism. It

seems likely that the primary photoprocess is not cleavage of the metal-halogen bond

but rather a charge transfer from Rh(III) to titania. This affords an electron in the

conduction band and a Rh(IV) species located within the band gap at about 1.43 V.

4.C-H Activation through Catalytic Photosulfoxidation of Alkanes

69

4. C-H Activation through Catalytic Photosulfoxidation of Alkanes

4.1 Introduction

The name sulfoxidation is referred to the concerted action of sulfur dioxide and

oxygen on n-paraffins (alkanes) or cyclo-paraffines to produce sulfonic acids. The

general reaction can be written as follow

R-H +SO2 +½O2 → RSO3H (4.1)

Sulfoxidation was discovered in Germany by C. Platz in 1940 irradiating with UV-

light a mixture of n-paraffins, sulfur dioxide and oxygen.139 This reaction represents a

rare example of an industrially applied process of C-H bond activation.140 It found

immediately great interest because it constitutes an easy way to produce straight chain

alkanesulfonates (SAS) which are applied as effective surfactants, good wetting agents

and emulsifiers.141 From then on, many processes were developed (see Paragraph

4.1.2) in order to reduce the use of UV-light or the formation of byproducts such as

sulfuric acid. But many of these methods have not satisfying yields or are quite

expensive (γ-ray) or use toxic sensitizers (Hg-sensitized process).

Drawing on these considerations and on the actual growing interest on green

chemistry, we decided to investigate new ways for obtaining alkanesulfonates which

do not require UV irradiation and produce less by-products.

In this chapter we report on the first catalytic photosulfoxidation of alkanes. This

reaction does not require UV lamps and toxic sensitizers, but only a non-toxic

semiconductor powder inducing alkane functionalization through visible light

excitation.


70

4.1.1 Industrial Importance of Photosulfoxidation

Owing to the increase in the demand of detergents, the reactions producing wash

active sulfonate (WAS) have achieved great significance in the last decades.

The current industrial method employing concentrated H2SO4 (oleum) for

manufacturing widely used linear alkylbenzene sulfonates (LABS) does not enable

sulfonation of saturated aliphatic hydrocarbons to produce saturated alkane sulfonates

(SAS). The reason lies in the inertness of alkanes and in the significant low solubility

of sulfuric acid in alkanes.

SAS have significant advantages over LABS.142 Though the detergent action is

comparable, SAS fulfil biodegradable criteria better than LABS. They are better

soluble in water and for this reason more suitable for liquid formulations of detergents.

Furthermore, the raw materials for SAS are alkanes available at cheaper rates and the

direct exploitation of them represents one of the major challenges in chemistry as

pointed out in the introductive part of this dissertation.

4.1.2 Industrial Processes

4.1.2.1 Mechanism of Industrial Sulfoxidation

The primary reaction steps of this rare alkane functionalization consist of UV-

excitation of SO2 to its triplet state via intersystem crossing143 followed by hydrogen

abstraction from the alkane producing an alkyl radical. An alternative C-H bond

cleavage mechanism by energy transfer is unlikely since the energy of the first excited

singlet state of SO2 is less than 380 kJ mol-1, whereas a C-H bond dissociation requires

about 400 kJ mol-1.


71

Figure 4.1: Scheme of the excitation of SO2 to 3SO2 which then drives the sulfoxidation.

Subsequent addition reactions with SO2 and O2 generate an alkylpersulfonyl

radical which in turn produces another alkyl starter radical and the persulfonic acid

(Fig. 4.2). Fragmentation and hydrogen abstraction (Eqs. 4.2 and 4.3) afford the

alkanesulfonic acid.144,145

Figure 4.2: Mechanism of the industrial sulfoxidation.

SO2 + hν

HSO2· O2

3SO2

RH

R·

SO2

RSO2·

RSO2-OO· RH

RSO2-OOH


72

RSO2-O-O-H → RSO2-O· + OH· (4.2)

RSO2-O· + R-H → RSO3H + R· (4.3)

R-H + OH· → R· + H2O (4.4)

According to this reaction scheme, photosulfoxidation is a photoinduced radical

chain reaction and therefore proceeds without further irradiation in the case of lower

alkanes (<C10) devoid of impurities. In the case of long unbranched alkanes of

insufficient purity termination steps like radical dimerization and radical-radical

recombination dominates and the reaction requires permanent irradiation. However,

addition of radical initiators or promoters like acetic anhydride again induces a chain

reaction.144,146 In general regioisomeric alkyl radicals are formed in the hydrogen

abstraction step except in the case of adamantane photosulfoxidation in the presence of

hydrogen peroxide affording regioselectively 1-adamantanesulfonic acid.147

The more interesting paraffin-mixture to produce WAS is the C8-C22 fraction. This

mixture is called Mepasin and is obtained from deep dehydrogenation of Kogasin II

(bp 230°C – 320°C), a hydrocarbon mixture achieved directly from the Fischer-

Tropsch synthesis.141

Not only aliphatic hydrocarbons can be sulfoxidized but also substituted

hydrocarbons such as alkylchloride, carbonic acids, esters, nitriles, alcohols and ethers.

Nevertheless, sulfoxidation of these products is not preparatively applied. Also low

molecular paraffins in the gas form such as butane can be sulfoxidized in solvents like

CCl4.144

Olefins, aromatic hydrocarbons and branched alkanes like 2,3-dimethylbutane are

not easily sulfoxidized and moreover are potential inhibitors for this reaction. In the

case of 2,3-dimethylbutane the abstraction of a tertiary hydrogen atom during the

chain reaction is more favorable and the resulting radical is too stable to propagate the

chain reaction. The inhibitory action of olefins and aromatic hydrocarbons was


73

explained by Calvert et al.148 They showed that these compounds react 100 times faster

than paraffins with 3SO2 but do not lead to sulfoxidation products.

Theoretically all expected isomers can be found in a sulfoxidation mixture.

However, in a hydrocarbon secondary carbon atoms are more reactive than the

primary atoms. The distribution of isomers has been determined for a few compounds

like n-hexane, n-heptane, and n-dodecane. In the case of n-heptane the ratio of the

relative reactivity between primary and secondary atoms turned out to be 1:30.149,150

In the sulfoxidation are also formed disulfonic acids. They are found in greater

amount than statistically expected. Probably, SO2 and O2 are better soluble in the

water phase (water is added in the light-water-process or formed through Eq. 4.4) and

further sulfoxidation occurs in greater extent. When the proportion of di- and poly-

sulfonic acids is higher than 13 %, the detergent properties of the sulfonates are greatly

diminished. For every 1% alkane conversion to sulfonic acid there is about 10% of di-

and poly-sulfonic acids formation. Nevertheless, higher proportion of di- and poly-

sulfonic acid conversion is avoided by limiting the alkane conversion to ca. 1%.144

Sulfoxidation is practically carried out in a continuous mode. According to the

stoichiometry of the reaction the optimal ratio of SO2/O2 would be 2:1, but since SO2

is 10-fold better soluble in the reaction mixture than O2, an equimolar flow of the two

gases is an essential prerequisite for a reasonable yield. The products have to be

continuously extracted from the reactor. Separation steps become necessary and the

unreacted reagents are recycled and fed into the reactor (vide infra).144

Sulfoxidation is carried out under UV irradiation in the case of light-water-

process,151 acetic anhydride-process,152 and chlorine-process153 or without irradiation

by using initiators like peracids, organic peroxides, ozone152 or γ-rays.154

4.1.2.2 Light-Water-Process

In this process light acts as reaction initiator and water as both reactant and solvent

to extract the product. The reaction is carried out continuously in a cylindrical reactor

into which the mercury medium pressure lamp is immersed. The reason of the


74

continuous mode operation is principally to avoid following two problems: during the

reaction the medium becomes turbid and the sulfonic acids which are not very soluble

in the alkane separate at the bottom of the reactor and in these conditions di- and poly-

sulfoxidation occurs more rapidly. Extracting the products with water limits further

sulfoxidation. Furthermore sulfonic acidssticks to the wall surrounding the light source

forming tarry deposits which block the passage of light.

The alkylpersulfonic acid formed according to the mechanism depicted in Fig. 4.2

does not generate new radicals according to Eq. 4.2 but reacts instantly with SO2 and

H2O to give sulfonic acid and sulfuric acid as shown in Eq. 4.5.

RSO2OOH + SO2 + H2O → RSO3H + H2SO4 (4.5)

Therefore, formation of starter radicals is inhibited in the presence of water and the

radical chain reaction can proceed only under permanent irradiation.

The optimal reaction-temperature is between 30 – 38 °C. Above 40 °C the yield

decreases quickly.

The light-water-process allowed sulfoxidation of high-molecular aliphatic

hydrocarbons and represents the best sulfoxidation process with regarding to yield and

quality of the product. In fact, in this case the ratio between mono- and poly-

sulfoxidized products is 9:1 whereas in the other processes it is much smaller.

Furthermore, because of the immediate and direct transformation of persulfonic acid

into the product, the formation of by products is very low.

The method of operation consists of five phases:

REACTION: The reactor is fed continuously with mepasin and water and from its

bottom a gas mixture of SO2 and O2 in the ratio 1:2 is introduced. The reaction is

carried out under pressure (normally 5 atm) and at temperature of about 30°C. The

circulating gases and powerful stirrers ensure intensive mixing in the reactor. 60kW

mercury arc lamps are used as the light source to initiate and maintain the reaction.


75

FIRST SEPARATION: in this step the unreacted mepasin is separated from the

aqueous phase containing sulfonic acid, sulphuric acid and a small amount of mepasin.

The unreacted mepasin is refluxed again into the reactor and the aqueous phase

undergoes the second separation.

SECOND SEPARATION: The former aqueous phase is preheated and fed in a second

fractionating column. The lower phase contains aqueous sulphuric acid (22%), the

upper phase is cooled and transferred to the neutralization step.

NEUTRALIZATION. The necessary amount of NaOH is fed in order to obtain

sulfonate and sulphate.

PURIFICATION: in the last step the remaining mepasin is removed from the

neutralized extract at 200°C in vacuo through an evaporator obtaining sodium

sulfonate, water and mepasin.

4.1.2.3 Acetic Anhydride Process

During the second world war was desired a process which could be carried out

without permanent irradiation because of the problems related with glass production

transport and installation. The acetic anhydride process turned out to be a suitable

alternative to the light-water-process.

This process is based on the experimental observation that the persulfonic acid can

react with acetic anhydride to give a stable and isolable alkyl-persulfonyl-acetic

anhydride according to the following equation

RSO2OOH + (CH3CO)2O → RSO2OOCOCH3 + CH3COOH (4.6)

The mixed anhydride decomposes in the presence of SO2 and H2O forming two

radicals (Eq. 4.7) which in turn through further H-abstraction react to sulfonic acid and

acetic acid (Eqs. 4.3 and 4.8).

RSO2OOCOCH3 → RSO2O· + CH3COO· (4.7)


76

CH3COO· + RH → CH3COOH + R· (4.8)

The reaction proceeds in this way after an initial imput (UV-light, ozone, γ-rays)

without further irradiation, although acetic anhydride has to be introduced

continuously into the reactor (the concentration of acetic anhydride should remain at

about 2.5 %).

As observed in the light-water-process, also in this case the product should be removed

continuously from the reactor. The reaction step is divided in two part: in the first

reactor (40°C) the mixed anhydride is formed which after an appropriate reaction time

is pumped in the second reactor (60°C) containing water or diluite acetic acid but no

more acetic anhydride. The persulfonic acids formed in the first step is now

transformed into sulfonic acid. The subsequent separation and purification steps occur

similarly to the light-water-process.

This process solves the problem of permanent irradiation and produces sulphuric acid

in a smaller extent. On the other hand, a higher amount of by-products is formed and

the produced acetic acid should be eliminated because it is not desirable for the

detergent properties of the sulfonate. These problems make the industrial scale

operation expensive and complicated.

4.1.2.4 Other Processes

OZONE PROCESS: Sulfoxidation of paraffins may be performed in the dark by

bubbling through them SO2 and ozone-containing oxygen. The product yield is

proportional to the amount of ozone introduced. Irradiating a paraffin suspension with

UV light and bubbling SO2 and ozone leads to a decrease in yield as compared to the

case in which oxygen is introduced.

CHLORINE PROCESS: Weghofer et al. developed a system in which the starter

radical is Cl· produced through UV irradiation of Cl2 (2%). Cl· abstracts a hydrogen

atom from the alkane and the reaction proceeds similarly to the other processes.


77

Cl2 2Cl· (4.9)

Cl· + RH → HCl + R· (4.10)

γ-RADIATION: Black and Baxter from the company ESSO in USA found that γ-

radiation from a Cobalt-60 source promotes sulfoxidation of alkanes. There are several

advantages:

(i) the presence of water is not necessary and therefore the separation step is simple

and less expensive.

(ii) only a relatively low intensity power source is required.

The disadvantage of this system is that the yield of di- and poly-sulfonic acid is very

high (up to 40% of the total mixture of sulfonic acid).

PERACIDS AND PEROXIDES: Saturated linear chain peracids such as peracetic acid

and its homologues, aromatic peracids or persulfonic acids are known to be good

initiators of sulfoxidation. However, it is required that the initiators are continuously

added during the reaction.

Organic peroxides are another important class of initiators although their utilization

demands a higher reaction temperature, which is more risky on the industrial scale. It

has been found that cyclohexanepersulfonyl peracetate, which decomposes at around

70°C, initiates sulfoxidation effectively. As already mentioned this type of initiator is

generally formed in situ when acetic anhydride is present during the sulfoxidation.

MERCURY PHOTOSENSITIZED PROCESS: A rare example for a sensitized

process (not industrially applied) is the mercury photosensitized sulfination of alkanes

with SO2 (Crabtree et al.155) producing initially sulfinic acids (RSOOH) and sulfinic

esters which have to be further oxidized to sulfonic acids by hydrogen peroxide. The

mechanism of this process is summarized in the following reactions.

Hg + hν (254 nm) → Hg* (4.11)

h ν


78

SO2 → SO2* (4.12)

RH + SO2* → R· + HSO2· (4.13)

SO2 + R· → RSO2· (4.14)

RSO2· + (H·, R·) → RSOO-(H, R) (4.15)

RSO2· + RH → RSOOH + R· (4.16)

RSOOH → RSO3H (4.17)

Since a low pressure Hg lamp was used for the irradiation, it is unlikely that direct

absorption by SO2 could be occurring at 254 nm (the most intense Hg line) because of

the small absorption coefficient of SO2 at that wavelength (ε = 1.1 atm-1 cm-1);

therefore, all the light should be absorbed by Hg. SO2 traps then R· (Eq. 4.14) and by

successive addition of H· or R· (Eq. 4.15) leaves to the sulfinic acid and sulfinic ester,

respectively. The latters have to be further oxidized to sulfonic acids by hydrogen

peroxide (Eq. 4.17).

4.1.3 State of Knowledge before this work

One of the most successful approach to visible light active titania photocatalysts,

was the synthesis of PtCl4 modified titania, carried out in our research group (see also

Chapter 3). Its great activity in degradation of various pollutants suggested to

investigate whether this novel catalyst possessed also photocatalytic activity in alkane

sulfoxidation. Since it was known that visible irradiation of TiO2-O-PtCl4 produced Cl·

and OH· it seemed likely that these radicals should be able to abstract hydrogen from

alkanes and therefore initiate sulfoxidation.

Hg*

H2O2


79

In 2002 first investigations on sulfoxidation were started in a system similar to that

described in the light-water-process. It consisted of two liquid phases (water and n-

heptane) in which PtCl4 modified titania was suspended and a mixture of SO2/O2 was

continuously bubbled through the suspension.156 The system was efficiently mixed by

magnetic stirring and irradiated with visible light. Unfortunately, the first positive

results could not be reproduced and no sulfonic acid could be detected anymore. In

fact, the catalyst was not suitable and the presence of water in this heterogeneous

system inhibited the reaction.

After that disappointing result also the sulfoxidation of solid alkanes was

attempted. The alkane chosen was adamantane for its high symmetry and because its

UV and thermal sulfoxidation were well known in literature. Adamantane was

dissolved in methanol, and after addition of titania the mixture was stirred under

SO2/O2 atmosphere and irradiated by visible light.157 However, the reported formation

of adamantane sulfonic acid could not be reproduced. It was therefore one aim of this

work to find out what the reason for the irreproducibility could be.


80

4.2 Results and Discussion: Reaction in Liquid Alkanes

During attempts to repeat the reported photosulfoxidation of adamantane in

methanol as a solvent it turned out that not the alkane but the alcohol reacted with

sulfur dioxide. To avoid the necessity of any solvent we changed to liquid alkanes like

n-heptane and cyclohexane.

Attempts to sulfoxidize olefins such as cyclohexene and cyclopentene by the

present method failed. The reaction mixure in a very few minutes became brownish

and the oil obtained after concentration in vacuo contained black viscous residues

suggesting olefin polymerization.

When a suspension of a titania powder in n-heptane was irradiated with visible

light (λ ≥ 400 nm) under an atmosphere of SO2/O2 = 1:1 (v/v), the formation of n-

heptanesulfonic acid was observed (Tab. 4.1).

N° Photocatalyst ri [mmol l-1 h-1]

1 Titanhydrat (A) 3.5

2 TiO2 (Hombikat, A) 5.0

3 TiO2 (R) 6.0

4 TiO2 (P25, A+R) 7.5

5 [TiO2]OPtCl4 (A) 0.0

6 [TiO2]ORhCl3 (A) 3.5

7 TiO2-C, TiO2-N (A) 3.5

Tab. 4.1: Initial rate ri of n-heptanesulfonic acid in the presence of different TiO2 photocatalysts. A and R denote anatase and rutile, respectively.

Only traces of sulfonic acid were observable in the absence of titania.


81

Initial product formation rates were 3.5 mmol/l.h and 5.0 mmol/l.h for the anatase

materials Titanhydrat and Hombikat, respectively, whereas for rutile and the mixed

phase powder P25 (75% anatase / 25% rutile) values of 6.0 mmol/l.h and 7.5 mmol/l.h

were observed. These surprisingly high rates for rutile may be due to the reported

better adsorption of SO2 as compared to anatase and to the fact that the rutile powder

employed (kindly provided from Prof. T. Egerton) had an unusual high specific

surface area (140 m2 g-1).

Out of the modified titania powders (entries 5-7), which are all good photocatalysts

in 4-chlorophenol visible light oxidation,117,133,158,159only the titania-chlororhodate

complex and carbon- or nitrogen-modified titania exhibited moderate rates of 3.5

mmol/l.h. The platinum chloride modified TH was totally inactive under these reaction

conditions. The same reaction was also carried out with cyclohexane. In the following,

if not otherwise specified, we will report on the n-heptane sulfoxidation taking into

account that similar results are obtained also for cyclohexane. The successful

sulfoxidation of this second alkane ensures the general applybility of the method.

4.2.1 Product Characterization

In a standard experiment, after 5 h of irradiation time the photocatalyst was filtered

through a micropore filter and the filtrate was concentrated in vacuo. The slightly

yellow, oily residue was dissolved in 3 ml of methanol and analyzed by HPLC with

indirect photometric chromatography (IPC) (for experimental details see Paragraph

4.4).

4.2.1.1 Principle of IPC and Measurements

Indirect photometric liquid chromatography (IPC)160,161 is the name given to a

technique which uses a UV-absorbing counter-ion in an ion-exchange mode with an

UV detector to determine UV-transparent ionic species. IPC, as described in detail by

Small and Miller, was first used for the determination of inorganic ions and then also

for quaternary ammonium salts (Later Larson and Pfeiffer), alkylamines and

alkanolamines.


82

Alkyl sulfonates are also UV-transparent compounds. Larson reported in 1985 on

their detection by means of this method. As already mentioned this approach is based

on the use of UV-absorbing eluents, made so by including in the eluent light absorbing

ions of the same charge as the ions to be separated. These ions have a dual role: (1) of

selectively displacing the sample ions from the chromatographic column; (2) of

revealing the sample ions to be separated. The appearance of sample ions in the

effluent is signaled by dips or troughs (negative peacks) in the baseline absorbance of

the effluent as the transparent sample ions substitute for the light-absorbing displacing

ions in the column.

Consider an ion exchange column which has been pumped and equilibrated with

an electrolyte UV-absorbing denoted as Na+E- so that the sites in the column are

occupied by E-. If the feed concentration of the eluent is maintained constant, a

detector placed at the outlet of the column would reveal a steady level of Na+ and E-

(Fig. 4.3).

Figure 4.3: Schematic representation of the principle of indirect photometric detection. A: before injection of the sample, B: after injection of the sample.(Taken from ref.161)

When a sample electrolyte denominated as Na+S- is injected, the anion S- will

generally retarded by the stationary phase and will exit at a characteristic elution time

depending on capacity of the exchanger, concentration of the solution and affinity of

the stationary phase for S- relatively to E-. A direct detection would show the

Abs

orba

nce

(a.u

.)

Abs

orba

nce

(a.u

.)

Elution Volume Elution Volume


83

concentration of S- rise and fall in a positive peak. Nevertheless, according to the

principle of electroneutrality and equivalence of exchange, the appearance of S- must

be accompanied from a concerted and equivalent change in E-, since the concentration

of sodium counter-ions is fixed. It therefore follows that the concentration of S- can be

indirectly (hence the name of this technique) monitored by continuously monitoring

the level of eluent ion E- and the troughs generated in the base line absorbance as

transparent sample ions elute.

Notable advantages of this technique are its single column simplicity, its

applicability to a wide range of ionic species and inherent great sensitivity.

In our case the role of E- is played by a hydrogenphthalate ion and of S- by the

sulfonic acid salt. At the beginn the chromatogram presents an off-scale deflection.

This is due to the ion exchange displacement of hydrogenphthalate by the injected

sample anions as a whole. After 2 minutes the equilibrium is again established and the

straight baseline shows at about 4 minutes a negative peak, the area of which is

proportional to the concentration of the alkylsulfonate contained in the sample.

Through a calibration curve we obtained the values summarized in the Table 4.1.

Abs

orba

nce

(a.u

.)

t / min0 1 2 3 4 5 6 7 8 9 10

0

0.2

0.4

0.6

0.8

-0.2

-0.4

-0.8

Figure 4.4: HPLC-chromatogram of the reaction product obtained in the photosulfoxidation of n-heptane (−) and of authentic 1-heptanesulfonic acid sodium salt (---). The presence of small amounts of 2-heptanesulfonic acid and other regioisomers cannot be excluded.


84

The dashed line represents a chromatogram of authentic 1-heptanesulfonic acid

whereas the solid line depicts the product obtained in the sulfoxidation of n-heptane.

The correspondence of the two peaks evidences the production of sulfonic acid.

4.2.1.2 IR Spectra and Amount of Sulphate

To check the nature of the product, the oily residue was neutralized with an

aqueous NaOH solution. After drying, the resulting salt was washed several times with

diethylether and analyzed by IR spectroscopy.

Fig. 4.5 compares IR spectrum of authentic heptanesulphonic acid sodium salt

(dashed line) with that of the product in the sulfoxidation of n-heptane (solid line). The

two spectra are almost identical, especially in the fingerprint region. The two bands at

1180 and 1055 cm -1 indicate, respectively, the asymmetrical and symmetrical

stretching of S=O bond. The bands in the regions between 2960 - 2850 and 1460 –

1380 cm-1 are typical for the aliphatic chain. We attribute the band at 620 cm-1 to the

stretching of C-S bond. There are no prominent peaks in the region of wavenumber

2600 cm-1, therefore there are no traces of sulphenic acids (S-H).

Figure 4.5: IR spectra (KBr) of the sodium salt of the reaction product obtained in the photosulfoxidation of n-heptane (−) and of authentic 1-heptanesulfonic acid sodium salt (---).

0 1000 2000 3000 400010

20

30

40

50

T %

Wavenumber / cm-1


85

Similar considerations could be made in the following figure illustrating the IR

spectra of authentic cyclohexane sulfonic acid sodium salt (dashed line) and the

product of the photocatalytic sulfoxidation of cyclohexane (solid line).

Figure 4.6: IR spectra (KBr) of the sodium salt of the reaction product obtained in the photosulfoxidation of cyclohexane (−) and of authentic cyclohexanesulfonic acid sodium salt (---).

Sulphite

Comparisons with sulphite IR spectra (not showed) suggest no traces of sulphite.

Probably in the presence of irradiated TiO2 they are totally oxidized to sulphate. An

easy experiment was performed in order to test the presence of sulphite ions. To an

aqueous solution of the oily residue, adjusted with NaOH to pH neutral, was added

Ba(OH)2 and a white precipitate formation was observed. Adding HCl solves the

eventually present BaSO3 but not BaSO4. After separation of the white powder, the

filtrate was added with hydrogen peroxide but no further precipitation was observed.

Thus, no evidence of sulphite ions was found in the reaction product.

0 1000 2000 3000 4000

10

20

30

40T

%

Wavenumber / cm-1


86

Sulphate

Comparison of the IR spectra of sodium sulphate with the sodium salt of the

reaction product and n-heptane gives some evidence for traces of sulphate (Fig. 4.7).

The small shoulder at 1126 cm-1 in the spectrum of the reaction product coincidizes

with the intense peak of sodium sulphate (spectrum a). This interpretation in

corroborated by ion chromatographic analysis of the reaction product exhibiting a

sulphate peak at the retention time of 13.6 minutes. From its concentration and from

the concentration of the sulphonic acid as measured by IPC analysis, a ratio of

sulphonic acid to sulphate of 30 to 1 was calculated.

0 1000 2000 3000 400010

20

30

40

50

60

70

80

90

100

T%

Wavenumber [cm-1]

c)

b)

a)

Figure 4.7: IR spectra of a) sodium sulphate, b) sodium salt of the reaction product and c) n-heptane.

As mentioned in the introductory part of this chapter, the separation of sulphuric

acid in the light-water-process represents a cost demanding step. The very small


87

amount of sulphate produced in the present visible light sulfoxidation is therefore

remarkable.

4.2.1.3 Elemental Analysis

Elemental analysis of the oily residue obtained after filtration and concentration in

vacuo of the reaction suspension afforded 43.06% C, 8.59%H, 17.74% S. These

values, within experimental error, are in good correlation with the theoretical values

(46.6% C, 8.8% H, 17.7% S) and suggest that this sulfoxidation affords mono-

heptanesulfonic acid.


88

4.2.2 Dependence on Photocatalyst Concentration

The concentration of P25 was varied in order to achieve optimal light absorption

with a minimum amount of catalyst.

0 1 2 3 4 5

0

10

20

30

40

Pro

duct

con

c. [m

M]

Catalyst conc. [ g/l ]

Figure 4.8: Dependence of the product concentration on the catalyst concentration obtained after 5 h irradiation.

After the concentration of 2 g/l a plateau is reached and a further increase in

catalyst concentration does not induce increase the sulfonic acid concentration. The

high photocatalyst concentration of 2 g/l ensures complete light absorption in each

experiment and therefore the initial rates in Tab. 4.1 (calculated from the product

concentration at 5 h irradiation time) are comparable.


89

4.2.3 Deactivation and Regeneration of the Photocatalyst

Fig. 4.9 illustrates the dependence of sulfonic acid concentration on the irradiation

time. Six standard reactions are shown, each performed with a different irradiation

time. The yield increases with increasing irradiation time up to 7 h, when a plateau is

reached. About at the same time, the initially perfect suspension separated into a clear

solution and a sticky catalyst layer, adhering to the bottom of the reactor. The color of

the catalyst has slowly changed to grey-yellowisch. Similar deactivation was found for

all catalysts summarized in Tab. 4.1.

0 2 4 6 8 10

0

5

10

15

20

25

30

Pro

duct

con

c. [m

M]

T im e / h

tota l deactivation

Figure 4.9: Dependence of sulfonic acid concentration on the irradiation time.

The catalyst could be fully reactivated by the following procedure: after 10 h

irradiation (to ensure total deactivation), n-heptane was removed from the reaction

mixture in vacuum, giving a thick slurry containing deactivated TiO2 and sulphonic

acid. Washing ultrasonically with methanol, changed the colour of the photocatalyst to

white. Then, the photocatalyst was separated from the methanolic phase by

centrifugation. The methanol solution was analyzed by IPC, while the TiO2 was dried

in vacuum at 40°C and reused for another reaction. This procedure was repeated three

times and the yield was almost the same after each successive reaction as shown in

Fig. 4.10. This indicates the catalytic nature of this photosulfoxidation.


90

0

10

20

30

40

hνhν

c(1)

/ m

M

hν

R R R

0 10 0 10 0 10

Time / h

Figure 4.10: Sequential photosulfoxidation of n-heptane. λirr ≥ 400 nm. R = regeneration.

This observation suggested that the reaction is inhibited by strong product

adsorption and that washing desorbs the sulfonic acid. Accordingly, no product

formation was observable when heptane sulfonic acid was added to the suspension

prior to irradiation.

A similar deactivation and activation was observed by Shang et al.162 during photo-

oxidation of sulfur dioxide in the presence of gaseous n-heptane and oxygen at UV-

irradiated titania powder at room temperature. In the n-C7H16/O2/TiO2 system, no

catalyst deactivation was observed, while for SO2/O2/TiO2 and n-C7H16/SO2/O2/TiO2,

the photocatalytic activity of TiO2 powder showed decreasing and eventually no

activity after used consecutively. Sulfur trioxide and sulfuric acid are supposed to be

the poisoning species. In this gas phase reaction presence of water vapor does not

influence the reactivity.


91

However, in our case product formation was inhibited when small amounts of

water like 0.3 vol% were present in the suspension. This may be due to blocking of the

reactive surface centres for heptane oxidation by preferential adsorption.

The influence of water will be discussed in detail in Paragraph 4.4, describing the

acetic acid system.

4.2.4 Surface Modifications of the Catalyst

In order to improve the reaction we tried to modify the surface of the catalyst in

two ways:

1. Enhance visible light absorption

2. Improve the dispersion of the catalyst (polar) in the suspension (apolar).

Many techniques are suitable in order to improve the visible light absorption as

already mentioned in the introductory part of this dissertation. Since in our group

surface modifications through transition metal salts and organic nitrogen- or carbon-

compounds have been extensively investigated, similar attempts were made in this

work.

Besides the rhodium159 and platinum modified titania61, we accomplished also

titania modification by RuCl3 · xH2O and IrCl3 · 3H2O with a similar method as

described for rhodium in Chapter 3. In all case we observed a new absorption in the

visible spectral region.


92

300 400 500 600 7000,00

0,05

0,10

0,15

0,20

0,25

F(R

∞)

Wavelength / nm

a b c d

Figure 4.11: Diffuse reflectance spectra of 4.0%Pt(IV)/TH (a), 2.0%IrCl3/TH (b), 4.0%RuCl3/TH (c), and 2.0%RhCl3/TH (d). The Kubelka-Munk function, F(R∞), is used as the equivalent of absorbance.

However, out of the mentioned catalysts only the rhodium modification was active

in sulfoxidation, whereas the other catalysts were not stable under the given

experimental conditions as could be noticed by colour losses during the reaction and

absence of sulfonic acid formation.

The well known carbon133 and nitrogen158 modified titania extend as well the onset

of light absorption of titania to the visible region (Fig. 4.12). They induced sulfonic

acid production (as summarized in Tab. 4.1) but to smaller extent than unmodified

titania.


93

400 500 6000,0

0,1

0,2

0,3

0,4

0,5

0,6

F(R

∞)

Wavelength / nm

a bc

Figure 4.12: Diffuse reflectance spectra of TH (a), TiO2-C (b), and TiO2-N (c). The Kubelka-

Munk function, F(R∞), is used as the equivalent of absorbance.

A better dispersion of the polar catalyst in a non polar solvent can be achieved by

making the solid surface more hydrophobic, i.e. by substituting the OH groups on the

surface of titania with non polar groups. We synthesized three different types of

hydrophobic materials:

1. Silylated TiO2

2. Phosphated TiO2

3. Fluorinated TiO2

The synthesis procedure of these powders is described in the experimental part. In

the case of silylated P25, although the contact between catalyst and alkane could be

optimized, the yield of the sulfoxidation of n-heptane decreased of about 70%.

Phosphate ions have been shown to adsorb strongly to the surface of TiO2.163 The

binding of anions can be related to the electrostatic interaction with the surface,

depending on the point of zero charge and to exchange reactions with the surface

hydroxyl groups. In the photocatalytic degradation process, these ions may either

block active sites or compete with organic contaminants for oxidizing radicals during


94

the photocatalysis process. Abdullah et al.164 showed that the presence of phosphate in

solution reduced the rate of titania photocatalyzed oxidation of model organic

contaminants by as much as 70%. Similarly, using phosphated P25 as a catalyst of

sulfoxidation, reduces the product yield by about 90%.

Since the flatband potential and the bandgap energy of titania does not change after

silylation or phosphatation, these results could be explained by blocking of the active

sites due to the surface modification.

Fluoride ions represent the sole exception in this context. Surface fluorination of

TiO2, strongly modifies its surface properties.165-168 The formed ≡Ti-F species

dominate at acidic pH, with an almost complete displacement of surface –OH groups

at pH 3.7. The effect of such adsorption has not been completely clarified yet. TiO2

fluorination in aqueous media has been reported to either increase or decrease the rate

of photocatalytic reaction involving different substrates. The results of a decreased

reaction rate have been related to the inhibition of hole transfer oxidation, consequent

to the hindered adsorption of substrates and to the reduced IFET rates due to the strong

electronegativity of fluorine. At the same time cases of enhanced activity were

explained by the generation of “bulk” hydroxyl radicals which are expected to be

stronger oxidants than OH radicals bound to the surface of the unmodified TiO2.

Macyk et al.165 observed that surface modification with fluoride ions or silyl groups

induced UV light photoactivity of TiO2 toward cyanuric acid degradation. They

attribute these results to the formation of highly oxidizing singlet oxygen. In fact,

surface modification enhances the energy transfer pathway and suppresses the

interfacial electron transfer which lead to formation of OH radicals.

In our case using fluorinated TiO2 enhanced the reaction rate by about 35%. The

reason of this is not easily understandable. In fact, the flatband potential of titania is

the same also after fluorination. The hypothesis of an energy transfer and consequent

singlet oxygen formation seems unlikely, because the silylated samples do not show

similar behaviour. Furthermore, formation of product was inhibited in the presence of

methanol which scavenges effectively OH radicals but not singlet oxygen. Perhaps a


95

higher availability of “bulk” OH radicals and a better contact between substrate and

catalyst could explain these results.

4.2.5 Interaction between SO2 and TiO2

The interaction between SO2 and TiO2 is well documented in literature162,169-174 for

at least three reasons:

• SO2 is one of the most common catalyst poisons.

• The Claus reaction is a well known process for the recovery of sulfur from

acidic gases containing hydrogen sulfide. In the modified Claus reaction, H2S is first

oxidized to SO2 and then the catalytic reaction between SO2 and the rest of H2S takes

place to produce sulfur and water according to Eq. 4.18:

2H2S + SO2 → 2H2O + 3/x Sx (4.18)

In practice γ-Al2O3 has long been used as the catalyst for the Claus reaction.

However, it was often found that it is readily deactivated due to the formation of

sulfate or strongly adsorbed species on the catalytic surface.175 Moreover small

amounts of oxygen may lead to a dramatic decrease of the catalytic activity. Since

1980, it has been increasingly reported that TiO2 based catalysts show superior

catalytic properties for this important reaction.176

• SO2 is one of the most important air pollutants and a very corrosive

molecule.

On the other hand, TiO2 is a widely used catalyst support for many technological

applications. Therefore, knowledge of the interaction, adsorption and desorption

processes of SO2 on TiO2 is of basic importance.

The results of previous work on this topic are somewhat contradictory, with one

study reporting little reaction and another evidencing formation of a surface sulphite-


96

and/or sulphate-like species.170 Moreover the experimental conditions are very

different and particularly the treatment of the used TiO2 (annealing, reduced surface,

single crystal, etc.) often makes the analysis scarcely applyable to the real catalytic

materials.

Temperature programmed desorption (TPD) and temperature programmed

electronic conductivity (TPEC) suggest a strong interaction between TiO2 and SO2 for

TiO2 annealed at 500°C.169

Adsorption of SO2 was also analyzed on a TiO2 (110) single crystal at 120 K by

means of core level synchrotron radiation photoemission.177 The authors concluded

that surface sites most relevant for adsorption are: oxygen in plane and bridging

oxygen atoms.

Figure 4.13: Schematic representation of the TiO2 (110) surface termination including the main adsorption sites: the white, grey, and black spheres represent oxygen atoms in plane, bridging oxygen atoms, and Ti atoms, respectively. V represents an oxygen vacancy.

The bridging oxygen atoms are involved in the formation of sulphite-, sulphate-

like species. The sulfur atom can also take the place of an oxygen vacancy, being

incorporated into the lattice, giving very strong interactions with the nearest oxygen

atoms. For temperatures around 120 K a sulphate phase coexist with the sulphite

species and adsorbed molecules. There is no indication of order in these processes.

Similar results were found for room temperature adsorption.

V


97

Yanagisawa173 reported of oxygen exchange between SO2 adsorbate and TiO2

surfaces. Adsorption processes of 18O-enriched SO2 were analyzed on vacuum-

annealed powders with temperature-programmed desorption, gas analysis, and Auger

electron spectroscopy. At room temperature a fair amount of oxygen exchange

between SO2 and lattice oxygens takes place on the surface.

A very interesting study for the aim of this dissertation is reported from Shang et

al.162 on the deactivation and regeneration of TiO2 nanoparticles in three photocatalytic

oxidation systems: n-C7H16, SO2 and n-C7H16/SO2 carried out at room temperature, in

a gas-solid phase batch reactor under UV irradiation. It was obwerved that the

oxidation of n-heptane was inhibited by the presence of SO2. Inhibition was

accompanied by oxidation of SO2 to SO3 adsorbed on the catalyst surface.

Furthermore, the catalyst acquired a yellow color and also further oxidation of SO2

was inhibited. This indicates that the adsorbed products block the active sites of TiO2.

Poisoning species are proposed to be SO3, sulfuric acid, as well as others not identified

organic byproducts. When the adsorbed products were removed by sonicating in water

or methanol, the deactivated catalyst was regenerated. Although this work does not

have any synthetic aspects and was performed under different experimental conditions,

it may enlighten some aspects of the photosulfoxidation discussed in this dissertation.

In order to investigate the poisoning species responsible of the deactivation process

and the related changes on the surface properties of titania we performed XPS

measurements and photo-electromotiv force experiments (PEMF).

4.2.6 XPS Results

4.2.6.1 XPS Principles

X-ray photoelectron spectroscopy (XPS) is a quantitative spectroscopic technique

that measures the binding energy of an electron as a function of the type of atom and

of its chemical environment.


98

When an atom is irradiated with monochromatic X-rays, electrons will be removed

(photoelectric effect). XPS spectra are obtained by irradiating a material with a beam

of aluminium or magnesium X-rays while simultaneously measuring the kinetic

energy (Ekin) and number of electrons that escape from the top 1 to 10 nm of the

material being analyzed. All of the deeper photo-emitted electrons, which were

generated as the X-rays penetrated 1–5 micrometers of the material, are either

recaptured or trapped in various excited states within the material.

Because the energy of a particular X-ray wavelength equals a known quantity, we

can determine the electron binding energy of each of the emitted electrons by using the

Eq. 4.19, based on the work of Rutherford (1914)

BE = hν - Ekin – Φ (4.19)

where BE is the energy of the electron emitted from one electron configuration

within the atom, hν is the energy of the X-ray photons being used, Ekin is the kinetic

energy of the emitted electron as measured by the instrument and Φ is the spectrometer

work function which can be compensated artificially. To count the number of electrons

at each Ekin value, with the minimum of error, XPS must be performed under ultra-

high vacuum conditions.

A typical XPS spectrum is a plot of the number of electrons detected versus the

binding energy of the electrons detected. Each element produces a characteristic set of

XPS peaks at characteristic binding energy values that directly identify each element

that exists in or on the surface of the material being analyzed. These characteristic

peaks correspond to the electron configuration of the electrons within the atoms, e.g.,

1s, 2s, 2p, 3s, etc.. To generate atomic percentage values, each raw XPS signal must be

corrected by dividing its signal intensity (number of electrons detected) by a "relative

sensitivity factor" (RSF) and normalized over all of the elements detected.

To fit experimental XPS peaks or to deconvolve multiple peaks typically means to

compare them with a theoretical model curve composed of model peaks. For modeling


99

the peak shapes we used a Gaussian-Lorentzian combination model. This method is

based on the least-square estimates for calculating the χ2 value as a degree of the

difference between the experimental and fitted curve. An iteration process is

performed until the χ2 value converges.

4.2.6.2 XPS Spectra

The nature of the surface species derivating from the adsorption of SO2 on TiO2

were investigated by means of XPS spectroscopy.

Since the deactivated catalyst even after drying in vacuum remained sticky, we

could not perform any XPS analysis in these conditions.

We observed that the regeneration procedure with dichloromethane instead of

methanol did not restore the activity of the catalyst which after washing maintained its

grey-yellowish color and remained at the flask bottom without generating a good

suspension in n-heptane. However with this procedure the catalyst was suitable to

enable the XPS analysis. The XPS spectrum of this sample is shown in Fig. 4.14b.

In order to get a deeper comprehension of the interaction between SO2 and TiO2

we report in Fig. 4.14a the XPS spectrum of P25 exposed for three days to SO2

atmosphere.


100

164 166 168 170 172

Inte

nsity

(a.u

.)

Binding Energy (eV)

A

B

C

a)

166 168 170 172

Inte

nsity

(a.u

.)

Binding Energy (eV)

B

Db)

Figure 4.14: a) XPS spectrum of P25 exposed for three days to SO2 atmosphere; b) XPS spectrum of the deactivated P25 washed with dichloromethane

The broad peak in spectrum 4.14a has a maximum at about 168.5 eV and two

shoulders at 167.1 eV and 169.7 eV. In the case of the deactivated P25 the peak

maximum is shifted to 169.3 eV and curve fitting gives the best result assuming the

presence of two peaks at binding energy of 168.7 eV (B) and 169.6 eV (D).

In the literature it is known that sulphur dioxide upon interaction with titania leads

to physisorption and chemisorption affording sulphur trioxide and sulphate exhibiting

binding energes from 164 to 169 eV.171 We therefore assigne the peak at 167.1 eV (A)

to adsorbed SO2, the peak at 168.5 eV (B) to adsorbed sulphur trioxide and the peak at

169.7 eV (C) to sulphate. Since the deactivated P25 powder was washed with

dichloromethane, the absence of an SO2 signal in Fig. 4.14b becomes understandable.

The higher intensity of the B component in Fig. 4.14a suggests that the SO3-like

species is the most probable surface intermediate in the SO2 adsorption process.


101

4.2.7 PEMF Results

4.2.7.1 PEMF Basics

To understand the deactivation process and in which extent the adsorbed species

on the deactivated titania influence the electron transfer, we performed photo-

electromotive force (PEMF) experiments. This technique provides informations about

the natural behaviour of photogenerated electron-hole pairs, since the measurements

are carried out without a contact electrode and without any external electric field and

therefore charge carrier concentration gradients and/or internal space charges are the

sole driving forces for PEMF generation.178 PEMF is sensitive to all factors

influencing the mobility of the charge carriers, like traps or structural changes. Fig.

4.15 depicts the principle of PEMF measurements.

Figure 4.15: Principle of PEMF measurement for an n-type semiconductor. (1) transparent NESA glass electrode, (2) insulating foils, (3) sample, (4) metal electrode.(Taken from ref..178)

Consider a single crystal illuminated from one side. Charge separation occurs and

the light intensity decreases exponentially with the penetration distance into the crystal

according to the Lambert-Beer law.


102

A (λ) = log I0 / I = ε (λ) c d (4.20)

where A(λ) is the absorbance, I0 the incident light intensity, I the light intensity at the

exit of the crystal, ε(λ) the absorption coefficient and d the path length.

The gradient of light absorption creates a charge carrier concentration gradient

which is the driving force of the charge carrier diffusion into the bulk. If electron and

holes have different mobilities, a spatial charge separation takes place and an internal

electric field between the illuminated and the dark side of the sample arises. This field

can be measured as an electric potential difference, the so called Dember potential or

photo-electromotive force. The same effect is observed if the single crystal is replaced

by a polymer film containing a semiconductor fine powder.

The maximum Dember potential179 can be expressed as

where kb is the Boltzmann constant, T the absolute temperature, e the charge of

electron, μh and μe the mobility of holes and electrons, respectively. Eq. 4.21 is valid

only for life times of the charge carriers longer than the light pulse and a high light

intensity is required to fill all the traps with charge carrriers. Therefore a laser is used

for flash illumination.

The most important kinetic parameters in the PEMF measurements are the

following:

• Sign of the signal: in an n-type semiconductor the electrons are more mobile

than the holes hence they will induce a positive charge on the dark electrode

leading to a positive voltage. Analogously, in a p-type semiconductor the

Umax (λ) = kb T

e

μe - μh

μe + μh

ε(λ) d (4.21)


103

holes are the more mobile charge carriers and hence the signal shows a

negative sign.

• Decay rate constant k: Decay curves can be fitted by assuming the presence

of a fast and a slow recombination reaction, both obeying a first order law

(Eq. 4.22). The fast process is connected with the rate constant k1 and can be

assigned to surface recombination whereas the slow process (rate constant

k2) corresponds to bulk recombination.

• Maximum Dember voltage: this value increases with increasing number of

charge carriers generated by the laser flash. It could be seen as a relative

measure of the efficiency of the charge separation.

In this work the kinetics of the PEMF signals was evaluated by using the

biexponential model according to Eq. 4.22. The experimental values of both partial

PEMFs U01 and U0

2 are related to the maximum value U max through Eq. 4.23.

U (t) = U1° exp (-k1t) + U2° exp (-k2t) (4.22)

Umax = U1° + U2° (4.23)

4.2.7.2 PEMF Measurements and Discussion

The following measurements were performed on P25 samples before and after

sulfoxidation reaction. They were embedded as fine powders in a polymer film (see

Experimental Part).

The photoelectrical characteristics of the catalyst P25 before and after the reaction

were explored by PEMF measurements.

Fig. 4.16 illustrates the PEMF signal recorded in microsecond time range of the

deactivated P25 after the reaction (a) and of pure P25 (b). Fig. 4.17 shows the PEMF

signal recorded in millisecond time range. The relevant curve parameters are

summarized in Tab. 4.2.


104

0.0 0.5 1.0 1.5 2.0 2.5

0

10

20

30

40

U /

mV

Time / µs

a)

b)

Figure 4.16: PEMF signals of the deactivated P25 (a) and of the pure P25 in μs time range.

0 50 100 150 200

0

20

40

60

U [m

V]

Time [ms]

a)

b)

Figure 4.17: PEMF signals of the deactivated P25 (a) and of the pure P25 in ms time range.


105

Tab 4.2: Maximum photovoltage Umax, partial photovoltages U01 and U0

2, life time of the charges on the surface (τ 1) and on the bulk (τ 2) for pure and deactivated P25.

In the case of pure P25 we observe a positive PEMF signal in the shorter time

range (μs). On the millisecond time range, at the beginn, the recorded decay-process of

PEMF signal is positive. Such signal is typical for an n-type semiconductor such as

pure TiO2.

On the other hand, the deactivated P25 exhibits a negative Dember voltage (visible

only in the microsecond time range), revealing that its photoelectrical behaviour is

typical for a p-type semiconductor. This conclusion is independent from the extent of

the potential being related only with the sign of the signal.

A similar change from n-type to p-type was also observed in the photocatalytic

oxidation of SO2 on TiO2 as indicated by surface photovoltage spectra (SPS) of TiO2

before and after the reaction.

The fact that Umax value for the reference P25 is about 20 times higher than Umax

value for the deactivated sample, reveals that after the reaction the efficiency of the

charge separation decreases dramatically.

Furthermore, the decay of the PEMF signal for the deactivated sample is about 5

times faster than that for P25. It means that in the former case the photogenerated

charges are suddenly trapped (in few microseconds) into states situated at the surface

of the catalyst. In fact, in the case of the deactivated sample the values of U02 and τ2,

P25 Umax [mV] U10 [mV] U2

0 [mV] τ1 [ms] τ 2 [ms]

pure

59.2 ± 2.8

138.2 ± 7.0

-79.0 ± 6.6

16.8

30.3

deactivated

-2.8 ± 0.3

-3.5 ± 0.8

-

5.8 · 10-4

-


106

which refer to the photoelectrical properties in the bulk, are very small and hence

surface recombination dominates. This conclusion is in agreement with the hypothesis

of a surface poisoning of the catalyst during the sulfoxidation.

The PEMF signals as a function of the number of laser pulses for P25 are

illustrated in Fig. 4.18 and 4.19 and the kinetic parameters are summarized in Tab. 4.3.

0,0 0,5 1,0 1,5 2,0 2,5

0

10

20

30

40

50

1.flash 2.flash 3.flash 4.flash 5.flash

U [m

V]

Time [µs]

Figure 4.18: Microsecond PEMF signals of the pure P25 as a function of the number of laser pulses.


107

0 50 100 150 200-20

0

20

40

60

80

1.flash 2.flash 3.flash 4.flash 5.flashU

[mV]

Time [ms]

Figure 4.19: Millisecond PEMF signals of the pure P25 as a function of the number of laser pulses.

Table 4.3: Maximum photovoltage Umax, partial photovoltages U10 and U2

0, life time of the charges on the surface (τ1) and on the bulk (τ2) for pure P25 as a function of the number of laser pulses.

Flash Umax [mV] U10 [mV] U2

0 [mV] τ 1 [ms] τ 2 [ms]

1 62.0 136.5 -74.5 16.2 30.7

2 63.6 139.6 -76.0 16.5 30.9

3 62.9 135.9 -73.0 16.2 30.6

4 64.1 162.2 -98.1 17.1 28.7

5 66.7 155.5 -88.8 16.6 29.5


108

Fig. 4.20 illustrates the PEMF signals in function of the number of laser pulses for the

deactivated sample. The kinetic parameters are summarized in Tab. 4.4.

0.0 0.5 1.0 1.5 2.0 2.5

-3

-2

-1

0

1.flash 2.flash 3.flash 4.flash 5.flash

U [m

V]

Time [µs]

Figure 4.20: Microsecond PEMF signals of the deactivated P25 as a function of the number of laser

pulses.

Table 4.4: Kinetic parameters of the PEMF signals as a function of the number of laser pulses for the deactivated P25.

flash Umax [mV] τ [ms]

1 -3.0 1.09 ·107

2 -2.9 1.06 ·107

3 -2.9 1.16 ·107

4 -3.1 1.04 ·107

5 -3.1 1.03 ·107


109

We did not observe neither for the reference nor for the deactivated P25 a PEMF

signal change with the number of laser flashes in the μs and in the ms time range. It

means that also for the deactivated sample in 120 seconds (time between two

consecutive flashes) there are no charge effects, suggesting that the energy traps are

not extremely deep.

In conclusion, during the photocatalytic sulfoxidation P25 changes its

photoelectrical behaviour from n-type to p-type and the charge carrier lifetime

decreases from 16.8 to 5.8 · 10-4 ms.

According to the XPS and PEMF results we attribute the type change to the

presence of oxidized sulphur compounds (mainly SO3) strongly adsorbed on the

surface. Being in a high oxidation state, they can act as efficient electron traps making

the holes the majority charge carriers and titania achieves a p-type character.

To sum up, the chemical deactivation process is due mainly to strong preferential

product adsorption on the TiO2 surface, so that further reagent adsorption is hindered:

in addition, the change from n- to p-type may decrease the efficiency of IFET

reactions.


110

4.2.8 Reaction Mechanism

Fig. 4.21 summarizes absorption spectra of P25 and SO2 dissolved in n-heptane

(saturated solution). From this and the experimental fact that photosulfoxidation is

initiated by visible light (λ ≥ 400 nm) one must conclude that a previously unknown

species must be responsible for the observed reaction.

300 400 500

0

2

4

Abs

orba

nce

/ a.u

.

λ / nm

a

b

Figure 4.21: Absorption spectra of P25 (a) and SO2 dissolved in n-heptane (saturated solution) (b).

Since only the modified titania powders (Table 1, entries 5-7) are able to absorb

visible light, it seemed likely that the unmodified materials may form a charge-transfer

complex with sulfur dioxide. As depicted in Fig. 4.22, visible light excitation of the

charge-transfer complex affords a conduction band electron [TiO2(e−)] and an

adsorbed sulfur dioxide radical cation. Oxygen reduction by [TiO2(e−)] produces

superoxide whereas the adsorbed sulfur radical cation may oxidize the alkane to the

alkyl radical and a proton (see Paragraph 2.3). Superoxide may also generate an alkyl

radical through protonation by adsorbed water or surface OH groups to the

hydroperoxyl radical103 and subsequent hydrogen abstraction from the alkane.


111

Figure 4.22: Mechanistic scheme of the visible light photosulfoxidation.

The alkyl radical thus produced is expected to initiate a radical chain reaction as

formulated for the stoichiometric UV-photosulfoxidation (Fig. 4.2).

Accordingly, experiments of flatband potential determination of P25 with visible

light (λ ≥ 400 nm) failed due to the absence of light absorption of titania

In summary, this novel visible light induced C-H activation can be classified as a

photocatalysis type B reaction, extending the previously known two-substrate addition [20] to the present three-substrate addition scheme A + B + C = D.

Many blank tests were performed:

• In the absence of TiO2 or using SiO2 instead of TiO2, only traces of product

could be detected.

• In the absence of SO2 no product formation was observed.

• In the absence of O2 (degassing by freeze-thaw cycles) only traces of

product were detected.

• In the dark or with irradiation at λ ≥ 455 nm the reaction does not proceed.

[TiO2(e-)---SO2+·]

[TiO2---SO2]

O2

RH RH

R· R· + H+

Vis


112

4.2.8.1 Evidences of Radical Chain Reaction

As discussed in the introduction the UV induced photosulfoxidation is a radical

chain reaction. Since the mechanism proposed for the present visible light induced

reaction (Fig. 4.22) generates also an alkyl starter radicalthe following experiments

were performed.

• Isopropanol

Complete inhibition was observed in the presence of only 10 vol% of isopropanol,

which should be much faster oxidized than the alkane and which is also an efficient

OH radical scavenger (the reaction rate of isopropanol with OH radical is extimated

1.9 · 109 M-1 s-1).180

• Sulfur hexafluoride

Sulfur hexafluoride181,182 is known to be one of the most inert inorganic

molecules. Its high electron affinity, however, has rendered it an extremlely valuable

specific electron scavenger. Nevertheless, its use as electron scavenger in aqueous

solution has been limited from its low solubility in water and from the fact that OH

radicals are also involved in its reaction path.

SF6 in water reacts with an electron with a rate constant of 1.65 · 1010 M-1 s-1

SF6 + e- → ·SF5 + F- (4.24)

The SF5 radical formed oxidizes water to an OH radical (with a lower reaction

rate of 6.3 · 103 M-1 s-1) giving also sulfur tetrafluoride and fluoride.

·SF5 + 2H2O → ·OH + F- + H3O+ + SF4 (4.25)


113

The sulfur tetrafluoride then is hydrolyzed according to

SF4 + 9H2O → SO32- + 4F- + 6H3O+ (4.26)

and SO32- is subsequently oxidized to SO4

2- by species such as ·OH, ·SF5 , H2O2,

so that the general equation can be written as

SF6 + e- 6F- + SO42- + 7H3O+ (4.27)

In our system the amount of water is negligible and the solubility of SF6 in n-

heptane is obviously higher than in water, SF6 being an apolar molecule. These

considerations justified the use of SF6 as electron scavenger in our system.

Adding 35 ml of gaseous SF6 at 1 atm into the reactor strongly inhibited the

reaction. SF6 competes with oxygen as electron scavenger, hindering the reductive

path of the process.

• Dimerization and dehydrogenation products

A mutual formation of dimerization and dehydrogenation products in the

sulfoxidation of cyclohexane was investigated through mass spectroscopy analysis.

In Fig. 4.23 we can observe the signals at m/z 165 and 81 attributed to the

fragments formed by cleavage of cyclohexylcyclohexane and cyclohexene,

respectively. These compounds must be produced if the reaction mechanism proceeds

through alkyl radical formation.

H2O, ·OH


114

Figure 4.23: Mass-spectrum of the filtered suspension after photocatalytic sulfoxidation of cyclohexane. The signals at m/z 81, 83, and 165 are attributed to the fragments formed by cleavage of cyclohexene, cyclohexane, and cyclohexylcyclohexane, respectively.

• Hydroquinone

Hydroquinone is an excellent radical scavenger belonging to the class of

polyphenolic antioxidants.183 In general, free radical scavenging activity of

polyphenols (ArO-H) is characterized by their hydrogen atom donating ability to

scavenge the radicals (R·)

ArO-H + R· → RH + ArO· (4.28)

The ability to donate a hydrogen atom is mainly governed by the O-H bond

dissociation enthalpy (BDE) which in the case of hydroquinone assumes a value of

329.2 kJ mol-1. The smaller the BDE the greater is the free radical ability of the

antioxidant.


115

Tab. 4.5 sums up the experiments carried out with hydroquinone.

Tab. 4.5: Summary of the experiments carried out with hydroquinone.

When the radical scavenger hydroquinone was present in the reaction mixture

already before irradiation no product formation was observed.

When after 2 h of irradiation, corresponding to a concentration of 20 mM of

heptane sulfonic acid, irradiation was stopped and the reaction mixture left under

continuous stirring for three days in the dark at room temperature, product formation

continued affording 50 mM of heptane sulfonic acid. However, when the radical

scavenger hydroquinone184 was present during the dark phase, sulfonic acid production

did not continue.

In other words the radical chain reaction, once initiated, proceed without further

irradiation consuming the gases presents in the mixture. This finding is of great

importance for an industrial application.

System conditions Product concentration

TiO2 - SO2 n-heptane

2 h irradiation IPC-analysis performed just after

irradiation.

20 mM


2 h irradiation IPC-analysis performed three days after

stirring in the dark.

50 mM

TiO2 - SO2 n-heptane - hydroquinone

5 h irradiation IPC-analysis performed just after

irradiation.

0


2 h irradiation Hydroquinone added after irradiation

IPC-analysis performed after three days.

29 mM


116

• OH radicals scavenging

A well known method to provide evidence of OH radical formation exploits the

selective oxidation of benzoic acid to salycilic acid through OH radicals.

Detection of salycilic acid can be easily carried out by fluorescence spectroscopy.

Unfortunately attempts of providing evidences of OH radical formation failed,

perhaps because of the multiphase complexity of the system. A signal at 420 nm could

be detected in the fluorescence spectra of the filtered reaction mixture, but from the

absence of a band at 360 nm one must conclude that the emission spectrum is not

generated from salycilic acid.

Figure 4.24: Fluorescence spectra of the filtered reaction mixture after 0, 1 and 2 hours. It is also showed a fluorescence spectrum of a solution of salicylic acid in n-heptane.

COOH COOH

OH

OH· (4.29)

350 400 450 500 5500

102030405060708090

100110

Inte

nsity

/ a.

u.

W avelength / nm

0h

1h

2h

salicylic acid in n-heptane


117

Although no traces of salycilic acid can be detected with this method, the presence

of OH radicals during the reaction cannot be excluded.

4.2.8.2 Evidence for Formation of a TiO2-SO2 CT Complex

Since P25, n-heptane and SO2 are not able to absorb visible light, it seemed likely

that sulfur dioxide may form a charge transfer complex with the titania surface.

In the following we provide some evidence for this hypothesis.

• Color test

Exposure of P25 to sulfur dioxide resulted in a yellowish coloration of the

powder. This color change implies that the system TiO2-SO2 is able to absorb visible

light although, in contrast to titania and SO2.

The same coloration was observed when small quantities of n-heptane were

present. The effect became even more noticeable when some drops of acetonitrile were

added. In fact, the presence of a polar solvent should stabilize the charge-transfer

complex.

• DRS spectra

In Fig. 4.25 are summarized the Kubelka-Munk functions of P25 (a) and of a

sample of P25 previously treated with SO2 (b).


118

Figure 4.25: Plots of Kubelka-Munk function vs. wavelength of P25 (a), P25/SO2 (b), and (c) = (b)- (a).

Substracting spectrum (a) from spectrum (b) clearly reveals a broad absorption

maximum indicating that a TiO2-SO2 CT complex absorbs in the range between 400

and 420 nm.

Generally, electron-richer adsorbates (like e.g. enediols185,186) display CT bands in

the visible region. The less electron-rich adsorbates such as thiocyanate showed the

corresponding absorptions in the UV region.187 The CT-band observed in our case

suggests that SO2 acts as electron donor. This is in contrast to the hypothesis of Yanxin

et al. suggesting an electron-accepting nature of SO2 when in contact with TiO2. These

findings suggest that designing materials which display a CT complex could represent

an easy way to make photocatalytic (chemical synthesis or pollutant degradation)

processes more selective and efficient.

410 415 420 425 430

0,000

0,003

0,006

0,009

F (R

∞ )

/ a.u

.

W avelength / nm

c

a b


119

• Photoelectrochemical measurements

In order to gain more information on the dynamics of the interfacial charge transfer

a set of photocurrent measurements was carried out in a HClO4 (0.1 M) electrolyte

containing two different reducing agents: SO2 and I- (see experimental part for details).

Fig. 4.26 shows a simplified scheme illustrating the processes leading to

photocurrent generation upon irradiation of a semiconductor powder (P25 in our case)

deposited onto an ITO-glass electrode.

Figure 4.26: Simplified scheme illustrating photocurrent generation upon irradiation of a powder

photocatalyst deposited on an ITO-glass electrode (adapted from ref.36). VB and CB represent valence and conduction band, respectively. All recombination pathways are omitted for the sake of clarity. For details see the text.

The electrodes used in this investigation consist of a porous network of particles on

ITO-glass easily penetrable by the electrolyte. Since the electron transfer from the

particles to the ITO layer can be assumed to occur readily because the Fermi level of

the anodically biased ITO layer lies well below the flat band potential of the titania, it

is the interfacial oxidation process which exerts a crucial influence on the photocurrent

response of the electrode. The electrode was irradiated from the back side (through the


120

ITO glass) with intermittent light at wavelength varying between 0 and 700 nm. UV

excitation affords an electron-hole pair (1). The electron should be easily transported

to the ITO glass support. The hole can either recombine directly with a conduction

band electron (primary recombination) or relax to an energy level lying close to the

valence band edge (2). From there it can again either recombine (secondary

recombination) or react with a reducing agent present in the electrolyte (3), whereby

only in the latter case a net photocurrent is observable. When no additional reducing

agent is added, the hole is supposed to oxidize water to oxygen. The photocurrent thus

observed, originates from oxidation processes at the contact surface between working-

electrode and electrolyte is called anodic photocurrent.

Fig. 4.27a shows the photocurrent action spectra of P25 in HClO4 1M solution

with (dashed line) and without (solid line) SO2 as reducing agent. The spectra in Fig.

4.27b were found analogously in LiClO4 1M with iodide as reducing agent.

320 340 360 380 400 420 4400

100

200

300

400

500

Cur

rent

den

sity

/ μA

cm-2

Wavelenght / nm

(a)


121

320 340 360 380 400 420 4400

100

200

300

400

500

Cur

rent

den

sity

/ μA

cm

-2

Wavelenght / nm

(b)

Figure 4.27: Photocurrent measured under intermittent irradiation (5s light, 5s dark) as a function of irradiation wavelength. (a): P25 coated electrode in HClO4 1M solution with (dashed line) and without (solid line) SO2 as reducing agent. (b) P25 coated electrode in LiClO4 1M solution with (dashed line) and without (solid line) iodide as reducing agent.

For TiO2 a typical behavior is observed with anodic photocurrent disappearing at

wavelength above 400 nm, the edge of the light absorption for titania.

The addition of a reducing agent increases the photocurrent. In fact, the reacting

holes can escape recombination more easily since the oxidation of iodide or of SO2 is

thermodynamically more favorable than water oxidation which is a very slow reaction

and requires a potential of about 2.0 V vs. NHE at pH 7. Thus, in absence of I- or SO2,

the photogenerated holes preferentially undergo fast recombination. This is also

evident observing the clearly different photocurrent transients in absence and presence

of added reducing agents. In the former case, after the initial rise of photocurrent

immediately after switching on the light, a rapid exponential decay is observed. Such a

shape of photocurrent transient is a typical fingerprint of surface recombination

processes.

The intensity of the photocurrent signal is obviously proportional to the number of

electrons injected into the ITO glass. A ratio between the intensity of this signal in the


122

presence and absence of reducing agent affords information about the surplus of

electrons generated in the presence of SO2 or iodide. Fig. 4.28 illustrates the

dependence of this ratio on the wavelength of irradiation for iodide and SO2.

Figure 4.28: Surplus of electron generated in the presence of SO2 or iodide with respect to the case without reducing agent, as a function of the wavelength of the irradiation.

In the UV range of irradiation SO2 and iodide are both oxidized from the light

generated holes and we observe an increase of electrons with respect to the case

without SO2 or I-. In the case of SO2 this increment is accented due to the greater

reducing power of SO2. However, the trend in this region is quite the same for SO2

and I-, demonstrating a similar electron transfer process for this two reducing agents.

From 400 nm onwards this ratio for SO2 becomes much greater than the

corresponding ratio for I-. This means that in this region a further surplus of electron

was detected compared to the sole oxidation of SO2. Since, to our knowledge, iodide

does not form any CT complex with titania, we attribute this result to the above

320 340 360 380 400 420 4401

2

3

4

5

6

7

8

9

10

Δ w

ith x

/Δ w

ithou

t x

Wavelenght / nm

X = SO2

X = I-


123

mentioned TiO2-SO2 CT complex whose visible light excitation generates additional

electrons in the conduction band.

4.2.9 Oxidation Products

With the sole exception of Takahara et al.188 who reported on the photooxidation of

cyclohexane in solutions containing hydrogen peroxide and titania particles under

visible light, photocatalytic oxidation of alkanes to the corresponding alcohols and

ketones occurs in oxygenated suspensions of TiO2 only under UV light irradiation.

However, in our system formation of oxidation products is also expected since the

visible light generated alkyl radical R· should undergo a fast reaction with oxygen.

Surprisingly, in the system cyclohexane/SO2/O2/P25 no traces of alcohol or ketone

or CO2 could be detected in the gas phase by gas chromatography. However, in the

liquid phase very small amounts of cyclohexanone and cyclohexanol were observable.

After 2 h of irradiation (λ ≥ 400 nm) about 6 x 10-3 mmol of sulfonic acid, 60 x 10-3

μmol of cyclohexanone and negligible quantities of cyclohexanol were produced.

These results show that the oxidation products are really minor byproducts of this

surprisingly selective alkane activation.

Why is sulfoxidation such a preferential reaction path? Why is the first step after

radical formation addition of SO2 to form RSO2· and not addition of O2 to form RO2· ?

The answer can be given by discussion of some basic kinetic aspects.

The rate constants of the reaction of alkyl radicals with SO2 and O2 in the gas

phase have been measured by Good and Thynne189. They found that the rate constant

with oxygen is about 70 time faster than with sulfur dioxide.

In our photosulfoxidation system these reactions probably occur on the titania

surface. As mentioned before SO2 forms a CT-complex with titania and therefore one

expects that its surface concentration should be much higher than that of oxygen.

Therefore, the rate of reaction with an alkyl radical should become faster as compared

to the addition reaction with O2.


124

4.2.10 Preparative Synthesis of Heptanesulfonic Acid Sodium Salt.

In order to demonstrate the practical aspects of this new sulfoxidation, we

performed the reaction in a 200 ml immersion lamp apparatus. Through the cooling

jacket a solution of 1 M NaNO2 was circulated as a cut-off filter to ensure irradiation

at λ ≥ 400 nm. More details about the reaction system are reported in the experimental

section.

After 5 h irradiation time and 5 days stirring in the dark it was possible to isolate

600 mg of heptane sulfonic acid sodium salt.

4.3 Results and Discussion: Reaction in Acetic Acid

In this dissertation we focused our interest on the role of the solvent choosen to

carry out the reaction. Many solvents have been tested but only acetic acid turned out

to be suitable for the sulfoxidation.

Tetrahydrofuran, methanol and butanol reacted under the given conditions and

competed with the sulfoxidation of n-heptane being more reactive molecules.

Tetrachloromethane did not show react in a blank test without alkane (other conditions

being equal) but in its presence sulfoxidation was inhibited. Probably, because of its

electron scavenging properties, it competed with oxygen inhibiting the reductive path

of the reaction.

On the other hand, blank tests without alkane showed that acetic acid was stable

under the reaction conditions and we went on to investigate possible advantages of this

solvent. Surprisingly, n-heptane photosulfoxidation proceeded without any noticeable

catalyst deactivation. The catalyst remained in suspension even after 15 hours of

irradiation and no color change of the powder was noticeable. This finding justified

further efforts for investigation of this system.


125

4.3.1 System Description and Product Characterization

When a suspension of titania in a mixture of n-heptane/acetic acid under an

atmosphere of SO2/O2 (1:1) was irradiated with visible light (λ ≥ 400 nm), formation

of heptanesulfonic acid was observed. Increasing the n-heptane concentration induced

a strong rate increase (Fig. 4.29). Since already at a 1:1 (v/v) mixture of n-

heptane/acetic acid sufficient reaction rates are obtained, all the following experiments

were performed at this ratio.

0 20 40 60 80 100

r i [m

M /

h]

% n-heptane

0

2

4

6

Figure 4.29: Dependence of the initial rate of heptanesulfonic acid formation on the concentration of n-heptane in acetic acid

After 5 h of irradiation (where not otherwise specified) the reaction mixture was

filtrated and concentrated in vacuum. The oily residue was diluited in methanol and

analyzed by means of IPC. To ensure formation of heptanesulfonic acid, the reaction

product was isolated and identified by IR spectroscopy.

Fig. 4.30 shows a comparison between IR spectra of the authentic heptane sulfonic

acid (a), of the product obtained in the presence of acetic acid (b) and of the product

obtained in the absence of acetic acid (c).


126

0 500 1000 1500 2000 2500 3000 3500 400005

101520253035404550556065707580

Tran

smitt

ance

[%]

Wavenumber [ cm-1]

a)

b)

c)

Figure 4.30: IR spectra of the authentic heptane sulfonic acid (a), of the product obtained in the presence of acetic acid (b) and of the product obtained in the absence of acetic acid (c).

The most important signals are present in each spectrum. We noticed in the case of

the acetic acid system, a more accented shoulder of sulphate at 1126 cm-1.

As already mentioned, at the ratio n-heptane/acetic acid = 1 no changes in the

character of the suspension or in the color of titania was noticeable. However, when

the ratio was equal to 9 the same deactivation process as in pure n-heptane was

observed. We hoped that in the presence of acetic acid a simultaneous regeneration

process could occur. We therefore followed formation of heptanesulfonic acid as

function of irradiation time (Fig. 4.31).


127

-2 0 2 4 6 8 10 12 14 16-505

101520253035

c (H

SA

) [m

M]

Time / h

Figure 4.31: Dependence of formation of heptanesulfonic acid (HSA) on irradiation time in acetic acid/n-heptane = 1:1 (v/v).

Surprisingly, although no visible changes were observable in the suspension, after

6 h product formation reached a plateau as in the case without acetic acid. Because of

the evidence of a still good suspension and of a white color of the powder, a strong

product adsorption cannot be given as the only explanation. Furthermore, the

hypothesis of lack of reagents (limitant reagent) during the reaction could be excluded.

The reason of this finding may be understood taking in consideration the interaction

between acetic acid and titania.

4.3.2 Acetic Acid Adsorption at TiO2

Acetic acid adsorption on TiO2 is a well known topic in literature190-194 and was

extensively studied in the solid/gas and solid/liquid regime. Photocatalytic oxidation of

acetic acid can be achieved under UV irradiation in the presence of titania. The α-

carbon leads to formation of CO2 without forming any long-lived intermediates, while

the β-carbon forms CO2 through metoxy, formaldehyde, and formate. Water is also

produced both during the degradation process and the adsorption process.194

Acetic acid adsorbs both molecularly and dissociatively as acetate on TiO2. The

latter results in titanium surface complexes containing a bidentate or monodentate

acetate ligand (structures I and II, Fig. 4.32).


128

Figure 4.32: Adsorption of acetic acid onto titania. (Taken from ref.193)

FTIR spectra from acetic acid–TiO2 interaction show two bands at 1552 and 1445

cm-1, attributed to acetate νas and νs vibrations, respectively. Different authors have

indicated that a separation between these two bands, Δν of 80 – 90 cm-1, corresponds

to a statistically greater amount of structure I, while Δν = 140 – 160 cm-1 indicates that

structure II is more probable on the titania surface. However, Δν assumes normally

intermediate values indicating a mixture of both structures. 193

Acetic acid complexation requires the presence of Lewis acidic (Ti4+) and basic

(O2-, OH-) centres:

CH3COOH + Ti4+ + [O2-]surf → CH3COO -······ Ti4+ + OH- (4.30)

CH3COOH + Ti4+ + [OH-]surf → CH3COO -······ Ti4+ + H2O (4.31)

In order to investigate the role of the adsorbed acetate on titania in the

sulfoxidation, we synthesized three acetic acid-modified titania samples starting from

different amounts of acetic acid in water during the adsorption step: 25%Ac-TiO2,

50%Ac-TiO2, 100%Ac-TiO2. The synthesis procedure is described in the experimental

part.

If a standard sulfoxidation reaction in pure n-heptane was carried out in the

presence of acetic acid modified titania samples (100%Ac-TiO2), no changes were


129

noticed in the yield of the reaction indicating that the plateau reached in Fig. 4.31

cannot be attributed to the influence of adsorbed acetate species.

4.3.3 Influence of Water

In most photocatalytic systems at least traces of water are necessary to observe the

desired reaction. The reason for this may be the formation of intermediate OH radicals

as very reactive intermediates and a positive influence of water on the interfacial

electron transfer. In the photooxidation of pollutants both enhancement and inhibition

of the degradation rate can be caused by water vapor. Water vapor inhibited the gas

phase degradation of ethylene195 whereas enhanced the gas phase photooxidation of

cyclohexane196 and toluene197 but no significant effects on benzene198 oxidation were

noted.

Another parameter that has to be evaluated studying the influence of water is the

hydrophobicity of the organic substrate. In this case water can totally solvate TiO2

inhibiting the contact between the titania particles and the organic substrate.This effect

is already noticed in the system without acetic acid where amounts of water like 0.3 %

(V/V) totally inhibited the sulfoxidation of n-heptane.

Taking into consideration that the acetic acid adsorption process affords one

molecule of water for every acetic acid molecule adsorbed, the amount of water in this

way produced on the titania surface should be enough to generate a hydrated TiO2

surface and to block further product formation.

In order to check this hypothesis a suspension of P25 in a solution n-heptane/acetic

acid (1:1) was stirred three days in the dark. Then the mixture was transferred in the

reactor and SO2 and O2 were metered into it. After 5 h of visible irradiation the powder

was still white and well suspended but only negligible amounts of sulfonic acid could

be detected. This suggests that the water produced in the acetic acid absorption process

inhibited photosulfoxidation.

We tried to remove water traces from the reaction mixture during the reaction

using zeolites, MgSO4 or CaSO4 but unfortunately these attempts failed because the


130

reaction was strongly influenced by these salts. Further investigation in order to afford

water separation and simultaneously a protective effect on the catalyst could open new

perspectives for this reaction.

4.4 Experimental part

4.4.1 Materials

Titanhydrat (Kerr-McGee Pigments, 300 m2/g) and P25 (Degussa, 50 m2/g) were

used as received. We are thankful to Prof. T. Egerton for a sample of high surface area

rutile (140 m2/g). [TiO2]OPtCl4,61 TiO2-C,133 TiO2-N,158 were prepared according to

literature and have surface areas of 260, 160, and 170 m2/g, respectively. Adamantane

(Acros), n-heptane (Fischer) and cyclohexane (Acros) and other compounds

mentioned were used as received.

4.4.2 Standard Photosulfoxidation

30 mg of the titania powder were suspended in 15 ml of n-heptane or cyclohexane

in a 20 ml solidex glass cuvette and sonicated for 15 min. Thereafter 60 ml of a 1:1

(v/v) gaseous mixture of O2 and SO2 were added by a syringe. Irradiation was

performed with an Osram XBO 150 W xenon arc lamp, (Io (400 nm – 520 nm) = 2 x

10-6 Einstein s-1 cm-2) installed in a light condensing lamp housing (PTI A1010S) on an

optical train. A cut–off filter of λ ≥ 400 nm was placed in front of the cuvette.


131

Figure 4.33: Spectral intensity distribution of the 150 W XBO lamp.

Figure 4.34: Experimental set up for standard sulfoxidation experiments. A: power supply, B: xenon-arc lamp with water cooling, C: IR filter, D: cut-off filter λ ≥ 400 nm, E: solidex glass cuvette, F: magnetic stirrer.

A

E

F

D

C B


132

The suspension was stirred magnetically. After 5 h of irradiation time the

photocatalyst was filtered through a micropore filter (Whatman 0.45 µm) and the

filtrate was concentrated in vacuo. The slightly yellow, oily residue was dissolved in 3

ml of methanol and analyzed by HPLC (IPC)160,161 or neutralized through NaOH to

obtained the salt which was washed several times with diethyl ether. Adamantane was

photosulfoxidized analogously by employing 136.2 mg (1 mmol) of adamantane in 15

ml of acetic acid.

Preparative isolation of heptanesulfonic acid was achieved using a 200 ml

immersion lamp set up (Fig. 4.35).

Figure 4.35: Immersion lamp set up to preparative production of heptanesulfonic acid. A: bubbling

of SO2 and O2 through water, B: immersion lamp reactor, C: cooling system, D: power supply, E: NaNO2 1M solution, F: peristaltic pump.

The mixture was irradiated through a 100 W tungsten-halogen lamp (the spectrum

is showed in Fig. 4.36). Visible light irradiation (λ ≥ 400 nm) was ensured through

circulating of a 1 M NaNO2 solution through the cooling jacket of the lamp. This

solution was continuously cooled and reintroduced with a peristaltic pump into the

jacket. The gases were metered into the reaction suspension after bubbling through

A

D

B

CF

E


133

water to control the flow rate. After 5 h irradiation the suspension was stirred

magnetically during 5 days and the product was isolated as described in the case of the

standard reaction.

Figure 4.36: Spectrum of the 100 W tungsten-halogen lamp employed in the immersion apparatus.

4.4.3 Instruments and Methods

• UV-Vis spectroscopy

Shimadzu UV – 3101 PC UV-Vis-NIR Scanning Spectrophotometer, Quarz

cuvette with d = 1 cm.

• Diffuse reflectance spectroscopy

Shimadzu UV-2401 PC UV-Vis Recording Spectrophotometer equipped with a

diffuse reflectance accessory. The samples were spread over a BaSO4 pellet after

this was used for measuring the background reflectance. The reflectance was

converted by the instrument software to F(R∞) according to the Kubelka-Munk

theory.

• IR spectroscopy

Perkin Elmer 16 PC FT – IR


134

• Elemental analysis

Carlo Erba Elemental Analyser Model 1108

• Specific surface

Gemini 2370, surface calculated according to the Brunauer-Emmet- Teller theory

• Ion chromatography

Dionex – 120, Ion Pac AS 14 column, conductivity detector, eluent NaHCO3 /

NaCO3 = 0.001 / 0.0035 M

• Gas Chromatography

Shimadzu GC-17A gas chromatograph. Column: Supelcowax TMC 30m, d: 0.54

mm - 1μ. Carrying gas: N2

• Mass spectroscopy

JEOL JMS 700 (EI 70 eV, FD 2kV)

• XPS

XPS spectra were recorded using a Phi 5600 ESCA instrument (pass energy of

46.95 eV, Al std, 300.0 W, 45.0°). The binding energy reference was taken as the

C1s peak from carbon contamination of the samples at 284.8 eV. Fitting of the

experimental XPS data was done after a background correction using the Shirley

method.199 For modeling the peak shapes, Gaussian-Lorentzian combinations were

used.

• Fluorescence spectroscopy

Perkin Elmer LS 50B luminescence spectrometer

• HPLC with IPC

SCL 10 AVP system controller, SIL -10A autosampler, SP10AVP model UV

detector, Column (250 x 4.6 mm I.D.) filled with Partisil 10 SAX (Whatman) which is

a strong anion exchanger. Water-acetonitrile (60/40, v/v) with 0.01 M potassium

hydrogenphthalate as UV absorbing counter ion was employed as eluent. Detections


135

were made at 304 nm and the pH value of the eluent was 5.8. Detector SP10AVP

model UV detector (304 nm).

• PEMF measurements

One hundred milligrams of the powder were dispersed by ultrasonics in 3 g of a

10% solution of polyvinylbutyrale in 1, 2-dichloroethane. The suspension was placed

on a glass plate (47 cm2 in area). After drying in a solvent atmosphere, the polymer

film was removed from the glass support. The remaining solvent was removed from

the film under vacuum. The resulting sheet had a thickness of 60 to 80 µm and exhibits

total light absorption in the UV.

The cell for PEMF measurements (Fig. 4.15) is constructed like a capacitor with

the sample as dielectric layer. The sample is flash-illuminated by a nitrogen laser type

“PNL 100” (LTB Lasertechnik Berlin, λflash = 337 nm, τ1/2 = 0.3 ns, 2.7 x 10-3 quanta

per flash).

The resulting PEMF is measured without any galvanic contact because there are

insulating foils between sample sheet and electrodes. This prevents charge injection

from the electrodes into the sample. The preamplifier has an impedance of about 1TΩ

and the PEMF measurements take place without any external electric field.

All experiments were carried out in air under normal pressure and at 298 K. The

signals were recorded in two different time range:

- till 2,5 μs after flash in order to record the process of build-up and of the fast

decay.

- till 200 ms after flash in order to record the process of slow decay.

In all the cases the signal of the first laser flash was recorded. After that, sampling

of 4 signals was done in order to explore the charge effects.

For sampling experiments the time distance between two flashes was 120 s. All the

measured values and curves showed are an average of three independent

measurements.


136

• Photoelectrochemical measurements

Photoelectrochemical experiments were performed with a tunable monochromatic

light source provided with a 1000 W Xenon lamp and a universal grating

monochromator Multimode 4 (AMKO, Tornesch, Germany) (Fig. 4.37).

Figure 4.37: Photoelectrochemical set-up used for photocurrent measurements (taken from ref.36).

Figure 4.38: Schematic view of key steps in electrode preparation(taken from ref.36): ITO glass is

placed between two pieces of normal glass (a) and the glass edges are covered with a scotch tape (b); powder suspension is dropped onto the first glass (c) and smeared regularly with a glass rod (d); after removal of two side glasses (e) and drying the ITO electrode is covered with an aluminum foil and another piece of glass (f) and then pressed using an IR pressing tool (g); the electrode is then contacted with a copper wire using a conductive tape (h).


137

For photocurrent measurements electrodes consisting of a porous nanocrystalline

film deposited on ITO-glass were prepared according to ref. 36(Fig. 4.38). The

conductive ITO-glass substrate (Präzision Glas & Optik, Iserlohn, Germany, sheet

resistance of ~10 Ω/sq.) was first cut into 2.5 × 1.5 cm pieces and then subsequently

degreased by sonicating in acetone and boiling NaOH (0.1M), rinsed with

demineralized water, and blown dry in a nitrogen stream. A suspension of 200 mg of

P25 TiO2 in 1 ml of ethanol was sonicated for 20 minutes and then deposited onto the

ITO glass by doctor blading using a scotch tape as frame and spacer. The electrodes

were then dried at 100 °C, covered with aluminum foil and a glass plate, pressed for 3

minutes at a pressure of 200 kg/cm2 using an IR pressing tool (Paul Weber, Stuttgart,

Germany) according to a procedure similar to that described in literature200. Such a

procedure yields a ~ 2.5 μm thick opaque and slightly translucent layer of TiO2 having

an excellent mechanical stability.

Figure 4.39: A photoelectrochemical cell used for photocurrent measurements viewed from the side (left) and from the front (right) (taken from ref.36).


138

The electrochemical setup consisted of a BAS Epsilon Electrochemistry

Potentiostat (BAS, West Lafayette, USA) and a three-electrode cell using a platinum

counter electrode and an Ag/AgCl (3 M KCl) reference electrode (Fig. 4.39). During

photoelectrochemical measurements the electrodes were pressed against an O-ring of

an electrochemical cell leaving a working area of 0.636 cm2. The photocurrent

experiments were carried out in a 0.1 M HClO4 solution when SO2 was used as

reducting agent, in a 0.1 M LiClO4 solution when iodide was used as reducting agent.

Nitrogen was passed through the electrolyte prior to the experiment whereas it was

supplied only to the gas phase above the electrolyte during the experiments. In the

experiment with iodide the KI concentration was 0.1 M. Presence of SO2 in the

electrolyte was achieved considering that SO2 is the stable form of S (IV) in acidic

solutions. After bubbling nitrogen in the electrolyte, stoichiometric amount of Na2S2O5

was introduced into the solution. Under these conditions this species is immediately

converted to SO2 according to Eqs. 4.32 and 4.33.

Na2S2O5 + H2O → 2NaHSO3 (4.32)

HSO3- + H+ → H2O + SO2 (4.33)

The bubbling was replaced with a nitrogen flow above the solution in order that

SO2 did not escape too much from the solution.

The wavelength dependence of photocurrent was measured at a constant potential

of 0.5 V vs. Ag/AgCl. The electrodes were irradiated from the back-side (through the

ITO glass) with light and dark phases of 5 and 10 s, respectively.


139

4.4.4 Surface Modification of Titania

• Metal salts modification

The ruthenium and iridium modified titania were obtained with a similar procedure

as described in Chapter 3 for the rhodium modification. The starting metal salts

were RuCl3 · xH2O and IrCl3 · 3H2O, respectively.

• Synthesis of 100%Ac-P25

1g of P25 was stirred during 12 h in the dark in 20 ml of glacial acetic acid. The

suspension was dried in vacuum at RT for 90 min.

The resulting powder was washed 9 times until the pH value of the washings

reached a constant value of 4.2. The powder was dried again at room temperature

under vacuum and finally ground. Analogous procedure was carried out to synthesize

25% Ac-P25 and 50%Ac-P25. In these cases P25 was stirred during 12 h in the dark in

a solution of water/glacial acetic acid 25% and 50% (v/v) respectively.

• Synthesis of Silylated P25

In a round bottom flask were mixed 1g of P25, 50 ml of n-hexane, 2 ml of

triethylamine and 7 g of C16H32Si(OCH3)3 (Wacker Silan 250, 2VP). The suspension

was boiled at 90° C for 24 h. The solid phase was separated from the liquid by

centrifugation and washed three times with n-hexane. The powder was dried at room

temperature in vacuum and finally ground.

The silylation process is described in Eqs. 4.34 and 4.35.

[Ti]OH + NEt3 → TiO- + HNEt3+ (4.34)

[Ti]O- + R-Si(OCH3)3 + HNEt3+→ [Ti]-O-Si(OCH3)2-R + CH3OH + NEt3 (4.35)


140

The base triethylamine reacts with the acidic groups on titania giving the strong

base TiO- which in turn attacks the silicon atom under formation of an oxygen bridge

between Ti and Si. Concomitant elimination of CH3OH reforms triethylamine.

Similarly, the remaining methoxy groups (OCH3) can be also replaced by other TiO-

groups.

• Synthesis of phosphated P25

A suspension of 1g P25, 2.76 g NaH2PO4 in 50 ml water was mixed during 12

hours. After centrifugation, the powder was washed three times with 50 ml H2O, dried

under vacuum at room temperature and finally ground.

• Synthesis of fluorinated P25

Fluorinated P25 was prepared by the impregnation method. 1g of P25 was

impregnated in 50 ml of 4% NaF solution at pH 3.2 (pH adjusted by HNO3) for 2 days

at room temperature. The powder separated by centrifuge was washed two times with

water at pH 3.2, dried in vacuum at 80°C for 2h and finally ground.

5. Summary

141

5. Summary

C-H bond activation and functionalization of alkanes is one of the major

challenges in chemistry. A rare example of an industrially applied process is the

photosulfoxidation of liquid alkanes by sulfur dioxide and oxygen in the presence of

UV light (Eq. 5.1).

R-H +SO2 +1/2 O2 + hν → RSO3H (5.1)

In the case of linear C16-20 chain alkanes the resulting alkanesulfonic acids are used

as biodegradable surfactants. The primary reaction steps of this rare alkane

functionalization consist of UV-excitation of SO2 followed by hydrogen abstraction

from the alkane producing an alkyl radical. Subsequent addition reactions with SO2

and O2 generate an alkylpersulfonyl radical which in turn produces another alkyl

starter radical and the persulfonic acid. Fragmentation and hydrogen abstraction afford

the alkanesulfonic acid. In general regioisomeric alkyl radicals are formed in the

hydrogen abstraction step. A rare example for a sensitized process is the mercury

photosensitized sulfination of alkanes with SO2 producing initially sulfinic acids

(RSOOH) and sulfinic esters which have to be further oxidized to sulfonic acids by

hydrogen peroxide. All the reactions mentioned above occur only upon excitation of

sulfur dioxide or mercury with UV light. In this dissertation we report on the first

catalytic photosulfoxidation of alkanes. This reaction does not require UV lamps and

toxic sensitizers, but only a non-toxic semiconductor powder inducing alkane

functionalization through visible light excitation.

In a preceeding dissertation157 some basic features of this novel reaction were

investigated, however reproducibility problems and basic mechanistic questions

required a more detailed investigation of this photocatalytic C-H activation reaction.

5. Summary

142

n° Photocatalyst ri [mmol l-1 h-1]

1 Titanhydrat(A) 3.5


3 TiO2 (R) 6.0

4 TiO2 (P25, A+R) 7.5



7 TiO2-C, TiO2-N (A) 3.5

Table 5.1: Initial rate ri of n-heptanesulfonic acid in the presence of different TiO2 photocatalysts. A = anatase, R = rutile.

When a suspension of a titania powder in n-heptane was irradiated with visible

light (λ ≥ 400 nm) under an atmosphere of SO2/O2 = 1:1 (v/v), the formation of n-

heptanesulfonic acid (1) was observed (Tab. 5.1). Only traces of the sulfonic acid were

observable in the absence of titania. Initial product formation rates were 3.5 mmol/l.h

and 5.0 mmol/l.h for the anatase materials Titanhydrat and Hombikat, respectively,

whereas for rutile and the mixed phase powder P25 (75% anatase/25% rutile) values of

6.0 mmol/l.h and 7.5 mmol/l h were observed. Out of the modified titania powders

(entries 5-7), which are all good photocatalysts in 4-chlorophenol visible light

oxidation, only the titania-chlororhodate complex and carbon- or nitrogen-modified

titania exhibited moderate rates of 3.5 mmol/l.h.

Under the given experimental conditions, formation of 1 stopped after 6 h of

irradiation time. However, separating the catalyst powder and washing with methanol

restored the activity. Repeating this procedure three times, the photocatalyst still

retained its original activity (Fig. 5.1). This observation suggested that the reaction is

inhibited by strong product adsorption and that washing desorbs the sulfonic acid.

Accordingly, no product formation was observable when 1 was added to the

suspension prior to irradiation. Fuhermore, during the photocatalytic sulfoxidation P25

5. Summary

143

changes its photoelectrical behaviour from n-type to p-type as indicated by time-

resolved photovoltage measurements; this may decrease the reaction efficiency.

Product formation was also inhibited when small amounts of water like 0.3 vol% were

present in the suspension. This may be due to blocking the reactive surface centres for

heptane oxidation by preferential product adsorption.

0

10

20

30

40

hνhν

c(1)

/ m

M

hν

R R R

0 10 0 10 0 10

Time / h

Figure 5.1: Sequential photosulfoxidation of n-heptane. λirr ≥ 400 nm. R = regeneration.

When after 2 h of irradiation, corresponding to a concentration of 15 mM of 1,

irradiation was stopped and the reaction left for three days in the dark at room

temperature, product formation continued affording 50 mM of 1. However, when the

radical scavenger hydroquinone was present during the dark phase, sulfonic acid

production did not continue.

These findings suggest that also this novel photosulfoxidation is a radical chain

reaction. However, in this case the alkyl starter radical is generated not via UV

excitation of sulfur dioxide but through visible light absorption of the TiO2/n-

heptane/SO2/O2 system. Since only the modified titania powders (Tab. 5.1, entries 5-7)

5. Summary

144

are able to absorb visible light, it seemed likely that the unmodified materials may

form a charge-transfer complex with sulfur dioxide. In fact, exposure of P25 to sulfur

dioxide resulted in a yellowish coloration of the powder originating from a broad

absorption maximum in the diffuse reflectance spectrum at 410-420 nm. Accordingly,

a preliminary mechanism for alkyl radical generation is proposed as schematically

depicted in Fig. 5.2. Visible light excitation of the charge-transfer complex affords a

conduction band electron [TiO2(e−)] and an adsorbed sulfur dioxide radical cation.

Oxygen reduction by [TiO2(e−)] produces superoxide whereas the adsorbed sulfur

radical cation may oxidize the alkane to the alkyl radical and a proton. Superoxide

may also generate an alkyl radical through protonation by adsorbed water or surface

OH groups to the hydroperoxyl radical and subsequent hydrogen abstraction from the

alkane.

Figure 5.2: Proposed mechanism of photocatalytic sulfoxidation of alkanes with visible light.

The alkyl radical thus produced is expected to initiate a radical chain reaction as

formulated for the stoichiometric UV-photosulfoxidation. In agreement with the

proposed mechanism is the complete inhibition observed in the presence of only 10

vol% of isopropanol, which should be much faster oxidized than the alkane and which

is also an efficient OH radical scavenger.

[TiO2(e-)---SO2+·]

[TiO2---SO2]

O2

RH RH

R· R· + H+

Vis

5. Summary

145

In summary, this novel visible light induced C-H activation can be classified as

a photocatalysis type B reaction, extending the previously known two-substrate

addition to the present three-substrate addition scheme A + B + C = D.

Surprisingly other expected by-products like ketones, alcohols and CO2 are

formed, if at all, only in negligible amounts. In the case of cyclohexane

photosulfoxidation only about 0.1% of cyclohexanone (relative to cyclohexanesulfonic

acid) were observed. As mentioned before SO2 forms a CT-complex with titania and

therefore one expects that its surface concentration should be much higher than that of

oxygen. Therefore, the rate of reaction with an alkyl radical should become faster as

compared to the addition reaction with O2. This could be the reason for the unexpected

selectivity of this reaction.

6. Zusammenfassung

146

6. Zusammenfassung

Die Aktivierung und Funktionalisierung von C-H Bindungen gehört zu den großen

Herausforderungen der Chemie. Das seltene Beispiel eines bereits industriell

angewandten Prozesses ist die Sulfoxidation flüssiger Alkane durch Schwefeldioxid

und molekularen Sauerstoff in Gegenwart von UV-Licht (Gl. 6.1).

R-H +SO2 +1/2 O2 + hν → RSO3H (6.1)

Im Falle von C16-20 Alkanen finden die daraus resultierenden Alkansulfonsäuren

als bioabbaubare Waschmittel Verwendung. Die primären Reaktionsschritte dieser

Alkanfunktionalisierung bestehen aus der UV-Anregung von SO2, dessen

Triplettzustand mit dem Alkan unter H-Abstraktion ein Alkylradikal erzeugt.

Anschließende Additionsreaktionen mit SO2 und O2 führen zu einem

Alkylpersulfonylradikal, welches über eine H-Abstraktion zum Alkylradikal und zur

Alkanpersulfonsäure führt. Nachfolgende Fragmentierung und Wasserstoffabstraktion

ergeben schließlich die Alkansulfonsäure und weitere Starterradikale. Im Allgemeinen

entstehen im ersten H-Abstraktionsschritt regioisomere Alkylradikale. Ein sehr

seltenes Beispiel für eine photosensibilisierte Variante ist die

Quecksilbersensibilisierte Sulfinierung von Alkanen mit SO2 zu Sulfinsäuren

(RSOOH) und Sulfinsäureestern. Nachfolgende Oxidation mit Wasserstoffperoxid

führt ebenfalls zu Alkansulfonsäuren. Alle oben erwähnten Reaktionen erfordern die

UV-Anregung von Schwefeldioxid oder Quecksilber. In dieser Dissertation berichten

wir dagegen über die katalytische Photosulfoxidation von Alkanen, die mit sichtbarem

Licht abläuft. Sie benötigt keine UV-Lampe oder einen toxischen Photosensibilisator,

sondern lediglich ein nicht-toxisches Halbleiterpulver als heterogenen

Photokatalysator. Manche grundlegenden Aspekte dieser neuen Reaktion wurden in

einer vorherigen Dissertation157 untersucht. Trotzdem traten Reproduzierbarkeit

6. Zusammenfassung

147

Probleme und mechanistische Fragen auf, die nach einer detalierteren Untersuchung

der photokatalytischen Aktivierung der C-H Bindung verlangten.

n° Photocatalysatoren ri [mmol l-1 h-1]

1 Titanhydrat (A) 3.5


3 TiO2 (R) 6.0

4 TiO2 (P25, A+R) 7.5



7 TiO2-C, TiO2-N (A) 3.5

Tabelle 6.1: Anfangsgeschwindigkeit ri der Bildung von n-Heptansulfonsäure (1) in Gegenwart verschiedener Titandioxidphotokatalysatoren. A und R bezeichnen Anatas- und Rutilmodifikationen.

Wurde eine Suspension von Titandioxid in n-Heptan unter einer Atmosphäre von

SO2/O2 = 1:1 (v/v) mit sichtbarem Licht (λ ≥ 400 nm) bestrahlt, ließ sich die Bildung

von n-Heptansulfonsäure (1) nachweisen (Tab. 6.1). Nur Spuren dieses Produkts

entstanden, wenn Titandioxid weggelassen wurde. Die Anfangsgeschwindigkeiten der

Produktbildung betrugen für die Anatasmodifikation Titanhydrat und Hombikat 3.5

mmol/l.h bzw. 5.0 mmol/l.h, während Werte von 6.0 mmol/l.h und 7.5 mmol/l.h für

Rutil bzw. das gemischtphasige P25 (75% Anatas/25% Rutil) erhalten wurden. Von

den modifizierten Titandioxiden (Tab. 6.1, Zeilen 5-7), die alle gute

Photokatalysatoren für die vollständige Oxidation von 4-Chlorophenol mit sichtbarem

Licht sind, [9-12] induzierten lediglich der Titandioxid-Chlororhodatkomplex sowie

Kohlenstoff- und Stickstoff-modifiziertes Titandioxid moderate

Reaktionsgeschwindigkeiten von 3.5 mmol/l.h. Unter den gegebenen experimentellen

6. Zusammenfassung

148

Bedingungen kam die Reaktion nach 6 h Belichtungszeit zum Erliegen. Wurde

allerdings anschließend das Katalysatorpulver abgetrennt und mit Methanol

gewaschen, konnte seine ursprüngliche Aktivität wieder hergestellt werden. Auch nach

dreimaliger Wiederholung dieser Prozedur änderte sich die Anfangsgeschwindigkeit

kaum (Abb. 6.1). Diese Befunde deuteten auf das Vorliegen einer Inhibierung durch

Produktadsorption, die durch Waschen mit Methanol wieder aufgehoben werden kann.

Dementsprechend entsteht keine weitere Sulfonsäure, wenn diese vor

Belichtungsbeginn zur Suspension zugesetzt wurde. Des Weiteren ändern sich bei der

photokatalytischen Sulfoxidation, die photoelektrische Eigenschaften des P25, von n-

type zum p-type (PEMF Ergebnisse). Dieser Vorgang kann die Effizienz der IFET

Reaktion verringern.

Die Bildung von n-Heptansulfonsäure wurde auch inhibiert, wenn die Suspension

geringe Mengen an Wasser (0.3vol%) enthielt. Dies könnte auf Blockierung der n-

Heptan adsorptionszentren durch bevorzugte Wasseradsorption zurückzuführen sein.

0

10

20

30

40

hνhν

c(1)

/ m

M

hν

R R R

0 10 0 10 0 10

Time / h

Abbildung 6.1. Sequentielle Photosulfoxidation von n-Heptan. λirr ≥ 400 nm. R = Katalysatorregenerierung.

6. Zusammenfassung

149

Wurde die Reaktion nach 2 h Belichtungszeit gestoppt, entsprechend der Bildung

einer 15 mM Lösung von (1), und die Suspension anschließend drei Tage im Dunkeln

aufbewahrt, stieg die Konzentration an (1) auf 50 mM. War dagegen während dieser

Dunkelphase der Radikalfänger Hydrochinon anwesend, konnte keine weitere

Sulfonsäurebildung beobachtet werden.

Obige Ergebnisse deuten darauf hin, daß auch diese neuartige

Photosulfoxidation eine Radikalkettenreaktion ist. Jedoch entsteht das

Alkylstarterradikal nicht durch UV-Anregung von Schwefeldioxid, sondern durch Vis-

Anregung des TiO2/n-Heptan/SO2/O2 - Systems. Da aber nur die modifizierten

Titandioxide (Tabelle 1, Zeilen 5-7) sichtbares Licht absorbieren können, erschien es

möglich, daß die unmodifizierten Titandioxide (Tabelle 1, Zeilen 1-4) mit

Schwefeldioxid einen im Sichtbaren absorbierenden Charge-Transferkomplex bilden.

In der Tat tritt eine schwache Gelbfärbung auf, wenn P25 mit Schwefeldioxid in

Kontakt gebracht wird. Eine entsprechende CT-Bande taucht im Diffusen

Reflexionsspektrum bei 410 - 420 nm auf. Auf Grund dieser Beobachtungen wird der

in Abbildung 6.2 skizzierte Mechanismus vorgeschlagen.

Abbildung 6.2: Mechanistischer Vorschlag zur lichtinduzierten Bildung von Alkylradikalen.

[TiO2(e-)---SO2+·]

[TiO2---SO2]

O2

RH RH

R· R· + H+

Vis

6. Zusammenfassung

150

Vis-Anregung des CT-Komplexes ergibt ein Elektron im Leitungsband

[TiO2(e−)] und ein adsorbiertes Schwefeldioxidradikalkation. Das erstere reduziert

Sauerstoff zu Superoxid, das zweite kann das Alkan unter Deprotonierung zum

Alkylradikal und zu einem Proton oxidieren. Nach Protonierung von O2− durch

Oberflächenhydroxylgruppen entsteht das Hydroperoxylradikal, welches über eine H-

Abstraktion ebenfalls das Alkylradikal erzeugen kann. Die auf beiden Wegen

erhaltenen Alkylstarterradikale reagieren weiter wie bei der stoichiometrischen UV –

sulfoxidation beschrieben.

Im Einklang mit diesem mechanistischen Vorschlag ist die vollständige

Inhibierung durch 10vol% Methanol, welches viel schneller als das Alkan oxidiert

werden sollte und außerdem noch ein effizienter OH-Radikalfänger ist.

Zusammenfassend läßt sich diese durch sichtbares Licht induzierte C-H –

Aktivierung als “Halbleiterphotokatalyse Typ B” klassifizieren. Sie erweitert den

bisher als Zweikomponentenaddition definierten Reaktionstyp um eine dritte

Komponente.

Andere erwartete Neben-produkte wie Ketone, Alcohole und CO2 entstehen, wenn

überhaupt, nur in vernachlässigbaren Mengen. Bei der Cyclohexan-photosulfoxidation

entsteht nur 0.1% Cyclohexanon (auf 100% Cyclohexansulfon-Säure). Wie schon

erwähnt, bildet SO2 mit Titandioxid einen im Sichtbaren absorbierenden Charge-

Transferkomplex. Auf Grund dessen kann man erwarten, dass die SO2 Konzentration

an der Katalysator Oberfläche viel grösser ist als die von O2. Daher sollte die Addition

des Alkylradicals schneller sein als die Addition von Sauerstoff. Dies kann der Grund

für die Selektivität dieser Reaktion sein.

151

7. References

(1)Labinger, J. A.; Bercaw, J. E. Nature 2002, 417, 507.

(2)Shilov, A. E.; Shul'pin, G. B. Chem. Rev. 1997, 97, 2879.

(3)Bard, A. J.; Whitesides, G. M.; Zare, R. N.; McLafferty, F. W. Acc. Chem. Res. 1995, 28, 91.

(4)Arndtsen, B. A.; Bergman, R. G.; Mobley, T. A.; Peterson, T. H. Acc. Chem. Res. 1995, 28, 154.

(5)Gol'dshleger, N. F.; Tyabin, M. B.; Shilov, A. E.; Shteinman, A. A. Zh. Fiz. Khim. 1969, 43, 2174.

(6)Gol'dshleger, N. F.; Es'kova, V. V.; Shilov, A. E.; Shteinman, A. A. Zh. Fiz. Khim. 1972, 46, 1353.

(7)Rudakov, E. S.; Zamashchikov, V. V.; Belyaeva, N. P.; Rudakova, R. I. Zh. Fiz. Khim. 1973, 47, 2732.

(8)Tret'yakov, V. P.; Arzamaskova, L. N.; Ermakov, Y. I. Kinet. Katal. 1974, 15, 538.

(9)Cooper, T. A.; Waters, W. A. Journal of the Chemical Society [Section] B: Physical Organic 1967, 687.

(10) Garnett, J. L.; Long, M. A.; Peterson, K. B. Aust. J. Chem. 1974, 27, 1823.

(11) Grigoryan, E. A.; D'Yachkovskii, F. S.; Mullagaliev, I. R. Dokl. Akad. Nauk SSSR 1975, 224, 859.

(12) Bergman, R. G. Nature 2007, 446, 391.

(13) Chen, H.; Schlecht, S.; Semple, T. C.; Hartwig, J. F. Science 2000, 287, 1995.

(14) Crabtree, R. H. Journal of the Chemical Society, Dalton Transactions 2001, 2437.

(15) Crabtree, R. H. Chem. Rev. 1995, 95, 987.

(16) Shilov, A. E.; Shteinman, A. A. Acc. Chem. Res. 1999, 32, 763.

(17) Jones, W. D.; Feher, F. J. J. Am. Chem. Soc. 1984, 106, 1650.

(18) Fokin, A. A.; Schreiner, P. R. Adv. Synth. Catal. 2003, 345, 1035.

(19) Fokin, A. A.; Schreiner, P. R. Chem. Rev. 2002, 102, 1551.

(20) Mayer, J. M. Acc. Chem. Res. 1998, 31, 441.

(21) Adam, W.; Curci, R.; Edwards, J. O. Acc. Chem. Res. 1989, 22, 205.

(22) Rajeshwar, K. Encyclopedia of Electrochemistry 2003, 6, 1.

(23) Tan, M. X.; Laibinis, P. E.; Nguyen, S. T.; Kesselman, J. M.; Stanton, C. E.; Lewis, N. S. Prog. Inorg. Chem. 1994, 41, 21.

(24) Cox, P. A. The Electronic Structure and Chemistry of Solids, 1987.

(25) Finklea, H. O.; Editor Studies in Physical and Theoretical Chemistry, 55: Semiconductor Electrodes, 1988.

(26) Pankove, J. I. Optical Processes in Semiconductors, 1971.

(27) Kasap, S. O. Principles of Electronic Materials and Devices, 3rd ed.; McGraw-Hill New York, 2006.

152

(28) Dung, D.; Ramsden, J.; Graetzel, M. J. Am. Chem. Soc. 1982, 104, 2977.

(29) Asahi, R.; Morikawa, T.; Ohwaki, T.; Aoki, K.; Taga, Y. Science 2001, 293, 269.

(30) Mardare, D.; Tasca, M.; Delibas, M.; Rusu, G. I. Appl. Surf. Sci. 2000, 156, 200.

(31) Tanaka, K.; Capule, M. F. V.; Hisanaga, T. Chem. Phys. Lett. 1991, 187, 73.

(32) Maruska, H. P.; Ghosh, A. K. Solar Energy 1978, 20, 443.

(33) Gerischer, H.; Heller, A. J. Electrochem. Soc. 1992, 139, 113.

(34) Bickley, R. I.; Gonzalez-Carreno, T.; Lees, J. S.; Palmisano, L.; Tilley, R. J. D. J. Solid State Chem. 1991, 92, 178.

(35) Hurum, D. C.; Agrios, A. G.; Gray, K. A.; Rajh, T.; Thurnauer, M. C. J. Phys. Chem. B 2003, 107, 4545.

(36) Beranek, R. PhD thesis, Friedrich - Alexander Universität, 2007.

(37) Butler, M. A.; Ginley, D. S. J. Electrochem. Soc. 1978, 125, 228.

(38) Morrison, S. R. Electrochemistry at Semiconductor and Oxidized Metal Electrodes; Plenum Press: New York, 1980.

(39) Paz, Y. C. R. Chim. 2006, 9, 774.

(40) Sclafani, A.; Herrmann, J. M. J. Phys. Chem. 1996, 100, 13655.

(41) Kormann, C.; Bahnemann, D. W.; Hoffmann, M. R. Environ. Sci. Technol. 1991, 25, 494.

(42) Hoffmann, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W. Chem. Rev. 1995, 95, 69.

(43) Cundall, R. B.; Rudham, R.; Salim, M. S. J. Chem. Soc., Faraday Trans. 1 1976, 72, 1642.

(44) Sawyer, D. T.; Gibian, M. J. Tetrahedron 1979, 35, 1471.

(45) Okamoto, K.; Yamamoto, Y.; Tanaka, H.; Tanaka, M.; Itaya, A. Bull. Chem. Soc. Jpn. 1985, 58, 2015.

(46) Lettmann, C.; Hildenbrand, K.; Kisch, H.; Macyk, W.; Maier, W. F. Appl. Catal., B 2001, 32, 215.

(47) Goto, H.; Hanada, Y.; Ohno, T.; Matsumura, M. J. Catal. 2004, 225, 223.

(48) Zang, L.; Macyk, W.; Lange, C.; Maier, W. F.; Antonius, C.; Meissner, D.; Kisch, H. Chem.-Eur. J. 2000, 6, 379.

(49) Anpo, M. Pure Appl. Chem. 2000, 72, 1787.

(50) Peter, L. M.; Wijayantha, K. G. U.; Riley, D. J.; Waggett, J. P. J. Phys. Chem. B 2003, 107, 8378.

(51) Qian, X.; Qin, D.; Bai, Y.; Li, T.; Tang, X.; Wang, E.; Dong, S. J. Solid State Electrochem. 2001, 5, 562.

(52) Vogel, R.; Hoyer, P.; Weller, H. J. Phys. Chem. 1994, 98, 3183.

(53) Ho, W.; Yu, J. C. J. Mol. Catal. A: Chem. 2006, 247, 268.

(54) Tristao, J. C.; Magalhaes, F.; Corio, P.; Sansiviero, M. T. C. J. Photochem. Photobiol., A 2006, 181, 152.

(55) Macyk, W.; Stochel, G.; Szacilowski, K. Chem.-Eur. J. 2007, 13, 5676.

(56) Kroeze, J. E.; Koehorst, R. B. M.; Savenije, T. J. Adv. Funct. Mater. 2004, 14, 992.

153

(57) Winnischofer, H.; Formiga, A. L. B.; Nakamura, M.; Toma, H. E.; Araki, K.; Nogueira, A. F. Photochem. Photobiol. Sci. 2005, 4, 359.

(58) Ingrosso, C.; Petrella, A.; Curri, M. L.; Striccoli, M.; Cosma, P.; Cozzoli, P. D.; Agostiano, A. Appl. Surf. Sci. 2005, 246, 367.

(59) Iliev, V. J. Photochem. Photobiol., A 2002, 151, 195.

(60) Ehret, A.; Stuhl, L.; Spitler, M. T. J. Phys. Chem. B 2001, 105, 9960.

(61) Burgeth, G.; Kisch, H. Coord. Chem. Rev. 2002, 230, 41.

(62) Kim, S.; Choi, W. J Phys Chem B FIELD Full Journal Title:The journal of physical chemistry. B 2005, 109, 5143.

(63) Gubin, S. P.; Shul'pin, G. B. Chemistry of Complexes with Metal-Carbon Bonds, 1984.

(64) Dawoodi, Z.; Green, M. L. H.; Mtetwa, V. S. B.; Prout, K. J. Chem. Soc., Chem. Commun. 1982, 802.

(65) Hall, C.; Perutz, R. N. Chem. Rev. 1996, 96, 3125.

(66) Halpern, J. Inorg. Chim. Acta 1985, 100, 41.

(67) Janowicz, A. H.; Bergman, R. G. J. Am. Chem. Soc. 1982, 104, 352.

(68) Ghosh, C. K.; Rodgers, D. P. S.; Graham, W. A. G. J. Chem. Soc., Chem. Commun. 1988, 1511.

(69) Ghosh, C. K.; Graham, W. A. G. J. Am. Chem. Soc. 1989, 111, 375.

(70) Michos, D.; Luo, X. L.; Crabtree, R. H. Inorg. Chem. 1993, 32, 2118.

(71) Sakakura, T.; Sodeyama, T.; Tokunaga, Y.; Tanaka, M. Chem. Lett. 1988, 263.

(72) Maguire, J. A.; Boese, W. T.; Goldman, A. S. J. Am. Chem. Soc. 1989, 111, 7088.

(73) Spillett, C. T.; Ford, P. C. J. Am. Chem. Soc. 1989, 111, 1932.

(74) Sakakura, T.; Abe, F.; Tanaka, M. Chem. Lett. 1991, 297.

(75) Itagaki, H.; Einaga, H.; Saito, Y. Journal of the Chemical Society, Dalton Transactions: Inorganic Chemistry (1972-1999) 1993, 1689.

(76) Sakakura, T.; Sodeyama, T.; Sasaki, K.; Wada, K.; Tanaka, M. J. Am. Chem. Soc. 1990, 112, 7221.

(77) Khusnutdinov, R. I.; Shchadneva, N. A.; Dzhemilev, U. M. Izvestiya Akademii Nauk SSSR, Seriya Khimicheskaya 1991, 2897.

(78) Horvath, I. T.; Cook, R. A.; Millar, J. M.; Kiss, G. Organometallics 1993, 12, 8.

(79) Stahl, S.; Labinger, J. A.; Bercaw, J. E. Angew. Chem., Int. Ed. 1998, 37, 2181.

(80) Luinstra, G. A.; Wang, L.; Stahl, S. S.; Labinger, J. A.; Bercaw, J. E. J. Organomet. Chem. 1995, 504, 75.

(81) Periana, R. A.; Taube, D. J.; Gamble, S.; Taube, H.; Satoh, T.; Fujii, H. Science 1998, 280, 560.

(82) Watson, P. L. J. Am. Chem. Soc. 1983, 105, 6491.

(83) Muzart, J. Chem. Rev. 1992, 92, 113.

(84) Shul'pin, G. B.; Druzhinina, A. N.; Nizova, G. V. Izvestiya Akademii Nauk SSSR, Seriya Khimicheskaya 1990, 212.

(85) Shul'pin, G. B.; Kats, M. M. React. Kinet. Catal. Lett. 1990, 41, 239.

154

(86) Bakke, J. M.; Braenden, J. E. Acta Chem. Scand. 1991, 45, 418.

(87) Tenaglia, A.; Terranova, E.; Waegell, B. Tetrahedron Lett. 1989, 30, 5271.

(88) Denisov, E. T.; Khudyakov, I. V. Chem. Rev. 1987, 87, 1313.

(89) Hsu, Y. F.; Yen, M. H.; Cheng, C. P. J. Mol. Catal. A: Chem. 1996, 105, 137.

(90) Felzenstein, A.; Goosen, A.; Marsh, C.; McCleland, C. W.; Van Sandwyk, K. S. S. Afr. J. Chem. 1989, 42, 143.

(91) Dugmore, G. M.; Powels, G. J.; Zeelie, B. J. Mol. Catal. A: Chem. 1995, 99, 1.

(92) Goldstein, A. S.; Drago, R. S. Inorg. Chem. 1991, 30, 4506.

(93) Ellis, P. E., Jr.; Lyons, J. E. Chromium-haloporphyrin complexes for the oxidation of butane to methyl ethyl ketone; (Sun Refining and Marketing Co., USA). Application: US, 1990; pp 4 pp.

(94) Barton, D. H. R.; Taylor, D. K. Stereoselective Reactions of Metal-Activated Molecules, Proceedings of the Symposium, 2nd, Wuerzburg, Germany, Sept. 21-23, 1994 1995, 3.

(95) Attanasio, D.; Suber, L.; Shul'pin, G. B. Izvestiya Akademi Nauk, Seriya Khimicheskaya 1992, 1918.

(96) Chambers, R. C.; Hill, C. L. Inorg. Chem. 1989, 28, 2509.

(97) Shul'pin, G. B.; Kats, M. M. Zhurnal Obshchei Khimii 1989, 59, 2738.

(98) Shul'pin, G. B.; Druzhinina, A. N. Izvestiya Akademii Nauk SSSR, Seriya Khimicheskaya 1989, 1185.

(99) Shul'pin, G. B.; Druzhinina, A. N.; Nizova, G. V. Izvestiya Akademii Nauk SSSR, Seriya Khimicheskaya 1991, 2463.

(100)Nizova, G. V.; Shul'pin, G. B. Zhurnal Obshchei Khimii 1990, 60, 2124.

(101)Shul'pin, G. B.; Nizova, G. V. React. Kinet. Catal. Lett. 1991, 45, 7.

(102)Shul'pin, G. B.; Nizova, G. V.; Kats, M. M. Zhurnal Obshchei Khimii 1990, 60, 2730.

(103)Carp, O.; Huisman, C. L.; Reller, A. Prog. Solid State Chem. 2004, 32, 33.

(104)Pichat, P. Catal. Today 1994, 19, 313.

(105)Parker, V. D. J. Am. Chem. Soc. 1992, 114, 7458.

(106)Boarini, P.; Carassiti, V.; Maldotti, A.; Amadelli, R. Langmuir 1998, 14, 2080.

(107)Almquist, C. B.; Biswas, P. Appl. Catal., A 2001, 214, 259.

(108)Cermenati, L.; Dondi, D.; Fagnoni, M.; Albini, A. Tetrahedron 2003, 59, 6409.

(109)Ohno, T.; Mitsui, T.; Matsumura, M. J. Photochem. Photobiol., A 2003, 160, 3.

(110)Macyk, W.; Kisch, H. Chem.-Eur. J. 2001, 7, 1862.

(111)Wang, C. M.; Mallouk, T. E. J. Phys. Chem. 1990, 94, 4276.

(112)Cunningham, J.; Al-Sayyed, G. J. Chem. Soc., Faraday Trans. 1990, 86, 3935.

(113)Theurich, J.; Lindner, M.; Bahnemann, D. W. Langmuir 1996, 12, 6368.

(114)Kisch, H.; Macyk, W. ChemPhysChem 2002, 3, 399.

(115)Kisch, H.; Zang, L.; Lange, C.; Maier, W. F.; Antonius, C.; Meissner, D. Angew. Chem., Int. Ed. 1998, 37, 3034.

(116)Zang, L.; Lange, C.; Abraham, I.; Storck, S.; Maier, W. F.; Kisch, H. J. Phys. Chem. B 1998, 102, 10765.

155

(117)Macyk, W.; Burgeth, G.; Kisch, H. Photochem. Photobiol. Sci. 2003, 2, 322.

(118)Cundari, T. R.; Moody, E. W. THEOCHEM 1998, 425, 43.

(119)Kortuem, G.; Braun, W.; Herzog, G. Angew. Chem. 1963, 75, 653.

(120)Edreva- Kardjieva, R. Bulg. Chem. Commun. 1992, 25, 166.

(121)Tauc, J.; Grigorovici, R.; Vancu, A. Physica Status Solidi 1966, 15, 627.

(122)Alimarin, I. P.; Shlenskaya, V. I.; Efremenko, O. A. Zh. Neorg. Khim. 1970, 15, 1040.

(123)Szacilowski, K.; Macyk, W.; Stochel, G. J. Mater. Chem. 2006, 16, 4603.

(124)Nozik, A. J.; Memming, R. J. Phys. Chem. 1996, 100, 13061.

(125)Gerischer, H. Z. Phys. Chem. 1960, 26, 223.

(126)Gerischer, H. Electrochim. Acta 1990, 35, 1677.

(127)Memming, R.; Moellers, F. Ber. Buns. Ges. Phys. Chem. 1972, 76, 475.

(128)Ward, M. D.; White, J. R.; Bard, A. J. J. Am. Chem. Soc. 1983, 105, 27.

(129)White, J. R.; Bard, A. J. J. Phys. Chem. 1985, 89, 1947.

(130)Finlayson, M. F.; Wheeler, B. L.; Kakuta, N.; Park, K. H.; Bard, A. J.; Campion, A.; Fox, M. A.; Webber, S. E.; White, J. M. J. Phys. Chem. 1985, 89, 5676.

(131)Roy, A. M.; De, G. C.; Sasmal, N.; Bhattacharyya, S. S. Int. J. Hydrogen Energy 1995, 20, 627.

(132)Bolts, J. M.; Wrighton, M. S. J. Phys. Chem. 1976, 80, 2641.

(133)Sakthivel, S.; Kisch, H. Angew. Chem., Int. Ed. 2003, 42, 4908.

(134)Sakthivel, S.; Kisch, H. ChemPhysChem 2003, 4, 487.

(135)Choi, W.; Termin, A.; Hoffmann, M. R. J. Phys. Chem. 1994, 98, 13669.

(136)Carraway, E. R.; Hoffman, A. J.; Hoffmann, M. R. Environ. Sci. Technol. 1994, 28, 786.

(137)Hoffman, A. J.; Carraway, E. R.; Hoffmann, M. R. Environ. Sci. Technol. 1994, 28, 776.

(138)Kormann, C.; Bahnemann, D. W.; Hoffmann, M. R. Environ. Sci. Technol. 1988, 22, 798.

(139)Platz, C.; Schimmelschmidt, K. DRP 735 096. In (I.G. Farben Industrie), 1940.

(140)Ullmann's Encyclopedia of industrial chemistry Wiley-VCH: Weinheim; Vol. A 25,772.

(141)Orthner, L. Angew. Chem. 1950, 62A, 302.

(142)Asinger, F. Paraffins, Chemistry and Technology, 1968.

(143)Ogata, Y.; Izawa, Y.; Tsuda, T. Tetrahedron 1965, 21, 1349.

(144)Ramloch, H.; Taeuber, G. Chemie in unserer Zeit 1979, 5, 157.

(145)Graf, R. Justus Liebigs Ann. Chem. 1952, 578, 50.

(146)Braun, M.; Maurette, M. T.; Oliveros, E. Photochemical Technology 1991, 354.

(147)Smith, G. W.; Williams, H. D. J. Org. Chem. 1961, 26, 2207.

(148)Sidebottom, H. W.; Badcock, C. C.; Calvert, J. C.; Rabe, B. R.; Damon, E. K. J. Am. Chem. Soc. 1971, 93, 3121.

(149)Asinger, F.; Geiseler, F.; Eckoldt, H. Chem. Ber. 1956, 89, 1037.

156

(150)Asinger, F.; Fell, B.; Pottkaemper, S. Chem. Ber. 1964, 97, 3092.

(151)Schimmelschmidt, K.; Cramer, G.; Graf, R. DRP.-Anm. J 68893. In Farbenindustrie A.G., 1941.

(152)Orthner, L.; Wagner, H.; Gruschke, H. DRP.-Anm. J 69611. In I. G. Farbenindustrie A.G., 1941.

(153)Weghofer, H. Fette und Seifen 1952, 54, 260.

(154)Roesinger, S. Chem. Ing. Tech. 1970, 42, 1236.

(155)Ferguson, R. R.; Crabtree, R. H. J. Org. Chem. 1991, 56, 5503.

(156)Burgeth, G. PhD thesis, Friedrich - Alexander Universität, 2002.

(157)Ramakrishnan, A. phD thesis, Friedrich - Alexander Universität, 2006.

(158)Sakthivel, S.; Janczarek, M.; Kisch, H. J. Phys. Chem. B 2004, 108, 19384.

(159)Dai, Z.; Burgeth, G.; Parrino, F.; Kisch, H. Journal of Organometallic Chemistry, in press 2008.

(160)Larson, J. R. J. Chromatogr. 1986, 356, 379.

(161)Small, H.; Miller, T. E., Jr. Anal. Chem. 1982, 54, 462.

(162)Shang, J.; Zhu, Y.; Du, Y.; Xu, Z. J. Solid State Chem. 2002, 166, 395.

(163)Nelson, B. P.; Candal, R.; Corn, R. M.; Anderson, M. A. Langmuir 2000, 16, 6094.

(164)Abdullah, M.; Low, G. K. C.; Matthews, R. W. J. Phys. Chem. 1990, 94, 6820.

(165)Janczyk, A.; Krakowska, E.; Stochel, G.; Macyk, W. J. Am. Chem. Soc. 2006, 128, 15574.

(166)Mrowetz, M.; Selli, E. Phys. Chem. Chem. Phys. 2005, 7, 1100.

(167)Minero, C.; Mariella, G.; Maurino, V.; Vione, D.; Pelizzetti, E. Langmuir 2000, 16, 8964.

(168)Minero, C.; Mariella, G.; Maurino, V.; Pelizzetti, E. Langmuir 2000, 16, 2632.

(169)Chen, Y.; Yi, J.; Li, W.; Jin, R.; Tang, S.; Hu, W. Catal. Today 1999, 50, 39.

(170)Thornton, G. Vacuum 1992, 43, 1133.

(171)Sayago, D. I.; Serrano, P.; Bohme, O.; Goldoni, A.; Paolucci, G.; Roman, E.; Martin-Gago, J. A. Phys. Rev. B: Condens. Matter Mater. Phys. 2001, 64, 205402/1.

(172)Jing, L.; Xin, B.; Yuan, F.; Wang, B.; Shi, K.; Cai, W.; Fu, H. Appl. Catal., A 2004, 275, 49.

(173)Yanagisawa, Y. Appl. Surf. Sci. 1997, 115, 377.

(174)Casarin, M.; Ferrigato, F.; Maccato, C.; Vittadini, A. J Phys Chem B FIELD Full Journal Title:The journal of physical chemistry. B 2005, 109, 12596.

(175)Chang, C. C. J. Catal. 1978, 53, 374.

(176)Thierry, D.; Robert, V. GB 2112597A, 1981.

(177)Sayago, D. I.; Serrano, P.; Bohme, O.; Goldoni, A.; Paolucci, G.; Roman, E.; Martin-Gago, J. A. Surf. Sci. 2001, 482-485, 9.

(178)Schiller, M.; Muller, F. W.; Damm, C. J. Photochem. Photobiol., A 2002, 149, 227.

(179)Karl, N. Organic semiconductors, 1974; Vol. 14.

(180)Docherty, K. S.; Ziemann, P. J. Aerosol Sci. Technol. 2003, 37, 877.

(181)Asmus, K. D.; Fendler, J. H. J. Phys. Chem. 1968, 72, 4285.

157

(182)Asmus, K. D.; Gruenbein, W.; Fendler, J. H. J. Am. Chem. Soc. 1970, 92, 2625.

(183)Thavasi, V.; Leong, L. P.; Bettens, R. P. A. J. Phys. Chem. A 2006, 110, 4918.

(184)Ishii, Y.; Matsunaka, K.; Sakaguchi, S. J. Am. Chem. Soc. 2000, 122, 7390.

(185)Wang, Y.; Hang, K.; Anderson, N. A.; Lian, T. J. Phys. Chem. B 2003, 107, 9434.

(186)Rajh, T.; Chen, L. X.; Lukas, K.; Liu, T.; Thurnauer, M. C.; Tiede, D. M. J. Phys. Chem. B 2002, 106, 10543.

(187)Kamat, P. V. Langmuir 1985, 1, 608.

(188)Takahara, Y. K.; Hanada, Y.; Ohno, T.; Ushiroda, S.; Ikeda, S.; Matsumura, M. J. Appl. Electrochem. 2005, 35, 793.

(189)Good, A.; Thynne, J. C. J. Trans. Farad. Soc. 1967, 63, 2720.

(190)Muggli, D. S.; Falconer, J. L. J. Catal. 1999, 187, 230.

(191)Bahnemann, D. W.; Kholuiskaya, S. N.; Dillert, R.; Kulak, A. I.; Kokorin, A. I. Appl. Catal., B 2002, 36, 161.

(192)Wang, C.-y.; Groenzin, H.; Shultz, M. J. J. Am. Chem. Soc. 2005, 127, 9736.

(193)Arana, J.; Rodriguez, J. M. D.; Diaz, O. G.; Melian, J. A. H.; Rodriguez, C. F.; Pena, J. P. Appl. Catal., A 2006, 299, 274.

(194)Backes, M. J.; Lukaski, A. C.; Muggli, D. S. Appl. Catal., B 2005, 61, 21.

(195)Yamazaki, S.; Tanaka, S.; Tsukamoto, H. J. Photochem. Photobiol., A 1999, 121, 55.

(196)Einaga, H.; Ibusuki, T.; Futamura, S. Environ. Sci. Technol. 2004, 38, 285.

(197)Martra, G.; Coluccia, S.; Marchese, L.; Augugliaro, V.; Loddo, V.; Palmisano, L.; Schiavello, M. Catal. Today 1999, 53, 695.

(198)Fu, X.; Zeltner, W. A.; Anderson, M. A. Appl. Catal., B 1995, 6, 209.

(199)Jo, M. Journal of Surface Analysis 1999, 5, 106.

(200)Lindstrom, H.; Magnusson, E.; Holmberg, A.; Sodergren, S.; Lindquist, S.-E.; Hagfeldt, A. Sol. Energy Mater. Sol. Cells 2002, 73, 91.

C-H Activation through Photocatalytic Sulfoxidation of … · C-H Activation through Photocatalytic...

Documents

Transcript of C-H Activation through Photocatalytic Sulfoxidation of … · C-H Activation through Photocatalytic...