19._tesi_vincenzo_vaiano

HETEROGENEOUS

PHOTOCATALYTIC

SELECTIVE OXIDATION OF

CYCLOHEXANE

Vincenzo Vaiano

Unione Europea UNIVERSITÀ DEGLI STUDI DI SALERNO

Fondo sociale europeo Programma Operativo Nazionale 2000/2006

“Ricerca Scientifica, Sviluppo Tecnologico, Alta Formazione”

Regioni dell’Obiettivo 1 – Misura III.4

“Formazione superiore ed universitaria”

Department of Chemical and Food Engineering

Ph.D. Course in Chemical Engineering

(IV Cycle-New Series)

HETEROGENEOUS PHOTOCATALYTIC

SELECTIVE OXIDATION OF CYCLOHEXANE

Supervisor Ph.D. student Prof. Paolo Ciambelli Vincenzo Vaiano

Scientific Referees

Prof. Roger I. Bickley

Prof. Leonardo Palmisano

Prof. Salvatore Vaccaro

Ph.D. Course Coordinator

Prof. Ernesto Reverchon

To my grandfather

Acknowledgments

Con questa attività di ricerca ho potuto fare un’importante esperienza di vita arricchita e impreziosita dal rapporto quotidiano con persone che mi hanno aiutato moltissimo a superare le difficoltà che incontravo.

Sono tante le persone a cui devo il raggiungimento di questo traguardo, persone che nei modi più disparati mi hanno aiutato, spronato, consigliato...e senza dilungarmi troppo nel cercare le parole migliori per ringraziarle, lo faccio.

Un ringraziamento importante va al Prof. Paolo Ciambelli per il sostegno

dimostratomi fin dall’inizio e per l’attenzione con la quale ha seguito e reso possibile la realizzazione di questo lavoro.

Un ringraziamento particolare va al Prof. Vincenzo Palma e alla Dott.

Diana Sannino, mai paghi nel profondermi approvazione e indispensabili consigli, a cui devo il raggiungimento di questa importante meta.

Voglio inoltre ringraziare il Prof. Salvatore Vaccaro e il Prof. Leonardo

Palmisano per la loro utile collaborazione.

Ringrazio gli Ingegneri Massimo Ricciardi, Antonietta Maria Manna, Emma Palo, Arianna Ruggiero, Maria Sarno, Giuseppa Matarazzo, Paola Russo e Caterina Leone che mi hanno fatto capire quanto sia importante collaborare in piena serenità, instaurando un rapporto di amicizia.

Vorrei inoltre ringraziare il Prof. Roger I. Bickley per l’ospitalità

offertami presso l’Università di Bradford e per i suoi preziosi suggerimenti. Un ringraziamento va al Prof. Eric M. Gaigneaux e al Dott. Frédéric Dury

che hanno realizzato le misure XPS presenti nel lavoro. Ma ringrazio soprattutto mia madre, mio padre, mia sorella Maria Rita e

mia zia Betta per avermi sostenuto in ogni occasione.

Publications list

1. P. Ciambelli, D. Sannino, V. Palma, V. Vaiano; Photocatalytic

selective oxidation of cyclohexane on Mo-exchanged ferrierite. Proc. of the 6nd National Congress on Science and Technology of Zeolites, September 20-23, 2003, Vietri sul Mare, Italy, p.125-126.

2. P. Ciambelli, D. Sannino, V. Palma, V. Vaiano; Gas-solid

photocatalytic oxidative dehydrogenation of cyclohexane to benzene

on MoOx/TiO2. Proc.of the 14nd National Congress on Catalysis, June 6-10, 2004, Lerici, Italy, p.24-25.

3. P. Ciambelli, D. Sannino, V. Palma, V. Vaiano; Heterogeneous

photocatalytic oxidative dehydrogenation of cyclohexane over

supported Mo oxide catalysts. Proc.of the 13nd International Congress on Catalysis, July 11-16, 2004, Paris, France, p.191.

4. P. Ciambelli, D. Sannino, V. Vaiano, V. Palma, S. Vaccaro;

Catalysed photooxidative dehydrogenation of cyclohexane to

benzene in a fluidized bed reactor. Proc.of the 7nd World Congress of Chemical Engineering, July, 10-14, 2005, Glasgow, Scotland, http://chemengcongress.somcom.co.uk/.

5. P. Ciambelli, D. Sannino, V. Palma, V. Vaiano; Cyclohexane

photocatalytic oxidative dehydrogenation to benzene on sulphated

titania supported MoOx; Stud. Surf. Sci. Catal. 155 (2005) 179-187.

6. P. Ciambelli, D. Sannino, V. Palma, V. Vaiano; Photocatalysed

selective oxidation of cyclohexane to benzene on MoOx/TiO2; Catal. Today. 99 (2005) 143-149.

7. P. Ciambelli, V. Palma, D. Sannino, S. Vaccaro, V. Vaiano; Selective oxidation of cyclohexane to benzene on molybdena-titania

catalysts in fluidized bed photocatalytic reactor. THE TOCAT 5 Conference on Advanced Catalytic Science and Technology, July, 23-28, 2006, Tokyo, Japan, submitted.

I

Contents

I Introduction................................................................................. 1

I.1 Heterogeneous photocatalysis.............................................. 2

I.1.1 Basic Principles........................................................... 3

I.2 Fundamental Engineering Aspects....................................... 7

I.2.1 Types of Radiation Sources ......................................... 8

I.2.2 Photoreactor Materials and Geometries..................... 9

I.2.3 Efficiency of Heterogeneous Photocatalytic Systems 10

I.2.4 Modeling of Photocatalytic Reactors ........................ 11

I.3 Aim of the Work ................................................................ 12

II State of the art....................................................................... 15

II.1 Photocatalysis as potential technology for synthesis of

chemicals from hydrocarbon feedstocks.................................................. 15

II.1.1 Photocatalytic oxidation of cyclohexane................... 16

II.1.2 Oxidative dehydrogenation of cyclohexane............... 17

II.2 Influence of catalyst surface acidity on photocatalytic

performances ............................................................................................ 18

II.3 Photocatalytic fluidized bed reactors ................................. 19

III Experimental Results: Photocatalytic oxidation of

cyclohexane on MoOx/Al2O3 ...................................................................... 23

III.1 Samples preparation....................................................... 23

III.2 Catalysts characterization .............................................. 23

II

III.2.1 ICP-MS...................................................................... 24

III.2.2 Thermal analysis (TG-MS) ........................................ 24

III.2.3 Micro Raman spectroscopy ....................................... 27

III.2.4 FT-IR Spectroscopy ................................................... 28

III.2.5 N2 adsorption measurements| .................................... 28

III.2.6 TPD investigations .................................................... 29

III.3 Laboratory apparatus for catalytic test........................... 29

III.3.1 Feed section............................................................... 30

III.3.2 Reaction section......................................................... 31

III.3.3 Analysis section ......................................................... 32

III.4 Photocatalytic tests conditions and typical trend ........... 33

III.4.1 Thermodynamic analysis ........................................... 36

III.5 Results and discussion ................................................... 37

III.5.1 Specific surface area and Chemical analysis ............ 37

III.5.2 Thermal analysis........................................................ 38

III.5.3 FT-IR spectroscopy.................................................... 40

III.5.4 Micro Raman spectroscopy ....................................... 42

III.5.5 Photocatalytic activity tests ....................................... 44

IV Experimental Results: Photocatalytic oxidation of

cyclohexane on zeolites supported MoOx ................................................. 47

IV.1 Samples preparation....................................................... 47

IV.2 Thermal analysis ............................................................ 48

IV.3 N2 adsorption measurements.......................................... 51

IV.4 FT-IR spectroscopy........................................................ 52

IV.5 Photocatalytic activity tests............................................ 54

V Experimental Results: Photocatalytic oxidation of

cyclohexane on MoOx/TiO2........................................................................ 59

V.1 Effect of molybdenum loading........................................... 59

V.1.1 Samples preparation.................................................. 59

V.1.2 Specific surface area and Chemical analysis ............ 59

III

V.1.3 Thermal analysis ....................................................... 61

V.1.4 FT-IR spectroscopy ................................................... 64

V.1.5 Micro Raman spectroscopy ....................................... 66

V.1.6 Photocatalytic activity tests....................................... 67

V.1.7 Effect of light intensity............................................... 73

V.2 Influence of sulphate content ............................................. 74

V.2.1 Samples preparation.................................................. 74

V.2.2 Specific surface area, chemical analysis and sample

acidity ................................................................................... 75

V.2.3 Thermal analysis ....................................................... 76

V.2.4 FT-IR spectroscopy ................................................... 77

V.2.5 Micro Raman spectroscopy ....................................... 79

V.2.6 Photocatalytic activity tests....................................... 79

V.3 Photocatalytic oxidative dehydrogenation of cyclohexane:

reaction mechanism ................................................................................. 84

V.3.1 Photocatalytic oxidation of cyclohexanol ................. 84

V.3.2 Photocatalytic oxidation of cyclohexene ................... 86

V.3.3 Role of the sulphate in the reaction mechanism........ 90

VI Photocatalytic flat-plate reactor.......................................... 93

VI.1 Experimental set up apparatus and photocatalytic tests

conditions ....................................................................................... 93

VI.2 Photocatalytic activity tests ........................................... 94

VII Photocatalytic fluidized bed reactor ................................... 99

VII.1 Photocatalytic fluidized bed reactor design ................... 99

VII.2 Preliminary results ....................................................... 102

VII.3 MoOx supported on TiO2/Al2O3 sample ...................... 106

VII.3.1 TiO2-Al2O3 sample preparation .............................. 107

VII.3.2 Results..................................................................... 108

VII.4 Photocatalytic fixed and fluidized bed reactor

comparison ..................................................................................... 109

IV

VII.5 Effect of light intensity ................................................ 119

VII.6 Photocatalytic oxidation of cyclohexane on sulphated

MoOx/γ-Al2O3 catalysts.......................................................................... 122

VII.6.1 Samples preparation ............................................... 122

VII.6.2 Specific surface area and Thermal analysis ........... 122

VII.6.3 Photocatalytic activity tests .................................... 124

VII.7 Photocatalytic oxidative dehydrogenation of ethylbenzene

to styrene ..................................................................................... 127

VII.7.1 Experimental results ............................................... 128

VIII Conclusions ..................................................................... 131

IX References............................................................................ 133

V

Index of figures

Figure 1 Schematic representation of a photocatalytic process..................... 2

Figure 2 Change in the electronic structure of a semiconductor compound as

the number N of monomeric units increases from unity to clusters of

more than 2000...................................................................................... 4

Figure 3 Illustration of the main processes occurring on a semiconductor

particle under electronic excitation....................................................... 5

Figure 4 Emission spectrum of a 40W actinic lamp ....................................... 9

Figure 5 Schematic representation of the modeling of a photocatalytic

reactor ................................................................................................. 11

Figure 6 Schematic representation of light scattering in (A) film fixed bed,

(B) granular fixed bed and (C) fluidized bed ...................................... 20

Figure 7 TGAQ500 Thermogravimetric Analyzer........................................ 25

Figure 8 SDTQ600 Simultaneous DSC/TGA................................................ 25

Figure 9 Pfieffer Vacuum Benchtop Thermostar mass spectrometer ........... 25

Figure 10 TG and DTG curves ..................................................................... 26

Figure 11 Dispersive MicroRaman (Invia, Renishaw) ................................. 27

Figure 12 Bruker IFS 66 FT-IR spectrophotometer ..................................... 28

Figure 13 Costech Sorptometer 1040........................................................... 29

Figure 14 Laboratory apparatus for catalytic test ....................................... 30

Figure 15 Mass flow controller .................................................................... 31

Figure 16 Annular gas-solid photocatalytic fixed bed reactor..................... 32

VI

Figure 17 Outlet reactor concentration (a.u.) of cyclohexane, oxygen

benzene and cyclohexene and (ppm) of carbon dioxide as a function of

run time................................................................................................ 34

Figure 18 Photocatalysed cyclohexane oxy-dehydrogenation to cyclohexene

and benzene ......................................................................................... 35

Figure 19 Grafic interface of Gaseq 0.74 program...................................... 36

Figure 20 Effect of temperature on cyclohexane conversion, carbon dioxide

and benzene selectivity ........................................................................ 37

Figure 21 Evolution of the weight loss during decomposition of ammonium

heptamolybdate together with DTG and DSC signals......................... 39

Figure 22 Thermogravimetric analysis of 8MoγAl catalyst after calcination

............................................................................................................. 40

Figure 23 IR spectrum of 8MoγAl obtained by subtracting the spectrum of

γAl ........................................................................................................ 41

Figure 24 IR spectra of αAl and 2MoαAl .................................................... 42

Figure 25 Raman spectra of γAl and 8MoγAl............................................... 43

Figure 26 Raman spectra of αAl and 2MoαAl ............................................. 44

Figure 27 Cyclohexane conversion on 8MoγAl (8Mo) and 2MoαAl (2Mo) as

a function of illumination time............................................................. 45

Figure 28 Selectivity to benzene cyclohexene and CO2 on 8MoγAl (8Mo) and

2MoαAl (2Mo) as a function of illumination time............................... 46

Figure 29 Ferrierite structure ...................................................................... 47

Figure 30 Thermogravimetric analysis of AFer and HFer........................... 49

Figure 31 Thermogravimetric analysis of NaY ............................................ 50

Figure 32 Thermogravimetric analysis of HY .............................................. 51

Figure 33 FT-IR spectra of ΗFer 5MoAFer and 20MoΑFer ....................... 52

Figure 34 FT-IR spectra of NaY and 20MoNaY ........................................... 53

Figure 35 FT-IR spectra of HY and 20MoHY .............................................. 54

VII

Figure 36 Outlet reactor concentration of cyclohexane and carbon dioxide

as a function of run time ...................................................................... 55

Figure 37 Outlet reactor concentration (a.u.) of benzene and cyclohexene on

5MoAFer as a function of run time ..................................................... 56

Figure 38 Cyclohexane conversion on 5MoAFer and 20MoAFer as a

function of illumination time ............................................................... 56

Figure 39 Benzene selectivity on 5MoAFer and 20MoAFer as a function of

illumination time.................................................................................. 57

Figure 40 Cyclohexene selectivity on 5MoAFer and 20MoAFer as a function

of illumination time.............................................................................. 57

Figure 41 Variation of specific surface area with MoO3 loading ................ 60

Figure 42 TG-MS results on DT2 sample..................................................... 61

Figure 43 TG-MS results on 2MoDT2 sample ............................................. 62



Figure 46 FT-IR spectra of DT2, 2MoDT2, 4MoDT2 and 8MoDT2 ........... 65

Figure 47 Raman spectrum of DT2 sample.................................................. 66

Figure 48 Raman spectra of DT2, 2MoDT2, 4MoDT2, 8MoDT2 and MoO3

samples ................................................................................................ 67

Figure 49 Cyclohexane conversion on DT2 as a function of illumination

time ...................................................................................................... 68

Figure 50 Carbon dioxide concentration on DT2 as a function of


Figure 51 Cyclohexane conversion on MoDT2s as a function of illumination

time ...................................................................................................... 70

Figure 52 Selectivity to benzene on MoDT2s as a function of illumination

time ...................................................................................................... 71

Figure 53 Selectivity to CO2 on MoDT2s as a function of illumination time72

Figure 54 Effect of light intensity on cyclohexane conversion ..................... 74

Figure 55 FT-IR spectra of MoT0, MoT05, and MoT20 .............................. 78

VIII

Figure 56 Raman spectra of MoT0, MoT05 and MoT20 samples ................ 79

Figure 57 Cyclohexane conversion on T0, T05 and T20 as a function of


Figure 58 Carbon dioxide concentration on T0, T05 and T20 as a function of


Figure 59 Mechanism for oxidation of cyclohexane on titania .................... 81

Figure 60 Cyclohexane conversion on Mo/T0, Mo/T05 and Mo/T20 catalysts

as a function of illumination time ........................................................ 82

Figure 61 Benzene selectivity versus cyclohexane conversion on Mo/T0,

Mo/T05 and Mo/T20 ............................................................................ 83

Figure 62 Carbon dioxide concentration on 2MoDT2, 4MoDT2 and

8MoDT2 as a function of illumination time......................................... 85

Figure 63 Cyclohexene conversion on 2MoDT2, 4MoDT2 and 8MoDT2 as a

function of illumination time................................................................ 87

Figure 64 Benzene concentration on 2MoDT2, 4MoDT2 and 8MoDT2 as a

function of illumination time................................................................ 87


8MoDT2 as a function of illumination time......................................... 88

Figure 66 Mechanism for oxidation of cyclohexene on bare titania ............ 89

Figure 67 Photocatalytic flat-plate reactor .................................................. 94

Figure 68 Cyclohexane conversion on MoDT2s catalysts as a function of


Figure 69 Selectivity to benzene on MoDT2s catalysts as a function of


Figure 70 Selectivity to CO2 on MoDT2s catalysts as a function of


Figure 71 Photocatalytic fluidized bed reactor .......................................... 100

Figure 72 Schematic picture of the cyclone................................................ 101

Figure 73 Schematic picture of the experimental apparatus ...................... 102

IX

Figure 74 Outlet reactor concentration of benzene on 4MoDT2 catalyst as

function of illumination time ............................................................. 103


benzene and cyclohexene and as a function of run time.................... 104

Figure 76 Effect of reaction temperature on cyclohexane conversion and

benzene outlet concentration. UV sources: two eye mercury lamps,

125W.................................................................................................. 106

Figure 77 Comparison between cyclohexane conversion and benzene outlet

concentration obtained on Mo8 and Mo10. Reaction temperature:

120°C. UV sources: two eye mercury lamps, 125 W......................... 108

Figure 78 Cyclohexane conversion as a function of illumination time on

8MoDT2 catalyst using the fluidized bed and the annular fixed bed

reactor. .............................................................................................. 109

Figure 79 Comparison of benzene outlet concentration on 8MoDT2 between

fixed bed and fluidized bed reactor. .................................................. 110

Figure 80 Total carbon mass balance in the fixed bed reactor as a function

of illumination time............................................................................ 112

Figure 81 TG-MS results on 4MoDT2 catalyst after photocatalytic results in

the fixed bed reactor .......................................................................... 113

Figure 82 TG-MS results on 8MoDT2 catalyst after photocatalytic results in

the fixed bed reactor .......................................................................... 113

Figure 83 Outlet reactor concentration (MS signal) of cyclohexene, and

benzene and (ppm) of carbon dioxide (NDIR analyzer) as a function of

temperarure on 4MoDT2 after activity measurements in the fixed bed

reactor. .............................................................................................. 114

Figure 84 Outlet reactor concentration (MS signal) of SO2 as a function of

temperature on 4MoDT2 after activity measurements in the fluidized

bed reactor......................................................................................... 115

X

Figure 85 Difference of Mon+

contributions found by XPS for fresh 8MoDT2

catalyst and that recovered after photocatalytic test both in the fixed

bed reactor and in fluidized bed reactor ........................................... 116

Figure 86 Detailed XPS scan spectrum of S 2p on fresh 8MoDT2............. 117

Figure 87 Detailed XPS scan spectrum of S 2p on 8MoDT2 after

photocatalytic test in the fixed bed reactor........................................ 118

Figure 88 XPS scan spectrum of S 2p on 8MoDT2 after photocatalytic test in

the fluidized bed reactor .................................................................... 119

Figure 89 Schematic picture of the modified photocatalytic fluidized bed

reactor ............................................................................................... 120

Figure 90 Comparison between cyclohexane conversion obtained by using

two and four UV sources. Reaction temperature: 120 °C. ................ 121


two and four UV sources. Reaction temperature: 120 °C. ................ 121

Figure 92 Thermogravimetric curves of 8Mo2S, 8Mo4S and 8Mo6S catalysts

........................................................................................................... 123

Figure 93 Cyclohexane conversion and cyclohexene outlet concentration as

a function of SO42-

percentage. Reaction temperature: 120 °C. UV

sources: four eye mercury lamps, 125 W........................................... 124


a function of reaction temperature .................................................... 125

Figure 95 Arrhenius plot for the photoreaction of cyclohexane on 8Mo2S 126

Figure 96 Ethylbenzene conversion and styrene outlet concentration as a

function of illumination time.............................................................. 129

XI

Index of tables

Table 1 List of catalysts with their MoO3 nominal content .......................... 23

Table 2 Specific surface area and MoO3 amount of MoOx/Al2O3 catalysts . 38

Table 3 Characteristics of Na,K-Ferrierite.................................................. 48


Table 5 Microporous volume of zeolites based samples .............................. 51

Table 6 List of catalysts with their MoO3 nominal contents ........................ 59

Table 7 List of catalysts and their characteristics ....................................... 60

Table 8 Hydroxyls and surface sulphates density ........................................ 64


Table 10 List of catalysts and their characteristics ..................................... 75

Table 11 Hydroxyls and surface sulphates density ...................................... 77

Table 12 Geometrical parameters of the cyclone ...................................... 101

Table 13 Comparison between performance of 4MoDT2 and 8MoDT2

catalysts. Reaction temperature: 70 °C. UV sources:two blacklight

blue, 160 W, PHILIPS ....................................................................... 104

Table 14 Effect of UV sources on 8MoDT2 catalyst. Reaction temperature:

100°C................................................................................................. 105

Table 15 Emission characteristics in the UVA and UVB range................. 105


catalysts in the fixed bed reactor....................................................... 110


catalysts in the fluidized bed reactor ................................................. 111

XII

Table 18 Catalysts and their characteristics.............................................. 122

Table 19 Nominal and effective sulphate content....................................... 124

Abstract

Heterogeneous photocatalytic oxidation has been extensively studied for

environmental applications such as detoxification processes by total oxidation of organics at room temperature both in water and in air. Until now, few studies have been performed in order to obtain fine chemicals.

Starting from the discovery of the occurrence of the photocatalytic oxidative cyclohexane dehydrogenation to benzene and cyclohexene nearly at ambient temperature and pressure, this work has been dedicated to improve the knowledge of this new type of photocatalytic reaction by optimising catalyst formulation, photocatalytic process and operative conditions

Some process aspects have been developed to explore its potential application in an industrial process.

For this purpose, the selective photooxidation of cyclohexane has been investigated in an annular gas-solid continuous flow reactor on molybdenum based catalysts, by changing molybdenum loading, nature and surface characteristic of the support.

Among several supports, such as zeolites, α and γ alumina and titania, highest photoactivity to dehydrogenated products was shown by MoOx/TiO2 catalysts.

In order to elucidate the effect of sulphate content on MoOx/TiO2 activity and selectivity, cyclohexane photocatalytic oxidation has been also studied on sulphated and unsulphated titania. These experimental results led to the formulation of a catalyst highly selective for benzene production, although with very low conversion of cyclohexane.

These experimental results have been confirmed employing a second type of fixed bed reactor, a plate photoreactor designed at Bradford University.

In order to improve cyclohexane conversion and to reduce the extent of the deactivation phenomena observed in fixed bed photoreactors, a photocatalytic fluidized bed reactor was designed and realized, improving the exposure of the catalysts to the light irradiation and assuring higher mass transfer rates. Since titania belongs to the group C of Geldart classification, it is not so easy to fluidize. Thus, to improve fluidization characteristics of

14

photocatalysts, new support obtained by sol-gel synthesis and physical mixture with A group solid have been employed.

Thermodynamic analysis performed for the cyclohexane oxidative dehydrogenation reaction indicated that temperature increase up to 330 °C gives increasing cyclohexane equilibrium conversion up to total conversion.

Experimental tests performed in the fluidized bed photoreactor by increasing temperature and light intensity, led to a significant increase in cyclohexane conversion of about ten times.

On the basis of the photocatalytic activity tests and physical and chemical characterization tests, a mechanism of oxidative dehydrogenation of cyclohexane to benzene and cyclohexene has been proposed.

Deeper knowledge of the reaction mechanism led to a formulation of innovative photocatalysts (sulphated MoOx/γ-Al2O3) selective for oxidative mono-dehydrogenation of cyclohexane.

Finally, the effectiveness of ethylbenzene to styrene conversion has been also checked.

I Introduction

Volatile organic compounds (VOCs) are widely used in (and produced

by) both industrial and domestic activities “Wolf et al. (1991)”. This extensive use results in their occurrence in aquatic, soil and atmosphere environments “Shah and Singh (1988)”. Many VOCs are toxic, and some are considered to be carcinogenic, mutagenic, or teratogenic “Wilkinson (1987)”. However, the most significant problem related to the emission of VOCs is centred on the possible production of photochemical oxidants, for example ozone and peroxyacetyl nitrate “Japer et al. (1991)”. Tropospheric ozone, formed in the presence of sun light from NOx and VOC emissions, is toxic to humans, damaging to crops and is implicated in the formation of acid rain “Cortese (1990), Fisherman (1991)”. Emissions of VOCs also contribute to localized pollution problems of toxicity and odor. Many VOCs are implicated in the depletion of the stratospheric ozone layer and may contribute to global warming.

As a result of all these problems, VOCs have drawn considerable attention in the last decade. Approximately 50% of the U.S. Enviromental Protection agency (EPA’s) list of priority pollutants is composed of VOCs. The Clean Air Act of 1990 calls for a 90% reduction in the emission of 189 toxic chemicals over the next 8 years, 70% of these being VOCs “Armor (1992)”. Therefore, there is currently a great deal of interest in developing processes which can destroy these compounds, and since a large number of the VOCs are oxidizable, chemical oxidation process can be looked upon as a viable method. The application of heterogeneous catalytic oxidation technology to air pollution control is well-established. Examples are automotive exhaust treatments “Armor (1992)” and catalytic incineration “Kosuko and Nunez (1990)”. In general, these catalysts operate at elevated temperatures.

Photocatalytic oxidation of organic compounds in the gas-phase using TiO2 as a catalyst appeared to be a promising process for remediation of air and ground-water polluted by VOCs. Heterogeneous photocatalysis using TiO 2 has several attractions:

Chapter I

2

(a) TiO2 is relatively inexpensive, (b) it dispenses with the use of other coadjutant reagents, (c) it shows efficient destruction of toxic contaminants, (d) it operates at ambient temperature and pressure and (e) the reaction products are usually CO2 and H2O, or HCl, in the case of chlorinated organic compounds (Figure 1).

Figure 1 Schematic representation of a photocatalytic process

I.1 Heterogeneous photocatalysis

The term “photocatalysis” is still the subject of some debate. For example, it is argued “Suppan (1994)” that the idea of a photocatalysed reaction is fundamentally incorrect, since it implies that, in the reaction, light is acting as a catalyst, whereas it always acts as a reactant which is consumed in the chemical process. In reality the term “photocatalysis” is in widespread use and is here to stay; it is not meant to, nor should it ever be used to, imply catalysis by light, but rather the “acceleration of a photoreaction by the presence of a catalyst”. The term “photoreaction” is sometimes elaborated on as “photoinduced” or “photoactivated” reaction, all to the same effect.

The above definition of “photocatalysis” includes the process of “photosensitization”, a process by which a photochemical alteration occurs in one chemical species as a result of the initial absorption of radiation by another chemical species called the photosensitizer. It follows from the above that heterogeneous photocatalysis involves photoreactions which occur at the surface of a catalyst. If the initial photoexcitation process occurs in an adsorbate molecule, which then interacts the ground state of the

Introduction

3

catalyst substrate, the process is referred to as a “catalysed photoreaction”. If, on the other hand, the initial photoexcitation takes place in the catalyst substrate and the photoexcited catalyst then interacts with the ground state adsorbate molecule, the process is a “sensitized photoreaction”. In most cases, heterogeneous photocatalysis refers to semiconductor photocatalysis or semiconductor-sensitized photoreactions.

I.1.1 Basic Principles

The overall process of semiconductor-sensitized photoreactions can be summarized as follows:

+−

≥

++ → DADAtorsemiconduc

Elight bg

where Ebg is the band gap energy of the semiconductor. If, in the absence of semiconductor and light of energy greater than or equal to Ebg, ∆G0 for reaction (1) is negative, the semiconductor-sensitized photoreaction is an example of photocatalysis “Bard (1980)”. Alternatively, if ∆G0 for reaction (1) is positive, the semiconductor-sensitized photoreaction is an example of photosynthesis “Bard (1980)”.

As indicated in Figure 2, for many compounds, as the number N of monomeric units in a particle increases the energy necessary to photoexcite the particle decreases. In the limit when N>>2000, it is possible to end up with a particle which exhibits the band electronic structure of a semiconductor, as illustrated in Figure 2, in which the highest occupied band (the valence band) and lowest unoccupied energy band (the conductance band) are separated by a bandgap Ebg, a region devoid of energy levels in a perfect crystal.

Chapter I

4

N=1 N=2 N=6 N>>2000

Figure 2 Change in the electronic structure of a semiconductor compound

as the number N of monomeric units increases from unity to clusters of more

than 2000

The basic principles of heterogeneous photocatalysis can be summarized shortly as follows “Litter (1999)”. A semiconductor (SC) is characterized by an electronic band structure in which the highest occupied energy band, called valence band (vb), and the lowest empty band, called conduction band (cb), are separated by a bandgap, i.e. a region of forbidden energies in a perfect crystal. When a photon of energy higher or equal to the bandgap energy is absorbed by a semiconductor particle, an electron from the vb is promoted to the cb with simultaneous generation of a hole (h+) in the vb. The electron in the cb and the h+ in the vb can recombine on the surface or in the bulk of the particle in a few nanoseconds (and the energy dissipated as heat) or can be trapped in surface states where they can react with donor (D) or acceptor (A) species adsorbed or close to the surface of the particle. Thereby, subsequent anodic and cathodic redox reactions can be initiated (Figure 3). The energy level at the bottom of the cb is actually the reduction potential of photoelectrons and the energy level at the top of the vb determines the oxidizing ability of photoholes, each value reflecting the ability of the system to promote reductions and oxidations.

Introduction

5

Figure 3 Illustration of the main processes occurring on a semiconductor

particle under electronic excitation

The flatband potential, Vfb, locates the energy of both charge carriers at the SC–electrolyte interface, and depends on the nature of the material and the system equilibria “Serpone (1997)”. From a thermodynamic point of view, adsorbed couples can be reduced photocatalytically by cb electrons if they have redox potentials more positive than the Vfb of the cb, and can be oxidized by vb holes if they have redox potentials more negative than the Vfb of the vb. The efficiency of a photocatalyst depends on the competition of different interface transfer processes involving electrons and holes and their deactivation by recombination “Serpone (1997), Hoffman et al. (1995), Fox and Dulay (1993), Linsebigler et al.(1995)”. The position of the flatband of an SC in solution follows a Nernstian pH dependence, decreasing 59mV per pH unit “Ward (1983)”, and consequently, the ability of electrons and holes to enact redox chemistry can be controlled by changes in the pH.

By using titania as semiconductor, the heterogeneous photocatalytic process is a complex sequence of reactions that can be expressed by the following set of simplified equations:

Chapter I

6

A great discussion exists nowadays about the oxidative pathway, which

could be performed by direct hole attack or mediated by HO● radicals, in their free or adsorbed form. The oxidative pathway leads, in many cases, to complete mineralization of an organic substrate to CO2 and H2O. Generally, A is dissolved O2, which is transformed in superoxide radical anion (O2

●−) and can lead to the additional formation of HO●:

The acceptor A can also be a metal ion species having a convenient redox

potential to be transformed into a different oxidation state:

Some oxide and chalcogenides have enough bandgap energies to be

excited by UV or visible light, and the redox potentials of the edges of the vb and cb can promote a series of oxidative or reductive reactions.

From the available semiconductors, ZnO is generally unstable in illuminated aqueous solutions, especially at low pH values, and WO3, although useful in the visible range, is generally less photocatalytically active than TiO2. Among others, CdS, ZnS and iron oxides have been also tested. However, and without any doubt, TiO2 is so far the most useful material for photocatalytic purposes, owing to its exceptional optical and electronic properties, chemical stability, non-toxicity and low cost.

TiO2 exists in three main crystallographic forms, brookite, anatase and rutile and it is the most widely used semiconductor. The energy bandgaps of anatase (3.23 eV, 384 nm) and rutile (3.02 eV, 411 nm) combine with the vb positions to generate highly energetic holes at the interface, giving rise to

Introduction

7

easy oxidation reactions. Anatase has been found, in most of the cases, to be photocatalytically more active than rutile. The most popular commercial form of TiO2 is produced by the German company Degussa under the name P-25; this sample contains around 80% anatase and 20% rutile and possesses an excellent activity. Recently, values for the Vfb of the cb and vb of Degussa P-25 have been calculated as −0.3 and +2.9V (pH = 0), respectively “Martin et al.(1994)”.

With regards to kinetic aspects, the rate of heterogeneous photocatalytic processes depends on the global “reaction resistance” but also on the concentration of photoproduced electron-hole pairs “Palmisano and Sclafani (1997)”. The concentration of these latter depends on the intensity of the radiation of suitable energy impinging on the system and on their recombination rate. When the maximum concentration of the pairs has been achieved (steady state), the rate depends on several factors, such as electronic, chemical and morphological properties of semiconductor, presence of additives in the reacting system, donor acceptor and acid-base properties of the solution and of the solid, temperature (high temperatures generally lead to higher rates because they provoke a more frequent collision between the substrate and the SC) and pressure. The adsorption of reagents, the charge transfer from and to reagents and the desorption of the products of the photoreaction are essential kinetic steps and their role should be evaluated every time as in catalysis. For example an augmentation to the rate of oxygen photo-uptake by a TiO2 surface can be created if the surface is pre-exposed to the vapour of a strongly adsorbing species, such as an aliphatic alcohol (ROH) “Bickley (1997)”. Alcohols adsorb very strongly up to monolayer coverage on TiO2 and the adsorbed species (e.g., RO- (ads)) can themselves act as very efficient hole traps. In these circumstances the rate of uptake of oxygen by the surface under band-gap energy illumination is greatly enhanced and attains a pseudo-zero order character. In these cases the surface becomes covered progressively with the oxidised form of the alcohol (ethanal when ethanol is used; propanone when propan-2-ol is used).

Accordingly, a study of strongly preadsorbed organic species in the presence only of a pure gaseous species such as O2 (g) is strictly a precursor to sustained photocatalytic oxidation where, with the availability also of excess of vapour or liquid phase reagent, a photocatalytic oxidation can occur with the progressive consumptions of both O2 (g) (or O2 dissolved) and the organic reagent e.g., propan-2-ol. In many examples, the kinetic of the photocatalytic reaction are controlled uniquely by the photo-electronic events induced in the solid in the copresence of the reacting partners.

I.2 Fundamental Engineering Aspects

Photocatalytic reactions are the result of the interaction of photons having the appropriate wavelength with a semiconductor. When the arriving light

Chapter I

8

has an energy equal or greater than the semiconductor band gap, radiation is absorbed and electrons are moved from the valence band to the conduction band giving rise to the formation of electron–hole pairs. These charge carriers can migrate to the catalyst surface in competition with an exothermic and normally fast recombination reaction. When they reach the semiconductor surface they may, once more recombine, or participate in successive chemical reactions “Cassano and Alfano (2000)”. The main components of a photocatalytic process are indeed the photoreactor and the radiation sources “Augugliaro et al. (1997)”. For thermal and catalytic processes the parameters that affect reactors performance are:

1. the mode of operation; 2. the phases present in the reactor; 3. the flow characteristics; 4. the needs of heat exchange; 5. the composition and the operative conditions of the reacting

mixture. For selecting the type of heterogeneous photoreactor additional

parameters must be considered since photons are the primary source for the occurrence of photoreaction. The selection of the construction material for the photoreactor must be generally done in order to allow the penetration of radiation into the reacting mixture. The choice of the radiation source must be made by considering that the absorbed radiation energy should be equal to or higher than the band gap.

I.2.1 Types of Radiation Sources

Different types of lamps allow generation of radiation with different ranges of wavelengths. As reported before, the choice of a particular lamp is made on the basis of the reaction energy requirements “Augugliaro et al. (1997)”. There are four many types of radiation sources:

a. arc lamps; b. fluorescent lamps; c. incandescent lamps; d. lasers.

In arc lamps the emission is obtained by a gas activated by collisions with electrons accelerated by an electric potential. Typical activated gases are mercury and/ or Xenon vapours.

In fluorescent lamps the emissions is obtained by exciting an emitting fluorescent substance, deposited in the inner side of a cylinder, by an electric discharge occurring in the gas filling the lamps. Generally these lamps emit in the visible region, but the “actinic” type ones have emission in the near UV-region. Of course the emission spectrum depends on the nature of the mixture of fluorescent substances user. Their power is quite small (up to 150W). In incandescent lamps, the emission is obtained by heating at very

Introduction

9

high temperature suitable filaments, of various nature, by current circulation. A typical emission spectrum of an actinic lamp is reported in Figure 4.

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

250 300 350 400 450

wavelenght, nm

sp

ec

tra

l ir

rad

ian

ce

, W

/m2

Figure 4 Emission spectrum of a 40W actinic lamp

I.2.2 Photoreactor Materials and Geometries

In order to allow the penetration of radiation into the photoreactor, the photoreactor wall between the lamp and the reacting mixture must be made of a material transparent to the radiation. If near-UV radiation is utilised by the photoreacting mixture, Pyrex glass can be used but if UV radiation is needed, quartz glass must be used. It must be noted that, by increasing the thickness of photoreactor walls, the light transmission decreases so that the size of reactor and its operative conditions (temperature and pressure) are not independent variables. In order to utilise all the radiation emitted by the lamp, the transmission or absorption of radiation from photoreactor walls is avoided by wrapping them with reflecting surfaces such as aluminium foils.

For photoreactors, the geometry and the spatial relation between reactor and light source are most important. In fact geometry plays an important role in determining reaction yields as well as reactor operability. The geometrical configuration of a photoreactor is usually chosen to obtain the maximum benefit from the pattern of irradiation taking into account shape and cooling requirements of available commercial lamps useful for wavelength required by the reaction.

Chapter I

10

The most used photoreactor geometries are: a. cylindrical; b. parallel plate; c. annular.

The irradiation may be normal or parallel to the reactor surface. In selecting the reactor geometry configuration, it is necessary to determine the optical path of the light which will be obtained within the reactor. It is, in fact, the most important factor affecting the absorption of radiation by the reaching mixture and therefore it determines the efficiency of the photoprocess.

I.2.3 Efficiency of Heterogeneous Photocatalytic Systems

In heterogeneous photocatalysis, quantum yield (Φ) has come to define the number of molecules converted per unit time relative to the number of photons absorbed per unit time as follows:

timeunit

photonsabsorbed

timeunit

moleculesdtransforme

=Φ (1)

In a heterogeneous photocatalytic process, where a solid is involved, the rates must be referred to the active sites. The difficulty of determining the number of the active sites is generally overcome by using the BET surface area of the particle instead of the active sites; the implicit assumption of this simplification is that the number of active sites is proportional to the surface area. For the occurrence of a photoreaction it is necessary that photons of suitable energy are absorbed by the semiconductor, which usually is in the form of polycrystalline porous particle. A first quantity that must be measured is therefore the rate of photons absorption (rpa), defined as:

areasurfacetime

photonsabsorbedrpa

×= (2)

Once radiation is absorbed, electron-hole pairs are generated if the photon energy equals or exceeds the semiconductor band-gap energy. The pairs can generate thermal energy (recombination), determine a lattice conversion (photocorrosion) or be trapped by suitable surface species and initiate a reaction sequence with adsorbed surface molecules (photocatalysis). On this ground a second quantity that must be measured is the specific reaction rate (srr), defined as:

Introduction

11

areasurfacetime

moleculesreactedsrr

×= (3)

The Φ values may be determined, knowing the rpa and srr parameters, by the following equation:

rpa

srr=Φ (4)

I.2.4 Modeling of Photocatalytic Reactors

The modeling of photocatalytic reactors requires a complex analysis of the radiation field in the photoreactor “Cassano et al. (1995)”. This analysis, linked to the modeling of the fluid-dynamics and the reaction kinetics, results in integro-differential equations which almost invariably require demanding numerical computations. Further advances of photocatalytic oxidation on an industrial scale will be facilitated by the availability of simpler mathematical models that retain the essential elements of a rigorous model and that can be easily used for scale-up and design.

Figure 5 shows a schematic representation of the modeling of a photocatalytic reactor. The development of a reactor model requires the inclusion of a number of sub-models. These are a radiation emission model, a radiation absorption-scattering model, a kinetic model and a fluid-dynamic model.

Radiation emission model

Radiation absorption-scattering model

LVRPA Kinetic model Material balance

Fluid-dynamic model

Reactor model

Figure 5 Schematic representation of the modeling of a photocatalytic

reactor

The central aspect of the modeling procedure is the calculation of the Local Volumetric Rate of Photon Absorption (LVRPA) at each point of the reaction space, which requires solving the radiative transfer equation (RTE) in the reaction space. Due to the complex nature of radiation scattering, this results in a set of integro-differential equations, which require demanding numerical computational efforts. In practice, combining a simplified radiation emission model of the light source with a simplified radiation absorption-scattering model in the reaction space and performing a radiation

Chapter I

12

balance in the reaction space can simplify the RTE. The above scheme assumes that the “useful photons” of a given

photocatalytic reactor, i.e. those photons with energy higher than the band-gap of the semiconductor photocatalyst, are absorbed by the solid photocatalytic particles only. This assumption removes the interdependence of the progress of the reaction and the attenuation of the radiation because the absorbing species do not undergo changes in concentration. Consequently, the incident radiation flux becomes a function of the reactor position only and can be obtained independently of the information provided by the material balance equation.

Once the LVRPA has been calculated, this is normally substituted into the kinetic equation and into the material balance equation which when solved with suitable boundary conditions, yields the concentration of a generic substrate at the reactor outlet.

Three approaches have been proposed in the literature for the calculation of the LVRPA: 1) The “rigorous approach” which involves the mathematical solution of the RTE, although its integro-differential nature makes this approach considerably complex “Cassano et al. (1995)”; 2) The “numerical approach” which involves the Monte Carlo simulation of the radiation field in the photoreactor, a simple but also a computationally demanding procedure “Pasquali et al. (1996)”; and 3) the “simplified approach” which models the radiation field in the photoreactor using two-flux “Raupp et al. (1997), Brucato and Rizzuti (1997)”, or six-flux “Li Puma et al. (2004)” radiation absorption-scattering models. The two and six-flux models yield a sensible representation of the LVRPA in the reaction space and allow a considerable simplification of the mathematical model. Finally the scattering properties of the photocatalyst, and the geometrical configuration of the photoreactor, determine in large extent the degree of complexity of a mathematical model “Li Puma et al. (2005)”.

I.3 Aim of the Work

Heterogeneous photocatalysis has been widely investigated as a novel technique for environmental applications such as detoxification processes both in water and in air. In contrast, very few studies have been conducted concerning photocatalysis as potential technology for the synthesis of chemicals from hydrocarbon feedstocks. Particularly, photocatalysed selective oxidation of hydrocarbons such as cyclohexane with gaseous oxygen is still a challenge.

Concerning this last aspect, in this work photocatalytic selective oxidation of cyclohexane on molybdenum based catalysts has been investigated by changing Mo loading, nature and surface characteristic of the support.

Introduction

13

The photocatalytic activity and selectivity of MoOx supported catalysts on TiO2, α− and γ−Al2O3, have been determined with an annular gas-solid continuous flow reactor. Photocatalytic test were also carried out on MoOx supported NaY, HY zeolites and Ammonium ferrierite (AFer).

In order to elucidate the effect of sulphate content on MoOx/TiO2 activity and selectivity, cyclohexane photocatalytic oxidation has been also studied on sulphated and unsulphated titania.

Activity and selectivity of MoOx/TiO2 catalysts have been also measured by a photocatalytic plate reactor.

Finally, a photocatalytic fluidized bed reactor has been designed and realized in order to improve both exposure of the catalysts to light irradiation and a good contact between reactants and catalyst.

II State of the art

II.1 Photocatalysis as potential technology for synthesis of

chemicals from hydrocarbon feedstocks

Over the past 10 years, the semiconductor TiO2 (Ebg = 3.2 eV) as a

photocatalyst has become the focus of numerous studies owing to its attractive characteristics and applications in the treatment of environmental contaminants both in water streams than in air streams “Yu et al. (1997)”. Complete mineralization of a broad variety of organic compounds containing unsaturated bonds by TiO2 photocatalysis has been reported “Gratzel. (1983)”.

The photodecomposition of chlorinated “Philips and Raupp (1992), Liu et

al. (1997)” and fluorinated “Ohtani et al. (1990)” organics, hydrocarbons “Fu et al. (1996)”, aldehydes, ketones, alcohols “Blake and Griffin (1988), Sauer and Ollis (1996), Peral and Ollis (1992)”, aromatics “Fu et al. (1995)” or trichloroethylene “Dibble and Raupp (1990), Dibble and Raupp (1992)” over UV-irradiated titanium dioxide is well documented in literature. A drawback of these processes is the formation of partial oxidation products, sometimes resulting in increased toxicity of the treated stream.

Few studies have been conducted on photocatalysis as potential technology for synthesis of chemicals from hydrocarbon feedstocks.

Significant examples of photocatalytic processes employed for synthetic purposes are oxidation and reduction processes, isomerization reactions, C-H bond activations, and C-C and C-N bond-reforming reactions “Maldotti et

al. (2002)”. For these purposes heterogeneous photocatalysis processes seem to be advantageous since the use of solar light as a reagent in oxidative catalysis is particularly relevant to realizing innovative and economically advantageous processes for conversion of hydrocarbons into oxygenates and, at the same time, to move toward a “sustainable chemistry” that has a minimal environmental impact. The second key reagent employed in the oxygenation processes considered herein in the molecule of O2. With regard to this last aspect, it is important the search for new catalysts capable of

Chapter II

16

inducing the oxofunctionalization of hydrocarbons represents a major target from the synthetic and industrial points of view “Roby and Kingsley (1996)”. On the basis of pure thermodynamic considerations, most organic compounds are not stable with respect to oxidation by O2. There are however, kinetic limitations. The activation of both O2 and the organic substrate may be achieved by photochemical excitation with light of the visible or of the near-ultraviolet regions (λ > 300 nm). Moreover, selectivity is a key issue in the catalysis of fine chemistry. To purpose this objective, all the steps of the oxidation process must be optimized. For example the use of heterogeneous and organized systems is a suitable way to control efficiency and selectivity of catalytic processes through the control of the microscopic environment surrounding the catalytic center. In particular the nature of the reaction environment may affect numerous physical and chemical functionalities of the photocatalytic system such as the absorption of light, the generation of elementary redox intermediates, the rate of competitive chemical steps, and the adsorption-desorption equilibria of substrates, intermediates, and final products.

In the seventies pioneer works reported photo-oxidation of alkanes and alcohols in heterogeneous gas-solid reactors “Walker et al. (1977), Djeghri and Teichner (1980), Bickley et al. (1973), Formenti et al. (1971)”. Alkanes have been oxidised on anatase TiO2 achieving 75% selectivity to acetone at 3% isobutane conversion, 30 % selectivity to butanone from n-butane, but complete oxidation from 1-butene and 2-butene. In more recent works, few studies employed gas-solid reactors.

Ethanol was selectively converted to acetaldehyde and formaldehyde on TiO2. Muggli and Falconer showed that selectivity of ethanol photocatalytic oxidation is changed as effect of titania sites poisoning by adsorption of reactants “Muggli and Falconer (1998)”. This seems suggest a way to design a selective catalyst for photocatalytic oxidation.

The studies of Frei on photocatalysed oxidation of hydrocarbons in zeolite cages indicate a potential way to achieve high selectivity in gas-solid systems. Very impressive results (100% selectivity to aldehydes and ketones from the corresponding alkanes, cycloalkanes, alkenes, aromatics) are overviewed “Blatter et al. (1998)”.

II.1.1 Photocatalytic oxidation of cyclohexane

The oxidation of cyclohexane to form cyclohexanone is significantly important reaction since cyclohexanone is an intermediate material to ε-caprolactone which is a raw material for nylon synthesis. The oxidation process of cyclohexane to produce cyclohexanone has been industrialized over cobalt-base catalyst with oxygen above 150 °C under high pressure “Davis and Kemp (1991)”. To make the reaction condition milder, new reaction system has been sought out. Hydrogen peroxide is often used as an

State of the art

17

oxidizing reagent to achieve the mild reaction condition “Carvalho et al. (1999), Armengol et al. (1999)”. Nevertheless, realization of oxidation by molecular oxygen is strongly desired.

The most studies of cyclohexane photocatalytic oxidation deal with slurry systems. Cyclohexanol, cyclohexanone, and polyoxygenates have been obtained on silica supported vanadia or polyoxytungstate catalysts “Molinari et al. (1998), Giannotti and Richter (1999)”. It is proposed that cyclohexanone is obtained from cyclohexane through a light induced radical mechanism resulting in the selective formation of ketone due to a confinement effect inside the zeolite cage “Blatter and Frei (1998)”.On titania cyclohexane photo-oxidation in dichloromethane leads to the formation of cyclohexene traces “Almiquist and Biswas (2001)”.

Gas-solid heterogeneous photocatalytic oxidation of cyclohexane and cyclohexene in humidified air was studied at 30 °C on P-25 titania powder “Einaga et al. (2002)”. Deep oxidation to CO2 was essentially obtained together with formation of carbon deposits on the surface, which is responsible for catalyst deactivation.

II.1.2 Oxidative dehydrogenation of cyclohexane

Potential useful intermediates and monomers for polyadditions (such as cyclohexene and cyclohexadiene), together with benzene can be produced by cyclohexane oxidative dehydrogenation reaction. Cyclohexane is available in large amounts in naphthas, so it can be recovered from them although with quite demanding procedures. By means of a refinery distillation tower (named benzene heartcut tower) where cyclohexane is present in the side fraction, a benzene-rich fraction (50% of benzene) is separated from a heavy gasoline bottom and a light gasoline head fraction “Blomberg et al. (2002)”. This side fraction is used to provide benzene for petrochemical processes. Due to the important role of benzene as an intermediate in petrochemistry, processes for conversion of low-value hydrocarbons into benzene, such as aromatization of light alkanes “Nishi et al. (2002)”, are under development. In this context, the conversion of cyclohexane, either recovered from the benzene heartcut tower side fraction, or still in mixture with benzene, can be a way to enhance the production of benzene and to fulfil the need for benzene in some cases.

The occurrence of oxidative dehydrogenation of cyclohexane to cyclohexene and benzene on MoO3 or MoO3/γ-Al2O3 catalysts in the temperature range 280-417 °C has been reported “Alyea and Keane (1996)”. Selectivity to dehydrogenated products is influenced by side-reactions such as combustion or cracking, leading to a heavily selectivity decrease at increasing temperature. A recent paper investigates the thermal production of benzene from cyclohexane in vapor phase over several catalysts, such as V2O5/SiO2, V2O5-Nb2O5/SiO2, Ce-,V-, Fe-phosphates, H-ZSM5, Co-ZSM5

Chapter II

18

“Panizza et al. (2003)”. It was found that gas-phase reaction occurs above 277 °C and selectivity to benzene was positively influenced by the presence of water vapor.

Very recently, heterogeneous oxidative dehydrogenation of cyclohexane to benzene was found for the first time to occur under photocatalytic conditions. Initial cyclohexane conversion of 38% and maximum selectivity to benzene of 41% were achieved on a MoOx/TiO2 catalyst in the presence of gaseous oxygen at temperature of 35 °C under UV irradiation in a gas-solid continuous flow reactor. The main by-product was carbon dioxide “Ciambelli et al. (2002)”.

II.2 Influence of catalyst surface acidity on photocatalytic

performances

Titanium dioxide is widely used in heterogeneous photocatalysis as a

semiconductor photocatalyst because of its long-term stability, no toxicity and good, often the best, photocatalytic activity. However, TiO2 is active only under ultraviolet (UV) light, so in recent years, transition metal ions doping has been widely performed by chemical synthesis and other methods, in order to improve photoactivity “Legrini et al. (1993)”. Karakitsou and Verykios (1993) showed that doping with cations having a valence higher than +4 can increase the photoactivity, whereas Mu et al. (1989) reported that doping with trivalent or pentavalent metal ions was detrimental to the photoactivity even in the UV region. Furthermore, according to a systematic study on the photoactivity and transient absorption spectra of quantum-sized TiO2 doped with 21 different metals the energy level and d-electron configuration of the dopants were found to govern the photoelectrochemical process in TiO2 “Yamashita et al. (1999)”. Even though the effects of metal doping on the activity of TiO2 have been a frequent topic of investigation, it remains difficult to make general conclusion.

The presence of sulphate and metal oxides as dopant of titania has been reported to enhance the photooxidation reactivity of several organic compounds. It was shown “Yu et al. (2002), Muggli and Ding (2001)” that TiO2 treatment with sulphuric acid could increase its photoactivity, although the mechanism was not clarified. Kozlov et al. (2003) reported that the treatment of TiO2 with sulphuric acid enhances the photocatalytic activity in acetone oxidation by 20–30%. Colon et al. (2003), studying the photocatalytic degradation of phenol on sulphated TiO2 in a slurry reactor, evidenced that the improvement of the photocatalytic activity is related to the optimisation of the redox step in the photocatalytic process instead of the acidity properties that could favour the adsorption of the organic substrate. Fu et al. (1999) studied the structure of SO4

2-/TiO2 and its activity for room temperature photocatalytic oxidation (PCO) of CH3Br, C6H6, and C2H4 in air. For catalysts calcined at 723 K, conversion of CH3Br over SO4

2-/TiO2

State of the art

19

was six times higher than over TiO2. Moreover, TiO2 deactivated faster than SO4

2-/TiO2; after 6 h of PCO, SO42-/TiO2 did not deactivate, whereas on TiO2

conversion decreased from 88 to 20% for C6H6 and from 60 to 12% for CH3Br. They concluded that the improved rate for SO4

2-/TiO2 was due to a greater surface area as well as a larger fraction of the anatase phase of TiO2, which is more active than rutile for PCO.

Doping TiO2 with metal oxides such as WO3, MoO3, and Nb2O5 increases surface acidity and PCO activity of TiO2. Cui et al. (1995) studied the activity of Nb2O5/TiO2 during PCO of 1,4-dichlorobenzene. They found that surface acidity increased and photocatalytic activity doubled when niobium oxide was deposited on titania up to a monolayer. However, higher niobium loading did not increase surface acidity and PCO photocatalytic activity due to Nb2O5 segregation. They concluded that the same structural feature that enhances surface acidity increases PCO rate. Similarly, the authors found correlation between PCO activity and acidity for TiO2 doped with WO3, MoO3, or Nb2O5 “Papp et al. (1994), Okasaki and Okuyama (1983), Lee et

al. (1997)”.

II.3 Photocatalytic fluidized bed reactors

For photocatalytic degradation processes, two methods of TiO2 application are favoured: (1) TiO2 suspended in aqueous media and (2) TiO2 immobilized on support materials. Fluidized beds are known as good chemical reactors due to excellent reactants contact, high mass and heat transfer rate and easy to control the reaction temperature. When a photocatalytic reaction takes place in a gas–solid reactor, it is necessary to achieve both exposures of the catalysts to light irradiation and a good contact between reactants and catalyst.

It is believed that fluidized bed can take advantages of better use of light, easy of temperature control, and good contacting between target compound and photocatalysts over slurry reactors or fixed bed reactors with immobilized TiO2.

Activity increase of fluidized bed is partially associated with higher light absorption due to utilization of scattered light by the catalyst. Figure 6 illustrates light scattering in the three types of beds typically used.

Chapter II

20

Figure 6 Schematic representation of light scattering in (A) film fixed bed,

(B) granular fixed bed and (C) fluidized bed

In the wavelength range of UVA light used, titania particles scatter a part of incident light. In the film bed, the scattered light never meets catalyst again. In granular fixed bed, some part of the scattered light meets the catalyst and is absorbed or scattered again. In fluidized bed, the probability for collisions of scattered light with titania particles is the highest “Vorontsov et al. (2000)”.

The fluidized bed of ultrafine particle of photocatalyst was applied to treat nitrogen oxides (NOx) by photocatalytic oxidation “Matsuda et al. (2001)”. Three different TiO2 particles with primary particle diameters of 7, 20 and 200 nm were used as the bed material. The fluidized bed of aggregates of 7 nm crystallites TiO2 exhibited high removal efficiency of NOx because of its large specific surface area. It was found that the amount of NOx removal is proportional to the specific surface area. Agglomerates of 7 and 20 nm particles appeared to be so hard that they were not destroyed during fluidization because of large adhesion forces of particles. In the case of 7 and 20 nm particle systems, the bed height was found to increase progressively with the increase in gas velocity, while the bed expansion is observed to level off for 200 nm particle system. The entrainment rates of 7 and 20 nm particle systems were found to be smaller than that of 200 nm particle system.

The effects of CuO loading on titania support, reaction temperature, and surface gas velocity on the photocatalytic reduction of NO have been determined in an annular flow type and a modified two-dimensional fluidized bed photoreactor. The optimum CuO loading was found to be 3.3 wt% and NO degradation conversion in the modified two-dimensional fluidized bed photoreactor was more than 70% at 2.5 times the minimum fluidization velocity, Umf “Lim et al. (2000)”. The decomposition of NO by photocatalysis increased with decreasing initial NO concentration and increasing gas-residence time and the reaction rate increased with increasing UV light intensity. The photocatalytic oxidation of ethanol vapour was investigated with an annulus fluidized bed reactor “Kim et al. (2004)” of

State of the art

21

silica gel powder coated TiO2 catalyst prepared by the sol-gel method. The UV lamp was installed at the center of the bed as the light source. It was found that at 1.2 Umf value, about 80% of ethanol with initial concentration.of 10000 ppm was decomposed and the increase of superficial gas velocity reduced the reaction rate significantly.

A loading type of Eu/Ti/Si catalyst was synthesized with Eu(NO3)3, Ti(OC4H9)4 and porous silica by sol-gel method “Ping et al. (2004)”. Benzene and its compounds were degraded in a three-phase fluidized bed photo-catalytic reactor and the reaction conditions for photo-catalytic degradation were investigated. The results indicated that comparing the binary catalyst Ti and Si and the triple compound catalyst of Eu/Ti/Si, the latter was more efficient. The photo-catalysis removal efficiency for benzene and its compounds would reach over 98%, and that was 10-20% higher than that as Ti/Si catalyst was used. Photocatalytic NH3 synthesis was successfully performed in a fluidized reactor of parallel-wall configuration by irradiating Fe-doped TiO2 with near UV light “Yue et al. (1983)”. The catalyst was prepared in such a way that good quality of fluidization was obtained. Mixing the catalyst with γ-alumina affected the fluidization behavior. NH3 production was increased when the catalyst was suitably fluidized because of enhanced utilization of light energy.

III Experimental Results:

Photocatalytic oxidation of

cyclohexane on MoOx/Al2O3

III.1 Samples preparation

All chemicals used in the experiments were HPLC grade obtained from Aldrich Co. α−Al2O3 (αAl), (Aldrich) and γ−Al2O3 (γAl), (Puralox SBA 150, SASOL S.p.A.), were impregnated with an aqueous solution of ammonium heptamolybdate (NH4)6Mo7O24·4H2O.

Powder samples were dried at 120 °C for 12 hours and calcined in air at 400 °C for 3 hours.

In Table 1 the list of prepared catalysts with their nominal MoO3 content is reported.

Table 1 List of catalysts with their MoO3 nominal content

Catalyst Nominal MoO3 content, wt%

αAl - 2MoαAl 2.0 γAl - 8MoγAl 8.0

III.2 Catalysts characterization

To characterize the samples studied in this work the following techniques were used:

• Inductively coupled plasma-mass spectrometry (ICP-MS); • Thermal analysis (TG-MS);

Chapter III

24

• Micro Raman spectroscopy; • Fourier Transform Infrared (FTIR) spectroscopy; • N2 adsorption at -196 °C to obtain specific surface area and

porosity characteristics; • Temperature programmed desorption (TPD).

III.2.1 ICP-MS

Inductively Coupled Plasma Mass Spectrometry (ICP-MS) is extensively used to detect and quantify the presence of the majority of the elements in the periodic table and is a very powerful tool for trace (ppb-ppm) and ultra-trace (ppq-ppb) elemental analysis. In ICP-MS, the plasma is formed from Argon gas and reaches very high temperatures of up to approximately 7000 K. The plasma is used to atomize and ionize the elements in a sample. The sample to be analysed is introduced into the plasma as a fine aerosol. As the sample aerosol passes through the plasma, it collides with free electrons, argon cations and neutral Argon atoms. The result is that any molecules initially present in the aerosol are quickly and completely broken down to charged atoms. The resulting ions are then passed through a series of apertures (cones) into the high vacuum analyzer. The isotopes of the elements are identified by their mass-to-charge ratio (m/z) and the intensity of a specific peak in the mass spectrum is proportional to the amount of that isotope (element) in the original sample. An ICP-MS Agilent 7500ce instrument was used for the analysis of Mo.

III.2.2 Thermal analysis (TG-MS)

The performances of samples as a function of temperature were determined by Air flow thermal analysis (TG-MS). The apparatus used were a TGAQ500 (Figure 7) thermogravimetric analyzer (TA Instruments) and a SDTQ600 (Figure 8) simultaneous DSC/TGA (TA Instruments). Both analyzers can be coupled to a Pfieffer Vacuum Benchtop Thermostar mass spectrometer (MS) (Figure 9)

Experimental Results: Photocatalytic oxidation of ciclohexane on MoOx/Al2O3

25

Figure 7 TGAQ500 Thermogravimetric Analyzer

Figure 8 SDTQ600 Simultaneous DSC/TGA

Figure 9 Pfieffer Vacuum Benchtop Thermostar mass spectrometer

Chapter III

26

TGAQ500 measures weight changes in a material as a function of temperature. The system works in a temperature range of 20-1000 °C, and weight variation resolution is 0.1µg . The sample, loaded in a crucible made of platinum and connected to the balance arm by a small hook, is progressively heated in the oven. A thermocouple controls the oven temperature and a second thermocouple reads the sample temperature. Sample pan loading and furnace movement are totally automated and there is a touch screen data display to change operating parameters. Typically measurements are carried out with 20 mg of sample in chromatographic air flow (60 Ncc/min) with a heating rate of 10 °C/min in the temperature range of 20- 800 °C.

The results are displayed as TG curves showing the mass variations as functions of temperature or time, and DTG curves showing the conversion rate (mass loss percentage per unit time) as functions of temperature or time.

Figure 10 contains the typical trends of the TG and DTG curves.

-1

0

1

2

3

DT

G (

%/m

in)

86

88

90

92

94

96

98

100

TG

(%

)

0 100 200 300 400 500 600

Temperature (°C)

Figure 10 TG and DTG curves

SDTQ600 provides a simultaneous measurement of weight change (TGA) and heat flow (DSC) on the same sample from ambient to 1500 °C. It features a proven horizontal dual beam design with automatic beam growth compensation, and the ability to analyze two TGA samples simultaneously. DSC heat flow data is dynamically normalized using the instantaneous sample weight at any given temperature. The sample is loaded in a crucible made of alumina and heated in the horizontal oven. There are two


27

thermocouples to control the oven temperature and the sample temperature. Measurements are carried out with about 30 mg of sample in chromatographic air flow (100 Ncc/min) with a heating rate of 10 °C/min in the temperature range of 20- 800 °C.

Pfieffer Vacuum Benchtop Thermostar mass spectrometer can measure the gas evolved from thermal analyzers up to 300 AMU. The evolved gases are introduced into a heated quartz capillary, which is extremely fine, in order to produce the necessary high vacuum, when the evolved gases enter the mass spectrometer. The heated capillary is necessary in order to prevent condensation of the hot gases on cold surfaces. The analysis of gases is performed by a very high sensitive quadrupole mass detector. The necessary high vacuum is obtained through 2 stages of vacuum pumps that are integrated into a compact housing. First stage is a rotary pump; second stage is a turbo molecular pump.

Both systems, the Mass Spectrometer and the Thermal Balance, are connected to a common PC for data acquisition.

III.2.3 Micro Raman spectroscopy

Raman spectroscopy is a technique for the identification and quantification of the chemical components of specimens.

When light is scattered by any form of matter, the energies of the majority of the photons are unchanged by the process, which is elastic or Rayleigh scattering. However, about one in one million photons or less, lose or gain energy that corresponds to the vibrational frequencies of the scattering molecules. This can be observed as additional peaks in the scattered light spectrum. The process is known as Raman scattering and the spectral peaks with lower and higher energy than the incident light are known as Stokes and anti-Stokes peaks respectively.

Laser Raman spectra of powder samples were obtained with a Dispersive MicroRaman (Invia, Renishaw) (Figure 11), equipped with 785 nm diode-laser, in the range 100-2500 cm-1 Raman shift.

Figure 11 Dispersive MicroRaman (Invia, Renishaw)

Chapter III

28

III.2.4 FT-IR Spectroscopy

This technique is based on the analysis of the absorption or of laser light by specific chemical bonds. Fourier Transform Infrared (FTIR) spectroscopy was performed in the 4000-400 cm-1 range with a resolution of 2 cm-1.

A Bruker IFS 66 FT-IR spectrophotometer, was used. Samples were diluted at 1 wt % in KBr. The mixture was ground and a transparent disk of 100 mg was prepared with a press in vacuum. Disks are introduced into the proper chamber and the scan is carried out at room temperature.

Figure 12 Bruker IFS 66 FT-IR spectrophotometer

The infrared spectrum includes all the radiation of wavelengths ranging from 0.1 to 1000 µm.

III.2.5 N2 adsorption measurements|

In order to study the porosity of catalysts powder, N2 adsorption measurement were carried out at -196 °C with a Costech Sorptometer 1040 (Figure 13).

The measurement was performed by continuous-flow method after sample pre-treatment at 150 °C for 2 h in He flow, in order to measure total specific surface area (via single and multi-point methods) and micropore volume (via micropore method).


29

Figure 13 Costech Sorptometer 1040

III.2.6 TPD investigations

Temperature programmed desorption (TPD) experiment of catalysts after activity measurements were carried out in N2 flow (500 Ncc/min) at atmospheric pressure in a quartz flow reactor, connected on-line with CO, CO2 (Uras 10E Hartmann & Braun) and with on-line quadrupole mass detector (MD800, ThermoFinnigan) in the range 20-500 °C, with an heating rate of 10 °C/min. For the analysis 1 g of sample was loaded into the microreactor.

III.3 Laboratory apparatus for catalytic test

Catalytic activity tests were performed using the laboratory apparatus shown in Figure 14. It consists of three sections:

• Feed section; • Reaction section; • Gas composition analysis section.

Chapter III

30

Figure 14 Laboratory apparatus for catalytic test

All the gas pipes (¼’’ e.d.) are of Teflon, connections are made with Swagelok union and two, three and four way Nupro valves. All the connections are in stainless steel to avoid any corrosion due to the presence of water.

III.3.1 Feed section

All the gases come from SOL SPA with a purity degree of 99,999%. Oxygen and nitrogen were fed from cylinders, nitrogen being the carrier

gas for cyclohexane (CH) and water vaporized from two temperature controlled saturators and by changing temperature and N2 flow, it is possible to obtain different concentrations in the reaction feed.

To feed an accurately controlled flow, Brooks measured flow controllers (MFC) are used, able to operate with a maximum pressure drop of 3 atm (Figure 15).


31

Figure 15 Mass flow controller

The working principle of the MFC is heat transport: the temperature difference in a capillary, where a part of the gas is split, it is measured. This temperature difference is proportional to the amount of heat adsorbed by the mass gas for the equation:

.∆T = K • Cp • Φm where: ∆T = temperature difference. Cp = specific heat of the gas. K = dimensional constant. Φm = mass flow. The instrument’s temperature detector produces an electrical signal from

0 to 5 V (c.c.); this signal is sent to the control unit (MFC C.U.) which converts the signal in volumetric flow. This control unit allows the mass flow of the gases to be regulated.

A rotameter for N2 is used for vaporizing water to feed to the reactor.

III.3.2 Reaction section

In the reaction section a system of valves allows the reactants to go to the reactor, and the products to the analysis section, or, in the bypass position, the reactants to the analysis section to verify the reactant composition.

The annular section of the fixed bed photocatalytic reactor (reactor volume: 7 l) was realised with two axially mounted 500 mm long quartz tubes of 140 and 40 mm diameter, respectively.

Chapter III

32

Gas In

Gas Out

Gas In

Gas Out

Figure 16 Annular gas-solid photocatalytic fixed bed reactor

The reactor was equipped with seven 40 W UV fluorescent lamps providing photons wavelengths in the range from 300 to 425 nm, with primary peak centred at 365 nm. One lamp (UVA Cleo Performance 40 W, Philips) was centred inside the inner tube while the others (R-UVA TLK 40 W/10R flood lamp, Philips) were located symmetrically around the reactor. Both photoreactor and lamps were covered with reflectant aluminum foils. In order to avoid temperature gradients in the reactor caused by irradiation, the temperature was controlled to 35 ± 2 °C by cooling fans.

The catalytic reactor bed was prepared in situ, by coating quartz flakes previously loaded in the annular section of a quartz continuous flow reactor with an aqueous slurry of catalysts powder. The coated flakes were dried at 393 K for 24 hours in order to remove the excess of physisorbed water. This treatment resulted in uniform coating well adhering to the quartz flakes surface.

III.3.3 Analysis section

The gas composition is determined by on line analysers connected to a PC for data acquisition. CO and CO2 concentration is measured by an on line non dispersive IR analyzer (Uras 10, Hartmann & Braun), working on the basis of specific adsorption of IR radiation (wavelength from 2 to 8 µm). Oxygen, cyclohexane and reaction products composition is determined by an on line quadrupole mass detector (MD800, ThermoFinnigan) that can


33

analyze the outlet reactor gas, introduced into a heated silica capillary, up to m/z = 800.

III.4 Photocatalytic tests conditions and typical trend

The amount of deposited catalyst, evaluated by weighing the reactor before and after the coating treatment was 20 g. Catalytic tests were carried out feeding 830 Ncc/min N2 stream containing 1000 ppm cyclohexane, 1500 ppm oxygen and adding 1600 ppm water to minimise catalyst photodeactivation “Einaga et al. (2002)”. Concerning this last aspect, several authors “Vorontsov et al. (2000), Martra et al. (1999), Marci’ et al. (2003), Augugliaro et al. (1999)” reported the positive influence of water on the oxidation rate of many hydrocarbons. In fact water increases photocatalytic activity because it is necessary for production of OH● radicals and for capture of photogenerated “Park et al. (1999)” holes enhancing the reaction between O2 and an electron promoted to the cb from the vb to form superoxide radical anion (O2

●−). A typical trend of photocatalytic test is reported in Figure 17 with

reference to a sulphated titania (2 wt % as SO3) supported MoOx catalyst with a nominal Mo load equal to 4.7 wt % (as MoO3).

Chapter III

34

0 50 100 150 200 250

time, min

MS

sig

na

l, a

.u.

0

10

20

30

40

50

60

CO

2,

pp

m

m/z = 84

m/z = 32

m/z = 78

m/z = 67

CO2, ppm

Lamp ON Lamp OFF


benzene and cyclohexene and (ppm) of carbon dioxide as a function of run

time

At the run starting time the nitrogen stream containing 1000 ppm cyclohexane, 1500 oxygen and 1600 water was passed through the reactor in the absence of irradiation at ambient temperature. Dark adsorption of cyclohexane is observed. Cyclohexane breakthrough time was about 10 minutes. Thereafter cyclohexane outlet concentration slowly increased to reach the inlet value after about 50 minutes, indicating that adsorption equilibrium of cyclohexane on the catalyst surface was attained. At that time the lamps were switched on: the cyclohexane outlet concentration immediately decreased to about 22% of the inlet value and then progressively increased with run time, reaching a steady state value corresponding to about 6% cyclohexane conversion after about 110 minutes. In the same figure the change of oxygen outlet concentration is also reported showing a general trend similar to that of cyclohexane. The analysis of products in the outlet stream disclosed the presence of benzene and (much lower amount) cyclohexene, as identified from the characteristic fragments m/z = 78, 77, 76, 74, 63, 52, 51, 50 (fragment 78 reported in Figure 17) and 82, 67, 54, respectively (fragment 67 reported in Figure 17), and of carbon dioxide, as detected by the NDIR analyser (Figure 17).


35

It is worthwhile to note that both breakthrough time of the products and change of concentration with time are different for the different products. Figure 17 shows that carbon dioxide is formed immediately after lamp switch on and reaches a maximum concentration value in about 30 minutes, slowly decreasing during the remaining run time. Instead, the outlet concentration of benzene progressively increases reaching a maximum value at greater time with respect to carbon dioxide. Then it decreases to a steady state value reached after about 220 minutes. A similar trend is shown by cyclohexene concentration, however the values are very much lower with respect to benzene.

In order to verify that cyclohexane was converted in a heterogeneous photocatalytic process, blank experiments were performed. A control test was carried out with the reactor loaded with uncoated quartz flakes. No conversion of cyclohexane was detected during this test, indicating the necessity of the catalyst for the observed reaction. A second test was performed with the catalyst loaded reactor, but leaving the lamps switched off even after cyclohexane adsorption equilibrium was reached. In these conditions the composition of the outlet reactor was identical to that of the reactor inlet, indicating that no reaction occurred in dark conditions.

These results confirm the occurrence of photocatalysed cyclohexane oxy-dehydrogenation to cyclohexene and benzene (Figure 18) together with deep oxidation to carbon dioxide.

Figure 18 Photocatalysed cyclohexane oxy-dehydrogenation to cyclohexene

and benzene

The catalytic performance was evaluated as: CH % conversion = 100·(moles of inlet CH – moles of outlet CH)/(moles

of inlet CH), BE % selectivity = 100·(moles of outlet BE)/(moles of inlet CH – moles

of outlet CH) CO2 % selectivity = 100·(moles of outlet CO2)/6·(moles of inlet CH –

moles of outlet CH), where CH is for cyclohexane and BE for benzene. Total carbon mass balance was evaluated by comparing the inlet carbon

as cyclohexane and the outlet carbon as the sum of unconverted cyclohexane and outlet benzene, cyclohexene and carbon dioxide.

Chapter III

36

After each photocatalytic test, catalyst was regenerated in air flow under UV irradiation for 16 h.

III.4.1 Thermodynamic analysis

Thermodynamic analysis has been carried out with Gaseq program that is written as a Microsoft Windows program with an easy graphic interface as shown in Figure 19.The basic principle of this program is the minimization of Gibbs free energy. Gaseq calculates chemical equilibrium in perfect gases.

Figure 19 Grafic interface of Gaseq 0.74 program

For calculations Gaseq uses the thermodynamic information on species that are provided in one or more files with the extension .tdd in the library.

A list of the types of calculation which can be performed is obtained by clicking the down arrow on the Problem Type box in the top left of the screen. The types available are:

� Equilibrium at defined temperature and pressure; � Adiabatic temperature and composition at defined pressure; � Equilibrium at defined temperature and constant volume;


37

� Adiabatic temperature and composition at constant volume; � Adiabatic compression/expansion; � Equilibrium Constant

For thermodynamic analysis the general reaction used is the following:

0.1C6H12 +0.15O2 +0.16H2O + 99.59N2→products

where the main products are C6H12, CO2, O2, N2, H2O, C6H6 and C6H10. The results of thermodynamic analysis have been reported in Figure 20 in

terms of cyclohexane conversion, benzene and carbon dioxide selectivity.

0

20

40

60

80

100

27 127 227 327 427 527 627

T, °C

co

nv

ers

ion

, %

0

20

40

60

80

100

120

se

lec

tiv

ity

, %

cyclohexane conversion

carbon dioxide selectivity

benzene selectivity

Figure 20 Effect of temperature on cyclohexane conversion, carbon dioxide

and benzene selectivity

Conversion increases with temperature and reaches about 100% at a temperature of about 330 °C. A similar trend is shown by benzene selectivity. Carbon dioxide selectivity decreases with temperature up to obout 0 % at a temperature of 327 °C. It is possible to see that benzene selectivity is higher than carbon dioxide selectivity for a temperature above 110 °C.

III.5 Results and discussion

III.5.1 Specific surface area and Chemical analysis

Chemical analysis (ICP-MS) was performed after microwave digestion (Ethos Plus from Milestone) of sample in H3PO4/HNO3 mixtures.

Chapter III

38

Table 2 contains the values of specific surface area and the results of chemical analysis (ICP-MS) of the samples as a percentage weight of MoO3.

Table 2 Specific surface area and MoO3 amount of MoOx/Al2O3 catalysts

Catalyst Nominal MoO3

content,

wt%

MoO3

by ICP analysis,

wt %

Specific

surface area

(BET), m2/g

αAl 0 - 10 2MoαAl 2.0 2.2 10 γAl 0 - 144 8MoγAl 8.0 7.6 147

It can be seen that surface areas of Mo loaded catalysts and unsupported aluminas are very similar. Moreover nominal MoO3 content is approximately equal to MoO3 evaluated by ICP-MS analysis. Considering 18% Mo loading as the upper for monolayer formation, as estimated by Inamura et al. (1998), a coverage degree of 42% was calculated for 8MoγAl, and. 108% was obtained for 2MoαAl “Imamura et al. (1998)”.

Specific surface area results together with MoOx coverage calculation suggest that molybdenum oxide is well dispersed on the alumina support and probably as a monolayer.

III.5.2 Thermal analysis

Measurements are carried out with SDTQ600 in air flow (100 Ncc/min) with a heating rate of 10 °C/min in the temperature range of 20- 700 °C.

Figure 21 shows the progression of the TG and DTG curves, together with the heat flow signal (DSC), during the decomposition of ammonium heptamolybdate used as precursor to impregnate alumina support.


39

-5.5

-0.5

DS

C,

W/g

-2

0

2

4

6

8

10

12

DT

G,

%/m

in

70

80

90

100

TG

, %

0 200 400 600

Temperature, °C

Figure 21 Evolution of the weight loss during decomposition of ammonium

heptamolybdate together with DTG and DSC signals

Up to 600 °C four decomposition steps at 127, 200, 295, and 385 °C, characterize the evolution of the weight loss, with continuous weight loss between the first and the third decomposition steps. The weight loss after the third decomposition step at 295 °C amounts to 17.1%, and after the fourth step at 385 °C, to 18.10%. The latter value is in agreement with the calculated weight loss of 18.48%, if MoO3 is assumed to be the decomposition product of the starting material. During the first decomposition step, an endothermic DSC signal can be distinguished. The third decomposition step also shows endothermic character, while the second and fourth steps are exothermic. With the exception of the first step, which begins with the evolution of water only, all steps are accompanied by the evolution of water and ammonia “Wienold et al. (2003)”.

Figure 22 displays the loss of mass with temperature during thermogravimetric analysis on 8MoγAl catalyst after calcination.

TG

DTG

DSC

Chapter III

40

-0.2

0.0

0.2

0.4

0.6

0.8

DT

G,

%/m

in

90

92

94

96

98

100

TG

,%

0 200 400 600 800

Temperature, °C

Figure 22 Thermogravimetric analysis of 8MoγAl catalyst after calcination

The sample showed an initial weight loss with a DTG peak at about 66 °C due to the adsorbed water loss. Above 200 °C up to about 870 °C no significant weight loss was observed. The TG results suggest that the dehydration and thermal decomposition of the precursor to form MoO3/γ−Al2O3 can be finished at 400 °C which is in agreement with the literature “Wienold et al. (2003)”. Similar results are obtained on 2MoαAl.

III.5.3 FT-IR spectroscopy

It is generally noted that Mo species formed on the surface of alumina are isolated etrahedral Mo, octahedral polymolybdate, and MoO3. Mo is present in the forms of heptamolybdate (Mo7O24

6-) and monomeric MoO42-. MoO4

2- is preferentially adsorbed on the basic hydroxyl groups present on alumina and, thus, isolated tetrahedral species are formed at very low Mo-loading “Jeziorowski and Knozinger (1979), Diaz and Bussel (1993)”. After consumption of the hydroxyl groups octahedral polymolybdate is produced “Jeziorowski and Knozinger (1979), Giordano et al.(1975), Wang and Hall (1980), Acro et al.(1992)”, and when the monolayer is completed Al2(MoO4)3 begins to be formed, followed by subsequent formation of MoO3 crystallites at higher Mo-loading “Medema et al.(1978)”. Spectroscopic techniques give more direct and useful information on the structures of Mo.

Figure 23 shows the IR spectrum of 8MoγAl measured by a transmission mode. The original spectrum has been subtracted from the spectrum due to γAl alone.

TG

DTG


41

550650750850950105011501250

Wavenumber, cm-1

Tra

ns

mit

tan

ce,

a.u

.

Figure 23 IR spectrum of 8MoγAl obtained by subtracting the spectrum of

γAl

Only one broad absorption band is observed at about 914 cm-1 in the Mo=O vibrational stretching region. This band is not due to MoO3 (bands at 599, 821, 873 and 995 cm-1) and is attributed to surface-bound Mo species, like octahedral (Oh) polymolybdate or isolated tetrahedral (Td) species. Since no band of bridged Mo–O–Mo species in the lower wavenumber region is present, this band is ascribed to isolated Td–Mo “Imamura et al. (1998)”.

IR spectra of αAl and 2MoαAl are contained in Figure 24.

Chapter III

42

400500600700800900100011001200

Wavenumber, cm-1

Tra

nsm

itta

nce,

a.u

.

αααα Al

2Moαααα Al

Figure 24 IR spectra of αAl and 2MoαAl

Visible IR Mo-related bands in the spectrum of 2MoαAl could not be obtained causing to the strong absorption by αAl, and no detailed information can be attained about MoOx dispersion.In addition, altough Mo species cover the support surface, sharp bands from MoO3 crystallites are not observed.

III.5.4 Micro Raman spectroscopy

Better evidence for Mo surface species was obtained by Micro Raman spectroscopy.

Raman spectra of γAl and 8MoγAl are reported in Figure 25.


43

150000

200000

250000

300000

350000

400000

400 600 800 1000 1200

Raman shift, cm-1

Co

un

ts

Figure 25 Raman spectra of γAl and 8MoγAl

It can be seen that γAl do not exhibit any Raman band due to the low polarizability of light atoms and the ionic character of the Al-O bonds “Wachs (1996)”. On 8MoγAl catalyst the main Mo=O band is at 949 cm-1. Other two bands are present as shoulders at 920 and 980 cm-1. The bands at 673, 822, and 999 cm-1 for MoO3 crystallites are absent.

Figure 26 shows the Raman spectra of αAl and 2MoαAl catalyst in the range 900-1100 cm-1.

γγγγAl

8MoγγγγAl

Chapter III

44

40000

60000

80000

100000

120000

900 950 1000 1050

Raman shift, cm-1

Counts

Figure 26 Raman spectra of αAl and 2MoαAl

Similar to γAl, also αAl do not present Raman signals. On 2MoαAl a band at 925 cm-1 and a main Mo=O band at 955 cm-1 are visible. There is also a band at 980 cm-1 present as a shoulder. Moreover the presence of MoO3 crystallites is not detected.

As several authors report “Wachs (1996), Bian et al. (1999)”, on Mo-alumina catalysts Mo=O bands near 950 cm-1 are characteristic of Mo=O stretching mode of polymolybdenyl species.

III.5.5 Photocatalytic activity tests

To better understand the contribution of the metal, preliminary tests were performed on γAl and on αAl. The obtained results evidenced that the both alumina supports didn’t exhibit any photoactivity.

Figure 27 and Figure 28 show the comparison between cyclohexane conversion and benzene, cyclohexene and CO2 selectivity as a function of illumination time on 8MoγAl (8Mo) and 2MoαAl (2Mo).

ααααAl

2MoααααAl


45

0

0.05

0.1

0.15

0.2

0.25

0.3

0 10 20 30 40 50

illumination time, min

cy

clo

he

xa

ne

co

nv

ers

ion

, %

8Mo

2Mo

Figure 27 Cyclohexane conversion on 8MoγAl (8Mo) and 2MoαAl (2Mo) as

a function of illumination time

Chapter III

46

0

20

40

60

80

100

120

0 10 20 30 40 50


se

lec

tiv

ity

, %

benzene 8Mo

cyclohexene 8Mo

CO2 8Mo

benzene 2Mo

cyclohexene 2Mo

CO2 2Mo

Figure 28 Selectivity to benzene cyclohexene and CO2 on 8MoγAl (8Mo) and

2MoαAl (2Mo) as a function of illumination time

Photocatalytic activity test showed that very low cyclohexane conversion, about 0.25%, was obtained on both catalysts (Figure 27), but the selectivity to CO2 was close to 100% (Figure 28). Only few ppm of cyclohexene and benzene were detected on 8MoγAl in the first two minutes of reaction (Figure 28).

IV Experimental Results:


cyclohexane on zeolites

supported MoOx

IV.1 Samples preparation

Zeolitic catalysts was prepared starting from Na,K-Ferrierite, with Si/Al ratio of 8.4 and NaY, HY (Engelhard). The characteristics of the Na, K-ferrierite are shown in Table 3.

Figure 29 Ferrierite structure

Chapter IV

48

Table 3 Characteristics of Na,K-Ferrierite

Bulk Density, g/cm3

0.40

Porous Diameter, Å

4.0

SiO2, dry wt %

84.9

Al2O3, dry wt %

8.6

Na2O, dry wt %

1.5

K2O, dry wt %

5.6

K2O/Al2O3

0.7

Na2O/Al2O3

0.28

SiO2/Al2O3

16.8

Ferrierite was ion exchanged to ammonium form with 1 M solution of ammonium nitrate in order to obtain the NH4 form (AFer). Powdered catalysts were prepared by wet impregnation of Afer, or HY or NaY with an aqueous solution of ammonium heptamolybdate (NH4)6 Mo7O24·4H2O, drying at 120 °C for 12 hours and calcination in air at 550 °C for 3 hours. Table 4 contains the list of samples prepared with indications of the nominal MoO3 load.



AFer - 5MoAFer 5.0 20MoAFer 20.0 NaY - 20MoNaY 20.0 HY - 20MoHY 20.0

IV.2 Thermal analysis

Thermogravimetric analyses were carried out on zeolites used as support.

Experimental Results: Photocatalytic oxidation of ciclohexane on zeolites supported MoOx

49

Figure 30 shows the TG and DTG curves of ferrierite in ammonium form (AFer) and of hydrogen ferrierite (HFer) obtained after calcination of AFer at 550 °C in air for 2 hours.

88

90

92

94

96

98

100

102

25 225 425 625 825Temperature, °C

TG

, %

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

DT

G,

%/m

in

Figure 30 Thermogravimetric analysis of AFer and HFer

As evidenced by the DTG minima, the two samples showed an initial weight loss at temperatures lower than 120 °C due to the adsorbed water loss. AFer sample showed a second weight loss in the 200 – 500 °C range, which is absent in HFER, due to ammonium decomposition. In both samples, another weight loss is present around 800 °C, due to water formed by hydroxyl condensation with the consequent collapse of the zeolite structure. Figure 31 shows thermogravimetry of NaY zeolite.

TG AFer

TG HFer

DTG AFer

DTG HFer

Chapter IV

50

72

76

80

84

88

92

96

100

25 225 425 625

Temperature, °C

TG

, %

-1.4

-1.2

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

0.2

DT

G,

%/m

in

Figure 31 Thermogravimetric analysis of NaY

The only pronounced peak on the sample maximises at 99 °C and is associated with the removal of water molecules present either in cages or channels of the zeolite.

Figure 32 shows the TG and DTG curves of HY zeolites.

TG

DTG


51

76

80

84

88

92

96

100

25 125 225 325 425 525 625 725

Temperature, °C

TG

, %

-1.6

-1.4

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

DT

G,

%/m

in

Figure 32 Thermogravimetric analysis of HY

It can be seen that there is only a DTG peak at around 80 °C caused by the loss of zeolite-adsorbed water.

IV.3 N2 adsorption measurements

Surface area and porosity characteristics were obtained by N2 adsorption at -196 °C. The microporous volume was obtained by the Dubinin method and its value for all the samples are shown in Table 5.

Table 5 Microporous volume of zeolites based samples

Catalyst Microporous volume, cm3/g

AFer 0.130 5MoAFer 0.045 20MoAFer 0.024 NaY 0.261 20MoNaY 0.084 HY 0.206 20MoHY 0.024

The results obtained evidenced that the addition of molybdenum oxide molybdenum on zeolite structure, leads to a microporous volume reduction

TG

DTG

Chapter IV

52

and in particular for MoOx/AFer samples, the microporous volume decreases with increase in Mo-loading.

IV.4 FT-IR spectroscopy

FT-IR spectra of HFer, 5MoAFer and 20MoAFer are contained in Figure 33. On all the samples, typical absorption bands of ferrierite in the range 400-1200 cm-1 are visible “Corbo et al. (1994), Pirone et al. (1996)”. Absorption at 1072 cm-1 is assigned to asymmetric stretching of SiO4, AlO4 tetrahedrons, while at 802 cm-1 to symmetric vibration. Bands at 594 and 533 cm-1 are related to double ring vibrations. Pore opening bands are at 461 and 437cm-1.

400500600700800900100011001200

Wavenumber, cm-1

Tra

ns

mit

tan

ce

, a

.u.

Figure 33 FT-IR spectra of ΗFer 5MoAFer and 20MoΑFer

Both 5MoAFer and 20MoAFer show a characteristic band at around 920 cm-1 probably due to Mo=O vibration of tetrahedral species MoO4

2- “Xu et

al. (1995)” and the intensity of this band increases with Mo content. On 20MoAFer catalyst additional bands at 963 cm-1 (assigned to polymolybdate) at 995 cm-1 and around 868 cm-1 (chracteristics of MoO3 crystallites) can be observed “Afanasiev et al. (1994)”.

FT-IR spectra of NaY and 20MoNaY are reported in Figure 34.

HFer

5MoAFer

20MoAFer


53

400500600700800900100011001200

Wavenumber, cm-1

Tra

nsm

itta

nce, a.u

.

Figure 34 FT-IR spectra of NaY and 20MoNaY

As it can be seen, with respect to unsupported NaY, FT-IR spectra of 20MoNaY show additional bands in the region 840-945 cm-1.due to the presence of Mo-species. It is possible to notice, at 921 cm-1, the same band present in the FT-IR spectrum of 20MoAFer catalyst, assigned to Mo=O vibration of MoO4

2- species between Mo and structural oxygen of the zeolite. Dimeric complexes in tetrahedral coordination show vibrational bands in the range 873–913 cm–1 and 919–943 cm–1 “Xu et al. (1995)”. Therefore, vibrational bands present around 881 cm–1, 894 cm–1, 906 cm–1 and 941 cm–1 are probably due to such species present on catalyst surface. Finally, the shoulder present at 868 cm-1 indicates the presence of MoO3 crystallites.

FT-IR spectra of HY and 20MoHY are reported in Figure 35.

NaY

20MoNaY

Chapter IV

54

400500600700800900100011001200

Wavenumber, cm-1

Tra

ns

mit

tan

ce

, a

.u.

Figure 35 FT-IR spectra of HY and 20MoHY

With respect HY, FT-IR spectrum of 20MoHY catalyst shows a shoulder around 929 cm-1 due to Mo=O vibration of MoO4

2- species. The other bands, present around 889 cm-1, 902 cm-1, 929 cm-1, are probably due to dimeric complexes in tetrahedral coordination. The shoulder at 954 cm-1 is assigned to Mo=O vibrational stretching. Finally the shoulder around 1006 cm-1 indicates the presence of MoO3 crystallites on catalyst surface.

For all prepared catalysts, FT-IR studies have evidenced penetrations of the Mo species into the zeolite cages during calcination, since it was detected the presence of Mo oxide species in tetrahedral coordination, characteristic of the migration of Mo species into zeolite “Corma et al. (1988)”.

IV.5 Photocatalytic activity tests

Cyclohexane and CO2 concentrations as functions of time during a photocatalytic test on AFer are reported in Figure 36.

HY

20MoHY


55

0

200

400

600

800

1000

1200

0 50 100 150 200

time, min

pp

m Cyclohexane

CO2

lamp on lamp off

Figure 36 Outlet reactor concentration of cyclohexane and carbon dioxide

as a function of run time

Loading 20 g of catalyst and feeding 1000 ppm cyclohexane and 3% oxygen in N2 (total flow rate: 500 Ncc/min) to the reactor in the absence of irradiation, adsorption of cyclohexane was observed . When the lamps were switched on, the cyclohexane outlet concentration immediately decreased of about 20% with respect to the inlet value and then progressively increased with run time, reaching a steady state value corresponding to about 15% of cyclohexane conversion after about 120 minutes with a CO2 production of 220 ppm and no partial oxidation product were detected.

Photocatalytic tests were also carried out on 5MoAFer and 20MoAFer. In these cases, loadings 20 g of catalyst and feeding 100 ppm cyclohexane, 1500 ppm oxygen and 1600 ppm water in N2 (total flow rate: 830 Ncc/min), in the steady state, cyclohexane conversion reached lower values (less than 1%), but the analysis of the reaction products disclosed the presence of benzene and cyclohexene, as identified from their characteristic fragments (Figure 37) with together carbon dioxide. These results demonstrate that photocatalysed cyclohexane oxy-dehydrogenation to cyclohexene and benzene occurs on molybdenum-supported ferrierite since AFer alone exhibits high activity only in total oxidation to carbon dioxide but is not active for conversion to benzene and cyclohexene.

Chapter IV

56

20 40 60 80 100 120 140 160 180 200

time, min

MS

sig

na

l, a

.u. m/z = 78

m/z = 67lamp on lamp off

Figure 37 Outlet reactor concentration (a.u.) of benzene and cyclohexene on

5MoAFer as a function of run time

The influence of Mo loading on AFer was evaluated and obtained results are summarized in Figure 38, Figure 39 and Figure 40.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0 10 20 30 40 50 60 70 80


cyclo

hexan

e c

on

vers

ion

, %

20MoAFer

5MoAFer

Figure 38 Cyclohexane conversion on 5MoAFer and 20MoAFer as a

function of illumination time


57

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60 70 80


be

nze

ne

se

lec

tiv

ity

, %

20MoAFer

5MoAfer

Figure 39 Benzene selectivity on 5MoAFer and 20MoAFer as a function of

illumination time

0

0.5

1

1.5

2

2.5

3

3.5

4

0 10 20 30 40 50 60 70 80


cyc

loh

ex

en

e s

ele

cti

vit

y,

%

20MoAFer

5MoAFer

Figure 40 Cyclohexene selectivity on 5MoAFer and 20MoAFer as a function

of illumination time

On both catalysts, after an illumination time of 80 minutes, cyclohexane conversion obtained was less than 1% (about 0.6 % on 5MoAFer and 0.2 % on 20MoAFer). It reached a maximum value after about 15 minutes and then decreased to a steady state value, evidencing a catalyst deactivation.

Chapter IV

58

Selectivity to benzene reached a very great steady state value (about 80%) on 20MoAFer while this value was 27 % on 5MoAFer. On 20MoAFer benzene selectivity showed a value higher than 90 % after about 10 minutes and then progressively decreased with illumination time. Instead on 5MoAFer, there was a progressive increase up to a steady state value. Similar behaviours were obtained for cyclohexene selectivity and its steady state value was very low for both catalysts (about 2 % on 5MoAFer and 2.5 for 20MoAFer). In summary, there was a decrease of cyclohexane conversion and an increase of benzene and cyclohexene selectivity by increasing Mo loading.

Finally, no photoactivity was obtained on HY, 20MoHY, NaY and 20MoNaY showing that it is important for system selectivity not only the presence of molybdenum on catalyst surface but also the type of zeolite used as support because it influences the distribution of Mo species on catalyst surface (as FT-IR spectra showed) and then its activity.

V Experimental Results:


cyclohexane on MoOx/TiO2

V.1 Effect of molybdenum loading

V.1.1 Samples preparation

TiO2 as anatase phase, containing 2 wt% sulphates (DT51, Rhone Poulenc) was used as support. The impregnation process was performed by means of aqueous solutions of ammonium heptamolybdate (NH4)6 Mo7O24·4H2O of different concentrations in order to obtain TiO2-based catalysts with different molybdenum percentages. Powder samples were dried at 120 °C for 12 hours and calcined in air at 400 °C for 3 hours.

The following table reports the list of catalysts with their nominal MoO3 contents.

Table 6 List of catalysts with their MoO3 nominal contents


DT2 - 2MoDT2 2.0 4MoDT2 4.7 8MoDT2 8.0

V.1.2 Specific surface area and Chemical analysis

Chemical analyses of Molybdenum load were performed after microwave digestion of catalyst in HNO3/HCl and HF/HCl mixtures. The list of samples with their surface areas and the results of chemical analyses in comparison

Chapter V

60

with nominal metal loadings with together theoretical MoOx surface coverage degree are reported in Table 7.

Table 7 List of catalysts and their characteristics

Catalyst Nominal

MoO3

content,

wt%

MoO3

by ICP

analysis,

wt %

Specific

surface

area

(BET),

m2/g

Theoretical

MoOx surface

coverage degree,

%

DT2 - - 71 - 2MoDT2 2.0 1.8 71 21 4MoDT2 4.7 4.4 68 54 8MoDT2 8.0 7.6 63 100

It can be seen that MoO3 loadings evaluated by ICP analysis are approximately equal to that one calculated for the impregnation process. MoOx coverage, calculated from the analysed MoO3 loading and assuming a monolayer capacity of 0.12% (w/w) for MoO3/m

2 “Ng and Gulari (1985), Kim et al. (1989)”, is 54% for 4MoDT2, 21% for 2MoDT2 and 100% for 8MoDT2.

Specific surface areas (S.S.A.) of the catalysts (m2/g) are plotted against measured MoO3 loading in Figure 41.

50

55

60

65

70

75

0 1 2 3 4 5 6 7 8MoO3, wt %

S.S

.A,

m2/g

Figure 41 Variation of specific surface area with MoO3 loading

Experimental results: Photocatalytic oxidation of cyclohexane on MoOx/TiO2

61

It can be seen that surface area of Mo loaded catalysts showed a decreasing trend with increasing Mo content. It is noted that it remained constant up to 1.8 MoO3 wt% and started to decrease with further increases in Mo-loading perhaps due to pore blockage.

V.1.3 Thermal analysis

Figure 42 shows TG-MS results on DT2 sample.

-0.1

0.0

0.1

0.2

DT

G, %

/min

95

96

97

98

99

100

101

TG

, %

0 200 400 600 800 1000

Temperature, °C

Figure 42 TG-MS results on DT2 sample

The first main step of weight loss below 150 °C is associated with hydration water desorption. The second step (present as a shoulder) that occurred up to about 390 °C, is related to the removal of OH- surface groups of titania. Morishige (1985) reported that TiO2 as the anatase phase undergoes the removal of most surface OH- groups in a temperature range extending up to 400 °C. The last weight loss, located at temperature higher than 600 °C, is attributed to the decomposition of sulphate species giving rise to gaseous SO3 as identified from its characteristic fragments m/z = 48 and 64.

Results of TG-MS analysis of calcined 2MoDT2, 4MoDT2 and 8MoDT2 catalysts are reported in Figure 43, Figure 44 and Figure 45.

TG

DTG

Chapter V

62

-0.2

0.0

0.2

0.4

0.6

DT

G,

%/m

in

95

96

97

98

99

100

TG

, %

0 200 400 600 800

Temperature,°C

Figure 43 TG-MS results on 2MoDT2 sample

TG

DTG


63

-0.2

0.0

0.2

0.4

0.6

DT

G,

%/m

in

90

92

94

96

98

100

102

TG

, %

0 200 400 600 800

Temperature, °C


-0.1

0.0

0.1

0.2

0.3

DT

G,

%/m

in

90

92

94

96

98

100

TG

, %

0 200 400 600 800

Temperature, °C


For all MoDT2s, two main complex stages of weight loss, respectively under and over 400°C, can be detected. The first main step (up to about

TG

DTG

TG

DTG

Chapter V

64

390°C) is associated to water and titania hydroxyl group removal, while the second main step is due to decomposition of different kinds of surface sulphates (as fragments m/z = 48 and 64 evidenced). Moreover it was found that samples didn’t show the weight loss due to the presence of ammonium ion. Sulphate stability is affected by the presence of the metal oxide on titania “Ciambelli et al. (1996)”. The thermogravimetric curves of DT2, 2MoDT2, 4MoDT2 and 8MoDT2 showed that the sulphate weight loss step in the catalysts containing molybdenum started at temperatures lower than in sulphated titania.

The hydroxyl groups and surface sulphates (as SO42-) density (evaluated

in the range 180-350°C and 400-800 °C respectively) are presented in Table 8.

Table 8 Hydroxyls and surface sulphates density

Catalyst Hydroxyls density,

mmol/m2

SO42-

density,

mmol/m2

DT2 0.028 0.0031 2MoDT2 0.015 0.0029 4MoDT2 0.018 0.0029 8MoDT2 0.017 0.0033

It is possible to observe that the introduction of Mo species on titania caused a decrease of hydroxyls density. All MoDT2s catalysts showed the same amount of hydroxyls/m2, despite the decrease of the specific surface area. This probably means that the introduction of molybdenum oxide on the titania surface occurred with the formation of new surface hydroxyls “Sorrentino et al. (2001)”. From the data reported in Table 8 it can also be seen that surface sulphates density is similar for all catalysts.

V.1.4 FT-IR spectroscopy

The FT-IR spectra as a function of Mo-loadings are shown in Figure 46.


65

700800900100011001200

Wavenumber, cm-1

Tra

ns

mit

tan

ce

, a

.u.

Figure 46 FT-IR spectra of DT2, 2MoDT2, 4MoDT2 and 8MoDT2

In the spectrograms, the 800–1200 cm−1 region is interesting as in this region characteristic bands due to Mo species can be detected. For pure titania, two broad and strong bands appear at 1049 and 1132 cm−1 which are attributed to S=O vibrations of the free sulphate groups “Samantaray and Parida (2001)”. FT-IR spectra of TiO2 supported catalysts showed Mo-related bands superimposed upon typical absorptions from the support. For all Mo catalysts no MoO3 crystallites sharp bands (Mo=O stretching vibrations at 992 cm-1 and bulk vibrations and 820 cm-1) are present “Matralis et al. (1995)”. A band at 954 cm-1 on 2MoDT2, at 957 cm-1 on 4MoDT and at 963 cm-1 on 8MoDT is observed. The bands near 960 cm-1 can be assigned to the stretching mode of molybdenyl species in hydrated form “Matralis et al. (1995), Lietti et al. (1996)”. In particular, they have been assigned to terminal Mo=O stretching of octahedral polymeric surface species “Matralis et al. (1995)”. The results indicate that the formation of

DT2

2MoDT2

4MoDT2

8MoDT2

Chapter V

66

octahedral polymeric molybdate is observed and the amount of this polymeric octahedral molybdate increases with increase in Mo-loading. Ng and Gulari (1985) have reported from their IR and Raman experiments that at lower loading, the majority of the surface molybdates are in tetrahedral coordination and at monolayer coverage, octahedrally coordinated polymeric surface species are formed. Beyond a monolayer, bulk molybdenum trioxide appears. Quincy et al. (1989) have also reported similar results.

V.1.5 Micro Raman spectroscopy

Raman spectrum of DT2 sample after calcination is presented in Figure 47.

0

50000

100000

150000

200000

250000

300000

100 200 300 400 500 600 700 800

Raman shift, cm-1

Co

un

ts

Figure 47 Raman spectrum of DT2 sample

The sample display bands at 144, 396, 514, 637 cm-1 with a very weak shoulder at 195 cm-1 due to the Raman active fundamentals of anatase “Alemany et al. (1995), Osaka et al. (1978)”. The supported molybdenum oxide phase on titania introduced a new Raman feature at Raman shift around 960 cm-1. The Raman spectra of the MoDT2s catalysts in comparison with DT2 and MoO3 spectra in the range 700-1200 cm-1 are shown in Figure 48.


67

200000

300000

400000

500000

600000

700000

800000

700 800 900 1000 1100 1200

Raman shift, cm-1

Co

un

ts

Figure 48 Raman spectra of DT2, 2MoDT2, 4MoDT2, 8MoDT2 and MoO3

samples

The weak band at 790 cm-1 is due to the first overtone of the 396 cm-1 band of TiO2 (anatase). MoO3 is characterized by the prominent peaks at 997 and 821 cm-1 which are attributed to the stretching vibrations of the terminal Mo=O and the brindging Mo-O-Mo bonds respectively “Hirata (1989)”. Supported molybdenum oxide phase on titania introduced a new Raman feature at Raman shift around 960 cm-1. In particular, Raman spectra of MoDT2s catalysts show a main band at 953 cm-1 on 2MoDT2 and at 958 cm-1 on 4MoDT2. On 8MoDT2 a complex band could be due to the overlapping of 956, 966, 978, 984 and 995 cm-1 peaks. The increasing in wavenumbers with increasing Mo loading has been attributed to higher polymerisation degree of Mo species “Cheng and Schrader (1979)”. The small peak at 995 cm-1 could indicate the incipient formation of segregated MoO3 crystallites. Therefore, the surface is mostly covered by polymeric octahedral species anchored at the surface for all MoDT2s catalysts.

V.1.6 Photocatalytic activity tests

Cyclohexane conversion and CO2 concentration obtained on DT2 catalyst are shown in Figure 49 and Figure 50 respectively.

DT2

2MoDT2

4MoDT2

8MoDT2

MoO3

Chapter V

68

0

5

10

15

20

25

30

0 10 20 30 40 50 60 70 80 90 100 110


cyc

loh

ex

an

e c

on

ve

rsio

n,

%

Figure 49 Cyclohexane conversion on DT2 as a function of illumination

time

Maximum cyclohexane conversion was about 25 %, decreasing to 3% in 30 minutes. A steady state condition was obtained after about 90 minutes of illumination with a conversion of approximately 2 %.


69

0

20

40

60

80

100

120

140

0 10 20 30 40 50 60 70 80 90 100 110


CO

2,

pp

m

Figure 50 Carbon dioxide concentration on DT2 as a function of

illumination time

Carbon dioxide was formed immediately after the lamps were switched on and reached a concentration of about 115 ppm after an illumination time of 110 minutes. CO2 and water were the only observed products and no other reaction products were detected. Thus, photocatalytic oxidation of cyclohexane on unsupported titania leads to a complete mineralization into CO2 and H2O without formation of by-products, according to the following reaction:

C6H12+9O2=6CO2+6H2O The comparison of cyclohexane conversion over 2MoDT2, 4MoDT2 and

8MoDT2 is shown in Figure 51.

Chapter V

70

0

5

10

15

20

25

30

35

40

45

50

0 10 20 30 40 50 60 70 80 90 100 110


cyclo

hexan

e c

on

vers

ion

, %

8MoDT2

4MoDT2

2MoDT2

Figure 51 Cyclohexane conversion on MoDT2s as a function of illumination

time

On all catalysts a maximum value was reached after about 5 minutes, then activity decreased approaching a steady state conversion. On 2MoDT2 maximum cyclohexane conversion was about 45 %, decreasing to about 20 % in 30 minutes. On 4MoDT2 maximum conversion was lower (about 21% after 8 minutes of illumination). It was about 7% after 30 minutes and 6% after 110 minutes. Increasing Mo loading up to 8 wt% MoO3 the initial maximum conversion was lower, about 15%, while steady state conversion was 2.3% after 30 minutes. Therefore the progressive coverage of the titania surface by MoOx species resulted in decreased initial and steady state cyclohexane conversions.

While on DT2 the only reaction product was CO2, all MoOx/TiO2 catalysts exhibited unexpected high selectivity to benzene. On 2MoDT2 selectivity to benzene was about 10% after 110 minutes. On 4MoDT2 (Figure 52) maximum selectivity to benzene reached 31%. On 8MoDT2 higher selectivity to benzene was observed (65% after 40 minutes). On all catalysts the presence of very low amounts of cyclohexene in the reaction products was detected (steady state selectivity was about 0.4% on 2MoDT2, 0.7% on 4MoDT2 and 1.5% on 8MoDT2).


71

0

10

20

30

40

50

60

70

0 10 20 30 40 50 60 70 80 90 100 110


be

nze

ne

se

lec

tiv

ity

, %

8MoDT2

4MoDT2

2MoDT2

Figure 52 Selectivity to benzene on MoDT2s as a function of illumination

time

Selectivity to CO2 on MoDT2s as a function of illumination time is reported in Figure 53.

Chapter V

72

0

5

10

15

20

25

30

0 10 20 30 40 50 60 70 80 90 100 110


CO

2 s

ele

cti

vit

y, %

8MoDT2

4MoDT2

2MoDT2

Figure 53 Selectivity to CO2 on MoDT2s as a function of illumination time

It can be seen that selectivity to CO2 decreased with increasing molybdenum loading. On 2MoDT2, 4MoDT2 and 8MoDT2 maximum selectivity to carbon dioxide was about 27 %, 8 % and 5 % respectively.

In addition, a catalyst containing 12 wt% of MoO3 nominal content was prepared. In this case, FT-IR and Raman spectra revealed the presence of MoO3 cristallytes on catalyst surface and no photocatalytic activity was obtained.

Photocatalytic activity tests on all MoDT2s catalysts evidenced that the presence of MoOx species on the surface of titania changes the selectivity of the catalyst with increasing molybdenum content indicating that the interaction between titania and supported molybdenum oxide plays an essential role in changing the catalyst selectivity. FT-IR and Raman spectroscopy data coupled with photocatalytic activity results showed that the selective formation of benzene is likely due to the presence of polymolybdate species supported on the titania surface. It could be argued that polymolybdate species poison unselective sites of the titania surface which would lead to total oxidation of cyclohexane.

Moreover, Figure 52 and Figure 53 clearly show the different shapes of selectivity curves relevant to benzene and carbon dioxide, respectively. Carbon dioxide is likely formed immediately after lamp switch on mostly on titania sites, while benzene formation seems to require the formation on intermediate surface species, whose generation should be related to some interaction between support and molybdenum oxide surface species. A possible reaction mechanism will be discussed.


73

V.1.7 Effect of light intensity

Light intensity is an important parameter to consider. Increasing the light intensity affects the rate of the reaction by increasing the number of charge carriers generated in the semiconductor. Most researchers have found different effects at different levels of light intensity “Rendel (1987)”:

• at low light intensities, the rate increases in proportion to the light intensity;

• at intermediate light intensities the rate only varies with the square root of intensity “Ollis (1991), Okamoto et al. (1985)”;

• at high light intensities the rate of photodegradation is independent of light intensity.

For photocatalysis, as light intensity increases the rate increases due to the increased number of oxidising species that are produced. The rate increases with light intensity (I) to a power n. At low light intensities the reaction rate increases directly in proportion to light intensity suggesting that few oxidising species are lost through recombination processes. At high light intensities, the reaction rate increases in proportion to I to the power of 0, i.e. the rate becomes independent of light intensity and the expected rate-limiting factor is mass transfer. At intermediate light intensities the rate only varies with the square root of intensity, “Okamoto et al. (1985), Kormann et

al. (1991)” and hence efficiency suffers. This was attributed by Egerton and King (1979) to energy wasting recombination reactions between electrons and holes. Increased intensity always results in an increase in the volumetric reaction rate until the mass transfer limit is encountered. However, once intermediate light intensities are reached any increase in I will not lead to a proportional increase in rate. The I1 to I0.5 rate transition is said to depend on the catalyst material “Matthews et al. (1990)”.

The dependence of the cyclohexane conversion on light intensity has been investigated on 4MoDT2 catalyst. The change of cyclohexane conversion in the steady state condition as a function of light intensity is shown in Figure 54.

Chapter V

74

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

0 50 100 150 200 250 300light intensity, W

cy

clo

he

xa

ne

co

nv

ers

ion

, %

Figure 54 Effect of light intensity on cyclohexane conversion

Cyclohexane was unconverted in the absence of light and its conversion increased up to 4 % in correspondence of a light intensity equal to 280W. In all cases selectivity to benzene and carbon dioxide was 27 % and 8 % respectively. The obtained results suggest that cyclohexane is not converted in absence of irradiation. Moreover the relationship is linear at low light intensities up to a certain point (intermediate light intensity) where the cyclohexane conversion starts to level off (high light intensities). This would suggest that the cyclohexane conversion is proportional to the light intensity up to a certain point where it becomes independent by light intensity.

V.2 Influence of sulphate content

V.2.1 Samples preparation

Three titanias were used as supports: two commercial titania samples (DT and DT51, Rhone Poulenc) with different sulphate contents (respectively 0.5 wt % and 2 wt %) and an ultrafine sulphate-free titania produced by laser-pyrolisis “Curcio et al. (1991)”. The samples are named, respectively, T0, T5 and T20 with reference to the sulphate content. MoOx-based catalysts


75

were prepared by wet impregnation of titania with aqueous solution of ammonium heptamolybdate (NH4)6Mo7O24·4H2O, followed by drying at 120 °C and calcination at 400 °C for 3 hours. The list of catalysts is reported in Table 9



T0 - MoT0 4.7 T05 - MoT05 4.7 T20 - MoT20 4.7

V.2.2 Specific surface area, chemical analysis and sample acidity

Table 10 contains the values of specific surface area, the results of chemical analysis of the samples, the theoretical MoOx surface coverage degree and point of zero charge (ZPC) values.

Table 10 List of catalysts and their characteristics

Catalyst ZPC

pH unit MoO3

by ICP

analysis,

wt %

Specific

surface

area (BET),

m2/g

Theoretical

MoOx surface

coverage

degree,

% T0 6 - 88 - MoT0 3.8 4.5 88 44 T05 4.7 - 67 - MoT05 1.8 4.2 68 50 T20 2 - 71 - MoT20 1.7 4.4 68 54

Chemical analysis showed that the Mo loading is similar for all supported catalysts whereas theoretical MoOx surface coverage increased with sample sulphate content. Moreover specific surface area data suggest that for MoT0 and MoT05 catalyst, Mo-species are well dispersed on titania surface; for MoT20 catalyst, there is a small decrease of S.S.A. value between catalyst and the relevant support.

Mass titration method was used to estimate the acidity of sample powders. The point of zero charge, which describes the acidity of oxide materials, may be measured using potentiometric titration, mass titration, or measurement of the wetting angles. The mass titration method of ZPC characterization was initially proposed by Noh and Schwarz (1989). Their

Chapter V

76

studies showed that when mass fraction of oxide powder is increased in aqueous electrolyte then the pH value of suspension approaches the ZPC value of oxide powder.

The limiting pH value is independent of the initial pH of the electrolyte “Žalac and Kallay (1992)”. The electrolyte is usually nitrogen-bubbled to prevent the influence of carbon dioxide on the pH of the electrolyte. The electrolyte usually contains only monovalent ions such as Na+, K+, and Cl−. It is assumed that the sample powder is pure and insoluble in the electrolyte used. If the sample powder is contaminated then it is possible that the impurities radically influence on the obtained pH values. It is also possible that there is not a limiting pH value at all “Noh and Schwarz (1989), Žalac and Kallay (1992)”. Acidic contaminations tend to decrease pH, whereas basic contaminations increase pH. Other mass titration experiment have been performed by Reymond and Kolenda (1999) and Žalac and Kallay (1992). Reymond and Kolenda (1999) studied ZPC of several pure and mixed oxides. The ZPC values obtained for pure oxides were in agreement with the values determined by other methods (e.g., potentiometric titration). For impure samples the obtained pH values indicated the chemical nature of impurities. Žalac and Kallay (1992) extended the idea of mass titration for contaminated samples. In addition, for ordinary mass titration, they performed traditional acid–base titration for sample suspensions, in which the solid powder content was so high that the limiting pH value had been achieved. They showed that using acid–base titration it is possible to determine both the ZPC of the sample powder and the nature and degree of contamination. In this study the mass titration studies were performed using procedures described in Noh and Schwarz (1989). Shorter stabilization times after each powder addition (2 hours in this study) were used to minimize possible dissolution of sample powders.

Table 10 shows that the value of ZPC of T0 is 6, according to the amphoteric character of anatase titania. The presence of sulphate increases surface acidity, leading to ZPC value of 4.7 on T05. As the sulphate load increases, ZPC decreases to 2. The comparison of ZPC values of catalysts and the relevant support indicates that the presence of molybdenum confers, in all cases, strongest acidity than that of the relevant support.

V.2.3 Thermal analysis

Hydroxyls and sulphates amount were evaluated by TG-MS analysis in the range 180-350°C and 400-800°C, respectively. The obtained results can be summarized in the following table in which there are the values of hydroxyls and sulphates (as SO4

2-) density on catalysts surface.


77

Table 11 Hydroxyls and surface sulphates density

Catalyst Hydroxyls density,

mmol/m2

SO42-

density,

mmol/m2

T0 0.052 - MoT0 0.022 - T05 0.034 0.00086 MoT05 0.019 0.00084 T20 0.028 0.0031 MoT20 0.018 0.0029

From the data reported in Table 11, it is possible to observe that MoOx/TiO2 catalysts have a lower hydroxyls density than the corresponding support, whereas the values of SO4

2- density are similar and increase with the surface sulphates amount.

V.2.4 FT-IR spectroscopy

In Figure 55 FT-IR spectra of MoT0, MoT05 and MoT20 in the range 700-1200 cm-1 are reported.

Chapter V

78

700800900100011001200

Wavenumber, cm-1

Tra

ns

mit

tan

ce

, a

.u

Figure 55 FT-IR spectra of MoT0, MoT05, and MoT20

FTIR spectra of TiO2 supported catalysts show MoOx species bands superimposed upon typical absorptions from the support. Sulphate absorption main bands are located at 1050 and 1132 cm-1 on T05 and, with enhanced intensity, on T20. For all Mo-based catalysts no evidence for MoO3 crystallites (sharp bands from Mo=O stretching vibrations at 992 cm-1 and bulk vibrations and 820 cm-1) are present “Maity et al. (2001)”. Bands at 947 cm-1 on Mo/T0, at 951 cm-1 on Mo/T05 and at 957 cm-1 on Mo/T20 are

MoT20

MoT0

MoT05


79

observed, assigned to terminal Mo=O stretching of polymeric surface species “Cheng and Schrader (1979)”.

V.2.5 Micro Raman spectroscopy

Raman spectra of MoT0, MoT05 and MoT20 are contained in Figure 56.

40000

50000

60000

70000

80000

90000

100000

110000

120000

130000

700 800 900 1000 1100 1200

Raman shift, cm-1

Co

un

ts

Figure 56 Raman spectra of MoT0, MoT05 and MoT20 samples

It can be seen that beyond the weak band at 790 cm-1 due to the first overtone of the 396 cm-1 band of titania, Raman spectra evidenced a main signal at 950 cm-1 on Mo/T0, at 956 cm-1 on Mo/T05 and at 958 cm-1 on Mo/T20, all characteristic of octahedral MoOx species. The increasing of wavenumber can be attributed to a higher degree of polymerisation of these last species on catalyst surfaces.

V.2.6 Photocatalytic activity tests

In Figure 57, cyclohexane conversion on T0, T05 and T20 as a function of illumination time is reported. Photocatalytic tests performed on these samples showed that cyclohexane conversion reached a maximum after about 5 minutes of illumination time and then decreased for all catalysts to reach a steady state conversion of about 2%, 5% and 8% on T2, T05 and T20, respectively.

MoT20

MoT0

MoT05

Chapter V

80

0

5

10

15

20

25

30

35

40

0 20 40 60 80 100


cy

clo

he

xa

ne

co

nv

ers

ion

, %

T0

T05

T20

Figure 57 Cyclohexane conversion on T0, T05 and T20 as a function of

illumination time

In Figure 58 CO2 production on T0, T05 and T20 as a function of illumination time is reported.

0

100

200

300

400

500

600

0 20 40 60 80 100


CO

2, p

pm

T0

T05

T20

Figure 58 Carbon dioxide concentration on T0, T05 and T20 as a function



81

For all catalysts, carbon dioxide was the only product detected in the gas phase and started forming immediately after lamp on, reaching steady state values (510 ppm, 335 ppm and 110 ppm on T0, T05 and T20 respectively) after about 110 minutes. Thus cyclohexane conversion and CO2 yield decreased with the sulphate content.

These last results may be explained taking into account the reaction mechanism of gas-solid photocatalytic decomposition of cyclohexane to CO2 on titania catalyst (Figure 59) reported by Einaga et al.(2002).

Figure 59 Mechanism for oxidation of cyclohexane on titania

Irradiation of TiO2 with UV light generates highly reactive electron-hole pairs. The hole subsequently oxidizes the surface hydroxyl groups to form the OH radicals “Einaga et al. (2002)”. The reactivity of OH radicals toward hydrocarbons has been well investigated “Seinfeld and Pandis (1997)”. The OH radicals abstract the H atoms of saturated C–H bonds of cyclohexane. The resulting intermediate radicals are subsequently oxidized by molecular O2. These intermediates are decomposed to CO2 via the subsequent oxidation processes. This proposed mechanism evidences that the OH radicals are the active species for the cyclohexane photocatalytic oxidation on titania. Moreover it has been reported that the catalytic activity of TiO2 could be maintained indefinitely under an abundance of water vapour since it can be adsorbed on the catalyst surface to form surface hydroxyl groups “Alberici and Jardim (1997), Obee and Brown (1995)”. The surface hydroxyl groups exhibit their important influence on the photoreaction process by trapping the charge transfer reaching the catalyst surface to produce very reactive surface hydroxyl radicals. From these considerations it is possible to observe that rehydroxylation process of the catalyst surface is essential for activity of titania toward cyclohexane total oxidation. Xie et al. (2004), studying gas-solid photocatalytic oxidation of heptane on sulphated and unsulphated titania in presence of water vapour, found that the presence of SO4

2− may be detrimental to the rehydroxylation ability of the catalyst. This last observation could explain the behaviour of cyclohexane conversion and CO2 production reported in Figure 57 and Figure 58 respectively.

Chapter V

82

0

5

10

15

20

25

30

35

40

45

0 50 100 150 200


cy

clo

he

xa

ne

co

nv

ers

ion

, %

MoT0

MoT05

MoT20

Figure 60 Cyclohexane conversion on Mo/T0, Mo/T05 and Mo/T20 catalysts

as a function of illumination time

Cyclohexane conversion on Mo/T0, Mo/T05 and Mo/T20 is shown in Figure 60. The analysis of the outlet stream disclosed the presence of benzene, cyclohexene (less than 1 ppm) and CO2 for all catalysts. A maximum value of conversion was reached after about 5 minutes, then activity decreased approaching a steady state conversion. On Mo/T0 the maximum cyclohexane conversion was about 40 %, decreasing to 17% in 25 minutes, less quickly with respect to Mo/T05 and Mo/T20. On Mo/T20 the initial maximum conversion was lower (about 21% after 8 minutes of illumination), reached 7% after 30 minutes and 4% after 220 minutes. On Mo/T05 the initial maximum conversion was higher with respect to Mo/T20 (about 23%), while steady state conversion was about 1% after 60 minutes.


83

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40

cyclohexane conversion, %

be

nze

ne

se

lec

tiv

ity

, %

Mo/T0

Mo/T05

Mo/T20

Figure 61 Benzene selectivity versus cyclohexane conversion on Mo/T0,

Mo/T05 and Mo/T20

In Figure 61 benzene selectivity as a function of cyclohexane conversion is reported. Low benzene selectivity and high selectivity to CO2 were observed on Mo/T0 in the whole conversion range. On Mo/T05 benzene selectivity ranged from 35 to about 70%. On Mo/T20 benzene selectivity reached values higher than 80 % (86% with a cyclohexane conversion of 6%). The dependence of benzene selectivity on cyclohexane conversion for all catalysts is less strong with respect to the usually found effect in hydrocarbon catalytic partial oxidation processes especially in the case of Mo/T20, for which the selectivity to benzene is very weakly decreasing with conversion. Therefore, the presence of sulphate species on titania surfaces enhanced the benzene yield more at higher sulphate contents.

In the section V.1, it was shown that polymolybdate species change the photoactivity of titania, the higher polymerisation favouring the formation of benzene. The FT-IR and Raman results showed that highly polymerised MoOx species are present on Mo/T20 catalyst that gives the best benzene yield. Even if, due to the presence of sulphate, the higher polymerisation degree could be associated to the lower surface area available to molybdenum oxide species deposition, leading to a higher surface density, the whole results suggest an active role of sulphate in promoting the selectivity to benzene. A possible role of the sulphate in the reaction mechanism will be discussed.

Chapter V

84

V.3 Photocatalytic oxidative dehydrogenation of cyclohexane:

reaction mechanism

Gas-solid partial photooxidation of paraffins on titanium dioxide under

UV irradiation lead to formation of ketones and aldehydes “Walker et al. (1977), Djeghri and Teichner (1980), Bickley et al. (1973), Formenti and Meriaudeau (1971)”. An intermediate formation of an alcohol by the addition of atomic oxygen to the paraffin was postulated. The alcohol may be either directly oxidized into an aldehyde or ketone or dehydrated into an olefin which, in turn, is oxidized at the ethylenic bond into aldehydes (primary carbon) or ketones (secondary carbon).

On the basis of this reaction mechanism, cyclohexene detected in the outlet stream could be product by a dehydration step involving cyclohexanol which may be formed by oxidation of cyclohexane. Because of the possibility of cyclohexanol undergoing this reaction (dehydration into cyclohexene), it was considered to investigated the photooxidation of cyclohexanol and cyclohexene in order to elucidate the mechanism of gas-solid photocatalytic oxidative dehydrogenation of cyclohexane.

Catalytic tests were performed on MoDT2s catalysts under the same conditions as for the cyclohexane.

V.3.1 Photocatalytic oxidation of cyclohexanol

Carbon dioxide produced by the photocatalytic oxidation of cyclohexanol is reported in Figure 62.


85

0

50

100

150

200

250

0 30 60 90 120


CO

2,

pp

m

2MoDT2

4MoDT2

8MoDT2


8MoDT2 as a function of illumination time.

Maximum value of CO2 concentration was reached after about 3 minutes, then it decreased approaching a steady state value. Carbon dioxide yield decreased with the increasing molybdenum content: after 120 minutes, it was 64 ppm, 36 ppm and 20 ppm on 2MoDT2, 4MoDT2, 8MoDT2 respectively and on all these catalysts no formation of benzene and cyclohexene was observed.

The obtained results evidenced that under UV irradiation of the catalysts, in the presence of oxygen, cyclohexanol was simply oxidized into CO2 showing that it is not an intermediate compound of cyclohexane oxidative dehydrogenation reaction. The cyclohexanol oxidation rate decreased with molybdenum content (Figure 62) probably because Mo-species poison sites of titania surface which lead to total oxidation of cyclohexanol.

Many authors showed that there is the formation of carboxylic acids as intermediates during the photooxidation of alcohols both in gas and in liquid phase “Guillard et al. (2002), Pillai and Sahle-Demessie (2002)”. These species are strongly adsorbed on titania surface “Ekstrom and McQuillan (1999)” and their photodegradation in presence of oxygen leads to carbon dioxide production “Jiang et al. (2004)”. In particular, under UV irradiation, CO2 is formed by ethanol through a sequence of two step including oxidation of the alcohol to form adsorbed acetic acid and its following decarboxylation with formation of CO2”Coronado et al. (2003)”.

On the basis of this last reaction mechanism, the general scheme of the photooxidation of cyclohexanol involving titania surface may be expressed in a schematic way as follows:

Chapter V

86

Cyclohexanol may be oxidized into strongly adsorbed intermediate

compounds (probably carboxylic acids) which, in turn, are decarboxylated into CO2 (which desorbs from catalyst surface) and others adsorbed intermediate compounds. These last compounds are then oxidized and decarboxylated by a series of consecutive steps which lead to formation of CO2 detected in the gas phase during photocatalytic oxidation of cyclohexanol (Figure 62)

V.3.2 Photocatalytic oxidation of cyclohexene

As cyclohexene has been identified as a principal reaction product and possible reaction intermediate of photooxidation of cyclohexane, photocatalytic oxidation of cyclohexene as the reactant was also considered in the same condition. Cyclohexene conversion (Figure 63) over the three molybdena systems yielded benzene (Figure 64) and carbon dioxide (Figure 65) as the only detected products.

Another adsorbed intermediate compound

-CO2

(g)

Another adsorbed intermediate compound

-CO2

(g)

Etc.

Adsorbed intermediate compounds

(probably carboxylic acids) Cyclohexanol (ads)

+O2

+O2


87

0

2

4

6

8

10

12

0 20 40 60 80 100 120


cy

clo

he

xe

ne

co

nv

ers

ion

, %

8MoDT2

4MoDT2

2MoDT2

Figure 63 Cyclohexene conversion on 2MoDT2, 4MoDT2 and 8MoDT2 as a


Cyclohexene conversion reached higher value (10% after 110 minutes) on 8MoDT2 with respect to 2MoDT2 (about 1%) and with respect to 4MoDT2 (about 8%).

0

20

40

60

80

100

120

0 20 40 60 80 100 120


ben

zen

e, p

pm

8MoDT2

4MoDT2

2MoDT2

Figure 64 Benzene concentration on 2MoDT2, 4MoDT2 and 8MoDT2 as a


Chapter V

88

0

10

20

30

40

50

60

0 20 40 60 80 100 120


CO

2,

pp

m2MoDT2

4MoDT2

8MoDT2


8MoDT2 as a function of illumination time

Figure 64 shows that benzene production increased with molybdenum loading. In particularly, after 110 min, benzene concentrations were 7 ppm, 80 ppm and 100 ppm on 2MoDT2, 4MoDT2 and 8MoDT2 respectively. Carbon dioxide production (Figure 65) showed the opposite trend: it was higher on 2MoDT2 (50 ppm after 110 min) with respect to 4MoDT2 (12 ppm), and with respect to 8MoDT2 (6 ppm).

The nature of products obtained shows that the reaction is not limited to a complete oxidation into CO2. In fact, cyclohexene itself is also involved in a photocatalytic oxidative dehydrogenation reaction giving benzene. It has been reported that cyclohexene is photooxidized into CO2 by using titania as photocatalyst “Einaga et al. (2002)”. Moreover a number of examples of oxidative photocatalysts are based on the use of polyoxometallates, thanks to their ability to undergo photoinduced electron transfers without irreversible modifications “Ermolenko and Giannotti (1996), Maldotti et al. (1996), Duncan and Hill (1997), Tanielian (1998)” and to control efficiency and selectivity of the photocatalytic processes “Mizuno and Misono (1998)”. In particular, Maldotti et al. (2003), showed the photoexcited decatungstate is able to initiate the selective oxidation of the cyclohexene through hydrogen abstraction in aqueous media. A key step in the photocatalytic cycle under aerobic conditions is the subsequent reoxidation of the decatungstate by O2.

Therefore carbon dioxide (Figure 65) detected during photocatalytic tests on cyclohexene is due to its total oxidation on bare titania whereas oxidative dehydrogenation to benzene, is probably catalysed by octahedral polymolybdate on catalyst surface as results obtained by increasing molybdenum content suggested (Figure 64).


89

From these considerations, it is possible to illustrate a probable mechanism of reaction involving both molybdate and titania.

Total oxidation of cyclohexene on bare titania may be occurred through the mechanism showed by Einaga et al. (2002) and reported in Figure 66.

Figure 66 Mechanism for oxidation of cyclohexene on bare titania

The hole generated by irradiation of catalyst oxidizes the surface hydroxyl groups to form the OH radicals which abstract the H atoms of saturated C–H bonds of cyclohexene. The resulting intermediate radicals are subsequently oxidized by molecular O2. These intermediates are decomposed to CO2 via the subsequent oxidation processes. Simultaneously, the photoexcited octahedral molybdate is able to initiate the oxidative dehydrogenation of adsorbed cyclohexene through hydrogen abstraction (eq. 5) to form benzene which desorbs from catalyst surface (eq. 6).

C6H10(ads) + 4Mo7O246-/TiO2 → C6H6(ads) + 4HMo7O24

6-/TiO2 (5)

C6H6(ads) → C6H6(g) (6)

O2-(lattice) + h+ → O- (7)

HMo7O246-/TiO2 + O- → Mo7O24

6-/TiO2 + OH- (8)

OH- + h+ → OH. (9)

OH. + OH. → H2O2 (10)

H2O2 → H2O + ½ O2 (11)

Photoreduced molybdate is than regenerated by O- (eq 8) which is formed by the reaction between lattice oxygen and positive hole (eq. 7). OH- formed by reoxidation of reduced molybdate is involved in a series of reactions

Chapter V

90

which lead to formation of water (eq. 9-11). In summary, the formation of benzene may be occur mainly according to a photocatalytic cycle that involves the molybdate both in the oxidized and in the photoreduced forms (Redox mechanism).

On the basis of these results and considerations, it is possible to hypothesize a mechanism of oxidative dehydrogenation of cyclohexane. Cyclohexane may be oxidized into CO2 and H2O because of unselective sites of titania according to mechanism showed in Figure 59 while Mo-species may be responsible of oxidative dehydrogenation of cyclohexane to cyclohexene and cyclohexene further oxy-dehydrogenation to benzene according to the following reactions.

TiO2 + hν → e- + h+ (12)

Mo7O246-/TiO2

+ hν → Mo7O246-/TiO2 (polymolybdate photoexcited)

(13)


6-/TiO2

(14)

C6H10(ads) → C6H10(g) (15)


6-/TiO2 (5)

C6H6(ads) → C6H6(g) (6)

HMo7O246-/TiO2 + O- → Mo7O24

6-/TiO2+ OH- (8)

OH- + h+ → OH. (9)

OH. + OH. → H2O2 (10)

H2O2 → H2O + ½ O2 (11)

V.3.3 Role of the sulphate in the reaction mechanism

A possible active role of sulphate in promoting the selectivity to benzene was suggested. Sulphate present on catalyst surface could realize a hydrogen


91

abstraction from an adsorbed cyclohexane molecule due to its strong basic properties but the results shown in the Figure 58 reveal that there is no formation of benzene over sulphated unsupported titania suggesting that sulphate could interact with octahedral molybdate but not with cyclohexane according to following reactions:

TiO2 + hν → e- + h+ (12)

Mo7O246-/TiO2

+ hν → Mo7O246-/TiO2 (polymolybdate photoexcited)

(13)

O2-(lattice) + h+ → O- (7)


6-/TiO2

(14)

C6H10(ads) → C6H10(g) (15)


6-/TiO2 (5)

HMo7O246-/TiO2 + O- → Mo7O24

6-/TiO2+ OH- (8)

HMo7O246-/TiO2 + SO4

2-/TiO2 → Mo7O246-/TiO2 + SO3

2-/TiO2 + OH.

(16)

OH. + OH. → H2O2 (10)

H2O2 → H2O + ½ O2 (11)

Probably surface sulphate- reoxidizes reduced molybdate (eq. 16) with together lattice oxygen (eq. 8) increasing the concentration of Mo7O24

6- which increases the rate of the reaction between molybdate and cyclohexane (eq. 14) improving the production of dehydrogenated compounds (cyclohexene and benzene).

VI Photocatalytic

flat-plate reactor

In recent years, a flat-plate photocatalytic reactor was considered as a

kind of efficient solar photoreactor for commercial application because this reactor has a high surface-area-to-volume ratio in the photoreactor and can treat the polluted air and water quickly and efficiently “Turchi et al. (1993)”.

Photocatalytic cyclohexane oxidative dehydrogenation tests were carried out on MoDT2s catalysts by using a photocatalytic flat-plate reactor in order to verify the possibility of obtaining the same reaction products with different configurations of reactor.

VI.1 Experimental set up apparatus and photocatalytic tests

conditions

The photocatalytic plate reactor was realised in steel with a quartz window (reactor volume: 0.7 l). All the gas pipes (¼’’ e.d.) are of Teflon, connections are made with Swagelok unions and two, and three way Nupro valves. All the connections are in stainless steel to avoid hydrocarbon adsorption on the walls and any corrosion due to the presence of water. Oxygen and nitrogen were fed from cylinders, nitrogen being the carrier gas for cyclohexane and water vaporized from two temperature controlled saturators and by changing temperature and N2 flow, it is possible to obtain different concentrations in the reaction feed. To feed a correct flow, flowmeters for N2 and O2 are used. The quartz window of the reactor was illuminated by an UV light source (MEDIUM PRESSURE MERCURY LAMP, 400W, PHILIPS) in a dark box. The temperature reaction was 120°C. In order to control the reaction temperature, a heater system was installed under the reactor. The gas composition was determined by on line quadrupole mass detector (Genesys) that can analyze the inlet and outlet reactor gas, introduced into two different heated capillaries made from silica. The mass spectrometer was connected to a PC for data acquisition. A

Chapter VI

94

schematic picture of the photocatalytic flat-plate reactor is reported in Figure 67.

Gas OutGas IN

Figure 67 Photocatalytic flat-plate reactor

An aqueous slurry of catalyst powder was used to coat a steel plate. The coated particles were dried at 90 °C for 8 hours in order to remove the excess of physisorbed water. This treatment resulted in uniform coating well adhering to the steel plate. The amount of deposited catalyst, evaluated by weighing the steel plate before and after the coating treatment was 1 g. The still plate was then located in the reactor. Catalytic tests were carried out feeding 830 Ncc/min N2 stream containing 1 % cyclohexane, 1.5 % oxygen and adding 1.6 % water to minimise catalyst photodeactivation. Lamp was switched on after complete adsorption of cyclohexane on catalyst surface.

VI.2 Photocatalytic activity tests

The analysis of the reaction products in the outlet stream disclosed the presence of benzene, cyclohexene and CO2. In order to verify that cyclohexane was converted in a heterogeneous photocatalytic process, blank experiments were performed. A control test was carried out with the reactor loaded with uncoated aluminium plate. No conversion of cyclohexane was detected during this test, indicating the necessity of the catalyst for the observed reaction. A second test was performed with the catalyst loaded reactor, without switching on the lamp even after establishing the cyclohexane adsorption equilibrium. In these conditions the composition of the outlet reactor was identical to that of the reactor inlet, indicating that no reaction occurred in dark conditions. The results indicate that the photocatalysed oxy-dehydrogenation of cyclohexane to cyclohexene and benzene together with deep oxidation to carbon dioxide occurs irrespective of the reactor configuration.

The comparison of cyclohexane conversion between 2MoDT2, 4MoDT2 and 8MoDT2 is shown in Figure 68.

Photocatalytic flat plate reactor

95

0

2

4

6

8

10

12

14

16

18

20

0 10 20 30 40 50 60 70


cyclo

hexa

ne c

on

vers

ion

, %

2MoDT2

4MoDT2

8MoDT2

Figure 68 Cyclohexane conversion on MoDT2s catalysts as a function of

illumination time

On all catalysts a maximum value was reached after about 3 minutes, then activity decreased approaching a steady state conversion. On 2MoDT2 maximum cyclohexane conversion was 18 %, decreasing to about 4.5 % in 10 minutes. On 4MoDT2 maximum conversion was lower (about 10%). It was about 2% after 20 minutes. Increasing Mo loading up to 8 wt% MoO3 the initial maximum conversion was lower, about 15%, while steady state conversion was 2.3% after 30 minutes. Therefore the progressive coverage of titania surface by MoOx species resulted in decreased initial and steady state cyclohexane conversion.

While on DT2 the only reaction product was CO2, all MoOx/TiO2 catalysts exhibited unexpected high selectivity to benzene (Figure 69). On 2MoDT2 selectivity to benzene was about 3% after 35 minutes. On 4MoDT2 maximum selectivity to benzene reached 14%. On 8MoDT2 higher selectivity to benzene was observed (34% after 15 minutes). On all catalysts the presence of very low amounts of cyclohexene in the reaction products were detected.

Chapter VI

96

0

5

10

15

20

25

30

35

40

0 10 20 30 40 50 60 70


ben

zen

e s

ele

cti

vit

y,

%2MoDT2

4MoDT2

8MoD2

Figure 69 Selectivity to benzene on MoDT2s catalysts as a function of

illumination time

Selectivity to CO2 on MoDT2s catalysts as a function of illumination time is reported in Figure 70.

0

2

4

6

8

10

12

0 20 40 60 80


CO

2 s

ele

cti

vit

y,

%

2MoDT2

4MoDT2

8MoDT2

Figure 70 Selectivity to CO2 on MoDT2s catalysts as a function of

illumination time

It can be seen that selectivity to CO2 decreased with molybdenum loading. On 2MoDT2, 4MoDT2 and 8MoDT2 steady state selectivities to

Photocatalytic flat plate reactor

97

CO2 were about 10 %, 4 % and 2 % respectively. The obtained results confirmed that the increasing presence of MoOx species on the titania increase the selectivity of the catalyst indicating that the interaction between titania and supported molybdenum oxide plays an essential role in changing the catalyst selectivity.

VII Photocatalytic

fluidized bed reactor

VII.1 Photocatalytic fluidized bed reactor design

A two dimensional fluidized bed photoreactor was designed in order to improve both exposure of the catalysts to light irradiation and a good contact between reactants and catalyst. The powder used for the design of the two dimensional photocatalytic fluidized bed reactor was α−Al2O3, (Aldrich) with a Sauter average diameter of 50 µm; its density is 3970 kg/m3. The equation 17 was used to compute the minimum fluidization velocity (Umf):

0.5Re (1135.7 0.0408 ) 33.7mf Ar= + ⋅ − (17)

where:

Re p f mf

mf

d Uρ

µ

⋅ ⋅= = Reynolds number at mimimum fluidization

velocity

3

2

( )p f s fd gAr

ρ ρ ρ

µ

⋅ ⋅ − ⋅= = Archimedes number

1

i

pi

p

dx

d

=

∑= Sauter average diameter

fluidizing gas densityf

ρ =

particle density of fluidizing solids

ρ =

minimum fluidization velocitymf

U =

Chapter VII

100

weight fraction of particles of size ii p

x d=

acceleration of gravityg =

viscosity of fluidizing gasµ =

The cross-section was evaluated by using a gas superficial velocity equal to 4*Umf.

A schematic picture of the photocatalytic fluidized bed reactor is reported in Figure 71

Gas IN

Gas OUT

Gas IN

Gas OUT

Figure 71 Photocatalytic fluidized bed reactor

The gas (flow rate: 830 Ncc/min) was introduced into the fluidized bed reactor with 40 mm x 10 mm cross section. The wall is made of pyrex-glass (2mm in thickness). A bronze filter (5 µm size) was used for gas feeding to provide uniform gas distribution. Its walls, 230 mm in height, were made of 2mm thick pyrex-glass.

During transient condition, some elutriation phenomena were observed. Thus in order to decrease the amount of transported particles, an expanding section (50 mm x 50mm cross-section at the top) and a cyclone, specifically designed, (Figure 72) are located on the top and at the outlet of the reactor respectively. The cyclone was designed by utilising the Lapple configuration “Lapple (1973)” and the diameter of its body was calculated by the following equation:

iV

QD

∗∗=

25.05.0 (18)

Photocatalytic fluidized bed reactor

101

where: Q = overall volumetric flow rate

Vi = cyclone inlet velocity = 2 m/s.

The other geometrical parameters (Table 12) of the cyclone were calculated by the relations of the Lapple configuration “Lapple (1973)”.

h

H

B

De

D

aS

b

Figure 72 Schematic picture of the cyclone

Table 12 Geometrical parameters of the cyclone

D, mm 7.45 a, mm 3.73 b, mm 1.86 B, mm 1.86 De, mm 3.73 h, mm 14.9 H, mm 29.8 S, mm 4.66

The fluidized bed reactor is illuminated by two UV light sources (BLACKLIGHT BLUE, 160W, PHILIPS or EYE MERCURY LAMP, 125W) in a dark box. In order to control the reaction temperature, a PID controller connected to a heater system was installed near the reactor.

Figure 73 reports a schematic picture of the experimental set up. Catalytic tests were carried out feeding 830 Ncc/min N2 stream containing 1000 ppm cyclohexane, 1500 ppm oxygen and 1600 ppm water.

D = cyclone body diameter

a = cyclone entrance length

b = cyclone entrance length width

B = solids eductor diameter

h = cyclone barrel length

H = cyclone height

De = cyclone outlet diameter

S = vortex tube length

Chapter VII

102

Figure 73 Schematic picture of the experimental apparatus

VII.2 Preliminary results

Titania is typically a solid of group C in the particle distribution proposed by Geldart (1973). This type of particles exhibit cohesive tendencies, and as the gas flow is further increased, usually “rathole”; the gas opens channels that extend from the gas distributor to the surface. If channels are not formed, the whole bed will lift as a piston. At higher velocities or with mechanical agitation or vibration, this type of particle will fluidize but with the appearance of clumps or clusters of particles. These phenomena play a negative role on particle light exposition. For testing Mo-titania catalyst, in order to obtain good fluidization, physical mixtures with α−Al2O3 at different percentages of Mo-titania catalysts were experimented. α−Al2O3 used in the reactor has physical characteristics of group A. When gas is passed upward through a bed of particles of groups A, B, or D, friction causes a pressure drop expressed by the Carman-Kozeny fixed-bed correlation. As the gas velocity is increased, the pressure drop increases until it equals the weight of the bed divided by the cross-sectional area. This velocity is called minimum fluidizing velocity, Umf. When this point is reached, the bed of group A particles will expand uniformly until at some higher velocity gas bubbles will form (minimum bubbling velocity, Umb). This process can avoid the appearance of clumps or clusters of titania particles and improve light exposure. In Figure 74 the outlet benzene


103

concentrations obtained by mixing 5g of 4MoDT2 with 40g of α−Al2O3 in comparison with that obtained mixing 10g of 4MoDT2 are reported.

0

5

10

15

20

25

30

35

40

45

50

0 10 20 30 40 50 60


be

nze

ne

, p

pm

5g 4MoDT2

10g 4MoDT2

Figure 74 Outlet reactor concentration of benzene on 4MoDT2 catalyst as


It can be seen that benzene production increased by increasing catalyst amount. It was about 8 ppm with 5g of 4MoDT2 and about 37 ppm with 10g of the same catalyst. The optimal mixture was composed of a physical mixture of 14g of catalyst and 63g of α−Al2O3.

In order to verify that cyclohexane was converted in a heterogeneous photocatalytic process, a control test was carried out with the reactor loaded with α−Al2O3 alone. No conversion of cyclohexane was detected during this test, indicating the necessity of the catalyst for the observed reaction. A second test was performed with the catalyst loaded in the reactor, without switching on the lamps, after cyclohexane adsorption equilibrium was reached. In these conditions the composition of the outlet reactor was identical to that of the reactor inlet, indicating that no reaction occurred in dark conditions.A photocatalytic test at 120°C was also performed by mixing 14g of DT2 with 63 g of α−Al2O3 in which no catalyst activity was found.

A typical trend of fluidized bed photocatalytic test is reported in Figure 75 with reference to 8MoDT2. When the lamps were switched on, the cyclohexane outlet concentration immediately decreased reaching a steady state value corresponding to about 4% cyclohexane conversion after about 10 minutes. In the same figure the change of oxygen outlet concentration is also reported showing a trend similar to that of cyclohexane. The analysis of products in the outlet stream disclosed the presence of benzene and

Chapter VII

104

cyclohexene, as identified from the characteristic fragments m/z = 78, 77, 76, 74, 63, 52, 51, 50 (fragment 78 reported in ) and 82, 67, 54, respectively (fragment 67 reported in Fig. 2). No presence of carbon dioxide was disclosed, as detected by the NDIR analyser. The outlet concentration of benzene progressively increased reaching a steady state value after about 50 minutes. A similar trend was shown by cyclohexene concentration, however the values are very much lower with respect to benzene. No deactivation of catalyst was observed during photocatalytic tests.

0 20 40 60 80 100 120 140

time, min

MS

sig

nal, a

.u.

m/z = 84

m/z = 32

m/z = 78

m/z = 67

Lamps on Lamps off


benzene and cyclohexene and as a function of run time

In Table 13 the comparison of photocatalytic performances in the fluidized bed reactor of 4MoDT2 and 8MoDT2 catalysts is summarised.


catalysts. Reaction temperature: 70 °C. UV sources:two blacklight blue,

160 W, PHILIPS

Catalyst cyclohexane

conversion,

%

benzene

selectivity,

%

carbon

dioxide

selectivity,

%

cyclohexene

selectivity,

%

4MoDT2 6 67 32.2 0.8

8MoDT2 4 99 0 1.0


105

On 4MoDT2 selectivity to benzene reached 67%, while that to carbon dioxide was 32.2 %. On 8MoDT2 about 100% selectivity to benzene was obtained and no formation of carbon dioxide was detected. With both catalysts the presence of a very low amount of cyclohexene in the reaction products was detected (selectivity was about 0.8% on 4MoDT2 and 1% on 8MoDT2). Therefore, the presence of MoOx species on the surface of titania increases the selectivity of the catalyst as the molybdenum content increases.

The effect of UV sources was evaluated on 8MoDT2 catalyst which was the most selective to benzene. In Table 14 the obtained results are reported.

Table 14 Effect of UV sources on 8MoDT2 catalyst. Reaction temperature:

100°C

UV source cyclohexane

conversion, %

benzene outlet

concentration, ppm

BLACKLIGHT BLUE, 160W 3 28.5

MERCURY LAMP , 125W 4.7 46.55

Cyclohexane conversion increased up to 4.7 % by using two mercury lamps with a power of 125W. In this case benzene outlet concentration was higher (46.6 ppm). In order to explain this last result, it is useful to examine the emission characteristics in the UVA and UVB range for both types of UV source (Table 15).

Table 15 Emission characteristics in the UVA and UVB range

UV source UVA

emission, W

UVB

emission, W

BLACKLIGHT BLUE, 160W 1.8 0.04

MERCURY LAMP , 125W 4.2 0.1

It can be seen that MERCURY LAMP has both UVA and UVB emission higher than BLACKLIGHT BLUE LAMP indicating a higher flux of photons emitted in the UVA region by the first UV source. This last aspect could explain the better catalytic performances obtained by using MERCURY LAMP.

The effect of reaction temperature on cyclohexane conversion and benzene production obtained by loading 14 g of 8MoDT2 catalyst mixed with 63 g of α−Al2O3 is shown in Figure 76.

Chapter VII

106

0

1

2

3

4

5

6

7

8

9

10

40 60 80 100 120 140 160 180

T, °C

cy

clo

he

xa

ne

co

nv

ers

ion

, %

0

10

20

30

40

50

60

70

80

be

nze

ne

, p

pm


benzene outlet concentration

Figure 76 Effect of reaction temperature on cyclohexane conversion and

benzene outlet concentration. UV sources: two eye mercury lamps, 125 W

Cyclohexane conversion and benzene production increased with increasing reaction temperature; moreover the presence of very low amounts of cyclohexene in the reaction products was detected. Moreover there is only a small increment of catalytic activity between 120 °C and 160 °C indicating that above a certain value of temperature, the reaction rate off and becomes independent of temperature. This limit depends on the geometry and on the working conditions of the photoreactor which probably ensures a total absorption of efficient photons.

VII.3 MoOx supported on TiO2/Al2O3 sample

An alternative to the physical mixing between titania based catalyst and α−Al2O3 is to realize TiO2-Al2O3 mixed oxide catalytic supports. In general, sol-gel methods have been preferred to produce such mixed oxide systems. When titanium and aluminium alkoxides are used as support precursors, small pore diameters of the order of 2-3 nm are obtained “Ramirez et al. (1993)”. Organic polymers were used to control textural properties of catalysts supported on single oxides “Basmadjian et al. (1962), Trimm and Stanislaus (1986)”, as well as in some mixed oxides systems such as Al2O3-


107

SiO2 “Mascia (1995), Snel (1984)”. The effect of the type of polymeric additive, its amount, molecular weight and sequence of addition of the reactants on the surface area and pore-size distribution in TiO2-Al2O3 mixed oxides were reported “Klimova et al. (1990)”. Wei et al. (1990) have examined the effect of the preparation method on the morphology of titania-alumina support. It is shown that the grafting technique gives the best dispersion of TiO2 on Al2O3.

Mixed oxides were used as catalyst in photocatalytic processes mainly for environmental applications. Two kinds of chemical vapour deposition approaches have been employed for the preparation of the photocatalyst of titanium dioxide supported on alumina “Zhang and Zhou (2005)”. One was simultaneous deposition and calcinations (one-step process); the other was preliminary gas impregnation of the support followed by a decomposition step (two-step process). The results of characterization indicated that the structure of the support was destroyed by a two-step process because of pore blocking. It was found that the one-step process resulted in a superior photocatalyst; this was attributed to higher external surface concentration and more perfect crystalline structure of TiO2. Titania, synthesized through a sol-gel procedure with acetylacetone chelating agent, was immobilized on γ-Al2O3 and deposited with photoreduced Ag “Chen et al. (2005)”. The prepared catalysts were characterized and applied for decolourization of methyl orange. The photoactivity of TiO2/γ-Al2O3 is affected by the H2O/Ti molar ratio applied in the sol-gel process. Incorporating photoreduced Ag to TiO2/γ-Al2O3 leads to reduction reaction of methyl orange in addition. to oxidation reaction, and yields a significant increment in decolorization efficiency. The preparation of two sets of polycrystalline photocatalysts prepared by supporting TiO2 (anatase) on TiO2 (rutile) or TiO2 (anatase) on Al2O3 is reported “Loddo et al. (1999)”. The powders were prepared by a wet impregnation method using titanium (IV) isopropylate. Both sets of TiO2 (anatase) supported samples resulted photoactive and the photoactivity toward 4-nitrophenol photodegradation increased by increasing the content of the anatase phase.

VII.3.1 TiO2-Al2O3 sample preparation

Titania-alumina support was prepared by dispersing DT2 powder in a boehmite sol (10 wt% of Condea Pural in bydistilled water, pH< 2 by HNO3). The sol was gelled by slight heating until it was too viscous to stir. The gel was thus dried at 120°C for 3 hours and then calcined at 500 °C for 2 hours. After calcinations the solid was crushed and sieved to recover 50 µm diameter particles. (NH4)6 Mo7O24·4H2O was used as MoO3 precursor to impregnate the alumina-TiO2 support (30wt% TiO2, 113 m2/g). Finally the powder was dried at 120 °C for 12 hours and calcined in air at 400 °C for 3

Chapter VII

108

hours. The calcined catalyst (10MoDTAl) contained 10 wt % of MoO3 as nominal loading.

VII.3.2 Results

In Figure 77 the comparison between the results obtained by mixing 14g of 8MoDT2 (Mo8) and 63 g of α−Al2O3 and by loading 30 g of 10MoDTAl (Mo10) are reported.

0

1

2

3

4

5

6

7

8

9

10

0 20 40 60 80


cy

clo

he

xa

ne

co

nv

ers

ion

, %

0

20

40

60

80

100

120

140

160

180

200

be

nze

ne

, p

pm

Cyclohexane conversion Mo8

cyclohexane conversion Mo10

benzene concentration Mo8

benzene concentration Mo10

Figure 77 Comparison between cyclohexane conversion and benzene outlet

concentration obtained on Mo8 and Mo10. Reaction temperature: 120 °C.

UV sources: two eye mercury lamps, 125 W

On 10MoDTAl cyclohexane steady state conversion was higher (about 9%) with respect to Mo8 (about 7%). On Mo8 benzene outlet concentration reached steady state value (70 ppm) after 10 min, while on 10MoDTAl it was higher (about 90 ppm). This result may come from the fluidization of a higher catalyst mass and consequently of a higher gas contact time together with better utilization of UV light than through α−Al2O3 diluent bed. Similar results have been reported by Lynette, and Gregory (1992) on silica support with respect γ Al2O3.


109

VII.4 Photocatalytic fixed and fluidized bed reactor comparison

Figure 78 shows cyclohexane conversion as a function of illumination time on 8MoDT2 catalyst using the fluidized bed and the annular fixed bed reactor.

0

2

4

6

8

10

12

14

16

18

0 20 40 60 80 100 120illumination time, min

cyclo

hexa

ne c

on

vers

ion

, %

fixed bed reactor


Figure 78 Cyclohexane conversion as a function of illumination time on

8MoDT2 catalyst using the fluidized bed and the annular fixed bed reactor.

In the fixed bed reactor, cyclohexane conversion reached a maximum value after about 5 minutes, then activity decreased approaching a steady state conversion. Maximum conversion was about 15 %, decreasing to 3% in 15 minutes. The steady state conversion was 2% after 30 minutes. The behaviour is completely different when the fluidized bed reactor is used. In this case, when the lamps were switched on cyclohexane conversion immediately increased reaching a steady state value corresponding to 4% cyclohexane conversion after about 10 minutes. This result evidences that the fluidization enhanced the cyclohexane conversion. In Figure 79 the comparison of benzene outlet concentration on Mo8 between fixed bed and fluidized bed photoreactor is reported.

Chapter VII

110

0

5

10

15

20

25

30

35

40

45

0 20 40 60 80 100 120


be

nze

ne

, p

pm fixed bed reactor


Figure 79 Comparison of benzene outlet concentration on 8MoDT2 between

fixed bed and fluidized bed reactor.

It can be seen that, by using the fixed bed reactor, the outlet concentration of benzene progressively increased reaching a maximum value of about 17 ppm after 37 minutes. Then it decreases to a steady state value (about 14 ppm) reached after about 80 minutes. Instead, in fluidized bed reactor, benzene outlet concentration progressively increased reaching a steady state value of 39 ppm (higher than that obtained in the fixed bed reactor) after about 20 minutes.

In Table 16 and Table 17 the comparison of photocatalytic performances in the steady state of 4MoDT2 and 8MoDT2 catalysts in the fixed and fluidized bed reactors is respectively reported.


catalysts in the fixed bed reactor


conversion,

%

benzene

selectivity,

%

carbon

dioxide

selectivity,

%

cyclohexene

selectivity,

%

4MoDT2 4 27 8 0.7

8MoDT2 2 65 5 1.5


111


catalysts in the fluidized bed reactor


conversion,

%

benzene

selectivity,

%

carbon

dioxide

selectivity,

%

cyclohexene

selectivity,

%

4MoDT2 6 67 32.2 0.8

8MoDT2 4 99 0 1.0

By using the fixed bed reactor, cyclohexane conversion was 4 % on 4MoDT2 and 2 % on 8MoDT2. On 4MoDT2 selectivity to benzene reached 31%, while selectivity to carbon dioxide was 8%. On 8MoDT2 higher selectivity to benzene was observed (65%) and selectivity to carbon dioxide was 5%. With both catalysts the presence of very low amounts of cyclohexene in the reaction products was detected (selectivity was about 0.7% on 4MoDT2 and 1.5% on 8MoDT2). In the fluidized bed, cyclohexane conversion was 6% and 4% on 4MoDT2 and 8MoDT2 respectively. On 4MoDT2 selectivity to benzene was 67%, while that to carbon dioxide was 32.2 %. On 8MoDT2 about 100% selectivity to benzene was obtained and no formation of carbon dioxide was detected. Moreover selectivity to cyclohexene was about 0.8% on 4MoDT2 and 1% on 8MoDT2.

The obtained results showed that both in the fixed bed reactor and in the fluidized bed reactor, cyclohexane conversion and selectivity to carbon dioxide decreased with increasing molybdenum content whereas selectivity to benzene and cyclohexene increased. Moreover with both catalysts, the performances obtained in the fluidized bed reactor are better than those obtained in the fixed bed reactor. Cyclohexane conversion and outlet benzene concentration characteristics suggested that there was no catalyst deactivation when the fluidized bed reactor was used. Selectivity data evidenced that in the fixed bed reactor, total carbon mass balance (evaluated by comparing the inlet carbon as cyclohexane and the outlet carbon as the sum of unconverted cyclohexane and outlet benzene, cyclohexene, carbon monoxide and carbon dioxide) was not close to 100% while the fluidized bed reactor it was almost complete. Carbon mass balance as function of illumination time during the catalytic tests in the fixed bed reactor on 4MoDT and 8MoDT catalysts has been reported in Figure 80. The carbon mass balance is closed to about 99 % on 8MoDT and 90% on 4MoDT in steady state conditions.

Chapter VII

112

70

75

80

85

90

95

100

0 20 40 60 80 100


tota

l c

arb

on

ma

ss

ba

lan

ce

, %

8MoDT2

4MoDT2

Figure 80 Total carbon mass balance in the fixed bed reactor as a function


Figure 81 and Figure 82 show TG-MS results of 4MoDT2 and 8MoDT2 catalyst after photocatalytic tests.


113

-0.2

-0.1

0.0

0.1

0.2

0.3

DT

G,

%/m

in

96

97

98

99

100

TG

,%

0 200 400 600 800

Temperature,°C Figure 81 TG-MS results on 4MoDT2 catalyst after photocatalytic results in

the fixed bed reactor

-0.4

-0.2

0.0

0.2

0.4D

TG

, %

/min

95

96

97

98

99

100

TG

, %

0 200 400 600 800

Temperature, °C Figure 82 TG-MS results on 8MoDT2 catalyst after photocatalytic results in

the fixed bed reactor

Thermogravimetric analysis on 8MoDT2 and on 4MoDT2 showed the presence of carbonaceous species adsorbed on catalyst surface since there was the formation of carbon dioxide in the range 200-430 °C (as evidenced

TG

DTG

TG

DTG

Chapter VII

114

by carbon dioxide characteristic fragments m/z = 44) that could explain the carbon deficit checked during the reaction in the fixed bed reactor. Moreover an interesting result was the total absence of surface sulphates evidenced by the absence of SO2 (g) (fragments m/z = 48 and 64).

A literature summary of catalyst deactivation in gas-solid photocatalysis was reported by Sauer and Ollis (1996). They showed that deactivation is a very commonly observed phenomenon, especially for single pass flow reactor. By using the fixed bed reactor, tests performed on both catalysts revealed that rapid decays of activity occurrred in the first minutes on stream to reach a stable value for conversion under steady state conditions. This initial decrease of activity could be due to a poisoning of the surface by carbonaceous species. It can be supposed that the effect could be due either to a strong adsorption of the reactant or the products (benzene, CO2 and cyclohexene), or to an adsorption of carboxylate or other carbonaceous species yielding a blocking of a part of the surface sites.

In order to elucidate this last aspect, TPD coupled with Mass Spectrometer and CO, CO2 continuous analyzers, in nitrogen flow, of both catalysts after activity measurement in the fixed bed reactor in the range 20-500 °C were performed (Figure 83).

0

50

100

150

200

250

300

30 130 230 330 430Temperature, °C

CO

2, p

pm

MS

sig

na

l, a

.u.

CO2

m/z = 78

m/z = 67

Figure 83 Outlet reactor concentration (MS signal) of cyclohexene, and

benzene and (ppm) of carbon dioxide (NDIR analyzer) as a function of

temperarure on 4MoDT2 after activity measurements in the fixed bed

reactor.


115

Desorption of cyclohexene (m/z = 67) from 100 °C up to 210 °C, of benzene (m/z=78) from 160 °C to 260 °C, and of CO2 from about 150 °C up to 500 °C were observed. No others products were found. Similar results were obtained for 8MoDT2 sample. Therefore, part of products remained adsorbed on the surface and could be related to the observed deactivation. It must be also noted that deactivated samples did not appear brownished, as found by Einaga et al.(2002). They identified carbon deposits on TiO2 surface in heterogeneous photocatalytic decomposition of benzene, toluene, cyclohexane and cyclohexene, finding that deactivated TiO2 catalysts were photochemically regenerated in the presence of water vapour and the carbon deposits were decomposed to COx. Therefore preliminary characterization of deactivated catalysts does not give any evidence for carbon deposit formation.TG-MS analysis on 4MoDT2 and 8MoDT2 catalysts after activity measurements in fluidized bed reactor evidenced the presence of surface sulphate and the absence of adsorbed carbonaceous compounds. According to these results, TPD tests performed on same catalysts (with the same conditions described above), showed the only desorption of SO2 (m/z = 64) from 270 °C (Figure 84). Desorption of benzene, cyclohexene and CO2 was not detected

200 250 300 350 400 450 500

Temperature, °C

m/z

= 6

4, a.u

.

Figure 84 Outlet reactor concentration (MS signal) of SO2 as a function of

temperature on 4MoDT2 after activity measurements in the fluidized bed

reactor

These last results confirm the presence of sulphate on catalysts surface after photocatalytic tests in the fluidized bed reactor. The non-appearance of

Chapter VII

116

the reaction products on catalyst surface could explain the absence of catalyst deactivation observed in the fluidized bed reactor. Moreover TG-MS and TPD results suggest that the deactivation phenomenon in the fixed bed reactor is probably also due to the disappearance of sulphate species.

The surface chemical states of the samples were examined by X-ray photoelectron spectroscopy (XPS). This last type of analysis was performed with a Kratos Axis Ultra instrument working with a monochromatic Al Kα radiation. Mo 3d, O 1s and C 1s bands and survey spectra were recorded. Binding energies were calibrated by fixing the C–(C–H) contribution of C 1s at 284.8 eV. Further details on XPS experiments and data treatments are given elsewhere “Dury et al. (2003)”. Figure 85 summarizes the difference of Mon+ contributions found by XPS for fresh 8MoDT2 catalyst and that recovered after photocatalytic test both in the fixed bed reactor and in fluidized bed reactor.

0

20

40

60

80

100

120

fresh after fixed bed

reactor

after fluidized

bed reactor

%

Mo5+

Mo6+

Figure 85 Difference of Mo

n+ contributions found by XPS for fresh 8MoDT2

catalyst and that recovered after photocatalytic test both in the fixed bed

reactor and in fluidized bed reactor

Fresh 8MoDT2 exhibited mainly Mo6+ and a small amount of Mo5+ species. A surface oxidation of 8MoDT2 occurred during photocatalytic test in both reactors as an increase of the Mo6+ contribution is observed. Correspondingly the Mo5+ contribution decreased. Figure 86, Figure 87 and Figure 88 show the detailed scan spectra of S 2p obtained on fresh 8MoDT2 catalyst and that recovered after photocatalytic test both in the fixed bed reactor and in fluidized bed reactor.


117

Figure 86 Detailed XPS scan spectrum of S 2p on fresh 8MoDT2

Chapter VII

118

Figure 87 Detailed XPS scan spectrum of S 2p on 8MoDT2 after

photocatalytic test in the fixed bed reactor


119

Figure 88 XPS scan spectrum of S 2p on 8MoDT2 after photocatalytic test

in the fluidized bed reactor

XPS spectra reveal that sulphur is absent from the surface of sample recovered after test in the fixed bed reactor whereas its continuing presence is found with the sample recovered after test in the fluidized bed reactor. Therefore, taking into account that a higher state of oxidation was found on 8MoDT2 after test both in the fluidized bed reactor and in the fixed bed reactor with respect to fresh sample, the deactivation of the catalyst may be correlated with the sulphur disappearance since the sample used in fluidized conditions remains active and maintains its sulphur content at the surface.

VII.5 Effect of light intensity

As showed in the section VI.1.7, the experimental tests performed by increasing UV light intensity showed an improvement of catalytic activity in terms of cyclohexane conversion. Therefore the fluidized bed system has been modified by introducing the possibility of irradiating the catalyst with four UV sources simultaneously. The schematic picture of the modified photocatalytic fluidized bed reactor is reported in Figure 89.

Chapter VII

120

Gas OUT

Gas IN

Figure 89 Schematic picture of the modified photocatalytic fluidized bed

reactor

The catalytic results in terms of cyclohexane conversion and benzene outlet concentration obtained on 8MoDT2 by irradiating the reactor with two and four UV light sources (EYE MERCURY LAMP, 125W) are shown in Figure 90 and Figure 91 respectively.


121

0

2

4

6

8

10

12

14

16

0 10 20 30 40 50 60 70 80illumination time, min

cy

clo

he

xa

ne

co

nv

ers

ion

, %

4 UV sources

2 UV sources


two and four UV sources. Reaction temperature: 120 °C.

0

20

40

60

80

100

120

140

160

0 10 20 30 40 50 60 70 80


be

nze

ne

, p

pm

4 UV sources

2 UV sources


two and four UV sources. Reaction temperature: 120 °C.

It can be seen that cyclohexane conversion and benzene outlet concentration were about 7 % and 70 ppm respectively when 2 UV sources were used and they increased up to about 13 % and 130 ppm in correspondence of 4 UV sources. The obtained results confirmed that

Chapter VII

122

catalyst activity increases by irradiating the reactor with a greater number of UV sources. Consequently, photocatalytic activity tests were carried out by using the modified fluidized bed reactor.

VII.6 Photocatalytic oxidation of cyclohexane on sulphated

MoOx/γγγγ-Al2O3 catalysts

Since selectivity and yield to benzene increased with the sulphate content

of the support and no formation of benzene was found when molybdenum was supported on unsulphated α- or γ- alumina in the gas-solid fixed bed reactor, the photoactivity of MoOx/γ-Al2O3 catalysts at various sulphation degrees toward oxidative dehydrogenation of cyclohexane was determined in the fluidized bed photoreactor.

VII.6.1 Samples preparation

γ−Al2O3 (Puralox SBA 150, SASOL, 144 m2/g) was impregnated with an aqueous solution of ammonium heptamolybdate (NH4)6 Mo7O24·4H2O. Powder samples were dried at 120 °C for 12 hours and calcined in air at 400°C for 3 hours The calcined catalysts contained 8 wt % of MoO3 (8Mo) nominal loading. A second impregnation of 8Mo with an aqueous solution of ammonium sulphate was performed. Powder samples so obtained were dried at 120 °C for 12 hours and calcined in air at 300 °C for 3 hours.

VII.6.2 Specific surface area and Thermal analysis

The list of catalysts and their characteristics is reported in Table 18.

Table 18 Catalysts and their characteristics

Catalyst

MoO3

nominal

content, wt %

SO42-

nominal

content, wt %

Specific surface area

(BET),

m2/g

8Mo 8 - 144 8Mo2S 8 2.4 135 8Mo4S 8 4.8 114 8Mo6S 8 7.2 111

It can be seen that specific surface area decreased as the sulphate content increased. Thermogravimetric curves of 8Mo2S, 8Mo4S and 8Mo6S are reported in Figure 92.


123

70

75

80

85

90

95

100

105

0 200 400 600 800

T, °C

TG

, %

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

DT

G,

%/m

in

TG 8Mo2S

TG 8Mo4S

TG 8Mo6S

DTG 8Mo6S

DTG 8Mo4S DTG 8Mo2S

Figure 92 Thermogravimetric curves of 8Mo2S, 8Mo4S and 8Mo6S

catalysts

The first main step (up to about 250°C) is associated to water removal, while the second main step (up to about 380 °C) is due to decomposition of ammonia. The third complex stage of weight loss is associated to different kind of surface sulphates. DTG curve of powder 8Mo2S showed only one high temperature peak (centred at about 717 °C), probably associated to the more stable sulphate coordinated on alumina. For 8Mo4S other two overlapping peaks appeared, centred at 540°C and 610°C respectively, probably due to a different sulphate coordinated on alumina. Another DTG peak, centred at about 430°C, appeared. This could be attributed to sulphate interacting molybdate. This last peak was observed also on 8Mo6S catalyst. It can be seen that on this last catalyst, four overlapping high temperature peaks, centred at 523°C, 685°C and 745°C due to different kind of surface sulphates coordinated on alumina.

Effective surface sulphate content (as SO42-) was evaluated by TG-MS

analysis in the range of temperature in which the sulphate decomposition occurred. It can be seen that SO4

2- content measured by thermogravimetric analysis corresponds to that calculated for the impregnation process (Table 19).

Chapter VII

124

Table 19 Nominal and effective sulphate content

Catalyst SO4

2- nominal

content, wt %

SO42-

effective

content, wt %

8Mo - - 8Mo2S 2.4 2.6 8Mo4S 4.8 4.9 8Mo6S 7.2 7.3

VII.6.3 Photocatalytic activity tests

The reaction temperature was 120 °C and a physical mixture of 13g of each catalyst and 58 g of α−Al2O3 was prepared. In the case of 8Mo sample, no catalyst activity was found whereas on sulphated catalysts the analysis of products in the outlet stream disclosed only the presence of cyclohexene and no formation of benzene and CO2, showing that the only reaction in the system was the oxidative dehydrogenation of cyclohexane to cyclohexene.

The effect of catalyst sulphate content on both cyclohexane conversion and cyclohexene outlet concentration is shown in Figure 93

0

5

10

15

20

25

30

0 2 4 6 8SO4

2-,wt %

cyclo

hexan

e c

on

vers

ion

, %

0

20

40

60

80

100

120

cy

clo

he

xe

ne

, p

pm


cyclohexene outletconcentration


a function of SO42-

percentage. Reaction temperature: 120 °C. UV sources:

four eye mercury lamps, 125 W

As it can be seen, in absence of sulphate, there was found no catalyst activity. Cyclohexane conversion increased to about 10 % with increasing


125

SO42- load up to 2.4 wt % and decreased with a further increase in SO4

2- load. A parallel trend can be observed for cyclohexene outlet concentration behaviour. The effect of reaction temperature on cyclohexane conversion and cyclohexene outlet concentration on 8Mo2S photocatalyst is shown in Figure 94.

0

5

10

15

20

25

30

80 90 100 110 120 130

T, °C

cyclo

hexan

e c

on

vers

ion

,%

0

20

40

60

80

100

120

cy

clo

he

xe

ne

, p

pm


cyclohexene concentration


a function of reaction temperature

Results reveal that cyclohexane conversion and cyclohexene production increases with increasing reaction temperature. It has been reported that UV light irradiation is the primary source of electron-hole pairs that are essential to initiate the photocatalytic reaction because generally the band gap energy is too high to be overcome thermally “Lim et al. (2000)”. Therefore, it can be considered that the increase of cyclohexane conversion may be mostly due to the increasing collision frequency of the molecule with increasing reaction temperature.

The results reported above verified the possibility to produce a cyclic olefin from a cyclic paraffin with a selectivity to cyclic olefin of 100% by means a photocatalytic process. Light alkanes oxidative dehydrogenation (ODH) process is today a potential route to obtain olefin “Kung (1994), Cavani and Trifirò (1995)” in comparison to industrial processes such as dehydrogenation and steam-cracking. However, catalysts that can be

Chapter VII

126

economically competitive with performances of common catalysts have not yet been developed. The main problem with most of the redox-type catalysts studied in ODH is that olefin yields do not exceed typically 30%. The reason for the low yields is the effective activation of alkenes by these catalysts. In fact alkenes remain strongly adsorbed on catalyst surface, inducing a high rate for consecutive deep oxidation of the desired product “Kung (1994), Watling et al. (1996)”. Recently, a new type of catalyst formulation was reported to give high yields of olefin in the oxidative conversion of alkanes “Fuchs et al. (2001)”. In this case olefin yields could reach values above 50%. The composition of these catalysts is similar to that studied for methane oxidative coupling and ethane ODH “Conway et al. (1991)” and contained a basic oxide (such as MgO) mixed with rare-earth oxide and promoted by alkali metal (Li, Na) oxide and halogen (Cl, Br) “Herskowitz et al. (1996)”.

It has been reported that the oxidative dehydrogenation of alkanes such as propane “Boisdron et al. (1995)” followed first-order kinetics in paraffin concentration. Therefore apparent first-order rate constant (k) for photocatalytic oxidative dehydrogenation of cyclohexane to cyclohexene on 8Mo2S catalyst were calculated according to the following equation at different reaction temperatures:

( )[ ]xw

Qk −−⋅= 1ln (19)

where Q is the overall volumetric flow rate, w is the mass of catalyst and x is the cyclohexane conversion employed under steady state conditions “Katasanos (1985)”. The Arrhenius plot for temperature dependence of apparent kinetic constant is shown in Figure 95.

-11

-10

-9

-8

-7

-6

0.0025 0.0026 0.0026 0.0027 0.0027 0.0028 0.0028

1/T, K-1

ln k

Figure 95 Arrhenius plot for the photoreaction of cyclohexane on 8Mo2S


127

The apparent activation energy and the pre-exponential factor were found to be 26 kcalmol-1 and 25.65 Nm3g-1h-1 respectively. The value of pre-exponential factor is lower than that in the most of catalytic oxidative dehydrogenation reactions whereas the apparent activation energy is very similar to the activation energy required for the same reaction type “Comite et al. (2003)”.

VII.7 Photocatalytic oxidative dehydrogenation of ethylbenzene

to styrene

Styrene is one of the most important base chemicals in the petrochemical

industry “Yoo (1996)”. Styrene is industrially manufactured by the dehydrogenation process of ethylbenzene (EB) over iron-based catalysts. Process is operated by adding excess steam to EB in an adiabatic reactor under pressurized condition with the reaction temperatures of about 600 °C. The problems associated with the dehydrogenation process of ethylbenzene are follows: thermodynamic limitation, low conversion rate, high endothermic energy and deactivation of catalysts by coke formation “Cavani and Trifirò (1995)”. Developments are going on to increase the concentration level by removal of the reaction product hydrogen and thereby shifting the thermodynamic equilibrium towards higher conversions. As an alternative way, the oxidative dehydrogenation of ethylbenzene has been proposed to be free from thermodynamic limitations regarding conversion, operating at lower temperatures with an exothermic reaction “Hanuza et al. (1985)”. The systems most often studied for the oxidative process are based on promoted inorganic oxide catalysts. Murakami and co-workers have studied a SnO2-P2O5 catalyst and found it to have moderate activity greater than 30 % conversion with greater than 80 % selectivity at 550°C “Murakami et al. (1981)”. Their work focused particularly on the role that promoters play in modifying the acid-base properties of the catalysts. Emig and Hofmann have reported a ZrO2-P2O5 system which is slightly more active in the same temperature range-conversion up to 55% with over 80% selectivity “Emig and Hoffmann (1983)”. Vrieland has similarly worked with phosphate-promoted catalysts and has recently reported a cerium pyrophosphate system which exhibits 76% conversion with 90% selectivity at 605°C “Vrieland (1988)”. In addition to focusing on the acid-base properties of the system, the latter two studies also investigated the role of carbonaceous overlayers on the surface of the catalyst. It was determined that each of the various systems studied had an induction period during which a thin uniform layer of carbon built up on the surface. In another study, Cadus and coworkers similarly found carbon overlayers on a sodium-doped alumina catalyst “Cadus et al. (1988)”. They provide a mechanism to show how the carbonaceous species, not the organic oxide, brings about the catalytic transformation. The influence of molybdenum, chromium and

Chapter VII

128

cobalt on the oxidation state and redox processes in V-Mg-O catalyst was studied by means of X-ray diffraction and ESR spectroscopy. The results obtained were correlated with catalytic activity and selectivity. It is shown that molybdenum doping increased the selectivity and the chromium and cobalt doping improved the activity of the catalyst studied “Oganowski et al. (1996)”. Selectivity to styrene as high as 97% was reported on Sn-Ge mixed phosphate “Turco et al. (1990)”. Mixed zirconium-tin phosphates gave much higher ethylbenzene conversion with respect to pure Zr and Sn phosphates, the selectivity to styrene being likewise high “Bagnasco et al. (1991)”. This behaviour is a consequence of the higher surface area of mixed phases. No influence of chemical composition on catalytic activity was found. Surface acidity of medium-high strength plays a relevant role in the reaction through the formation of a catalytically active coke. However, the photocatalysed oxidative dehydrogenation of ethylbenzene to styrene has never been reported.

VII.7.1 Experimental results

Since an olefinic bond was obtained from a paraffinic bond by means a photocatalytic oxidative dehydrogenation process (as it was shown in the section VII.6.3), the possibility to form styrene from ethylbenzene was investigated. Photocatalytic activity tests were carried out on 8Mo2S sample by feeding 830 Ncc/min N2 stream containing 1000 ppm ethylbenzene, 1500 ppm O2 and 1600 ppm H2O. The reactor was illuminated by four UV light sources (EYE MERCURY LAMP, 125W). The reaction temperature and catalyst weight were 120°C and 14g, respectively. The only reaction product was styrene and no formation of CO2 was detected. Ethylbenzene conversion and styrene outlet concentration are reported in Figure 96.


129

0

5

10

15

20

25

30

0 20 40 60 80 100illumination time, min

eth

ylb

en

ze

ne

co

nv

ers

ion

, %

0

20

40

60

80

100

120

sty

ren

e,

pp

m

ethylbenzene conversion

styrene outlet concentration

Figure 96 Ethylbenzene conversion and styrene outlet concentration as a


The steady state value of ethylbenzene conversion was reached after about 25 minutes and its value was about 11%. Styrene outlet concentration was 110 ppm after 85 minutes of illumination and increased less quickly with respect to ethylbenzene conversion. Total carbon mass balance was closed to 100% and no catalyst deactivation was observed.

VIII Conclusions

The occurrence of photocatalysed heterogeneous oxidative

dehydrogenation of cyclohexane in mild conditions has been studied in a gas-solid annular photocatalytic fixed bed reactor in the presence of Mo-supported catalysts, obtaining interesting selectivity to benzene and cyclohexene.

Photocatalytic performances of molybdenum active phases loaded on several supports, such as zeolites, α and γ alumina and titania with different level of surface sulphates, have been deeply investigated. With respect to MoOx/AFer, MoOx/α−Al2O3, MoOx/γ−Al2O3, MoOx/NaY and MoOx/HY, highest photoactivity to dehydrogenated products was shown by MoOx/TiO2 catalysts.

The formation of polymolybdate species spread on anatase surface increased by increasing molybdenum load up to that corresponding to the monolayer formation. In this case, high selectivity to benzene was obtained, whereas titania alone was 100% selective to carbon dioxide. After monolayer formation, segregation of MoO3 cristallytes has been observed, with a significant loss in photooxidative dehydrogenation activity.

Photocatalytic activity tests performed on anatase titania with different sulphate content revealed that only deep oxidation of cyclohexane to CO2 occurred; steady state cyclohexane conversion decreasing with increasing sulphate content. By contrast photocatalytic selective cyclohexane oxidation to benzene and cyclohexene was observed on sulphated titania supported MoOx. The presence of sulphate species on the surface of titania enhanced benzene yield more with increases in the sulphate content up to 2 wt%. On the basis of these last results, the selective formation of benzene is due to the presence of polymolybdate species supported on the titania surface, but also in the presence of surface sulphates.

The obtained results evidenced that benzene selectivity can be tuned by catalyst formulation.

Photocatalytic activity tests performed on Mo-supported sulphated titania in a photocatalytic flat-plate reactor confirmed the results obtained with the annular photocatalytic fixed bed reactor.

Chapter VIII

132

In order to enhance the exposure of the catalysts to light irradiation, to improve the contact between reactants and catalyst, and to permit easier realisation of catalytic bed, a two dimensional fluidized bed photoreactor has been designed and realised. Strong improvements in catalyst light exposure and solid-gas mass transfer rates have been reached. As a result of these improvements, a significant reduction in the total volume of reactor has been achieved, passing from about 7 liters to about 0.1 liters with higher cyclohexane conversion in similar operative conditions. Moreover it is worth to note that by using fluidized bed photoreactor, catalyst deactivation phenomena, probably due to the reaction products strogly adsorbed on active sites, are absent.

In addition, a thermodynamic analysis of the oxidative dehydrogenation reaction has been performed, and results indicated that the increase of temperature up to 330 °C gave an increase in cyclohexane conversion. The effect of the light intensity and of the radiation wavelength was also optmized.

Physico-chemical characterization tests performed on catalysts after photocatalytic activity tests both in the fixed and in the fluidized bed reactors suggested that catalyst deactivation may be correlated also with the sulphur disappearance. In fact the samples used in fluidized conditions remain active and maintain their sulphur content at the support surface.

A mechanism for the catalytic photooxidative dehydrogenation of cyclohexane based on consecutive steps of oxy-dehydrogenation occurring on active sites in competition with total oxidation on bare titania has been proposed. This mechanism considers the oxidation of cyclohexane to cyclohexene and its further oxy-dehydrogenation to benzene occurring on molybdenum oxides active sites hypothesizing a detailed reaction network.

Deeper knowledge of reaction mechanism led to a formulation of innovative photocatalysts (sulphated MoOx/γ-Al2O3) selective for oxidative mono-dehydrogenation of hydrocarbons.

By changing the catalyst support and active species, it has been demonstrated that total selectivity to mono- or further dehydrogenated products can be obtained. The effectiveness has been verified in cyclohexane conversion to cyclohexene and in ethylbenzene oxy-dehydrogenation to styrene.

IX References

• Acro M. D., S. R. G. Garrazan, V. Rives, and J. V.Garcia-Ramos

(1992) J. Mater. Sci., 27, 5921. • Afanasiev P., C. Geantet, M. Breysse, G. Coudurier, J. C. Vedrine

(1994) J. Chem. Soc. Faraday Trans, 90, 193. • Alberici R.M., W.F. Jardim (1997) Appl. Catal. B: Enviromental,

14, 55. • Alemany L., L. Lietti, N. Ferlazzo, P. Forzatti, G. Busca, E.

Giamello, F. Bregani (1995) J. Catal., 155, 117. • Almquist C. B., P. Biswas (2001) Appl. Catal.A, 214, 259. • Alyea E.C., M.A. Keane (1996), J. Catal., 164, 28. • Armengol E., A. Corma, V. Fornés, H. Garc´ýa, J. Primo (1999)

Appl. Catal. A: General., 181, 305. • Armor J.N. (1992) Appl. Catal., B: Environ., 1, 221. • Augugliaro V., S. Cosuccia, V. Loddo, L. Marchese, G. Martra, L.

Palmisano, M. Schiavello (1999) Appl. Catal. B: Environ., 20, 15. • Bagnasco G., P. Ciambelli, M. Turco, A. La Ginestra, P. Patrono

(1991) Appl. Catal., 68, 69. • Bard A.J. (1980) Science, 207, 139. • Basmadjian D., G. N. Fulford, B. I. Parsons and D. S. Montgomery

(1962), J. Catal., 1, 547. • Bian G., Y. Fu, Y. Ma (1999) Catal. Today, 51, 187. • Bickley R. I., G. Munuera, F. S. Stone (1973) J. Catal., 31, 398. • Blake N.R. and G.L. Griffin (1988) J. Phis. Chem., 92, 5697. • Blatter F., H. Frei (1996) J. Am. Chem. Soc., 118, 6873. • Blatter F., H. Sun, S. Vasenkov, H. Frei (1998) Catal. Today, 41,

297. • Blomberg J., P.J. Schoenmakers and U.A. Th. Brinkman (2002) J.

Chromatogr. A, 972, 137. • Boisdron N., A. Monnier, L. Jalowlecki-Duhamel, Y. Barbaux

(1995) J. Chem. Soc. Faraday Trans., 91, 2899.

Chapter IX

134

• Brucato A., L. Rizzuti (1997) Ind. Eng. Chem. Res., 36, 4748. • Cadus L. E., L. A. Arrua, O. F. Gorriz, J. B. Rivarola (1988) Ind.

Eng. Chem. Res., 27, 2241. • Carvalho W.A., M. Wallau, U. Schuchardt (1999) J. Mol. Catal.,

144, 91. • Cassano A. E., O. M. Alfano (2000) Cat. Tod., 58, 167. • Cassano A.E., C.A Martin, R.J Brandi, O.M Alfano (1995) Ind. Eng.

Chem. Res., 34, 2155. • Cavani F., F. Trifido (1995) Appl. Catal. A: General, 133, 219. • Cavani F., F. Trifido (1995) Catal. Today, 24, 307. • Chen L. C., F.R. Tsai, C. M. Huang (2005) J. Photochem. Photobiol.

A: Chem, 170, 7. • Cheng C. P., G. L. Schrader (1979) J. Catal., 60, 276. • Ciambelli P., G. Lisanti, D. Sannino, V. Palma, The TOCAT 4 Pre-

Conference Symposium in Kobe – Catalysis for the Environment and New Energy Sources – July 12th, 2002 – Kobe, Japan

• Colón G., M.C. Hidalgo, J.A. Navìo (2003) Appl Catal B: Environ., 45, 39.

• Comite A., A. Sorrentino, G. Capannelli, M. Di Serio, R. Tesser, E. Santacesaria (2003) J. Mol. Catal., 198, 151.

• Conway S.J., D.J. Wang, J.H. Lunsford (1991) Appl. Catal. A: General, 79, L1–L5.

• Corbo P., M. Gambino, D. Sannino, P. Ciambelli (1994), Mater. Eng., 5, 237.

• Corma A., M.I. Vazquez, A. Bianconi, A. Clozza, J. Garcia, O. Pollota, J.K. Cruz (1988) Zeolites, 8, 464.

• Coronado J. M., S. Kataoka, I. Tejedor-Tejedor, M. A. Anderson (2003) J. Catal., 219, 219.

• Cortese A.D. (1990) Environ. Sci. Technol., 24, 442. • Cui K., K. Dwight, S. Soled, A. Wold (1995) Solid State Chem.,

115, 187. • Curcio F., M. Musci, M. Notaro, G. Quattrini (1991) Mat. Sci. Mon.,

66, 2569. • Davis D.D., D.R. Kemp (1991) in: J.I. Kroschwitz, M. Howe-

Grant(Eds.), Kirk–Othmer Encyclopedia of Chemical Technology, Vol. 1, Wiley, New York, p. 46

• Diaz A. L., M. E.Bussell (1993) J. Phys. Chem., 97, 470. • Dibble L.A., G.B. Raupp (1990) Catal. Lett., 4, 345. • Dibble L.A., G.B. Raupp (1992) Environ. Sci. Technol., 26, 492. • Djeghri N., S. J. Teichner (1980) J. Catal., 62, 99. • Duncan D.C., C.L. Hill (1997) J. Am. Chem. Soc., 119, 243. • Dury F., E.M. Gaigneaux, P. Ruiz (2003) Appl. Catal. A: General,

242, 187.

References

135

• Egerton A., C.J. King (1979) J. Oil Colour Chem. Assoc., 62, 386. • Einaga H., S. Futamura, T. Ibusuki (2002) Appl. Catal. B:

Enviromental, 38, 215. • Ekstrom G.N., A.J. McQuillan (1999) J. Phys. Chem. B, 103, 10562. • Emig G., H. Hoffmann (1983) J. Catal, 84, 15. • Ermolenko L.P., C. Giannotti (1996) J. Chem. Soc. Perkin Trans., 2,

1205. • Fisherman J. (1991) Chemosphere, 22, 685. • Formenti M., J. P. Meriaudeau, S. J. Teichner (1971) Chemtech., 1,

680. • Fox M.A., M.T. Dulay (1993) Chem. Rev., 93, 341. • Fu X., L.A. Clark, W.A. Zeltner and M.A. Anderson (1996) J.

Photochem. Photobiol. A., 97, 181. • Fu X., W.A. Zeltner and M.A. Anderson (1995) Appl. Catal. B:

Environ., 6, 209. • Fu X., Z. Ding, W. Su, Chin (1999) J. Catal., 20, 321. • Fuchs St., L. Leveles, K. Seshan, L. Lefferts, A.A. Lemonidou, J.A.

Lercher (2001) Top. Catal. , 15, 169. • Geldart D. (1973) Powder Technol., 7, 285. • Giannotti C., C. Richter (1999) Int. J. Photoenergy, 1, 1. • Giordano N., J. C. J Bart, A. Vaghi, A. Castellan, G. Martinotti

(1975) J. Catal., 36, 81. • Gratzel M. (1983), in: Energy Resources through Photochemistry

and Catalysis, Academic Press, New York,. • Guillard C., P. Thèron, P. Pichat, C. Pètrier (2002) Water Reasarch,

36, 4263. • Hanuza J., B. Jesovska-Trzebiatowska, W. Oganovski (1985) J. Mol.

Catal., 29, 109. • Herskowitz M., M. Landau, M. Kaliya (1996), PCT Int. Appl. WO.

9622161. A1. 960725. • Hirata T. (1989) Appl. Surf. Sci., 40, 179. • Hoffmann M.R., S.T. Martin, W. Choi, D. Bahnemann (1995)

Chem. Rev., 95, 69. • Inamura S., H.Sasaki, M. Shono, H. Kani (1998) J. Catal. 177, 72. • Japer S.M., T.J. Wallington, S.J. Rudy, T.Y. Chang (1991) Environ.

Sci. Technol., 25, 415. • Jeziorowski H., H. Knozinger (1979) J. Phys. Chem., 83, 1166. • Jiang D., H. Zhao, S. Zhang, R. Jhon (2004) J. Catal., 223, 212. • Karakitsou K.E., X. E. Verykios (1993) J. Phys. Chem., 97, 1184. • Katasanos N. A. (1985) J. Catal, 94, 376. • Kim D.S., Y. Kurusu, I.E.Wachs, F.D. Hardcastle, K. Segawa

(1989) J. Catal., 120, 325. • Kim M., W. Nam. , G. Y Han (2004). J. Chem. Eng. , 21, 721.

Chapter IX

136

• Klimova T., E. Carmona, J. Ramirez (1998) J. Mat. Sci. , 33, 1981. • Kormann C., D.W. Bahnemann, M.R. Hoffmann (1991) Environ.

Sci. Technol., 25, 494. • Kosuko M., C.M. Nunez (1990) J. Air Wastes Manage. Assoc., 40,

254. • Kozlov D., D. Bavykin, E. Savinov (2003) Catal Letters, 86, 169. • Kung H.H. (1994) Adv. Catal., 40, 1. • Lapple C. E. (1973) in: Air Pollution Enginering Manual, USEPA,

AP-40, 2nd Ed. • Lee W. I., G. J. Choi, Y. R. Do (1997), Bull. Kor. Chem. Soc., 18,

667. • Legrini O., E. Oliveros, A.M. Braun (1993) Chem.Rev. , 2, 671. • Li Puma G., J.N Khor, A. Brucato (2004) Environ. Sci. Technol., 38,

3737. • Li Puma G., Proc. of the 7th World Congress of Chemical

Engineering, 2005, Glasgow. • Lietti L., P. Forzatti, L.J. Alemany, G. Busca, E. Giamello, F.

Bregani (1996) Catal. Today, 29, 143. • Lim T.H., S.M. Jeong, S.D. Kim, J. Gyenis (2000) J. Photochem.

Photobiol. A: Chemistry, 134, 209. • Linsebigler A.L., G. Lu, J.T. Yates Jr. (1995) Chem. Rev., 95, 735. • Litter M. I. (1999) Appl. Catal., B: Environ., 23, 89. • Liu H., S. Cheng, J. Zhang, C. Cao, W. Jiang (1997) Chemosphere,

35, 2881. • Loddo V., G. Marci, C. Martin, L. Palmisano, V. Rives, A. Sclafani

(1999) Appl. Catal. B: Environmental, 20, 29. • Lynette A.D., B.R. Gregory (1992) Environ. Sci. Technol., 26, 492. • Lyons J. E., G. W. Paashall (1994) Catal. Today, 22, 313. • Maity S. K., M. S. Rana, S. K. Bej (2001) Appl. Catal. A: General,

205, 215. • Maldotti A., A. Molinari, P. Bergamini, R. Amadelli, P. Battioni, D.

Mansuy (1996) J. Mol. Catal., 113, 147. • Maldotti A., A. Molinari, R. Amadelli (2002) Chem. Rev., 102,

3811. • Maldotti A., R. Amadelli, I Vitali, L. Borgatti, A. Molinari (2003) J.

Mol. Catal., 703, 204. • Marcì G., M. Addamo, V. Augugliaro, S. Cosuccia, E. Garcìa

Lopez, V. Loddo, G. Martra, L. Palmisano, M. Schiavello (2003) J. Photochem. Photobiol. A: Chemistry, 160, 105.

• Martin S.T., H. Herrmann, M.R. Hoffmann (1994) J. Chem. Soc., Faraday Trans., 90, 3323.

• Martra G., S. Coluccia, L. Marchese, V. Augugliaro, V. Loddo, L. Palmisano, M. Schiavello (1999) Cat. Tod. ,53, 695.

References

137

• Mascia L. (1995) Trends Polym. Sci., 3, 61. • Matralis H., S. Theret, Ph. Sebastians, M. Ruwet, P. Grange (1995)

Appl. Catal. B: Enviromental, 5, 271. • Matsuda S., H. Hatano, A. Tsutsumi (2001) Chem. Eng Journal, 82,

183. • Matthews R.W., M. Abdullah, G.K.-C. Low (1990) Anal. Chim.

Acta, 233, 171. • Medema J., C. Van Stam, V. H. J. De Beer, A. J. A. Konings, and D.

C. Koningsberger (1978) J. Catal., 53, 386. • Mizuno N., M. Misono (1998) Chem. Rev., 98, 199. • Molinari A., A. Maldotti, R. Amadelli, A. Sgobino, V. Parassiti

(1998) Inorg. Chim. Acta, 272, 197. • Morishige K., F. Kamo, S. Ogawara, and S. Sasaki (1985) J. Phys.

Chem., 89, 4404 • Mu W., J.M. Herrmann, P. Pichat (1989) Catal. Lett., 3, 73. • Muggli D. S., J. L. Falconer (1998) J. Catal., 175, 213 • Muggli D.S. and L. Ding (2001) Appl. Catal. B: Environ., 32, 181. • Murakami Y., K. Iwayama, H. Uchida, T. Hattori, T. Tagawa (1981)

J. Catal, 71, 257. • Ng K.Y.S., E. Gulari (1985) J. Catal., 92, 340. • Nishi K., S. Komai, K. Inagaki, A. Satsuma and T. Hattori (2002)

Appl. Catal. A: Gen., 223, 187. • Noh J., J. Schwarz (1989) J. Colloid Interface Sci., 130,157. • Obee T.N., R.T. Brown (1995) Environ. Sci. Technol., 29, 1223. • Oganowski W., J.Hanuza, H. Drulis,W. Mista, L. Macalik (1996)

Appl. Catal. A: General, 136, 143. • Ohtani B., Y. Ueda, S. Nishimoto, T. Kagiya and H. Hachisuka,

(1990) J. Chem Soc. Perkin Trans., 11, 1972 • Okamoto K., Y. Yamamoto, H. Tanaka, M. Tanaka, A. Itaya (1985)

Bull. Chem. Soc. Jpn., 58, 2015. • Okasaki S., T. Okuyama (1983) Bull. Kor. Chem. Soc. Jpn., 56,

2159. • Ollis D.F.(1991) in: E. Pelizetti, E. Schiavello (Eds.), Photochemical

Conversion and Storage of Solar Energy, Kluwer Academic Publishers, Dordrecht, p. 593.

• Oshaka T., F. Izumi, and Y. Fujiki (1978) J. Raman Spectrosc., 7, 321.

• P. Ciambelli, M.E. Fortuna, D. Sannino, A. Baldacci (1996) Catal. Today, 29, 161.

• Palmisano L. and A. Sclafani (1997) in: M. Schiavello, Heterogeneous Photocatalysis, Wiley & Sons: London, p 109.

• Panizza M., C. Resini, G. Busca, E. F. Lòpez, and V. S. Escribano (2003), Catal. Lett., 89, L-4.

Chapter IX

138

• Papp J., S. Soled, K. Dwight, A. Wold (1994) Chem. Mater. 6, 496. • Park D., J. Zhang, K. Ikeue, H. Yamashita, & M. Anpo (1999) J.

Catal., 185, 114. • Pasquali M., F. Santarelli, J. F. Porter, P.L. Yue (1996) AIChE J.,

42, 532. • Peral J. and D.F. Ollis (1992) J. Catal., 136, 554. • Philips L.A. and G.B. Raupp J. Mol. Cat. (1992), 77, 297. • Pillai U. R., E. Sahle-Demessie (2002) J. Catal., 211, 434. • Ping C., N. Qian, Z. Ya-li (2004) CODEN: HYZHEN ISSN: 1001-

0017. Journal written in Chinese, 78, 61 • Pirone R., P. Ciambelli, G. Moretti and G. Russo (1996) Appl. Catal

B: Environmental, 8, 197. • Quincy R.B., M. Houalla, A. Proctor, D.M. Hercules (1989), J.

Phys. Chem., 93, 5882. • R. I. Bickley (1997) in: M. Schiavello, Heterogeneous

Photocatalysis, Wiley & Sons: London, p 87. • Ramirez J., L. Ruiz-Ramirez, L. Cedeno, V. Harle, M. Vrinat and M.

Breysse (1993) Appl. Catal., 93, 163. • Raupp G. B., J. A. Nico, S. Annangi, R. Changrani, R. Annapragada

(1997) AIChE J., 43, 792. • Rendell D (1987) in: Fluorescence and Phosphorescence, Analytical

Chemistry by Open Learning, Wiley, New York,. • Reymond J.P., F. Molenda (1999) Powder Technol., 103, 30. • Roby A., J. P. Kingsley (1996), Chemtech, 26, 39. • Samantaray S. K., K. M. Parida (2001) Appl. Catal. A: General, 211,

175. • Sauer M.L., and D.F. Ollis (1996) J. Catal., 158, 570. • Sauer M.L., D.F. Ollis (1996) J. Catal., 163, 215. • Seinfeld J.H., S.N. Pandis (1997) in: Atmospheric Chemistry and

Physics: from Air Pollution to Climate Change, Wiley/Interscience, New York,.

• Serpone N. (1997) J. Photochem. Photobiol. A: Chem., 104, 11. • Shah J. J, H.B. Singh (1988) Environ. Sci. Technol., 22, 1381. • Snel R. (1984) Appl. Catal., 12, 347. • Sorrentino A., S. Rega, D. Sannino, A. Magliano, P. Ciambelli, E.

Santacesaria (2001) Appl. Catal. A: General, 209, 45. • Suppan P. (1994) Chemistry and Light, Royal Society of Chemistry,

Cambridge. • Tanielian C., K. Duffy, A. Jones (1997) J. Phys. Chem. B, 101,

4276. • Trimm D. L. and A. Stanislaus (1986) Appl. Catal., 21, 215. • Turchi C, M. Mehos and J. Pacheco (1993), Design issues for solar-

driven photocatalytic systems, in Photocatalytic Purification and

References

139

Treatment of Water and Air, Ed by Ollis DF and Al-Ekabi H, Elsevier, Amsterdam, The Netherlands, pp 789–794

• Turco M., G. Bagnasco, P. Ciambelli, A. La Ginestra and G. Russo (1990) Stud. Surf. Sci. Catal. , 55, 327.

• Vorontsov A. V., E. N. Savinov, P. G. Smirniotis (2000) Chem. Eng. Sci., 55, 5089.

• Vrieland G. E (1988) J. Catal, 111, 1. • Wachs I. E. (1996) Catal. Today, 27, 437. • Walker A., M. Formenti, J. P. Meriaudeau, S. J. Teichner (1977) J.

Catal., 50, 237. • Wang L. and W. K. Hall (1980) J. Catal., 66, 251. • Ward M.D., J.M. White, A.J. Bard (1983) J. Am. Chem. Soc., 105,

27. • Watling T.C., G. Deo, K. Seshan, I.E. Wachs, J.A. Lercher (1996)

Catal. Today, 28, 139. • Wei B., Q. Xin, X.X. Guo, P. Grange and B. Delmon (1990), Appl.

Catal., 63, 305. • Wieinold J., R. E. Jentoft, T. Ressler (2003) Eur. J. Inorg. Chem., 6,

1058. • Wilkinson C. F. (1987) Environ. Sci. Technol., 21, 843. • Wolf K., A. Yazdani, P. Yates (1991) J. Air Wastes Manage. Assoc.,

41, 1055. • Xie C., Z. Xu, Q. Yiang, N. Li, D. Zhao, D. Wang, Y. Du (2004) J.

Mol. Catal., 217, 193. • Xu Y., Y. Shu, S. Liu, J. Huang, X. Guo (1995) Catal. Lett., 35, 233. • Yamashita H., Y. Ichihashi, M. Takeuci, S. Kishiguchi, M. Ampo

(1999) J. Synchrotron Rad., 6, 451. • Yoo J.S. (1996) Appl. Catal. A: General, 142, 19. • Yu J.C., J. Lin, R.W.M. Kwork (1997) J. Photochem. Photobiol. A:

Chem., 111, 199. • Yu J.C., J. Yu and J. Zhao (2002) Appl. Catal. B: Environ., 36, 31. • Yue P. L., F. Khan, L. Rizzuti (1983) Chem. Eng. Sci., 38, 1893. • Žalac S., N. Kallay (1992) J. Colloid Interface Sci., 149, 233. • Zhang X., M. Zhou, L. Lei (2005) Appl Catal., A: General., 282,

285.

19._tesi_vincenzo_vaiano

Documents

cyclohexane on

vincenzo palma

vaiano gas

cyclohexane supervisor

diana sannino

ringraziamento importante

salvatore vaccaro

cyclohexane over supported