Alma Mater Studiorum – Università di Bologna

DOTTORATO DI RICERCA IN

Chimica

Ciclo XXVII

Settore Concorsuale di afferenza: 03/C2

Settore Scientifico disciplinare: CHIM/04

THE SYNTHESIS OF MALEIC ANHYDRIDE:

STUDY OF A NEW PROCESS AND IMPROVEMENT
OF THE INDUSTRIAL CATALYST

Presentata da: Giulia Pavarelli

Coordinatore Dottorato Relatore

Prof. Aldo Roda prof. Fabrizio Cavani

Esame finale anno 2015

 Summary

Summary

1 Aim of work ..............................................................................5

2 Introduction ...............................................................................7
2.1 Maleic anhydride .............................................................................. 7
2.2 Maleic anhydride production .......................................................... 10
2.3 Commercial Maleic Anhydride Technologies ............................... 13
2.3.1 Fixed bed .................................................................................. 14
2.3.2 Fluidized bed ............................................................................ 15
2.3.3 Transported bed ........................................................................ 16
2.4 Vanadyl pyrophosphate catalyst ..................................................... 17
2.4.1 Synthesis of the precursor VOHPO4·0.5H2O........................... 18
2.4.2 Thermal treatment of the precursor VOHPO4·0.5H2O:
mechanism of transformation of VHP to VPP ..................................... 20
2.4.3 Activation of (VO)2P2O7 catalyst ............................................. 25
2.4.4 P/V ratio ................................................................................... 26
2.4.5 Vanadium oxidation state: V4+/V5+ ratio .................................. 27
2.4.6 Acidity of the catalyst .............................................................. 29
2.4.7 Redox properties and effect of the gas phase composition ...... 29
2.5 New catalyst systems ...................................................................... 30
2.5.1 Supported systems .................................................................... 31
2.5.2 Addition of dopants ................................................................. 32
2.6 Mechanism of n-butane oxidation to maleic anhydride ................. 34
2.6.1 Reaction scheme ....................................................................... 35
2.6.2 Nature of active sites ................................................................ 36

1
 Summary

2.7 Alternative reactant for the synthesis of maleic anhydride: 1-

butanol ......................................................................................................39
2.7.1 1-butanol ...................................................................................40
2.7.2 1-butanol to maleic anhydride: .................................................43
2.7.2.1 1-butanol to butenes ..............................................................46
2.7.2.2 Keggin-type polyoxomethalates ...........................................48

3 Experimental .......................................................................... 52
3.1 1-butanol to maleic anhydride ........................................................52
3.1.1 Reactants...................................................................................52
3.1.2 Catalyst preparation ..................................................................53
3.1.3 Lab-scale plant tests .................................................................54
3.1.4 DRIFT Spectroscopy-Mass Spectroscopy ...............................55
3.2 n-Butane to maleic anhydride .........................................................55
3.2.1 Lab-scale plant tests .................................................................55
3.2.2 Industrial catalyst for MA from n-butane ................................55
3.3 General arrangements .....................................................................57
3.3.1 Data elaboration: conversion, yield and selectivity .................60
3.3.2 Catalyst characterization ..........................................................61

4 Results and discussion ........................................................... 63

4.1 Synthesis of maleic anhydride from n-butane ................................63
4.1.1 Catalysts characterization .........................................................64
4.1.2 The effect of n-butane concentration .......................................65
4.1.3 Raman investigation of A catalyst: the effect of water ............72
4.1.4 Study of the ageing of industrial VPP catalyst.........................75
4.2 Synthesis of maleic anhydride from 1-butanol...............................85

2
 Summary

4.2.1 Dehydration of 1-butanol ......................................................... 85

4.2.1.1 Reactivity tests with supported Keggin-type polioxometalates
catalysts .............................................................................................. 85
4.2.2 Reactivity tests with VOPO4·2H2O catalysts........................... 98
4.2.3 Characterization of VOPO4·2H2O catalysts........................... 104
4.2.4 Reactivity tests with vanadyl pyrophosphate DuPont
catalyst…. ........................................................................................... 107
4.2.5 1-butanol oxidehydration to MA with vanadyl pyrophosphate
DuPont catalyst ................................................................................... 110
4.2.5.1 Thermal reactivity experiments: blank tests ....................... 110
4.2.5.2 The effect of temperature .................................................... 113
4.2.5.3 The effect of contact time ................................................... 116
4.2.5.4 The effect of oxygen partial pressure.................................. 117
4.2.5.5 The effect of 1-butanol concentration ................................. 119
4.2.5.6 1-butene oxidation to MA ................................................... 121
4.2.5.7 A DRIFTS study of the mechanism of 1-butanol
oxidehydration ................................................................................. 122
4.2.5.8 The reactivity of intermediates and the reaction mechanism:
the role of crotonaldehyde ............................................................... 129
4.2.6 Bio-butanol to maleic anhydride ............................................ 136
4.2.7 VPP DuPont catalyst characterization ................................... 144

5 Conclusions .......................................................................... 148

6 Acknowledgments ................................................................ 153

7 Ringraziamenti ..................................................................... 154

3
 Summary

8 References ............................................................................ 157

4
 Aim of work

1 Aim of work

Nowadays the industrial production of maleic anhydride (MA) is achieved principally

by the selective oxidation of n-butane, a process catalyzed by V/P/O-based materials,
namely the vanadyl pyrophosphate (VPP), a compound with formula (VO)2P2O7.
Although this process has been widely used since the ‘80s, the importance of MA as a
building block in the chemical industry is still largely recognized and this is why the
interest for both improving the technology and more deeply studying this reaction has
never ended. Within this framework, the first topic of my PhD thesis was investigated
in collaboration with a company specialized in the production of organic anhydrides
and their derivatives, Polynt SpA, with the aim of improving the performance of the
process for the selective oxidation of n-butane to maleic anhydride and comparing the
behavior of an industrial VPP-based catalysts (produced at an industrial scale by
Polynt) when utilized either in the industrial plant or in lab-scale reactor. In particular,
we investigated how the catalysts are affected by the reaction conditions (for example
the hydrocarbon concentration) and how the addition of a dopant (specifically, Nb)
could enhance the catalytic performance. Moreover, we monitored the ageing of the
catalyst and we correlated the gradual deactivation process with the modifications
occurring on the catalyst surface and in the bulk.
The second topic of my PhD thesis was produced within the Seventh Framework (FP7)
European Project “EuroBioRef”. The study was focused on a new route for the
synthesis of maleic anhydride starting from non-petroleum based reactants, in particular
from renewable raw materials. The project intended to utilize an alternative reactant
produced by fermentation of biomass, “bio-1-butanol”. In this field, my PhD was
dedicated to the investigation of three possible catalytic configurations for the
feasibility of the process; in particular, the process was divided into two main reactions:
the dehydration of 1-butanol to butenes and the selective oxidation of butenes to maleic
anhydride. We investigated the features needed to catalyze the reactions successfully
and we proposed different materials as catalysts, namely Keggin-type

5
 Aim of work

polyoxometalates, VOPO4∙2H2O and (VO)2P2O7. The reactivity of 1-butanol was tested

under different conditions, in order to optimize the performances and understand the
interaction between the alcohol and the catalyst surface. Then, we studied the key
intermediates in the mechanism of 1-butanol oxidehydration to MA, in order to
understand the possible reaction mechanism. Lastly, we compared the reactivity of the
chemical 1-butanol with the behaviour of different types of bio-butanols produced by
biomass fermentation.

6
 Introduction

2 Introduction

2.1 Maleic anhydride

Maleic anhydride (MA) is the acid anhydride of maleic acid and its formula is
C4H2O3. It is an organic compound with four atoms of carbon in a cyclic structure and
at room temperature is a white solid with an acrid odor (Figure 2.1).

Figure 2.1: Maleic anhydride structure

Physical and chemical characteristics of MA are reported in Table 2.1.

Molecular weight (g/mol) 98.057
Melting point (°C) 52.56
Boiling point (°C) 202
Density (g/cm3) 1.314
Specific gravity at 60°C (g/cm3) 1.31
Specific gravity, solid (g/cm3) 1.43
Viscosity at 70°C (mPa·s) 1.53
Vapor Pressure at 70°C (mmHg) 6.2
Vapor Density (air=1) 3.4
Flash point, closed cup (°C) 103
Flash point, open cup (°C) 102
Autoignition temperature (°C) 447
Flammable limits (%) Lower: 1.4 Upper: 7.1
Heat of combustion (Kcal/mol) 333.9
Heat of formation, solid (Kcal/mol) 112.2
Heat of fusion (Kcal/mol) 3.26
Heat of hydration to maleic acid (Kcal/mol) 8.33
Heat of Vaporization (Kcal/mol) 13.1
Specific Heat, solid (cal/g/°C) 0.285

7
 Introduction

Specific Heat, liquid (cal/g/°C) 0.396

Refractive index 1.515
Solubility in xylene at 30°C (g/L) 163.2
Solubility in water at 30°C (g/L) 572
Solubility in acetone at 30°C (g/L) 227
Solubility in benzene at 30°C (g/L) 439.4
Exposure limits (ppm) PEL: 0.25 REL: 0.25 TLV: 0.1

Table 2.1: Physical and chemical properties of MA [1, 2, 3]

MA is a very versatile molecule, that lends itself to many applications; indeed with
three active sites (two carboxylic groups and one double bond C=C) it is an excellent
joining and cross-linking material, that make MA a bifunctional and a platform
molecule. World production and consumption of maleic anhydride in 2010 were
reported to be approximately 1.7 million metric tons and its demand was expected to
grow on an average of 3-6 % by 2020 [4]. The main use of MA is as monomer in the
manufacture of polymers, representing more than 50% for the production of
unsaturated polyester resins and about 20% to produce alkyd resins. In the remaining
part MA is utilized in different end-use applications as intermediate molecule to
obtain other important organic compounds, such as 1,4 butandiol (BDO).

Figure 2.2: World consumption of MA by end use sector in 2013 [5]

MA could be transform to obtain special anhydrides, such as succinic anhydrides (by

hydrogenation) and tetrahydrophthalic anhydride (by Diels-Alder reaction with
butadiene). Furthermore, succinic ahydride is hydrolyzed to succinic acid and
converted to γ-butyrolactone (GBL) [6], tetrahydrofuran (THF) [7, 8] and 1,4

8
 Introduction

butandiol (BDO) [8, 9]. On the other hand, tetrahydrophthalic anhydride could be
hydrogenated to esahydrophthalic anhydride, that is used together with methyl-
esahydrophthalic as vulcanizer agent and as hardener in the production of epoxy
resins. MA is also an important intermediate in the fine chemical industry, particularly
in the manifacture of agricultural chemicals (pesticides) and lubrication oil additives.
It is also a component of several copolymers (MA-styrene, MA-acrylic acid) and
paints. Moreover, MA could be hydrolyzed to maleic acid and fumaric acid, which are
used as additives to adjust the acid flavor; fumaric acid is also utilized as drug in the
treatment of psoriasis and as intermediate in the synthesis of tartaric acid. Finally, MA
is utilized for the production of aspartic acid, an intermediate of aspartame
production. In Figure 2.3 the applications of MA are summarized:

Figure 2.3: Applications of MA [10]

9
 Introduction

2.2 Maleic anhydride production

The industrial production of MA is achieved either by the selective oxidation of

benzene or the selective oxidation of n-butane.
The selective oxidation of benzene (Figure 2.4) is the oldest process and even if it has
been replaced by n-butane, it continues to be used in Far East countries, namely in
China. The reactor technology is a fixed-bed configuration: the reaction occurs in gas
phase, at 400-450°C, in a multi tubular plug flow reactor [11].

Figure 2.4: Selective oxidation of benzene to MA

The catalyst is composed by V/Mo oxides (V2O5 and MoO3) supported on an inert
and high-conductive material, in order to better remove the heat of reaction. Reaction
gas passes through the tubes over the catalysts, the process pressure being adjusted to
an optimim value of 0.15 – 0.25 MPa. A considerable amount of heat is produced
during the reaction (ΔH = - 1875 kJ/mol) and the undesirable secondary reactions of
combustion to COx are strongly exothermic and ‘‘hot spots’’ of 340-500 °C can occur
on the catalyst surface [12]. The advantage of this process is that benzene conversion
is almost total (about 96%) and MA selectivity can reach 73%; because of its toxicity,
the unreacted benzene is catalytically incinerated. Indeed, benzene is classified as a
carcinogen so its emission are strictly controlled by environmental legislations. These
rules differ very much in various countries: nowadays the limit emission in air of
benzene in Italy is established by D. Lgs. 155/2010 and it is fixed at 5 μg/m3 per year
[13]. In Figure 2.5 the process of benzene oxidation to maleic anhydride is reported:
after reaction the gases are cooled by salts melts and crude MA is recovered by partial
condensation and by water scrubbing as maleic acid solution. Then, the water solution

10
 Introduction

is dehyderated and pure MA is obtained by a batchwise distillation (o-xylene is added

as entraining agent for the azeotropic distillation).

Figure 2.5: Plant for the production of maleic anhydride by benzene oxidation: a) reactor; b) salt bath
cooler; c) partial condense; d) acid scrubber; e) alkali scrubber; f) dehydration column; g) phase
separator

Since the 70’s n-butane has gradually been replaced benzene as reactant for the
synthesis of MA and nowadays approximately 80% of MA is produced starting from
n-butane (Figure 2.6). The major incentives of the C4 hydrocarbon are the lower cost
and the lower environmental impact, that render n-butane an inexpensive and non-
toxic feedstock.

Figure 2.6: Selective oxidation of n-butane to MA

11
 Introduction

A list of the advantages is reported below:

1. Cost: n-butane is cheaper than benzene (n-butane: 750 €/ton; benzene: 1000
€/ton);
2. Toxicity: n-butane is a non-toxic reagent, instead of benezene that is a proven
carcinogenic compound;
3. Atom economy (E factor): ratio between atoms of carbons of the target
product and the reactant; for n-butane the ratio is 1 (C4/C4), for benzene is
C6/C4 (2 atoms of carbon lose in CO2)
4. Separation and purification: starting from benzene many heavy compounds are
formed as by-products;
5. Heat removal: starting from n-butane the reaction is less exothermic (-1260
kJ/mol instead of -1875 kJ/mol [12]).

The success of the process was the the discovery of a suitable and highly selective
catalyst, the vanadyl pyrophosphate (VO)2P2O7 (VPP), that is one of the most
successful example of an industrial catalyst used for the synthesis of a bulk chemical
compound starting from an alkane.
Nevertheless, the selective oxidation of n-butane has many disadvantages, as all
alkane oxidation process:

1. Total oxidation: CO and CO2 are formed by parallel and consecutive

combustion reactions involving unreacted n-butane and products formed; in
order to limit the combustion the n-butane conversion has to be maintained
lower than 80%;
2. Exothermicy: the reaction is highly exothermic and run-away must be avoided
by an efficient heat removal;
3. Flammable mixtures: the composition of reactive mixture n-butane/air must be
controlled and maintained under its flammability limit, in order to guarantee
the safety of the process.

12
 Introduction

In order to have a successful industrial process, a simultaneous fine-tuning of the

catalyst and of the reactor technology is required [14].

2.3 Commercial Maleic Anhydride Technologies

Several reactor configurations have been developed for commercial maleic anhydride
production [14]:

1. Fixed bed: practised and licensed by Huntsman (formerly Monsanto)

Scientific Design and Technobell;
2. Fluidized bed: practised and licensed by Polynt (formerly Lonza), Ineos
(formerly BP) and BOC/Mitsubishi;
3. Transported bed: Circulating Fluidized Bed Reactor (CFBR), developed by
DuPont and recently abandoned [15].

In every kind of technology the basic plant configuration could be summarized in a

common scheme reported in Figure 2.7.

FEED preparation

REACTOR

RECOVERY ENERGY
crude MA recovery

PURIFICATION
crude MA

Figure 2.7: Basic MA plant configuration

13
 Introduction

MA plants are distinguished by type of reactor, method for the recovery of crude MA
and feed composition (percentage of n-butane in air). In Table 2.2 the most important
industrial technologies used for the synthsis of MA starting from n-butane are
reported.

Process Reactor Recovery Feed composition

ALMA (Polynt) Fluidized bed Anhydrous 3.6-5% n-butane
BOC/Mitsubishi Fluidized bed Aqueous 3.6-5% n-butane
Sohio-UCB Fluidized bed Aqueous 3.6-5% n-butane
Monsanto (Huntsman) Fixed bed Anhydrous 1.8% n-butane
Scientific Design Fixed bed Aqueous 1.8% n-butane
Technobell Fixed bed Aqueous 1.8% n-butane
Transported bed
DuPont Aqueous <4% of n-butane in air
(CFBR)

Table 2.2: Industrial technologies for the synthesis of MA from n-butane [16, 17, 18]

2.3.1 Fixed bed

In a fixed bed configuration the catalyst is loaded in pellets or in various shapes in a

multi tubular reactor containing several thousand reaction tubes, tipically either 3 or
6.5 m long and with 21-25 mm as diameter. Because of its explosion hazards n-butane
has to be fed in limited inlet concentrations, lower than 1.85% mol in air (lower
explosion limit). With this technology, feeding 1.7% n-butane in air at 400°C, the
highest MA yield obtained is around 60% and n-butane conversion is 80%.
The crude MA could be recovered by adsorption in water or in organic solvent. In
water MA is hydrolyzed to maleic acid, then the acid is dehydrated (at temperature
below 130°C to minimize the isomerization to fumaric acid) and finally crude MA is
purified by distillation. In the recovery with organic solvent o-xylene is tipically used
and in such way about 98% of the MA produced is recovered, avoiding fumaric acid
formation.
The advantages of fixed bed reactors are:

14
 Introduction

1. MA yield: it is higher than that one obtained by fluidized bed;

2. Larger plant: it is possible to build plants in parallel;
3. Retrofitting of benzene plants: outdates plants used for the synthesis of MA by
benzene route could be transformed in n-butane based plants.

2.3.2 Fluidized bed

Fixed bed reactor was the first technology used for the synthesis of MA starting from
benzene; for this reason it was applied also for the n-butane route and it has been
conventionally practiced since several decades. The main advantage is the possibility
to easly convert benzene based plant to n-butane based ones. However, fluidized bed
reactors offer many advantages:

1. Higher productivity: possibility to work with higher inlet n-butane

concentrations with low risk of explosion (due to quenching effects of
particles on free radicals in the bed);
2. More efficient heat removal: the formation of hot sposts is minimized and the
runaway combustion risk is suppressed;
3. Lower investment cost;
4. Lower cost of post-reactor treatment, because of the more concentrated
streams;
5. Efficient energy recovery: the process provides the export of high pressure
steam, working with an air rate lower than the fixed bed;
6. Shorter downtime for catalyst replacement: in the fluidized bed technology the
time spent to load and download the catalyst is shorter because it is replaced
by a continuous make-up of fresh catalyst.

One of the most advanced fluidized bed technology is the ALMA process (Alusuisse
Italia-Lummus Crest), licensed by Polynt SpA (formerly Lonza). In this process n-
butane and air are fed to a fluidized bed catalytic reactor working at 400-430 °C.
Cooling coils are merged in the bed to generate high-pressure steam. In the recovery
section, an organic solvent is used to remove the MA from the reaction effluent gas; a

15
 Introduction

conventional absorption/stripping scheme is used. Crude MA is refined by continuous

distillation to separate light compounds and heavy impurities. Tail gas is sent to an
incinerator, which converts residual hydrocarbon (and CO) and the developed energy
is recovered for producing extra steam. The n-butane in the feed composition is about
4% mol; conversion is typically 80-85%, with a molar yield to MA of about 50%
[17].
The bulk VPO catalyst is treated by spray-drying with a low amount of additives in
order to improve its mechanical resistance. To limit the abrasion of bulk catalyst
particles, different techniques can be employed:

1. Impregnation of the active components on an inert support characterized by

good fluidization properties and high attrition resistance;
2. Addition of additives;
3. Encapsulation of the active phase in a silica structure.

However fluidized bed reactor has also some disadvantages:

1. Lower MA selectivity, because of feed back-mixing;

2. High mechanical stress and abrasion of the catalyst;
3. Greater catalyst inventory;
4. More complex scale-up.

In order to maintain the necessary particle size distribution of the catalyst, an amount
of new catalyst is refilled by make-up.

2.3.3 Transported bed

Latest reactor technology is the circulating fluidized bed reactor (CFBR), developed
by DuPont [15], in which the VPP catalyst was circulated between two different
separated vessels, the regeneration and the reaction zones (oxidizing and reducing
zones). The main objective has been to maximize the utilization of selective catalyst
lattice oxygen: n-butane and oxygen are fed separately, so higher n-butane
concentrations (20%) could be fed, giving an enhancement of MA productivity. The

16
 Introduction

process could be schematically divided into two parts: (1) the riser reactor, where n-
butane is fed (without air) and where the catalyst transforms the hydrocarbon to MA
by means of lattice oxygen the, being itself consequently reduced; (2) the re-oxidation
part, where the catalyst is recovered by a cyclone and transported to the regenerator
reactor, where solely air is fed to restore the oxidation state of the catalyst. Then MA
is recovered using water; after the hydrolysis, aqueous MA and maleic acid are
reduced to tetrahydrofuran in a hydrogenation reactor. In CFBR technology the n-
butane conversion is about 50% and MA yield is about 37%. Because of the riser
reactor system, it has to be used a catalyst highly resistant to the attrition: indeed, VPP
was coated by silica, which gives a very high mechanical resistance and does not
cause any selectivity decreasing. The DuPont plant worked from 1996 to 2004 for
tetrahydrofuran (THF) production and in 2005 it was shut down.

2.4 Vanadyl pyrophosphate catalyst

The best-suited catalyst known for the selective oxidation of n-butane to MA is

vanadyl pyrophosphate (VO)2P2O7 (VPP), a bulk vanadium-phosphorus mixed oxide
with a particular crystalline structure [17, 19, 20, 21]. In the scientific literature many
different preparation methods are reported, however there is a general agreement on
these following common steps:

1. Synthesis of the catalyst precursor: vanadyl hydrogenphosphate emihydrated,

VOHPO4·0.5H2O (VHP); in the synthesis, the most common reactants used
are vanadium pentoxide V2O5 as source of V5+ and phosphoric acid H3PO4 as
source of P5+;
2. Thermal treatment (calcination) of the precursor to generate the active catalyst
(VO)2P2O7 (VPP);
3. Activation of the catalyst vanadyl pyrophosphate.

17
 Introduction

2.4.1 Synthesis of the precursor VOHPO4·0.5H2O

The method of synthesis of the catalyst precursor VOHPO4·0.5H2O consists of

refluxing a mixture of vanadium pentoxide and phosphoric acid in either an organic or
aqueous medium. In the reaction, V5+ is reduced to V4+ species and the reducing agent
is oxidised: this it could be hydrogen chloride (HCl) in aqueous solution or an organic
solvent, tipically alcohols such as isobutanol or benzyl alcohol. The catalyst precursor
precipitates and then it is further dehydrated to produce the active phase (VPP): the
reaction produce also water and the oxidative products derivated from the alchol.

Concerning the synthesis, at least three kinds of preparation could be distinguished:

VPA, VPO and VPD routes [22].
The VPA route is the oldest precursor synthesis made in aqueous medium with a
reducing agent like HCl or hydrazine [23, 24, 25]. In these strong acid conditions the
solid is soluble in the aqueous medium: at first it is solubilized as VOCl3 and then it
reduced to V2O4; the final VHP is obtained by solvent evaporation, together with an
amorphous phase, or after a crystallization process, adding water [19]. In

Figure 2.8 the scheme of the VPA method is reported.

Figure 2.8: Scheme of the VPA method

The VPO route is the synthesis with organic reductant (a single alcohol – generally
isobutanol - or a mixture of alcohols) and this is the most common preparation [26,
27]. In the organic medium V2O5 is solubilized forming vanadium alcoholates, which
are reduced by the organic solvent to V2O4. Thus, H3PO4 is added and it reacts with
V2O4 at the liquid-solid interface forming the VHP. The choise of the alchols greatly
affects the catalyst behaviour: in particular, the addition of glycols, like 1,2- ethandiol,
1,3-propandiol, 1,3-butandiol and 1,4-butandiol, has a considerable effect on the

18
 Introduction

morphological features of the vanadyl orthophosphate hemihydrate, because the

organic compounds are trapped in the structure [31]. In Figure 2.9 the scheme of the
VPO method is reported.

Figure 2.9: Scheme of the VPO method

Finally there is the VPD route, composed by two-steps, where the VHP is formed by
the reduction of dihydrate vanadyl phosphate (VOPO4∙2H2O) with alcohol [28, 29]. In
Figure 2.10 the scheme of the VPO method is reported.

Figure 2.10: Scheme of the VPD method

The synthesis of the precursor is a fundamental step in the overall process because it
strictly affects the catalytic behaviour of the active phase (VPP). For these reasons,
there are many scientific papers and patents dealing with the catalyst manifacture [30,
31, 32]. Indeed, one of the main innovations in the industrial synthesis of MA has
been the replacement of the corrosive HCl with organic reducing agent, beeing the
organic medium a solvent for the phosphoric acid but not for the V/P oxide. With the
catalyst synthetized by the VPO route an important improvement in terms of MA
yield was achieved, from 43% with the VPA route to 54% with VPO route [32].

Mainly the three methods differ for the morphology and the surface area of the
catalyst obtained. By the VPA route a catalyst with low surface area is formed (3-5
m2/g) and often an impurity phase VO(H2PO4)2 is co-produced. With the VPO
method, higher values of surface area (10 m2/g) a more defective structure are
possibile; the presence of defects is known to improve the catalyst performances in
terms of MA yield. These aspects are a consequence of alcohols molecules, retained

19
 Introduction

inside the VHP layers: these organic residues (“intercalated”) trapped into the
crystalline structure are responsible of a thin platelet structure. Moreover, the
precursors samples obtained by VPO method are less crystalline and preferentially
expose (001) planes, correspondent to (100) planes of vanadyl pyrophosphate, which
possess higher density of active sites [16, 31]. The VPD route has been attracting
attention recently: the morphology of the material is markedly different from those
synthetized via VPA and VPO method: the the particles have a “rosette” morphology
and an higher surface area (30 m2/g). This method permits a good control of the
precursor morphology but the low crystallinity and the dominance of non-selective
planes (220) make this route not so suitable; this is why the most used industrial
method of synthesis of VHP remains the VPO one.

Nevertheless, in the scientific literature other methods of preparation are reported

[22], differing for alternative V and P sources: i) the hydrothermal synthesis, using
V2O4 instead of V2O5, H3PO4 or H4P2O7 in an autoclave at 145°C); ii) NH4VO3, with
oxalic acid and H3PO4; iii) a mixture of VCl3 and V2O5, or even iv) vanadium metal
to reduce V2O5; finally, also v) V4O9 was utilized as a vanadium source.

2.4.2 Thermal treatment of the precursor VOHPO4·0.5H2O:

mechanism of transformation of VHP to VPP

The crystal structure of the precursor VOHPO4·0.5H2O is made of pairs of [VO6]

octahedra sharing a face, linked together by an hydrogenphosphate tetrahedral groups
(Figure 2.11). The molecule of water is bridge-connected and shared between the two
neighbor vanadium atoms, oriented in trans position to the atom of oxygen of the
vanadyl group (V=O): the water is responsible of the layered structure of the
precursor [33]. The layers are kept together in the c direction through strong hydrogen
bonds involving the water molecules and the P-OH groups: this relatively strong
network is reflected in the high temperature required for water loss of the precursor.
In this tridimensional structure interconnected by hydrogen bonds, the nature of the
organic alcohol used in the VPO preparation is strictly connected to the intercalation
properties (see previous paragraph): indeed, the alcohol could reduce distance

20
 Introduction

between the planes, allowing the formation of crystals with the predominant exposure
of (001) planes [34, 35, 36].

Figure 2.11: VOHPO4·0.5H2O structure [29]

After the synthesis of vanadyl hydrogen phosphate hemihydrate the material has to be
thermally treated to generate the active phase, (VO)2P2O7. This transformation
involves two molecules of water lost in two dehydration steps [29]; the reaction is
summarized in Figure 2.1.

2 VOHPO4∙0,5H2O → 2 VOHPO4 + H2O

2 VOHPO4 → (VO)2P2O7 + H2O

Figure 2.12: Scheme of the synthesis of VPP starting from the precursor VHP

The first dehydration forms an amorphous or microcrystalline compound, having still

typical functional groups of hemihydrate vanadyl acid orthophosphate; the second
step involves the condensation of orthophosphate groups to form the typical
pyrophosphate groups. In the traditional lab-scale synthesis the transformation of

21
 Introduction

VHP to VPP occurs by a calcination in air at 350°C, followed by a calcination in

nitrogen at 550°C. The thermal treatment of the precursor is a multi-steps method
composed by:

1. Drying at temperature under 300°C, in order to remove organic residues (in

case of VPO route) or Cl- ions (in case of VPA route) avoiding water removal;
2. Dehydration and/or calcination with formation of the active phase VPP.

Several methods of dehydration are reported in literature, namely differing for the
temperature and for the gas phase composition. A list of the most used treatment is
reported below.

1. In-situ dehydration (inside the reactor): at 280°C, feeding a lean n-butane/air

mixture at low residence time; temperature and reactive mixture are increased
until reaching the common reaction conditions (this procedure is completed in
one day);
2. Oxygen-free dehydration at high temperature (>400°C) followed by a
treatment with n-butane/air reactive mixture [29, 37];
3. Calcination at high temperature (>400°C) and then feeding of n-butane/air
reactive mixture [38, 39]
4. Hydrothermal treatment, composed by a flow of air and water at 275°C (pre-
calcination) and after nitrogen at 390°C [40]

Compared with the precursor, the active phase (VO)2P2O7 has a different crystal
structure; this was solved by Gorbunova et al. in 1979 [41] and later refined by
Nguyen et al. 1995 [42]. Vanadyl pyrophosphate is built of pairs of [VO6] octahedra
linked through a common hedge, forming double chains of octahedral connected by
the hedge-oxygen in the c direction and by pyrophosphate groups, with a
phosphorous-bridge bond (Figure 2.13). Along the chains, single bond V-O and
double bond V=O are alternate, respectively 1.60 Å and 2.30 Å long. The unit cell of
VPP is orthorhombic [43].

22
 Introduction

Figure 2.13: (VO)2P2O7 structure [29]

Several hypothesis have been proposed for the mechanism of transformation of VHP
to VPP, however all agree that vanadyl pyrophosphate retains the morphology of the
precursor [28, 44, 45]. The relation between the two crystalline structures indicates
that a topotactic transformation occurs, since the conversion takes place without
breaking V-O-P bonds, while weak V-OH2 and e P-OH2 bonds are broken and a
couple of [HPO4] unit condense into pyrophosphate groups. There mechanism was
studied by Torardi et al. [46] and by Bordes et al. [35, 47, 48, 49] and was confirmed
by in situ XRD, SEME-TEM and XPS experiments [45, 36, 50]. In the topotactic
transformation the face-shared octahedra are converted in edge-shared ones, resulting
in a small expansion (about 12%) of the a axis and with the inversion of the central
atom of phosphorous. The structural (001) basal planes and (220) planes of the
precursor are transformed, respectively, in the correspondent (200) parallel to the
basal planes and (024) planes of VPP, maintaining further the same relative
broadening. Because of this strong relationship, it is possible to control the structure
of vanadyl pyrophosphate by a fine-tuning of the emihydrate synthesis.

23
 Introduction

It is worth to note that physical and chemical characteristics of the final VPP catalyst
are affected by several parameters involving not only the precursor synthesis:

1. Temperature, time and thermal treatment conditions;

2. Reductant agent (VPO, VPA or VPD route)
3. Precursor morphology;
4. Addition of dopants;
5. Structural defects;
6. P/V ratio
7. Optimal V4+/V5+ ratio

In particular, the latter point is the result of a combination of effects involving the
precursor synthesis, the P/V ratio, the presence of dopants and the working conditions
of the catalyst, i.e the reactive n-butane/air mixture composition. Even though
vanadium is present as V4+ in stoichiometric vanadyl pyrophosphate, after the thermal
treatment of precursor some crystalline and amorphous vanadium phosphates different
from (VO)2P2O7 are formed [51, 52, 53]. Indeed, the calcination process could lead to
the formation of oxidized V5+ species, VOPO4 phase, and vanadyl metaphosphate
VO(PO3)2 phase, that may coexist together with VPP. For this reason in the scientific
literature there are several hypothesis concerning the identification of the real active
phase of the V/P/O based catalyst. Nevertheless, there is a general agreement on
considering that the active phase of the catalyst for the selective oxidation of n-butane
to maleic anhydride is a core of crystalline (VO)2P2O7 covered with little isolated
amount of VOPO4 phases [54, 55]. The relative amount of each compound mostly
depends on the preparation and the thermal treatment/calcination of the precursor.
Otherwise, while VPP catalyst is working in the reaction conditions, it may undergo
some surface transformations without substantial structural change, but however
playing an active roles in the catalytic cycle [56].

24
 Introduction

2.4.3 Activation of (VO)2P2O7 catalyst

After the thermal treatment (VO)2P2O7 does not possess yet the proper characteristics
of an optimal catalyst. In order to achieve stable catalytic performances, VPP has to
undergo to a time of equilibration under n-butane/air reaction mixture at 400°C for
several hours, typically 50-100 hours. The catalyst could be considered “equilibrated”
or “activated” when it has reached a steady behavior in terms of n-butane conversion
and MA yield/selectivity. During the equilibration time on stream, the chemical
structure of VPP gains stability thought a preferential exposure of (001) planes.
These are recognized to be the structure-sensitive planes for the formation of MA
[19], while the side faces give low selectivity because of the difficulty of the
reoxidation of vanadium atoms to V5+ in these planes [57, 58]. Typically an
equilibrated catalyst as an average oxidation state of vanadium between 4.00 and 4.04,
a surface area of 16-25 m2/g and a slight excess of phosphorous through a P/V ratio
between 1.00 and 1.05 [16]. The equilibration period could be more or less long
depending on the evolution of the n-butane conversion vs. time; generally the fresh
catalyst is more active but less selective than the equilibrated one and the initial high
conversion decreases during the equilibration time (with an increasing MA
selectivity). If the fresh catalyst is highly reduced (containing only V4+), it is
necessary an equilibration period longer than 100 hours (200-300h), while in case of
an highly oxidised sample (containing also V5+) at least 500 hours are needed to
activate vanadyl pyrophosphate. For this reason, the equilibration process inside the
reactor is very expensive in terms of commercial cost and for an industrial application
it has to be preferred a shorter process in which a stable, active and selective catalyst
is obtained directly into the reactor. This is why many patents concerning the
activation of the industrial VPP catalyst have been published, for example by Lonza
[59] and Hunsman [60]. The improvement could be achieved by a control of the
catalyst shape, forming the material into pellets or tablets before the calcination with
the combined addition of dopants: the final catalyst is has a P/V ratio in the range
1.03-1.10 and a surface area between 25-40 m2/g [60]. Another type of innovation
was realized by activating the catalyst directly inside the fluidized bed reactor: the

25
 Introduction

ageing was carried out under reaction conditions, with the advantages of obtaining an
activated catalyst even at the beginning, giving stable performances and a better
V4+/V5+ control [59].

2.4.4 P/V ratio

One of the most important parameter affecting the catalytic performance of vanadyl
pyrophosphate is P/V ratio, which is the atomic ratio between the V source and the P
source used in the preparation of the precursor. Several studies concerning this topic
are reported in literature [61, 62, 63] and the general agreement is that a slight excess
of phosphorous is necessary to optimize VPP catalytic performances in the selective
oxidation of maleic anhydride. This amount is proposed to affect V4+/V5+ ratio,
stabilizing V4+ phase: P/V ratio lower than 1.0 gives a catalyst active but not selective
because the oxidation takes place easier than in samples with a P/V ratio > 1.0; on the
other hand, catalysts with P/V ratio between 1.0 and 1.2 show a lower activity but an
higher selectivity to MA, because the oxidation of V4+ in (VO)2P2O7 is limited [16].
Generally on the catalyst surface the P/V ratio is higher than in the bulk [54, 64] and
this phosphorous enrichment on the surface layer is considered the responsible of the
selective oxidation to MA. However, too much P content leads to the formation of
VO(PO3)2 which is detrimental for the catalytic activity [65].
In conclusions, the variation of the P content modifies the redox properties of VPP
catalyst, affecting the rate of V4+ oxidation to V5+ and the rate of V4+reduction to V3+.
For this reason the redox properties of the catalyst are correlated also with the
reducing or oxidizing power of the reactive mixture: at low conversion (< 10%), when
the reactive mixture is more reducing, P-rich and P-poor catalysts behave in the same
way. However, at high conversion (> 80%), when the reactive mixture is more
oxidizing, P-poor catalysts form preferentially carbon oxides becoming less selective,
because of their increased ability to oxidize V4+ to V5+; on the other hand, P-rich
catalysts, even if they show low activity and low selectivity at low conversion, they
keep more stable performances, avoiding a strong decline of MA selectivity at high
conversion [63].

26
 Introduction

2.4.5 Vanadium oxidation state: V4+/V5+ ratio

The selective oxidation of n-butane to maleic anhydride is a redox reaction, where n-

butane is oxidized and oxygen is reduced apparently. In fact, it is the lattice oxygen of
the catalyst structure that is actually involved in the redox mechanism, while the
oxygen fed in the reactive mixture is necessary to restore the oxidised state of
vanadium and the redox cycle. This mechanism follows the Mars-Van Krevelen
scheme [66]: in the first step the lattice oxygen of the catalyst oxides the n-butane and
V4+ ions are reduced to V3+; in the second step, the catalyst is re-oxidized by
molecular O2 co-fed with the C4 hydrocarbon. Many species of V3+ and V5+ could be
present either as amorphous or crystalline phosphates or as defects of the vanadyl
pyrophosphate structure; the thermal treatment adopted for the precursor plays an
important role in determine the final ratio of these compounds.
For this reason, the oxidation state of V in VPP catalyst has an important role in the
catalyst behavior and its function has been debated for several years [67, 44, 68].
Volta et al. pointed out that the optimal V5+/V4+ ratio has to be equal to 0.25, so
defined and isolated amount of V5+ over a core of V4+ should be the key of the success
of the VPP catalyst [69, 70, 71]. Several authors have supported the hypothesis of the
strategic role of V5+ [72, 73, 74]: isolated V5+ sites interacting with VPP are able to
provide the optimal surface concentration of sites which permit the alkane activation
and O-insertion; on the other hand, if VOPO4 are formed as bulk phases, they become
detrimental for the catalyst selectivity [71]. Bordes have asserted that the two
oxidation states of V are related to two distinct steps: the activation of n-butane by
oxygen is associated with V4+ species, while the insertion of oxygen to form MA is
associated to V5+ species [75].
Several forms of VOPO4 phases have been identified: αI, αII, β, δ, ε, γ, ω and hydrated
phases VOPO4·nH2O [76, 77, 78, 79]. The δ and γ VOPO4 were the first structures
identified after the precursor treatment in O2 or air [80] and in used catalysts [81], but
they were finely described some years later, together with the other polymorphic and
hydrated phases [82, 83]. In general VOPO4 structures consist of [VO6] octahedra and
[PO4] tetrahedra held together by V-O bond and hydrogen bonds, only connected

27
 Introduction

through a corner (no edge-sharing, as VPP): the different allotropic forms are
distinguished by the arrangement of octahedral and tetrahedra inside the layer. For
example, αI-VOPO4, αII-VOPO4, VOPO4·2H2O and δ-VOPO4 have a tetragonal
structure, even if the δ one was supposed to be orthorhombic for many years and its
structure was resolved recently [82]. It is worth to note that δ-VOPO4 phase shows a
significant long V-O bond (3.1 Å) and an high O=V—O angle (168°), factors that are
supposed to be the causes of its activity in the selective cycle of MA production.
According to literature papers, the frequency of these phases in VPP-based catalysts
decreases along this series: αII, β > δ > γ, αI > ω >> ε.
Different is the debate concerning the role of V3+: the scientific literature generally
agrees about its negative effect on the selectivity, while the generation of a discrete
number of these species in the lattice of the VPP associated with anionic vacancies
have been proposed to play a positive role on catalytic activity [84, 85]. Generally,
the reduced vanadium species are generated during the thermal treatment of the
precursor: their concentration is function of the amount of organics trapped in the
precursor and of the heat treatment.

As reported in the previous paragraph in case of P/V ratio, also V4+/V5+ ratio is
function of the reaction conditions, i.e the reducing/oxidizing power of the gas phase
atmosphere. Mallada et al. analyzed the effect of hydrocarbon-rich and hydrocarbon-
lean conditions on the oxidative state of the catalyst surface [86]: in the first case, V3+
surface phases developed together with amounts of C deposits. The same conclusions
were reached by Volta et al. [87]: the reduction of V5+ and the formation of C deposits
are responsible for the decrease of the selectivity to MA. However, the deactivation is
not so rapid when the catalyst is pre-oxidized at mild temperature. One way to
overcome these limitations is to add promoters, which helps to maintain a higher V
oxidation state through either the formation of specific compounds or through the
mediation of V re-oxidation in the redox mechanism.

28
 Introduction

2.4.6 Acidity of the catalyst

Vanadyl pyrophosphate catalyst is considered a multifunctional system because it

possesses acid and redox properties both. The acidity occurs in Lewis acid sites (V4+)
and in Brønsted ones (P—OH). These sites were detected by IR experiments with
probe molecule, like NH3, pyridine and acetonitrile [88, 89, 90]. It has been
hypothesized that the two types of acidity played a different role in the reaction
mechanism: in particular, Lewis sites could take part in the hydrogen abstraction,
while the Brønsted sites could play a key role in the selective path to MA; the
presence of these letter sites was also confirmed by the P-enrichment on the catalyst
surface. A more thorough analysis was given by Centi et al. [91], who supposed that
Brønsted sites act as a combination of effects, namely through the stabilization of the
intermediates and of the adsorbed oxygen species, favoring the oxygen-insertion.
The acidity of VPP catalyst was also studied by density functional theory (DFT) [92,
93]: as expected, the terminal P—O oxygen species are the most basic surface
oxygen, in both P-poor and P-rich samples. Moreover, they are involved in the post-
activation step, i.e the selective oxidation to non-combustion products: this could be
considered another explanation of the positive effect of P-enrichment for promoting
selectivity to MA. DFT calculations also indicate that the P—O—V oxygen is as
nucleophic as the terminal P—O, explaining in that way its role in the rupture of C—
H bonds. In conclusions, it is possible to affirm that the specificity of VPP catalyst is
strictly related also to its acidic properties.

2.4.7 Redox properties and effect of the gas phase composition

The redox properties of the feed influence the surface oxidation state of the catalyst,
which greatly affects its behavior. The oxidation state of vanadium depends on (i)
redox potential of the feed, (ii) rate of reduction and (iii) rate of oxidation of the
catalyst. These parameters were studied in literature varying the feed composition, the
catalyst residence time, the temperature, the pressure and the reactor technology [86,
94, 95]. In general, it was confirmed that the key point for an active and selective
catalyst is to keep the surface at an optimized oxidation state: this could be achieved

29
 Introduction

not only by a fine-tuning of catalyst preparation, but also by regeneration procedures,

by supplying a proper oxygen content during the reaction, by addition of dopants [96].
Since the VPP performances are largely correlated to the operating conditions, also
the conversion of n-butane is a variable parameter to take into account. In fixed bed
reactors the mixture is typically 1-2% n-butane in air while in fluidized bed the
hydrocarbon is more concentrated, 4-5% n-butane in air: in the latter case it is
possible to recycle the unconverted reactant, reducing the operative costs and
increasing the productivity of the process. Otherwise, as the feed becomes more
reducing, a drop in MA selectivity and in n-butane conversion could happen: this
means that the hydrocarbon activation occurs through reaction with nucleophilic
lattice oxygen or adsorbed lattice oxygen. Since the oxygen partial pressure decreases,
the number of available surface O2- species, which should be supplied through the
inlet oxygen, becomes non insufficient and the MA selectivity tends to decrease as the
feed composition becomes more reducing. In conclusion, when the catalyst is
adequately oxidized, catalytic performances could be kept constant.

2.5 New catalyst systems

It is known that the maximum yield to MA is limited by:

1. Parallel reactions of n-butane combustion and oxidative degradation to acetic

and acrylic acids;
2. Consecutive reactions of combustion, becoming more relevant when the
alkane conversion reaches 70-80%. This has been attributed to local catalyst
overheating due to reaction exothermicity and to the poor heat-transfer
properties of the material.

In order enhance the catalytic performances, several pathways could be chosen:

1. Improvement of the heat removal: use of highly heat-conductive supports for

VPP catalyst;

30
 Introduction

2. Control of oxygen concentration in the feed, in order to maintain an high

C4/O2 ratio over the catalytic bed;
3. Addition of dopants.

2.5.1 Supported systems

In general, the supports employed for systems used in oxidation reaction should have
the following characteristics:

1. Good mechanical properties, to improve the attrition resistance of the catalyst;

2. High thermal conductivity, to avoid hot spots on the catalyst surface;
3. Low surface area, to avoid high residence times that favor the total oxidation
reactions of the reactant;
4. Chemical inertia to reactants and to the active phase, to avoid changes in its
morphology.

Initially, vanadyl pyrophosphate was supported on oxides materials such as TiO2 and
Al2O3 [97],:in the complex all VPP-supported systems showed worse performances
than the VPP-bulk systems, probably because in these systems the reaction does not
follow the Mars van Krevelen mechanism. In particular, the TiO2-supported showed
both activity and selectivity lower than commercial VPO catalysts. This behavior was
investigate by Overbeek at al. [98, 99], which suggested that the activity worsening of
titania-supported VPO catalysts was related to a strong interaction between VPP and
TiO2, and a different reducibility of VPP on support. Ruitenbeek et al. [100] observed
no changes in average vanadium oxidation state during equilibration, since they
concluded that the lattice VPP oxygen species were not involved in n-butane
oxidation and the reaction does not follow the Mars van Krevelen mechanism.
For the SiO2-supported systems a weaker interaction than the titania-ones was
observed. However, the silica-supported catalyst was less active, but more selective to
MA: this behavior was attributed to the non-reducible nature of silica support and to
the higher surface P/V ratio found. On the other hand, in the VPP-supported on Al2O3

31
 Introduction

an high affinity between Al and P limited the dispersion of active phase and
consequently the catalysts showed poor catalytic performances [101].
Recently, highly heat-conducting supports were studied, namely β-SiC, Si3N4 and BN,
respectively with thermal conductivity of 140-270 W·m-1·K-1, 6 W·m-1·K-1 and 31
W·m-1·K-1 [102]. This characteristic together with their relatively high surface area
(>20 m2/g for β-SiC, prepared via the “shape memory synthesis”) renders them
potential useful supports for oxidation catalyst. The advantages given by a better heat
transfer permits a significant gain in MA yield; moreover, the chemical inertness of
these supports did not modify the morphology of VPP active phase [103]. For
example, with β-SiC-supported VPO (30 wt.% of active phase) it was possible to
reach the best result in MA selectivity under hydrocarbon-rich conditions ever
reported: 54% of yield at 72% of conversion at 485 °C.

2.5.2 Addition of dopants

One of the most common route to improve the performance of a catalyst is the
addition of dopants. Since the ’80, together with the study of active phase, several
promoters elements for vanadyl pyrophosphate employed for the selective oxidation
of n-butane to MA were studied [104, 105, 106, 107, 108] and addressed in patent
literature [109, 196, 197]. A vast array of metals has been studied and this work was
summarized by Ye at al. [110, 111] in a list of the elements in order of activity to MA
(not the specific activity, the surface areas show a large variation 26-51 m2/g):

Zr>Ce>La>Fe>Co>Cu>Nb>Ti>Mo>Ca>Si>W>Ni>Ge>K

This work was continued by Hutchings [112] who plotted the activity of the dopants
in function of the surface area: the major part followed a linear correlation.
The scientific literature reported that Ce, Co, Ce, Cu, Fe, Hf, La, Mo, Nb, Ni, Ti, Zr
are elements commonly used to enhance activity: these cations are suggested to form
solid solutions with the catalyst, [(VO)xM1-x]2P2O7 (where M is the promoter metal)
causing defects on VPP lattice, which can function as active sites for the n-butane
oxidation. More in detail, the addition of Zr have been found to improve activity:
despite other metals, it is not incorporated into the VPP lattice, but it forms an

32
 Introduction

amorphous phase that probably is responsible of the activity improvement [113, 114].
Moreover, it was found that a low concentration of Zr permits to lower the
temperature required to reach the maximum MA yield; also, it has been proposed that
Zr creates acidic surface sites that prevent the desorption of the reaction intermediates
(butenes, butadiene and furan) facilitating the desorption of MA. The effect of
promoters was summarized by Ballarini et al. [115] and it is reported in Table 2.3.

Dopant/V ratio Promotional effect Reasons for promotion

C: 15  25%
Control of the optimal V5+/V4+ surface
S: 0  11%
Co, Co/V 0.77 ratio; stabilization of an amorphous
under hydrocarbon-rich
Co/V/P/O compound
conditions
C: 55  79%
Co, Co/V >2 S: 43  35% Optimal surface Lewis acidity
at 653K
C: 44  60%
Ce+Fe S: 63  66% Improvement of redox properties
in absence of O2
Increase of catalytic Fe replaces V4+ in (VO)2P2O7
Fe, Fe/V 0.08
activity Re-oxidation rate is increased
C: 22  73% Increase of surface area + increase of
Ga, Ga/V 0.10
S: 55  51% intrinsic activity (electronic effect)
C: 20  17% Increase of surface acidity promotes
Nb, 0.25 wt%
S: 35  53% desorption of MA
Nb concentrates at the surface, where
C: 58  75% defects are generated. Nb acts a n-type
Nb, Nb/V 0.01
S: 70 70% dope; development of a more oxidized
surface

Table 2.3: Summary of dopants for VPP catalyst: 2° column C = conversion and S = selectivity, for
un-doped (left values) and doped catalyst (right value) in selective oxidation of n-butane to MA.

33
 Introduction

In recent years, one of the most studied dopant is Nb: the great interest has leaded to
its successful implementation in industrial scale [109, 196] and its role in VPP
catalyst has been widely discussed in literature in the last two decades. Many
hypothesis have been suggested:

1. Increase of acidity associated with the generation of defects responsible for

C—H activation [116,117];
2. Improvement of MA selectivity by a structural incorporation within V
phosphate, preventing overoxidation of the products [118];
3. Shortening of the activation time of the catalyst;
4. Action as n-type dopant for the p-type VPP semiconductor [119];
5. Control of the V4+/V5+ ratio, by modification of the redox properties of
vanadium [96];
6. Formation of solid solution V1-xNbxOPO4 and mixed oxides [(VO)2-
x(Nb)x(P2O7)1-x(PO4)2x, which increase the conductivity and the charge transfer
properties of the catalyst [120].

2.6 Mechanism of n-butane oxidation to maleic anhydride

The selective oxidation of n-butane to maleic anhydride is a demanding reaction

because it involves the exchange of 14 electrons and 7 oxygens on the surface of the
catalyst, coproducing water. During the reaction, the catalytic redox couple of V is
reduced, while the n-butane is oxidized. At the steady state the rate of reduction of the
catalyst is equal to the rate of oxidation, rred=rox. The catalytic cycle follows the Mars-
Van Krevelen mechanisms, that is reported below:

reduction: R-CH + 2KO  R-C-O + H2O + 2K

oxidation: 2K + O2(gas)  2KO

where R-CH e R-C-O are respectively the reactant and the products, while KO e K are
the oxidized and reduced catalysts.

34
 Introduction

The selectivity to the target molecule is due to the participation of lattice oxygens,
while the oxygen co-fed is needed to refill the oxygen vacancies reproducing the
proper oxidation state of vanadium. These weakly adsorbed oxygen species are
strongly electrophilic and they are responsible of total oxidation [121].

2.6.1 Reaction scheme

Several hypothesis regarding the reaction mechanism were suggested, otherwise the
most proposed reaction scheme is represented in Figure 2.14: n-butane is transformed
into MA through different sequential steps of oxidative dehydrogenation and oxygen
insertion [16, 19, 122, 123, 124, 125].

Figure 2.14: Reaction intermediate steps of selective oxidation of n-butane to MA

1) n-butane + 1/2 O2  butenes + H2O (oxidative dehydrogenation)

2) butenes + 1/2 O2  butadienes + H2O (allylic H extraction)
3) butadienes + 1/2 O2  2,5-dihydrofurane (1,4-oxygen insertion)
4-5) 2,5-dihydrofurane +2 O2  MA + 2 H2O (o-allylic insertion lactone mediated)
6a) 2,5-dihydrofurane + 1/2 O2  furan + H2O (allylic extraction)
6b) furan + 3/2 O2  AM + H2O (electrophilic oxygen insertion)

This proposed mechanism was derived by experiments dealing with the isolation of
some intermediates at unusual reaction conditions (such as under high C4/O2 ratio or
at low residence time) [126, 127]; also, the intermediates butene, butadiene and furan
were identified by Rodemerck et al. under vacuum conditions [73].

35
 Introduction

In Figure 2.15 an alternative reaction scheme is reported: the activation of n-butane

occurs on the VPP surface with the production of olefins; then, the oxidation goes
through the formation of non-cyclic carbonylic compounds, which are subjcted to
oxygen-insertion and give MA finally.

Figure 2.15: Alternative reaction scheme for the synthesis of MA through carbonylic intermediates

The rate determining step is the activation of n-butane and its oxidative
deydrogenation to butenes. In order to obtain high MA selectivity, the rate of the
oxidative dehydrogenation has to be higher than the oxygen insertion one, because
olefins formation (butene and butadiene) has be favorited instead of molecules formed
through oxidative path, such as crotonaldehyde and methylvinylketone, which are
precursors of COX and lead to a drop of MA selectivity. However, recent
computational calculation have revealed that probably 1-butene does not interact with
VPP catalyst so its key role as first intermediate in the mechanism scheme has to be
revised [128, 129].

2.6.2 Nature of active sites

Vanadyl pyrophosphate catalyst is a complex system because of the great variety of

phases and species involved in the catalytic process. The multifunctionality is its
peculiarity, because both redox and acid sites are present: acidity is necessary to
activate alkane and to quickly desorb the acid products; at the same time the electrons
transfer needs metal atoms that have to be active in redox mechanism and in oxygen
insertion, avoiding over-oxidation reactions [89, 28, 130]. For this reason, only with

36
 Introduction

some isolated oxidizing centers the partial oxidation of the reagent could be reached,
and a proper V4+/V5+ ratio is needed to limit the consecutive oxidation reactions of the
intermediates and to allow their desorption from the catalyst surface.
It is worth to note that the complexy of VPP catalyst is reflected in its ability to be
active and selective not only in the synthesis of MA but also in different oxidation
reactions, involving alkanes, olefins, anhydrides and aromatics. A list of the reactions
catalyzed by VPP is reported in Table 2.4.

Type of reaction Reagent Product

Isobutyric acid Methacrilic acid
Ciclohexane Benzene
Oxidative dehydrogenation
Succinic anydride Maleic anhydride
Alkanes olefins
Olefins Diolefins
Allylic oxidation (H-abstraction)
2,5-dihydrofuran furan
Methacrolein Methacrilic acid
o-xylene Phtalic anhydride
Electrophilic O-insertion
Butadiene Furan
Furan Maleic anhydride

Table 2.4: Reactions catalyzed by VPP

The wide variety of reactions catalyzed by vanadyl pyrophosphate is due to the

different natures of its active sites; in Table 2.5 the active species and their functions
are summarized.

Active sites Main function

Lewis acidic sites (Vanadium) Alkane activation
Bröensted acidic sites (P—OH) H-abstraction and product desorption
One electron redox couple: V5+/V4+, V4+/V3+ Alkane activation
5+ 3+
Two electrons redox couple: V /V Several hypothesis
Bridged oxygen: V-O-V e V-O-P Oxidative dehydrogenation and

37
 Introduction

oxygen insertion
Terminal oxygen: (V=O)3+, (V=O)2+ Oxygen insertion
Adsorbed molecular oxygen: species 1-
Total combustion to COx
peroxo and 2-superperoxo

Table 2.5: Active sites in vanadyl pyrophosphate

1. Lewis acid sites: they are attributed to the presence of V ions coordinatively
unsatured in the VPP (100) planes and are identified by FT-IR absorption of
basic probe molecules. The main role is to permit the H-abstraction from n-
butane [16].

2. Bröensted acid sites: they were identified by FT-IR spectroscopy and

correspond to P-OH groups in the terminal position of phosphorus tetrahedra.
They act facilitating the H removal in the activation of C—H bonds, favoring
the H shift toward the H2O formation and making easy the maleic anhydride
desorption [131].

3. V5+ species: they are active in the oxygen insertion of activated hydrocarbons
producoing oxidized compounds (MA but also to COx). Its amount on the
surface may vary from 20 to 100%, depending on the several parameters (i.e
preparation method, dopants). A controlled amount of V5+ is necessary to
improve activity, but an excess of oxidized V is detrimental for selectivity (the
optimal V5+/V4+ ratio is about 0.25) [70, 71].

4. V4+ species: thery are responsible for n-butane activation; the (V=O)2+ species
are involved in the H-extraction from n-butane and in the allylic oxidation
[16]; some authors suggested that they act also in oxygen insertion on
butadiene with the formation of the 5-elements ring compounds [132].

5. V3+ species: their role is difficult to be clarified; the formation of V3+/P/O

compounds is detrimental for the selectivity, otherwise small amount of V3+

38
 Introduction

and its associated anionic vacancies could be positively affect the catalytic
activity [133, 134].
6. Bridge oxygens V-O-P: these species are involved in the oxidative
dehydrogenation of activated n-butane to butadiene; moreover they play a role
in the oxygen insertion reaction on the 5-elements ring compounds.

7. Adsorbed molecular oxygen: molecular oxygen is absorbed in 1-superoxo

and 2-peroxo species; they are responsible of over-oxidation reactions,
because of theier strong and non-selective nucleophilicity. They are also able
to interact with the V=O group, thus activating n-butane [135].

2.7 Alternative reactant for the synthesis of maleic

anhydride: 1-butanol

The industrial synthesis of maleic anhydride has been utilized for many years,
otherwise nowadays there is a significant interest dealing with the development of
new synthetic routes starting from non-petroleum based reactants. In this field, the
production of MA starting from renewable raw materials is demonstrated by recent
papers in which either 2,5-hydroxymethylfurfural or furfural are used as reactants for
liquid-phase or gas-phase oxidations. In particular in the scientific literature reported
these examples of selective oxidation: 2,5-hydroxymethylfurfural with air in liquid
phase with an homogeneous catalyst VO(acac)2 based [136]; furfural to maleic acid in
liquid phase, with a phosphomolibdic acid-based catalyst [137], or the same reaction
in gas phase, with a vanadium oxide catalyst supported on alumina [138]; in these
cases reported both furfural and 2,5- hydroxymethylfurfural are aldehydes obtained
by sugars rearrangement and transformation.
For the replacement of n-butane, it was suggested a reactant derived from biomass
fermentation, the simplest alcohol with four atoms of carbon, 1-butanol. This could be
easy produced by biomass fermentation and it could be a interesting molecule for the
synthesis of building blocks, because of its availability and its competitive cost.

39
 Introduction

The reaction of synthesis of maleic anhydride starting from 1-butanol is reported in

Figure 2.16.

Figure 2.16: Selective oxidation of 1-butanol to maleic anhydride

2.7.1 1-butanol

1-butanol is an organic compound liquid and colorless at room temperature. Its physic
and chemical characteristic are reported in Table 2.6.

Formula C4H9OH
Molecular weight (g/mol) 74.12
Melting point (°C) -89.3
Boling point (°C) 118
Flammable limits (%vol) Lower: 1.4 Upper: 11.3
Solubility in water (wt% at 30°C) 7.85

Table 2.6: Physical and chemical properties of 1-butanol [2]

1-butanol can be produced by means of various technologies [139]:

1. The most important industrial process is propylene hydroformylation (oxo

synthesis) with subsequent hydrogenation of the aldehyde formed, i.e, 1-
butyraldehyde (and minor amount of 2-methylpropanal as the by-product).

2. Reppe process, consisting in the direct hydroxycarbonylation of propylene into

1-butanol (with 2-methylpropanol as the by-product).

40
 Introduction

3. Aldol condensation of acetaldehyde into acetaldol in basic conditions, the

dehydration of acetaldol into crotonaldehyde in the presence of an acid, and –
lastly - the hydrogenation of the latter into 1-butanol. The source for the
process can be bio-ethanol, which can be (oxi)dehydrogenated into
acetaldehyde. The multi-step process can occur in a single step starting
directly from ethanol (which also serves as internal H-transfer reagent,
developing a MPV-type reduction), using a single bifunctional catalyst, and
proper reaction conditions (Guerbet alcohol [140, 141, 142, 143]).

4. Catalytic hydrogenation of CO into higher alcohols; nowadays, however, this

process has no commercial importance.

5. ABE process: biomass fermentation based on Clostridium acetobutylicum, a

bacterium leading to the formation of an ABE mixture (acetone, butanol,
ethanol). This process was largely used during the first half of the last century;
today there is a renewed interest in it, especially with the aim of producing 1-
butanol as a biofuel. With regard to the overall biomass use, 1-butanol has an
advantage over ethanol, as it can be produced from both five- and six-carbon
sugars with no organism modifications; this allows for the efficient use of
cellulose and hemicelluloses. Nowadays, various companies supply 1-butanol
produced by fermentation.

For this reason recently the ABE process has been revaluated and the research has
been focused on the improvement of the process by:

1. Replacement with cheaper raw arterials, for example lignocellulosic waste and
algae;
2. Genetic modification of the bacteria or discovery of new, more efficient
microorganisms and not sensible to alcohol concentration;
3. Implementation of the fermentation reactor, using continuous plants and
increasing the bacteria amount by supporting them on fibres and inorganic
materials able to give higher productivity;

41
 Introduction

4. Reducing of energy costs associated with separation and purification steps, by

realization of in-situ separation.

Recently, DuPont and BP have developed a new bio-reactor for the specific
production of bio-butanol as additive for gasoline, starting from beet; other research
are focused on new technologies for the production of alcohols from lignocellulosic
biomass, but in the latter case a pre-treatment and an hydrolysis steps (enzymatic or
chemical) are necessary in order to obtain sugars.

Bio-butanol could have different applications, for examples as fuel in the automotive
industry [144], as reactant in the chemical industry, and this is why it could play a key
role in the sustainable growth. Moreover, it gives many advantages, compared with
bio-ethanol and other bio-fuels [145]:

1. Higher energetic content than bio-ethanol (+30%) that makes it comparable

with gasoline;
2. Easy to transport in the existing oil pipeline: compared with bio-ethanol it has
a lower vapour pressure, it is less corrosive and less soluble in water;
3. Lower flammability limit: safer than other fuels;
4. Miscible in gasoline in many ratios.

1-butanol is a versatile molecule because it could be used to synthesize many

products; moreover, the availability and relatively low cost of 1-butanol make this
compound an interesting bio-platform molecule for the synthesis of chemicals.
Alcohols generally decompose by either dehydrogenation to produce a carbonyl or
dehydration to produce an alkene; dehydration and dehydrogenation routes are
competitive in the gas phase in the absence of O2 up to 500°C, while dehydrogenation
predominates at higher temperatures [146]. On dehydrogenating catalysts, 1-butanol
transforms into two main products, butyraldehyde and 4-heptanone (a symmetric
ketone: ketonisation reaction) [147] and smaller amounts of hydrocarbons, such as
butene, heptene, and 3-methylheptane, as well as of other oxygenated compounds,
such as 4-heptanol, dibutylether, and butylbutyrate. Butyraldehyde (like all aldehydes)

42
 Introduction

may undergo reactions such as Tishchenko dimerisation (to form butylbutyrate),

aldolisation, and ketonisation [148, 149, 150, 151]; however, deep dehydrogenation,
aromatisation and, lastly, carbonisation may also occur. Depending on the catalyst and
conditions used, other reactions may occur, such as either decarbonylation to yield
CO, H2 and C3H6, or demethylation to yield propanal and a CHx radical [152]; the
butenes formed by 1-butanol dehydration may decompose to produce propylene and a
CHx radical [153]. In the presence of oxygen, 1-butanol yields butyraldehyde by
means of oxidehydrogenation [152, 154]. Acid catalysts for 1-butanol dehydration
into butenes include heteropolycompounds [155, 156, 157], AlPO4 [158, 159],
sulphated alumina [160], hydroxyapatite [161], Ta-montmorillonite and Ta/Si mixed
oxides [162], silica-alumina [163], Al/P and Al/Ga/P oxynitrides [164].

2.7.2 1-butanol to maleic anhydride:

The synthesis of maleic anhydride from 1-butanol requires two reactive steps:

1. Dehydration of 1-butanol to butenes;

2. Selective oxidation of butenes to MA.

To catalyze these two steps two different catalytic properties are needed: the
dehydration requires an acid catalyst, while the oxidation needs specific redox
properties. The hypothetic reaction scheme for the synthesis of MA from1-butanol is
similar to that one proposed for the n-butane route and it is reported in Figure 2.17.

Figure 2.17: Reaction scheme for the synthesis of MA from 1-butanol

The difference is in the first reactive step, where in place of the oxidative
dehydrogenation of n-butane there is the dehydration of 1-butanol, giving the same

43
 Introduction

intermediates products, butenes. Because this reaction is formally divided in two

steps, it could be hypothesized three possible reactor configurations:

1. Two-reactors system: the reaction is carried out in two consecutive reactors:

in the first dehydration occurs with an acid catalyst and in the second one the
selective oxidation of butenes formed in the previous step is finished with a
redox catalyst. Keeping the two reactive steps separated, it is possible to
optimize them both in terms of technology (for example the dehydration is
endothermic and the oxidation is exothermic), catalyst design and operative
conditions, so theoretically, this should be the best configuration for the
optimization of the selectivity. Moreover, the purification and the separation
of the mixture of butenes produced could be realized after the first reactor,
becoming possible to feed the purified olefins to the second reactor. In that
way 1-butanol is fed only in the first reactor and air only in the second one;
however, higher investment costs are the disadvantage of two-reactors
configuration.

2. One-reactor system: the reactions take place in only one reactor, arranged
with two separated catalytic beds: the first is the acid catalyst for 1-butanol
dehydration, below the second redox catalyst is loaded to oxidize butenes.
Many papers concerning the selective oxidation of butenes to MA have been
published, namely dealing with Fe and Co molibdates-based catalyst [165,
166] and also with VPP, catalyst for MA from n-butane [166, 167]. Obviously
this configuration is less flexible than the previous one, because it is not
possible to diversify the control the operative conditions (temperature,
separation). Problematic is also the management of the catalytic beds, because
we have only one reactive mixture composed by 1-butanol and air: so the acid
catalyst has to be active even in presence of oxygen, while the redox catalyst
receiving the products flow coming out from the first bed must be resistant to
water poisoning effect (co-produced in large amount in the dehydration) and
it must not catalyze secondary reactions on un-reacted 1-butanol. The main

44
 Introduction

advantage of this configuration is the lower investment cost, thanks to a

single-two beds-reactor.

3. One-reactor and one-pot reaction: in this case a bi-functional catalysts is

required, in order to carry out with only one catalyst and in only one reactor
the one-pot reaction; this means that a catalyst with proper acidity and redox
properties has to be utilized. Disadvantages described in point 2 are the same
for this type of technology, however the advantages given by a one-pot
reaction are not only the reduction of the investment costs, but also the a
simpler plant management.

In Figure 2.18 the three reactor configurations described above are represented.

air 1-butanol/air 1-butanol/air

1-butanol 1-butene

water,
1-butanol,
others products products
products

Figure 2.18: Reactor configurations for the synthesis of MA from 1-butanol; 1) Two-reactors 2) One-
reactor and two catalysts, 3) One reactor and one-pot reaction

45
 Introduction

2.7.2.1 1-butanol to butenes

Dehydration of alcohols to olefins is an important and industrially demanded reaction,

since the main applications of the alkenes is in the industry of polymers. The
elimination of water and the formation of a double bond follow the well-known
Zaitsev rule, which affirms that the alkene formed in greatest amount derives from the
removal of the hydrogen from the β-carbon having the fewest hydrogen substituents;
this means that the most stable alkene is the most substituted one.

In general, the dehydration of alcohols requires an acid environment: indeed the

reaction could be catalyzed by a strong acid (H2SO4) in acqueos medium, by weaker
acid such as POCl3 [167] or by solid materials [168, 169, 170]. More in details, for 1-
butanol dehydration, both 1-butene and 2-butenes (cis and trans) are formed but the
main products are the second ones, since the Zaitsev rule is respected. Moreover, 1-
butene is the primary product and it is in equilibrium with cis 2-butene and trans 2-
butenes: these productis are involved in the equilibrium of thermic isomerization,
moved preferentially to 2-butenes.

In order to enhance the reactivity of 1-butanol towards dehydration, several acid

catalysts could be used and many examples are reported in literature:
polyoxometalates, silico-aluminates, alumino-phosphate and zeolites [159, 171, 172,
173]. Janik et al. [171] studied the reaction mechanism for the dehydration of 2-
butanol, in presence of Keggin type-polyoxometalates supported catalysts (POM).
The reaction goes through an equilibrium adsorption of the alcohol over the Bröensted
acid sites in POM surface, forming adsorbed species, which in a second time release
water in E1-type elimination mechanism. Similar to this, it is possible to propose an
analogous dehydration mechanism for 1-butanol; the scheme is reported in Figure
2.19.

46
 Introduction

Figure 2.19: Reaction mechanism of 1-butanol dehydration on POM supported catalysts

The first step of the mechanism is the adsorption of 1-butanol molecule on the catalyst
surface, which takes place through the establishment of hydrogen bonds between the
oxygen O of the alcohol and the acid sites; subsequently, the adsorbed specie is
protonated, causing the breaking of C—O bond, the formation of an alkoxy adsorbed
specie and the co-prodution of water. Lastly, the dehydration reaction is completed
with the desorption of water and the deprotonation of the alkoxy specie to butene.
Otherwise, the molecules of 1-butanol on the POM surface could also give a
intermolecular reaction, forming dimer species as n-butyl ether; these molecules,
occupying the acid sites in place of 1-butanol, are detrimental for the efficiency of the
catalyst and cause a slowing down of the rate of reaction. This means that the rate of
1-butanol dehydration depends on the ratio between the kinetic constant of the
elimination reaction (k2) and the kinetic constant of the formation of the dimer species
(k3) (see Figure 2.19). According with Matachowski et al. [174] (who have made in
evidence the same mechanism for the dehydration of ethanol on POM supported
catalysts) these two competitive reactions two different pathways:

47
 Introduction

1. Intramolecular reaction: the most favourite one, especially at high temperature,

leads to butenes;
2. Intermolecular reaction: favourite ad low temperature, it leads to n-butyl ether.

2.7.2.2 Keggin-type polyoxomethalates

Polyoxometalates (POM) [175] are special chemical compounds composed by oxygen

and metals of the first transition state, namely V, Nb, Ta, Mo, W, in their highest
oxidation state; they may contain also an heteroatom, typically P, As, Si, Ge. In other
words POM are metal oxides that may contain hundreds of atoms in the same cluster:
they can reach a very high nuclearity and may differ from each other through a wide
variety of shapes, size and composition. The most important POM structures are the
Keggin-type and the Dawson-type; not always a priori the POM structure could be
determined, because it depends on the preparation method. Characterizations and
study of POM properties started at the beginning of the last century, mainly by
spectroscopic (UV-Vis, IR, Raman), electrochemical and NMR analysis [176, 177].
The application of POM in both homogeneous and heterogeneous catalysis takes
advantage of the easy electrons exchange of these materials, that occurs without an
overall modification or decomposition of the chemical structure [178, 179]. Several
examples are reported in scientific literature: as acid catalysts for dehydration
reactions [171, 180], as redox “green” catalyst in organic-solvent-free systems [181,
182], as photosensitive materials acting as primer for radical polymerization reactions
[183].
Thanks to their peculiar properties, POM have shown interesting applications also in
medical and biological field as anticancer, antiviral and antibacterial [184].
Usually the synthesis of POM takes place in aqueous medium, but it could be realized
also in non-aqueous medium or at solid state [185, 186]. The synthesis consists in the
protonation of the oxometalates anion, giving the condensation of the tetrahedral
structures into a more complex polyanionic structure: the synthesis is regulated by a
proper choice of the reaction conditions, namely pH, temperature and concentration of

48
 Introduction

the reactants. Depending on the mechanism of the condensation occurred, two classes
of compounds could be formed:

1. Isopolyanions (IPA): the condensed units have the same structure and a
octahedral polyanion is formed; generally, the number n of units condensed iis
higher than 6. The reaction could be schematized as reported below:

nMO42- + 2mH+ ↔ MnO(4n-m)(2n-m)- + mH2O

2. Heteropolyanions (HPA): when the condensation of the units occurs around a

central metal atom or heteroatom, giving a compound with the general formula
XsMnOmy- (X = heteroatom, M = metal of the starting oxomethalate). These
clusters are associated with a number of protons (H+) that balance the negative
charge of the HPA, producing heteropolyacids and water molecules which
guarantee the cohesion of the crystalline structure. Moreover, it is commonly
accepted that the HPA formation follows a kinetic control [187]; their
structure is function of X/M ratio, so many types of HPA clusters could be
obtained, because of the wide variety of combination of metal, heteroatoms
and X/M ratios.

The most recurring structure is the Keggin-type cluster, typically of POM with
X/M=1/12 ratio, i.e. an arrangement of twelve metal atoms surrounding a central
atom. This structure was first proposed by Pauling [188] and later confirmed by
Keggin [189] in XRD experiments on the PW12O40-3: the heteroatom is linked to four
oxygen atoms, forming a tetrahedral structure, while each metal atom is connected
with six oxygen atoms, forming an octahedral. The condensation of three of these
units produces a trimetallic M3O13 group (see Figure 2.20); these trimetallic groups
are linked together surrounding the central heteroatom and giving the Keggin-type
POM structure with the following formula:

[(XO)4M12O36]n-

where X = heteroatom (for example P, As, Si, Ge), M = metal of the first transition
serie (for example V, Nb, Ta, Mo, W).

49
 Introduction

Figure 2.20: Trimetallic group (left) and Keggin-type POM (right) [175]

POM with Keggin-type structure have three symmetry axis, that means that different
rotations are allowed and different isomers are generated. Theoretically, five isomers
are possible, but only three have been synthesized, isolated and characterized.
Besides Keggin-type, another important primary structure of POM is the Dawson-
type, where X/M ratio is 2/18 and the general formula is X2M18O62n-.

In this thesis our interest was focused on the application of Keggin-type POM as
catalyst for the dehydration of 1-butanol to butenes, thanks to their acid properties.
POM totally-substituted and partially-substituted were applied: in particular in two
samples the protons of the heteropolyacids necessary for the charge balance were
substituted with an equal number of alkali cations (Cs+ and K+). As reported by
Corma et al. [190] this substitution modifies the secondary structure (i.e the union of
the Keggin units) without altering the primary structure. These clusters are
characterized by an increased surface area and a good thermal stability, while the
solubility in water becomes lower than the original heteropolyacid. The growth in
surface area is due to the formation of a microporous and well-defined structure,
which depends on the dimension of the alkali cation utilized in the substitution
mechanism. In particular, the microporosity is due to rotation and translation
phenomena of HPA: this generates a partial removal of the barrier inside the
interstitial holes of the original heteropolyacid and permits the formation of channels
between HPA and the cations utilized for the charge balance. These substituted-POMs

50
 Introduction

show an high catalytic activity for reactions requiring acid functionalities and it is
known that they exhibit a peculiar activity in some case higher than other acid
catalyst, such as zeolites, sulphated zirconia and Nafion-H [191].

51
 Experimental

3 Experimental

3.1 1-butanol to maleic anhydride

3.1.1 Reactants

Different types of 1-butanol were used, both chemically sourced (CB, purchased from
Aldrich) and bio-based, from different suppliers (BB1, BB2, and BB3: see Table 3.1).

Code pH Impurities

Chemical 5.8 1-propanol, ethanol, 2-butanol, n-butylether, acrolein, butanoic acid ethyl ester/propanoic
1-butanol acid propyl ester
Bio-1-butanol 5.3 After Claisen: 1-propanol, 1,1-dibutoxybutane, 2-methyl-1-butanol, butyraldehyde,
BB1 acrolein, benzaldehyde, 2-butanol, 2-methyl-1-propanol (isobutanol), propionaldehyde, 3-
methyl-1-butanol, butylacetate, butylformate, styrene, C21H15 N (tentative identification: 5H-
Naphtho[2,3-C]carbazole-5-methyl)
BB1-pur 5.3 As BB1, but with lower concentration: 1-propanol, 1,1-dibutoxybutane, butyraldehyde,
furan
BB1- 6.1 1,1-dibutoxybutane, 1-propanol, 2-butanol, 1-butanol 3-methyl, 1-propanol 2-methyl,
superpur methanol (all impurities are in lower concentrations than in BB2 – the purity is better
than for BB1-pur)
After Claisen : Acetone, 1-propanol, butanal, 2-butanol, 1-propanol 2-methyl, 1-butanol-2-
methyl, butylacetate, acetone dibutyl acetal, 1,1 dibutoxy butane. No carbazole.
BB2 6.3 Ethylpropionate, butylpropionate, 2-methyl-1-butanol, butylacetate, ethanol, 1-propanol,
butylacetate, 1,3 butanediol
BB3 6.1 n-butylether, 1,3 diazine, 1-propanol, butyrraldehyde, 2-butanol, 1-butanol 3-methyl,
1,1-dibutoxybutane, methanol, acetone, 2-propanol 2-methyl, 2-pentanol, 1,3 butanediol
After Claisen: 1-propanol, butanal, 2-butanol, 2-pentanol, 1-propanol 2-methyl, 1-butanol-2-
methyl, 1-butene-2-butoxy, 2-butene-1-butoxy, n-butyleter, butylacetate, 1,1 dibutoxy butane,
1,3-diazine (CAS 000289-95-2). No carbazole.

Table 3.1: Chemical impurities and pH present in chemical 1-butanol and in bio-1-butanols used for
reactivity experiments.. Main impurities are highlighted in bold.

Some bio-butanol were purified in order to lower the impurity content; the following
method was set up: 0.03 g of bleaching earth enriched with activated charcoal was

52
 Experimental

added to 10 g of BB1; the slurry was stirred for 5 h, at room temperature; then the
solid was separated by centrifugation. This “purified” BB is referred to as BB1-pur in
Table 3.1. As can be seen, the treatment was only partly successful, since the removal
of the more undesired contaminants was not complete. The treatment was repeated for
a longer time, and more specifically: 17 g of BB1 were treated with 0.06 g (0.35 wt%)
of bleaching earth enriched with activated charcoal, at room temperature for 24 h. The
suspension was then centrifuged at 1500 rpm for 15 min; the liquid was separated and
then centrifuged again for 15 min, and finally filtered with a filter-syringe (0.20
micron). This sample is referred to as BB2-superpur in Table 3.1. The procedure used
was efficient, because the sample was much purer than the fresh one, with remarkably
decreased concentrations for all impurities, and no residual amount of heavier by-
products (see Table 3.1).

3.1.2 Catalyst preparation

DuPont VPP catalyst

The catalyst was a vanadyl pyrophosphate delivered by DuPont, which showed the
best performance in n-butane oxidation to MA in a circulating-fluid-bed reactor.
Details concerning the characteristics of catalysts were reported in a paper by
Hutchings et al. [192].

Synthesis of VOPO4∙2H2O
The synthesis of VOPO4∙2H2O (VPD) was prepared suspending the desired amounts
of V2O5 (99% Sigma Aldrich) and H3PO4 (85% Sigma Aldrich) in water, according to
the procedure reported in the literature [193]. The phosphoric acid was added in large
excess (P/V= 4), since in a stoichiometric amount does not permit a complete
dissolution of V2O5 and gives an incomplete conversion of vanadium pentoxide.
The equipment used for VPD synthesis is reported in Figure 3.1.

53
 Experimental

Figure 3.1: Apparatus for lab-scale catalyst synthesis

The mixture was heated at reflux temperature (100°C) for 17 hours, under vigorous
stirring: the color changes from dark orange (V2O5) to bright yellow (VOPO4∙2H2O).
After the mixture filtering and washing, the solid obtained is dried for few hours at
100 °C.
The calcination was performed over the catalyst (pellet 30-40 mesh) directly inside
the lab-scale reactor at from room temperature up to 440°C; then in isothermal step at
440°C for 24 h.

3.1.3 Lab-scale plant tests

Catalytic tests were carried out in a quartz continuous flow reactor, by using an inlet
feed containing 1% 1-butanol in air, with a W/F = 1.33 g·s·ml−1. Products were
analyzed online by sampling a volume of the outlet gas stream which was injected
into a gas chromatograph equipped with two columns: HP-1 for the separation of 1-
butanol, C4 hydrocarbons, MA, formaldehyde, acetic acid, and acrylic acid, and
Carbosieve SII to analyse O2, CO, and CO2.

54
 Experimental

The catalysts used for the reactivity tests were loaded in the reactor in different ways:
DuPont VPP was used is powder (microsphere), VPD and VPD-Nb in pellets 30-40
mesh: for all the tests corindone was utilized as inert filler.

3.1.4 DRIFT Spectroscopy-Mass Spectroscopy

The sample was pre-treated at 450°C in a He flow (10 mL min-1) for 45 min, in order
to remove any possible molecules adsorbed on the material. Then the sample was
cooled down to the adsorption temperature (140, 300 or 360°C) and 1-butanol was fed
at 0.6 μl·min-1 until saturation was reached. Subsequently, He was left to flow until
weakly adsorbed ethanol was evacuated. In the case of the TPD experiments, the
temperature was raised at 10°C min-1 up to 400°C). The IR apparatus used was a
Bruker Vertex 70 with a Pike DiffusIR cell attachment. Spectra were recorded using a
MCT detector after 128 scans and 2 cm−1. resolution. The mass spectrometer was an
EcoSys-P from European Spectrometry Systems.

3.2 n-Butane to maleic anhydride

3.2.1 Lab-scale plant tests

The reactivity tests were carried out using A or X catalysts, in form of pellets sized in
30-40 meshes. The reactor was filled with an inert material, steatite, in order to avoid
side reactions in homogeneous phase. N-butane was not fed pure but it was furnished
as 10%vol mixed in He. The experiments were conducted either in hydrocarbon lean
or rich conditions, i.e 1.7% or 4%, varying the residence time from W/F = 1.33
g·s·ml−1 to shorter one. Typically, the catalytic tests were carried out in a temperature
range between 340°C and 440°C.

3.2.2 Industrial catalyst for MA from n-butane

In the study of the synthesis of MA from n-butane industrially-made catalysts were

utilized: for confidential agreements they have been renamed “A” and “X”.

55
 Experimental

The A catalyst has been developed by Lonza for the selective oxidation of n-butane to
MA in fluidized bed reactors [194]. This is a V/P/O based catalyst, typically produced
with the organic route (VPO): the precursor VHP is heated at high temperature
(350°C) obtaining a mixture of vanadyl pyrophosphate, amorphous phases and V3+ or
V5+ compounds; then the material is dried through a spry-drying method and calcined
at 425°C with a mixture of air and steam in fluidized bed calciner.
The thermal treatment described so called hydrothermal was developed in industrial
scale by Lonza and allowed the obtainment of a new catalytic system more efficient
than the that one prepared with a traditional thermal treatment. This was confirmed by
catalytic tests experiments carried out in the same reaction conditions: the A sample
showed not only better performances but also more stable during time on stream,
giving an global improvement in MA yield. [40]. More in details, the A catalyst
shows to be active already in the first hours of its time on stream, without requiring
the equilibration period commonly given to the traditional catalysts to obtain a proper
chemical modification and stabilization of the catalyst surface. This significant
improvement is due to the peculiar conditions of the hydrothermal treatment, where
the steam-rich atmosphere and the high pressure favor the formation of an higher
crystalline VPP structure, avoiding the development of amorphous phases and of
other undesired VPO-based phases. Moreover, this treatment favors the production of
low amount of δ-VOPO4, which is known to positively affect the catalytic behavior of
VPP [55]. These characteristics are translated in a more efficient catalysts, thanks to
the better performances and the shortening of the activation period in the reactive
mixture.

Some years later, a new catalytic system was developed, that we have renamed “X”.
This was synthesized following the same method of preparation and the same
hydrothermal treatment as the A one: the innovation was the introduction of Nb as
dopant. A controlled amount of Nb is added using NbCl5 as Nb source and leads to a
significant improvement of the catalytic performance of the VPP system. Many
papers dealing with the role of Nb have been published [116, 119, 120], and the

56
 Experimental

behavior of X catalyst catalyst agrees with the same consideration reported for Nb-
doped VPP.

3.3 General arrangements

Lab-scale plant tests

Catalytic tests were carried out in a continuous-flow fixed bed quartz reactor. The
system permits to change different reaction parameters: feed composition, residence
time and temperature. The lab-plant is schematized in Figure 3.2 for n-butane/air
configuration, and in Figure 3.3 for 1-butanol/ He or air configuration.

Key:
1 gas feed
10 2 mass flow meters
13 3 three ways valve
vent 4 on/off valve
5 bubble flow meter
11 6 four ways valve
7 manometer
9 9 8 reactor
8 7 9 reactor oven
10 reactor thermocouple
vent 11 T control
11 5 12 crystallizer
12 13 sampling point
14 bubbler
4 15 sampling heated valve
15 11 heated lines

13 2
vent
3 1
14 12
air
6
5 butane
4
He

Figure 3.2: Scheme of laboratory scale plant, for n-butane catalytic tests configuration

57
 Experimental

2 Key:
1 air/He feed
11 1-butanol 2 1-butanol feed
14 3 mass flow meters
4 three ways valve
5 on/off valve
12 6 bubble flow meter
7 four ways valve
10 10 8 manometer
9 8 vent 9 reactor
10 reactor oven
vent 11 reactor thermocouple
12 12 T control
13 13 crystallizer
5 14 sampling point
15 bubbler
16 12 6 16 sampling heated valve
heated lines

14 3
vent
4 1
15 13
air
7
6
5
He

Figure 3.3: Scheme of laboratory scale plant, for 1-butanol catalytic tests configuration

The apparatus could be divided in three main different parts:

1) Feed
2) Reaction
3) Downstream

1) Feed
Gases (n-butane, He, air) were fed to the reactor, in separated streams or
simultaneously, by three mass-flow meters ((2) in Figure 3.2; (3) in Figure 3.3. 1-
butanol is liquid at room temperature and it was fed by a syringe pump ((2) in Figure
3.3), properly calibrated for the desired quantity of liquid flow; it was vaporized in a
inlet line heated at 120°C. n-butane, air and He flows could be mixed together or kept
separated: then they are fed to a three-ways valve (3 or 4) and send either to a bubble
flow meter (5 or 6), in order to measure the inlet flow rate, or to the reactor, through a
four-ways valve (6 or 7). The switch of this four-ways commutation valve permits
two possible configurations: to feed simultaneously either n-butane/air to the reactor

58
 Experimental

and He to gas chromatograph through the by-pass line; the other possible
configuration is n-butane/air to the gas chromatograph and He to the reactor.

2) Reaction
The fixed bed reactor is a quartz tube, operating at atmospheric pressure. On the top
of the reactor (room temperature) a porous septum allows the sampling of the feed
inlet. The thermocouple is positioned in the inner part of reactor inside a thin quartz
tube at the level of the catalytic bed. The reactor is inserted inside a Lenton oven LFT
12/25/250 with inner diameter 2.5 cm and external diameter 25 cm.
The exit of the reactor emerging outside the oven is wrapped with a heater string kept
at 220°C, in order to avoid the crystallization of the products coming out from the
reactor.

3) Downstream
Coming out from the reactor, the flow of gas products is split: a part is sent to the on-
line analysis in the gas chromatograph by opening the on/off valve (5 or 6) and
connected to the FID detector; the other part flows to the vent, passing through a
crystallizer and then to a glass bubbler filled with acetone, in order to condense the
organic products not yet collected in the crystallizer. From the crystallizer septum, it
is possible to sample the incondensable gases and inject them in the gas
chromatograph. The total flow coming out from the lab-scale plant could be measured
by a bubble flow meter (5 or 6), positioned before to the vent.

Products Analysis
Reactants and products has been analyzed by a Varian CP-3380 gas chromatography,
equipped with the following columns and detectors:
1) semicapillar CPSil-5CB column (30m length; i.d. 0.53mm; stationary phase of
100% dimethylpolysiloxane with a thickness of 3.00 μm). By this column are
separated maleic anhydride, by-products (acetic acid, acrylic acid and other
light compounds e.g. butenes) and n-butane (in and out). The correspondent
detector is a FID.

59
 Experimental

2) packed Carbosieve SII column (2m length, stationary phase of active carbons
having dimensions of 80-100 mesh). In this column CO, CO2, H2O, O2 and N2
are separated and detected by a TCD.

The carrier gas used is He. The temperature ramp program of the oven is: 7 minutes at
40°C, then the temperature is increased up to 220°C with rate of 30°C/min, finally
isothermal step at 220°C for 10 minutes.
It is worth to note that the semicapillary column and FID detector are connected in on-
line analysis: the sampling loop valve (about 350 μl) is maintained at 200°C.
The incondensable products (CO, CO2, O2, N2) have been sampled with a gas syringe
(500 μl volume) through the porous septum positioned in the crystallizer or at the top
of the reactor; then the sampled gaseous mixture have been injected in the packed
column, linked to the TCD.

3.3.1 Data elaboration: conversion, yield and selectivity

By gas chromatograph analyses the percentage values of conversion, yield and

selectivity were determined, using the following equations:

n mols of converted reactant

Conversion   100
nmols of fed reactant

n mols of product/stoichiometric coeff.

Yield   100
n mols of fed reactant/stoichiometric coeff.

Yield
Selectivity   100
Conversion

 Yields
C balance   100
Conversion

60
 Experimental

3.3.2 Catalyst characterization

Raman spectroscopy
Raman analysis were carried out using a Renishaw Raman System RM1000
instrument, equipped with a Leica DLML confocal microscope, with 5x, 20x and 50x
objectives, video camera, CCD detector and laser source Argon ion (514 nm) with
power 25 mW. The maximum spatial resolution is 0.5 µm and the spectral resolution
is 1 cm-1. For each sample a wide number of spectra have been collected changing the
surface position. Generally, the parameters of spectrum acquisition are: 5
accumulations, 10 seconds, 25% of laser power to prevent sample’s damage, 50x
objective.

In-situ analysis
“In-situ” analysis were performed using a commercial Raman cell (Linkam
Instruments TS1500). The quantity of sample used for the analysis is about 5-10 mg.
The gas flow, fed from the beginning of the experiment, is about 10 ml/min; to get an
humid flow, air (or N2) was fed through a bubbler containing water, the temperature
of which was changed (partial pressure of steam from 0.03 to 0.10 bar). Spectra are
recorded at room temperature (rt), while increasing temperature (heating rate
generally of 100°C/min, except 10°C/min for NP4 and NP2 samples) up to the desired
one and during the isotherm period. The laser power used is 25% which permits a
good spectrum acquisition without damaging the sample; the acquisition parameters
are 10 accumulations, 10 s each one; the objective used is 20x.

X-Ray powder diffraction

The XRD measurements were carried out using a Philips PW 1710 apparatus, with Cu
Kα (λ = 1.5406 Å) as radiation source in the range of 5°<2<80°, with steps of 0.1
grade and acquiring the signal for 2 seconds for each step (phosphates samples) or in
the range of 10°<2<50°, with steps of 0.05 grade and acquiring the signal for 3
seconds (VPP samples). Reflects attribution was done by the Bragg law, using the d
value: 2d sen = n.

61
 Experimental

Specific surface area (BET single point)

The specific surface area was determined by N2 adsorption at 77K (the boiling T of
nitrogen), with a Sorpty 1750 Instrument (Carlo Erba). The sample was heated at
150°C under vacuum, to eliminate water and other molecules eventually adsorbed on
the surface. After this pre-treatment, sample was maintained at 77 K in a liquid
nitrogen bath, while the instrument slowly sent gaseous N2, which was adsorbed on
the surface. By BET equation it was possible to calculate the volume of monostrate
and finally the sample surface area.

62
 Results and discussion

4 Results and discussion

My PhD work was focused on the study of the synthesis of maleic anhydride by
means of the catalytic selective oxidation in gas phase of either n-butane or a new bio-
based reactant, 1-butanol.

Although the selective oxidation of n-butane has been one of the most studied topics
in catalysis since the 80’s, several patents and scientific papers still continue to be
published because of both the significant role of MA as a platform molecule and of
the difficulty in understanding the reactivity features of the industrial catalyst and the
parameters affecting its behavior [115, 195, 196, 197]. My research was aimed at
investigating the selective oxidation of n-butane to maleic anhydride with both an
industrial and an academic perspective at the same time, thanks to the collaboration
with the company Polynt SpA.

The second part of my work was focused on the study of a new route for the
production of MA, starting from a bio-based reagent, 1-butanol. This research took
place within the Seventh Framework Programme of European Union, with the aim of
replacing the petroleum-based synthesis of chemicals with alternative routes
integrated in the bio-refinery concept.

4.1 Synthesis of maleic anhydride from n-butane

The first part of my PhD work was focused on the synthesis of maleic anhydride from
n-butane in collaboration with Polynt SpA, an important Italian company specialized
in the production of organic anhydrides (and their derivatives) and of the industrial
catalysts utilized for their production at plant-scale. We carried out catalytic tests in
lab-scale reactor, using industrial catalysts delivered by Polynt; the aim was to
investigate and understand the catalytic behaviors of the industrial system, and

63
 Results and discussion

compare the lab-scale data with those obtained in the industrial plant. In fact, although
the lab-scale reactor used was a fixed-bed and the industrial is a fluidized-bed, the
study of the catalytic behaviors in a smaller scale is an useful way to find new factors
in order to improve the system and to achieve new goals for the industrial process.
We studied mainly two types of vanadayl pyrophosphate catalysts, renamed “A” and
“X” for confidential reasons (see Experimental); the work was divided in the
following parts:

1. Comparison between the catalytic performances of A and X catalysts under

either hydrocarbon-lean or hydrocarbon-rich conditions, which are
respectively the n-butane concentration used in fixed-bed (1.7%) and
fluidized-bed (4%).
2. Study of the effect of water over the A catalyst, by means of catalytic tests and
in-situ Raman analysis.
3. Study of the industrial ageing of the X catalyst, by means of catalytic tests and
Raman analysis.

4.1.1 Catalysts characterization

The industrial catalysts studied, A and X, are bulk-type, that means totally constituted
of active phase, without any support or diluting material. They differ in the
composition and in the method of preparation: the A catalyst is mainly made of VPP,
while the X one also contains Nb as a dopant. Both systems were industrially
synthesized by means of a patented method developed by Lonza [59, 40, 59]: this so
called “hydrothermal” route gives catalysts which are already activated, avoiding the
need for an equilibration period under reactive mixture.
These catalysts have been characterized by Raman spectroscopy, XRD and BET
surface area; the conditions are reported in the Experimental section. The Raman
analysis will be treated in the following section.

X-ray diffraction analysis

XRD analysis of A and X powder catalysts are reported in Figure 4.1.

64
 Results and discussion

Figure 4.1: XRD diffractograms of A and X catalysts.

It is evident that the diffraction patterns are attributed to (VO)2P2O7 crystalline phase:
in fact, even if the X catalyst is doped with Nb, its low amount could not be detected.
However, the X catalysts appears to be slightly less crystalline than the A, showing
broader lines.

BET surface area

The surface area of the catalysts has been measured by the BET method: for the A
catalyst it was 21 ± 2 m2/g, while for the X was 25 ± 2 m2/g.

4.1.2 The effect of n-butane concentration

The first aspect investigated was the different catalytic behavior of the undoped and
doped catalysts, under either hydrocarbon-lean or hydrocarbon-rich conditions.
Many hypothesis have been proposed in literature about the role of Nb in VPP
catalyst for the selective oxidation of n-butane to MA [96, 119], but these studies
were only limited to lab-scale samples. Our aim is to find scientific evidences of the
dopant effect of Nb also in industrial catalysts.

65
 Results and discussion

Catalytic tests were carried out in lab-scale reactor, using the same configuration for
both systems: the catalyst was loaded in pellet sized 30-40 mesh, using steatite as
inert filler for the void space. Identical were also the reaction conditions: 0.8 g of
catalyst, contact time W/F = 1,33 g∙s/ml, 36 ml/min as total flow. The experiments
were conducted in function of temperature, between 340°C and 420°C, which is the
typical range utilized at plant-scale. Two feed compositions were chosen: 1.7% n-
butane in air, that is the hydrocarbon-lean condition, normally utilized for fixed-bed
configuration, and 4% n-butane in air, that is the hydrocarbon-rich condition,
normally utilized for fluidized-bed configuration. The experimental results are
reported in Figure 4.2 and in Figure 4.3.

Cat A 1.7% n-butane in air

100 100
90 90

n-butane conversion (%)

80 80
Selectivities (%)

70 70
60 60
50 50
40 40
30 30
20 20
10 10
0 0
320 340 360 380 400 420 440
Temperature (°C)

Figure 4.2: Catalytic behavior of A catalyst in the selective oxidation of n-butane to MA in

hydrocarbon-lean condition as function of the temperature. Conditions: feed 1.7 mol% n-butane, 20%
O2, remaining: N2; W/F 1.33 g·s·mL-1 Symbols: n-butane conversion (); selectivity in MA (), COx
(), acids (acrylic and acetic) (), light ().

66
 Results and discussion

Cat A 4% n-butane in air

100 40
90

n-butane conversion (%)

80
30
Selectivities (%)

70
60
50 20
40
30
10
20
10
0 0
320 340 360 380 400 420 440
Temperature (°C)

Figure 4.3: Catalytic behavior of A catalyst in the selective oxidation of n-butane to MA in

hydrocarbon-rich condition as function of the temperature. Conditions: feed 4 mol% n-butane, 20% O2,
remaining: N2; W/F 1.33 g·s·mL-1 Symbols: n-butane conversion (); selectivity in MA (), COx (),
acids (acrylic and acetic) (), light ().

In general, in both conditions similar trends are shown: the conversion of n-butane
and the selectivity to COx both increase with temperature, while the selectivity to MA
decreases with temperature. However, the values of conversion and selectivity show a
significant difference in the two conditions, in particular a drop in activity, passing
from 80% to 20% of conversion shown at 420°C under hydrocarbon-rich conditions.
On the other hand, a higher n-butane concentration favors the selectivity, that
becomes slightly higher at low temperature, increasing from 67% to 76% at 340°C.
The lower activity shown under n-butane-rich conditions is attributable to both a
saturation of the catalyst surface (it is known that a limited number of active sites is
available on the surface of the VPP catalyst), and an over-reduction of V, which
makes the number of available oxidized V sites lower compared to more oxidizing
conditions (those which develop under n-butane-lean conditions).

67
 Results and discussion

The same experiments were carried out for the X catalyst, with the aim of
investigating the effect of Nb. The catalytic tests in hydrocarbon-lean and rich
conditions are reported in Figure 4.4 and in Figure 4.5, respectively.

The Nb-doped catalyst exhibits a similar behavior under n-butane lean atmosphere
compared to the A catalyst (indeed, a slightly lower selectivity is shown with the X
catalyst at 400°C, at ca 70% n-butane conversion). Nevertheless, when the
hydrocarbon fed is 4%, the MA selectivity is substantially stable in the entire range of
temperature, accompanied by an increasing trend of the conversion.

Cat X 1.7% n-butane in air

100 100
90 90

n-butane conversion (%)

80 80
Selectivities (%)

70 70
60 60
50 50
40 40
30 30
20 20
10 10
0 0
320 340 360 380 400 420 440
Temperature (°C)

Figure 4.4: Catalytic behavior of X catalyst in the selective oxidation of n-butane to MA in

hydrocarbon-lean condition as function of the temperature. Conditions: feed 1.7 mol% n-butane, 20%
O2, remaining: N2; W/F 1.33 g·s·mL-1 Symbols: n-butane conversion (); selectivity in MA (), COx
(), acids (acrylic and acetic) (), light ().

68
 Results and discussion

Cat X 4% n-butane in air

100 60
90

n-butane conversion (%)

50
80
Selectivities (%)

70
40
60
50 30
40
20
30
20
10
10
0 0
320 340 360 380 400 420 440
Temperature (°C)

Figure 4.5: Catalytic behavior of X catalyst in the selective oxidation of n-butane to MA in

hydrocarbon-rich condition as function of the temperature. Conditions: feed 4 mol% n-butane, 20% O2,
remaining: N2; W/F 1.33 g·s·mL-1 Symbols: n-butane conversion (); selectivity in MA (), COx (),
acids (acrylic and acetic) (), light ().

The comparison with the A catalyst highlights that the X catalyst is clearly more
active, but the selectivity to MA is lower at 340°C, and quite similar (ca 66%) at
400°C. Therefore, because of the increased conversion at 400°C, the yield to MA
obtained (i.e., at a temperature and conversion close to those used industrially), is
clearly better with the X doped catalyst than with the A undoped one. This fully
agrees with what shown industrially; since industrially usually the reaction
temperature is fixed, the Nb-doped VPP exhibits a superior catalytic performance.
These results confirm the industrial performance: the doped-VPP has better catalytic
performances than the undoped one, but this gain is evident only at hydrocarbon rich
conditions. In fact, if we compare the tests conducted at 1.7% n-butane, the addition
of Nb does not give any improvement neither in selectivity nor in activity; therefore,
under hydrocarbon-lean conditions, the two catalysts behave very similarly. This can
be explained by taking into account that under hydrocarbon-rich conditions the

69
 Results and discussion

catalyst surface works with a non-optimal V4+/V5+ ratio, with a relatively large
concentration of reduced sites compared to oxidized V sites; in fact, it is known that
this feature is greatly affected by the gas phase environment [86]. Under these
circumstances, the Nb dopant could play its role, by favoring the oxidation of V4+ to
V5+ thus allowing for a more favorable V4+/V5+ ratio even under highly reducing
conditions [96]. In particular, Nb5+ acts by favoring the formation of controlled
amounts of δ-VOPO4, the V5+ compound which is known to be responsible for the
selective formation of MA [55]. Summarizing, when the reaction environment is more
oxidizing (namely at high temperature when the conversion is the highest, and at low
n-butane content in feed), the formation of δ-VOPO4 occurs even in absence of Nb,
explaining why A and X catalysts show the same performance. On the other hand,
when the system works at 4% n-butane, the environment is more reducing and Nb
becomes essential in order to allow the in-situ generation of the δ-VOPO4 phase, that
is the reason why the X catalyst exhibits its best performances in this condition.
In conclusions, the experiments carried out with the A and X catalysts at low and high
n-butane concentration have led to a significant result, since it has made possible to
demonstrate the role of Nb in affecting the V4+/V5+ ratio also in the industrial VPP
catalyst.

After the catalytic tests, we carried out some in-situ Raman spectra of A and X
catalyst with the aim of investigating which are the conditions needed for the
formation of the δ-VOPO phase in the two catalysts.

The Raman spectra for the A and X catalysts are reported in Figure 4.6 and Figure
4.7, respectively.

70
 Results and discussion

•
a.u. • (VO)2P2O7
αI αI–VOPO4
δ δ–VOPO4

δ
αI αI
• • αI αI TIME
4h

3h

2h

1h

0h
rt

1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200

Raman shift (cm-1)

Figure 4.6: In-situ Raman spectra of A catalyst collected during time on-stream analysis. Feed: air and
steam. Temperature: 400°C.

•
a.u. • (VO)2P2O7
δ δ–VOPO4

δ δ

•
• δ
TIME

4h
3h
2h

1h

0h

rt

1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200

Raman shift (cm-1)

Figure 4.7: In-situ Raman spectra of X catalyst collected during time on-stream analysis. Feed: air.
Temperature: 400°C.

71
 Results and discussion

The Raman analysis shows that the formation of the δ-VOPO4 phase occurs in
presence of air (without steam) for the X catalyst, while for the A catalyst it is
necessary the co-feeding of steam (without steam there was no formation of δ-
VOPO4; spectra not reported). These results confirm our hypothesis dealing with the
role of Nb in the formation of the δ-VOPO4 phase: in fact, while in the doped sample
the bands attributable to δ-VOPO4 phase increased during time-on-stream, in the case
of the A catalyst only a slight growth of these bands could be seen, while being less
intense than bands attributable to the unselective αI-VOPO4 compound.

4.1.3 Raman investigation of A catalyst: the effect of water

The Raman analysis described in the previous paragraph have shown the key role of
water in the modification of the catalyst surface, in particular its correlation with the
formation of δ-VOPO4 phase. Since the selective oxidation of n-butane to MA gives
four molecule of water per molecule of n-butane, water is an important co-product of
the reaction which could alter the composition of the surface phases. V/P/O
compounds are known to have the tendency to undergo changes as a consequence
of slight variations in either temperature or composition of the atmosphere in the
reaction condition [65, 115]: for this reason the variation of steam partial pressure
was investigated by many authors, in particular it was examined in correlation with
the P/V ratio [104] and the surface acidity [192]. It is clear that the steam partial
pressure increases in parallel with the conversion of n-butane and should enhance the
hydrolysis phenomena. Caldarelli et al. [198] reported some reactivity tests carried out
by co-feeding 10% of inlet steam using zirconia supported vanadyl pyrophosphate
catalysts with different P/V ratio, 1.1 and 1.5: in both cases the addition of steam led
to a decrease in the selectivity to COx and an increase in the selectivity to maleic acid,
being MA hydrolyzed to its acid. Nevertheless, scientific literature reports that water
could also negatively affects the VPP catalyst, because when a relatively high water
partial pressure is present, the hydrolysis of the active phase could occur [199, 200,
201]. In particular, water could dissociatively adsorb breaking V-O-V, V-O-P or P-O-
P bonds and form V-OH and/or P-OH groups [202]: when the hydrolysis of V-O-P

72
 Results and discussion

occurs, isolated or oligomerized VOx species are generated (containing either V4+ or
V5+) surrounded by phosphate groups [VOx + (PO4)n]. These oxidized phases rich in
V5+ are more active than VPP but at the same time they favor the consecutive
reactions of combustion, causing a drop in MA selectivity. Other scientific papers
dealing with the effect of steam have been reported: for example steam could have a
reversible blocking effect, responsible of a decrease in activity and a simultaneous
increase in selectivity to MA [203, 204]; on the other hand, the hydrolysis of the
active phase could evolve with the formation of an amorphous and thin layer of VxOy
which gives an oxygen-enrichment of the VPP surface [199, 200, 201],
Our aim was to investigate the effect of steam over an industrial VPP catalyst, both in
terms of catalytic performances and surface modification; in particular, we
investigated the effect of water at intermediate temperature (360°C) using the A
catalyst in hydrocarbon-lean conditions (1.7% n-butane, W/F 1.33 g·s/ml, catalyst in
pellets 30-40 mesh, steatite as inert filler). The results obtained in the catalytic tests
with and without the co-feeding of steam are reported in Figure 4.8.

70

60
Conversion and selectivities (%)

50

40

30

20

10

0
Without H2O 3% H2O

Figure 4.8: Results of reactivity experiments with and without the steam co-feeding over the A
catalyst. Conditions: feed 1.7 mol% n-butane in air (17 % O2), remaining: N2 and He; temperature
360°C. Symbols n-butane conversion (); selectivity to MA (), COx (), acids (acrylic and acetic)
(), light ().

73
 Results and discussion

As shown in Figure 4.8, the steam co-feeding has a positive effect on the catalytic
behavior of the A catalyst: in fact, while it gives a slight decrease in n-butane
conversion, a significant improvement in terms of selectivity occurs, since COx are
reduced from 47% down to 19% and MA increases from 52% to 63%. In order to
understand the modifications occurred, we carried out in-situ Raman experiments
reproducing the effect of steam over the catalyst surface: we fed a mixture of air and
steam over the A catalyst at 360°C and we collected spectra in function of time-on
stream. Raman spectra are reported in Figure 4.9.

•
• = (VO)2P2O7
αI = αI–VOPO4
a.u.
δ = δ–VOPO4

δ
αI αI
• αI αI
• TIME
4h

3h

2h

1h

0h

rt
1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200

Raman shift (cm-1)

Figure 4.9: In-situ Raman spectra of A catalyst collected during time on-stream analysis. Feed: air and
steam. Temperature: 360°C

In-situ Raman experiments show that the co-feeding of water favors the formation of
δ-VOPO4 phase, whose relative amount increases during time. As previously
reported, δ-VOPO4 phase is poorly active but very selective to MA and this agrees
with the catalytic results: in fact, 3% of steam led to a slight decrease in n-butane
conversion and an improvement in selectivity to MA. This result allows us to state
that a proper partial pressure of steam fed over an industrial undoped vanadyl

74
 Results and discussion

pyrophosphate catalyst could favor the formation of limited and isolated δ-VOPO4,
enhancing the catalytic performances of the system.

4.1.4 Study of the ageing of industrial VPP catalyst

Since the industrial VPP catalyst is widely used in several technologies, one key point
to address is the mechanism of its ageing and regeneration. The first point to be
considered is the nature of the VPO phases that may coexist with VPP: different
crystalline, polymorphic, oxidized, reduced and hydrated phases have been observed
and studied [76, 77, 79, 80]. The ageing of the catalyst is strictly related to the
reactivity of these phases, since modifications occurring on the catalyst surface and in
the bulk greatly affect its behavior. First of all, the equilibrium shape of VPP is not
platy but mostly prismatic and as a consequence of this, one possible mechanism of
ageing involving the structure is the progressive increase of thickness of the VPP
crystals, that makes the VPP more prismatic along the time of reaction [80, 205]. As a
consequence, a decrease of the selective (100) face exposure occurs with a decrease of
MA selectivity. Another ageing mechanism involves the over-reduction of vanadium
sites, leading to the local formation of V3+ species: it is known that these sites are
formed as a result of the strongly reducing reaction conditions, especially at high
C4/O2 ratio [126]. V3+ species may have a positive effect in some cases, but could
also evolve towards the formation of detrimental compounds such as VPO4, V(PO3)3
and V4(P2O7)3.
Dealing with industrial VPP, a frequent cause for ageing is the presence of an excess
of phosphorus: even if P is fundamental to keep an optimal compromise between
activity and selectivity, preventing the over-oxidation of vanadium [16, 63],
phosphorus enrichment on the surface could lead to the formation of VO(PO3)2 which
is detrimental for the catalytic activity [65]. On the other hand, also a loss of P could
happen, a phenomenon particularly relevant in industrial plants, which could lead to a
worsening of performance; this explains why P is added as a make-up in the inlet
feed. In particular, the loss of P occurs upon the reoxidation of VPP: the
pyrophosphate group [P2O7] gives rise to two phosphate group [PO4], producing

75
 Results and discussion

[VOx] species [104]. In this process a key role is played by water: in fact, the
formation of four molecules of water per MA molecule (not accounting those
accompanying the formation of CO and CO2) leads to a substantial hydration of the
surface. This means that when VOPO4 is already present, it will transform into
hydrated phases VOPO4·nH2O if the reaction temperature is below 350°C; on the
other hand, too much water will produce the segregation of phosphorus, forming
strong phosphoric acid sites and [VOx] species. These oxidized phases rich in V5+ are
more active than VPP and give an increment in the conversion of n-butane; otherwise,
at the same time they favor the consecutive reactions of combustion, causing a drop in
MA selectivity.
Another way of ageing is related to the adsorption of products and reactants during
the reaction: this phenomenon could cause the deactivation of the catalyst, since some
carbon deposits could occupy the active sites.
Despite what we have just reported about ageing, VPP industrial catalyst is
considered one of the most long-running bulk catalyst [40]. Otherwise, several
regeneration treatments of VPO catalyst have been patented using different reagents,
consisting in a reduction treatment with n-butane [206], H2, CO, CH4 or H2S [207],
sulfur oxides [208] or H2O [209]; in particular, the treatment with n-butane patented
by Lonza [206] has the aim of restoring the average valence state of less than 4+. The
mechanism of deactivation during long-term runs, i. e. for an industrial catalyst, is
probably more complex than being related only to P loss. In fact, various aspects are
involved in the regeneration process:

1. Desorption of adsorbed organic species formed during reaction, which are

responsible of the inhibition of the active sites [63].
2. Surface reconstruction of the active phase.

In this field, our work was focused on the study of the evolution of a VPP industrial
catalyst (sample M) during a long-term run (i. e. some years of run in a fixed-bed
industrial-scale reactor) with the aim of investigating the deactivation factors affecting
both the catalytic activity and the catalyst surface modifications.

76
 Results and discussion

Catalytic tests in lab-scale reactor

First, the catalyst ageing was monitored by means of catalytic tests in lab-scale reactor
as function of the temperature, in the usual fixed-bed conditions: 1.7%mol n-butane,
17%mol O2, contact time W/F 1,33 g·s/ml, total flow 36 ml/min, 0.8 g of catalyst in
pellets sized 30-40 mesh, steatite as inert filler. We chose different catalyst samples as
representative of the increasing age of the catalyst: to simplify, we have renamed
them with increasing numbers, corresponding to the period elapsed in the reactor. The
results of catalytic tests are summarized in Figure 4.10 and Figure 4.11.

Ageing of the M catalyst

90
n-butane conversion (%)

80
70
60
50
40
30
20
10
0
330 340 350 360 370 380 390 400 410 420 430 440 450
Temperature (°C)

Figure 4.10: Ageing of different samples of M catalyst in the selective oxidation of n-butane to MA:
conversion of n-butane as function of the temperature Conditions: feed 1.7 mol% n-butane, 20% O2,
remaining: N2. Age of the catalyst: fresh (); 4 years (), 5 years (), 6 years (), 7 years (), 8
years (□).

77
 Results and discussion

Ageing of the M catalyst

80
75
MA selectivity (%)

70
65
60
55
50
45
40
0 10 20 30 40 50 60 70 80 90
n-butane conversion (%)

Figure 4.11: Ageing of different samples of M catalyst in the selective oxidation of n-butane to MA.
MA selectivity versus n-butane conversion. Conditions: feed 1.7 mol% n-butane, 20% O2, remaining:

N2. Age of the catalyst: fresh (); 4 years (), 5 years (), 6 years (), 7 years (), 8 years (□).

The results reported in Figure 4.10 and Figure 4.11 clearly show the deactivation
trend of the industrial catalyst after 8 years of work. The fresh catalyst gives the
highest values of n-butane conversion in the entire range of temperature examined and
the highest MA selectivity, leading to the best catalytic performance in terms of both
activity and selectivity. The worsening due the ageing is represented by a significant
drop in activity, since the conversion of n-butane passes from 80% at 420°C for the
fresh catalyst, to less than 40% at 440°C for the 8-years old catalyst. Worth of note,
the worsening of the catalytic performance is more pronounced in the second half of
catalyst life, as shown by the low values of conversion measured for the “oldest”
samples. Nevertheless, the ageing of the catalyst is not simply a gradual reduction of
the catalytic performance due to long-time runs, but it has to be investigated also in
terms of surface modifications occurred.

78
 Results and discussion

Characterization by Raman spectroscopy

It is known from scientific literature that the evolution of the vanadyl pyrophosphate
catalyst during reaction is affected by the P/V ratio [16] and by the reaction condition,
namely the hydrocarbon/O2 ratio [86]. For this reason, one of the possible causes for
catalyst ageing is the formation of VPO phases which worsen the activity and the
selectivity. Therefore, in order to understand the catalyst evolution, we analyzed the
different samples by Raman spectroscopy, identifying the VPO phases present on the
catalyst surface. The spectra are reported in the following figures (Figure 4.12).

•
a.u
• (VO)2P2O7

VO(PO3)2

VO(PO3)2 8 years

•
•
8 years
7 years
6 years
5 years
4 years
fresh

1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200

Raman shift (cm-1)

Figure 4.12: Raman spectra of the industrial catalyst, collected for studying the effect of ageing; the
sample are named with their age. Spectra collected over grey and pale-blue particles.

It is shown that the fresh catalyst only contains vanadyl pyrophosphate, while the
progressive ageing leads to the formation of vanadyl metaphophate VO(PO3)2, as
demonstrated by the increasing relative intensity of its characteristics bands at 692,
957, 1216 and 1255 cm-1. Worth of note, the spectra were collected over grey and
pale-blue particles: in particular, over the pale-blue particles the formation of the

79
 Results and discussion

metaphosphate was so relevant to be detected also as an isolated phase, as shown in

the “8-years” sample. Since the VO(PO3)2 phase is characterized by a P/V ratio twice
as much that one of the VPP, this is a clear indication that ageing occurs through an
excessive phosphorus surface enrichment, that could lead to the segregation of the
metaphosphate and cause the drop of the catalyst performance. As already mentioned,
the P/V ratio greatly affects the catalyst behavior and even if a slight excess of
phosphorus is essential to reach the optimal performance, a too high P/V ratio has a
negative effect and reduces the catalyst re-oxidation rate [47, 65, 76]. In fact, when
the phosphorus content increases, it is more difficult to keep the vanadium in its
higher oxidation state, so as a consequence the selectivity to MA decreases. In
particular, the scientific literature reports that VO(PO3)2 can be easily formed when
the P/V ratio is in the range between 1.6 – 2.0 [76] and that this compound is inactive
in n-butane oxidation [210]. Even if its action could prevent the oxidation of the V4+
over the selective (100) plane when the P/V ratio is in slight excess [211], on the other
hand VO(PO3)2 has an important detrimental effect on the VPP catalyst when it is
present in high amount; therefore, we can explain the drop in activity shown during
ageing with the formation of the inactive VO(PO3)2.

Raman investigation was continued over the oldest sample, namely over particles
showing different colors and particular aspects: the most significant spectra are
reported in Figure 4.13, Figure 4.14 and Figure 4.15.

80
 Results and discussion

•
• (VO)2P2O7
a.u
(VO)HPO4·0,5H2O
(VO)HPO4·4H2O
◊ αI-VOPO4

1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200

Raman shift (cm-1)

Figure 4.13: Raman spectra of the “8-years” industrial catalyst collected over brown particles.

a.u

β-VOPO4

1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200

Raman shift (cm-1)

Figure 4.14: Raman spectra of the “8-years” industrial catalyst collected over yellow particles.

81
 Results and discussion

a.u Carbon deposits

1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200
Raman shift (cm-1)

Figure 4.15: Raman spectra of the “8-years” industrial catalyst collected over black particles.

Raman analysis demonstrate that the ageing process is not only due to the formation
of the vanadyl metaphosphate, but it is a more complex process in which other
compounds are involved, since hydrated phases (VOHPO4·nH2O), β-VOPO4 and coke
deposits have been detected. In Figure 4.13 Raman spectra have evidenced the
presence of the catalyst precursor VOHPO4·0.5H2O and of the vanadyl
hydrogenphosphate tetrahydrated VOHPO4·4H2O together with the VPP phase. This
indicates that, even if VPP still represents the main VPO phase, the catalyst surface
has suffered a strong process of hydration. Reminding that the reaction produces four
molecules of water per molecule of n-butane, the steam co-produced has an important
effect on the surface modification of VPP occurring during long-time permanence in
the reactor, since it could lead to the hydration of the active phase, with the formation
of VHP and even of more hydrated phases, such as VOHPO4·4H2O [65, 212]. We can
hypothesize that the progressive ageing of the catalyst has compromised the
desorption ability of the material, being this hypothesis confirmed also by the
presence of coke deposits (Figure 4.15). Moreover, the Raman spectrum reported in
Figure 4.14 shows that the ageing is accompanied also by the formation of the

82
 Results and discussion

oxidized phase β-VOPO4. In theory, the orthophosphate VOPO4 compounds are

formed after the thermal treatment of the precursor and they could be transformed to
VPP by the reactive mixture during the activation period. However, the different
VOPO4 compounds have a different attitude to be reduced to VPP, depending on their
morphology and structures [80]. Differently from the other VOPO4 phases, the
formation of β-VOPO4 occurs at higher temperature and through a consecutive
reaction scheme:

δ-VOPO4  γ-VOPO4  β-VOPO4

δ-VOPO4 and γ-VOPO4 are obtained by dehydration of the precursor respectively at

450°C in air and 680°C in O2 and γ-VOPO4 is transformed into β-VOPO4 at 780°C
[205]. Moreover, also αI-VOPO4 could be transformed to β-VOPO4 at the high
temperatures typically met in the hot-spots zones of fixed-bed reactors [19]. In
general, the presence of β-VOPO4 is frequent in used catalysts that have become
inactive after long-time runs: in fact, one the reason of the loss of the catalytic
performance is due to the β-VOPO4 phase, the presence of which is known to reduce
both the activity and the selectivity. Moreover, because of its greater stability at high
temperature compared with the other VOPO4 forms, once generated it cannot evolve
to any other compound [80].
Lastly, over the surface of the “8-years” catalyst the presence of carbon deposits has
been detected (Figure 4.15). More in details, the D4 band was detected at 1200 cm-1,
that is due to the A1g symmetry stretching of the amorphous carbon: in particular this
signal is correlated to the stretching of C sp2 and sp3 of disordered graphitic structures
[213, 214].
We suppose the carbon deposits are generated as a consequence of a strong and
irreversible adsorption of the organic compounds on the catalyst surface, being the
desorption process compromised. This phenomenon generates a covering layer over
the catalyst surface that blocks the active sites, as demonstrated by the signals of the
VPP bands that could not be detected any more.
The formation of coke over solid catalysts is one of the most frequent causes of
catalyst deactivation in industrial processes using acid catalysts, for example zeolites

83
 Results and discussion

in FCC, Pt/Al2O3 in reforming, Pt/zeolite or NiMoS/zeolite in hydroisomerization and

hydrocracking processes. The deactivation is caused by the poisoning of the active
sites through the blockage of their pores: the organic molecules are retained on the
catalyst surface because of their strong adsorption and their low volatility [215, 216].
The mechanism of deactivation depends on the reaction temperature and on the nature
of the reactants: in particular, from n-alkane (like n-butane) there is generally no coke
formation at low temperature (<150°C), but it happens at higher temperatures
(>400°C). Since VPP is a bifunctional catalyst and possesses both acid and redox, the
acid sites could catalyze the condensation and rearrangement of the unsaturated
species, in our case the intermediates molecules involved in the scheme of reaction
from n-butane to MA, i.e., butenes and butadiene [16, 19]. If we remind that one of
the consequences of the ageing is the phosphorus-enrichment occurring on the catalyst
surface, this means that an increased acidity is localized on the surface and it could
favor the condensation reaction of the organic molecules adsorbed.
In conclusion, the ageing of the industrial vanadyl pyrophosphate catalyst leads to a
progressive worsening of the overall catalytic performances due to a complex process
characterized by a combination of effects:

1. Phosphorus-enrichment, represented by the formation of VO(PO3)2, an

unselective and detrimental phase.

2. Formation of unselective V5+ phase, namely β-VOPO4.

3. Modification of the acidity of the material and worsening of the desorption

properties, as shown by the formation of hydrated phases (VOHPO4·nH2O)
and of carbon deposits.

4. Strong adsorption of the organic products of reaction, that causes the

formation of coke and the blockage of the active sites.

84
 Results and discussion

4.2 Synthesis of maleic anhydride from 1-butanol

4.2.1 Dehydration of 1-butanol

The synthesis of maleic anhydride from 1-butanol can be schematized in two steps:

1. Dehydration of 1-butanol to butenes

2. Selective oxidation of butenes to maleic anhydride

These reactions need catalytic systems with different properties, in particular acidity
for the dehydration and redox property for the selective oxidation. At the beginning
of the research, the first step of dehydration of 1-butanol to butenes (1-butene, 2-
butenes cis and trans) was studied, with the aim of investigating and improve the
efficiency of this step (Figure 4.16). Solid catalysts with acid properties like Keggin-
type polyoxometalates, vanadyl pyrophosphate (VPP) and vanadyl phosphate
dihydrated (VOPO4·2H2O) were chosen; they were tested feeding 1% 1-butanol in
He, in gas phase lab-scale reactor, with residence time of 1.33 g·s/ml.

Figure 4.16: Scheme of dehydration of 1-butanol to 1-butene and its isomerization to cis and trans 2-
butenes.

4.2.1.1 Reactivity tests with supported Keggin-type polioxometalates

catalysts

The catalysts studied are made of 30%wt. of Keggin-type polyoxometalate (active

phase) supported on CARi-ACT Q15 (silica gel, 70%wt.); in the list below the
composition of the active phases is reported:

1. Cs2HPW12O40;
2. Cs3PW12O40;

85
 Results and discussion

3. K3HSiW12O40;
4. K4SiW12O40.

Catalysts were delivered by the team UCCS (Dr Benjamin Katryniok, Prof. Sebastien
Paul and Prof. Franck Dumeignil), Unité de Catalyse et de Chimie du Solide UMR
CNRS, Equipe VALBIO, University of Lille and École Centrale de Lille, within the
framework of the European Project EuroBioRef.

Catalytic tests were carrried out using the following conditions: 0.8 g of catalyst sized
in pellet 30-40 mesh, diluted with an inert filler (steatite); contact time 1.3 s and inlet
feed composition 1% 1-butanol in He, the latter with a flow rate of 36 ml/min.
It is worth noting that in completely salified polyoxometalates, like cat.
Cs3PW12O40and cat. K4SiW12O40, no acidity should be present, because of the lack of
protons in the structure; nevertheless, even if in the method of preparation a suitable
amount of alkaline cation is added to compensate for the negative charge of the
cluster, some traces of the previous non-substituted heteropolyacid still remain.

First of all, we carried out a blank experiments in order to study the reactivity of 1-
butanol in gas phase in presence of an inert solid, used as filler, without any catalyst.
The aim is to investigate if the filler has some effect on the reactivity of 1-butanol; the
experiment was conducted under the same conditions and reactor configurations used
for the catalytic test, but without catalyst. The results of the blank test are reported in
Figure 4.17.

86
 Results and discussion

Blank test steatite 1% butanol in He

60

Conversion and selectivities (%) 50

40

30

20

10

Figure 4.17: Blank test, 1% of butanol in He at 250°C over steatite. Symbols: 1-butanol conversion
(); selectivities in butyraldehyde (),1,4- butadiene (), 1-butene (), 2-buteni (), isobutanol
(), n-butyl ether ().

Analyzing the results reported, we can state that 1-butanol is a very reactive molecule,
being converted only by thermal effect even in absence of an acid catalyst.
Nevertheless, its transformation is not selective, because together with the
dehydration products (butenes), also isomerization (isobutanol), condensation (n-butyl
ether) and selective dehydrogenation (butyraldehyde) products are formed. Another
blank test was carried out at a lower temperature (150°C): the same 1-butanol
conversion was achieved, but n-butyl ether and isobutanol also formed. Therefore, in
order to enhance 1-butanol conversion and to improve its selectivity to butenes for the
two-steps synthesis of maleic anhydride, an acid catalyst has to be added in the
process.
First of all we tested the catalytic performances of the four supported-polyoxometalate
samples in the reaction of 1-butanol dehydration in function of time on stream, at a
constant temperature. The conditions chosen were: temperature of 250°C, contact
time 1.3 g·s/ml, feed composition 1% 1-butanol in He (flow rate of 36 ml/min and
0.8 g of catalyst in 30-40 mesh pellets. In Figure 4.18, Figure 4.19 and Figure 4.20 the

87
 Results and discussion

results of the catalytic tests for Cs2HPW12O40, Cs3PW12O40 and K3HSiW12O40 are
reported.

Cs2HPW12O40 1% butanol in He 250°C

100 100

1-butanol conversion (%)

80 80
Selectivities (%)

60 60

40 40

20 20

0 0
0 1 2 3 4 5 6 7
Time (h)

Figure 4.18: Reactivity of Cs2HPW12O40 catalyst in function of time on stream at 250°C. Symbols: 1-
butanol conversion (), selectivities in 1-butene (*), 2-buteni (), butyraldehyde (), 1,4-butadiene
(), light ().

Cs3PSiW12O40 1% butanol in He 250°C

100 100
1-butanolo conversion (%)

80 80
Selectivities (%)

60 60

40 40

20 20

0 0
0 1 2 3 4 5 6 7
Time (h)

Figure 4.19: Reactivity of Cs3PW12O40 in function of time on stream at 250°C. Symbols: 1-butanol
conversion (), selectivities in 1-butene (*), 2-buteni (), butyraldehyde (), 1,4-butadiene (),
light ().

88
 Results and discussion

K3HSiW12O40 1% butanol in He 250°C

100 100

1-butanol conversion (%)

Selectivities (%) 80 80

60 60

40 40

20 20

0 0
0 1 2 3 4 5 6 7
Time (h)

Figure 4.20: Reactivity of K3HSiW12O40 in function of time on stream at 250°C. Symbols: 1-butanol
conversion (), selectivities in 1-butene (*), 2-buteni (), butyraldehyde (), 1,4-butadiene (),
light ().

The behavior of the 3 samples examined is very similar; at 250°C, all of them convert
completely 1-butanol, with almost total selectivity to 1-butene and 2-butenes. There
are very small amounts of butyraldehyde and of an identified compound (maybe a
light hydrocarbon). Minor differences between the three samples concern the
selectivity to each of the butenes. In particular, the catalytic behavior of sample
Cs3PW12O40 is identical to that obtained with the silica-alumina catalyst tested in the
past; all samples (and the silica-alumina as well) however show a different behavior
compared to alumina. The ratio between the two butenes isomers is close to the
thermodynamic one [217]; this means that the isomerisation of the double bond in the
primary product of dehydration (1-butene) to yield 2-butenes is very quick.

Finally the results of the experiment with the K4SiW12O40 catalyst is reported in
Figure 4.21. As shown in the graph, it was the least active catalyst because even at
250°C it did not reach complete conversion, and also gave the lowest C balance with
more than 20% yield to missing C (1-butene yield 27%, 2-butenes yield 47%).
Despite all this, the conversion was stable during the experiment time (as it was for

89
 Results and discussion

the other samples, also, but in that case conversion was total); this indicates that the
formation of coke did not affect so much the active sites for 1-butanol dehydration.
Moreover, the catalyst showed a lower selectivity to 2-butenes, which also confirms
that the catalyst was less active, because at 250°C it did not reach equilibrium for
olefins isomerization; an alternative explanation is the selective poisoning of sites for
isomerization.

100

80
Conversion, selectivity %

butanal

light
60
butadiene

1-butene
40
2-butenes

20

0
0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 5 5,5 6 6,5
Time (h)

Figure 4.21: Reactivity of K4SiW12O40 in function of time on stream at 250°C Symbols: 1-butanol
conversion (), selectivities in 1-butene (*), 2-buteni (), butyraldehyde (), 1,4-butadiene (),
light ().

Afterwards, we tested the performance of the four samples in the dehydration of 1-

butanol in function of temperature, in order to evaluate any possible differences in
catalytic performances. The conditions chosen for 1-butanol dehydration were: feed
composition 1% 1-butanol in He (flow rate 36 ml/min), range of temperature variable
from 130 to 310°C, depending on 1-butanol conversion obtained with the catalyst
tested, contact time 1.3 s, 0.8 g of catalyst formed in pellets sized 30-40 mesh. First
we started tests at 310°C, then we decreased the temperature until 1-butanol
conversion was lower than 99% and finally we increased temperature until the
conversion was total. In Figure 4.22, Figure 4.23, Figure 4.24 and Figure 4.25 the

90
 Results and discussion

results of the catalytic tests in terms of conversion of 1-butanol and yield to the
products are reported.

In order to carry out the experiments in function of temperature, we had to regenerate

the catalyst because they had been already used for time-on stream tests. The
procedure adopted for regeneration was a treatment in flowing air at 350°C for 25 h
(air flow rate 36 ml/min for 0.3 g catalyst) in the glass reactor of the lab plant.

Figure 4.22: Catalytic test of Cs2HPW12O40 in function of temperature.

91
Results and discussion

Figure 4.23: Catalytic test of Cs3PW12O40 in function of temperature.

Figure 4.24: Catalytic test of K3HSiW12O40 in function of temperature.

92
 Results and discussion

Figure 4.25: Catalytic test of K4SiW12O40 in function of temperature.

It is possible to comment as follows:

1. The four samples examined show important differences in terms of activity: in

general, catalysts with Cs and P were more active than those with K and W.
With Cs2HPW12O40 total conversion of 1-butanol was shown at 170°C; this
was the most active sample. If we compare Cs2HPW12O40 with Cs3PW12O40,
and K3HSiW12O40 with K4SiW12O40, we notice that in both cases an increased
acidity (due to the presence of protons) led to higher activity at low
temperatures. Nevertheless, it is amazing that even with samples which do not
hold protons, still remarkable conversions were achieved at moderate
temperatures. Worth of note, blank experiments with the empty reactors gave
negligible 1-butanol conversion in this temperature range.

2. For all catalysts, the dehydration of 1-butanol led to the same products, and the
highest selectivity was shown for 1-butene and 2-butenes cis and trans. We
identified also the presence of small amounts of n-butyl ether: the formation
of this compound occurred in particular at low temperature, and this behavior

93
 Results and discussion

was observed with all samples. Moreover, there were very small amounts of
butyraldehyde and butadiene. These results agree with works in literature
[171, 218, 219], in particular it is known that Cs2HPW12O40 catalyst for
dehydration of ethanol leads to the related olefin (ethylene), aldehyde
(acetaldehyde) and ether (diethyl ether).

3. Together with yields values, the “yield to missing carbon” is also plotted, that
is the % lack in C balance. We think that this lack of C is due to a strong
adsorption of organic compounds occurring at low temperatures, whereas at
high temperature this was due to the formation of coke. In fact, all spent
catalysts were black after the reactivity test because of the formation of coke
on the surface. Therefore, the optimal temperature range may be between 170
and 200°C (in function of the catalyst type); lower temperatures lead to poor C
balance because of alcohol “irreversible” adsorption, higher temperatures
because of coke accumulation. Usually, in the intermediate range, the C
balance is as high as 85-90%, which means that 10% of the converted 1-
butanol is missing; this value is not far from the experimental error.

In order to confirm the alcohol adsorption at low temperature, a desorption test with
catalyst K3HSiW12O40 was carried out. After one day of reaction (dehydration of 1-
butanol, 1% 1-butanol in He – flow rate 36 ml/min – at 170 °C), we stopped the feed
and we kept the sample in flow of He still at 170°C, monitoring the evolution of
organics. We plotted the number of C moles in function of the desorption time: Figure
4.26 shows that after 20 hours there were still organic compounds leaving the catalyst
surface, namely 2-butenes, 1-butene and 1-butanol.

94
 Results and discussion

9,0E-09

8,0E-09

7,0E-09 1-butanol

6,0E-09 n-butil ether

Carbon (mol)

5,0E-09 butyarldehyde

4,0E-09 1-butene

3,0E-09 2-butenes

2,0E-09 total carbon

1,0E-09

0,0E+00
0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0
Time (h)

Figure 4.26: Absolute quantity of organic compounds desorbed during the K3HSiW12O40 catalyst
“cleaning” with He, at 170°C.

In particular, the value of carbon missing is greatly affected by temperature: it is more

relevant at low temperature, when 1-butanol conversion is relatively low, and at high
temperature, when the conversion is total.

In conclusions, the catalytic activity of the four polyoxometalates in the dehydration

of 1-butanol is strictly correlated with their acid strength: as reported in literature
[171], POM acidity could be determined by the deprotonation energy, representing the
cluster attitude as H+ donor. This energy, and consequently both the activity and the
selectivity, depends on the structure composition of the catalytic system: variation on
it could lead to important modification of the acidity without alteration of the
crystalline structure, being the latter a peculiar distinctive trait of Keggin-type POM.

An important key point of POM behavior is the nature of the central atom: a decrease
in the oxidation state leads to an increase on the number of protons necessary for the
overall charge balance of the structure and a consequent rise of the anionic charge.
This phenomenon produces an increase on the deprotonation energy and a reduction
of the acid power, resulting in a softening of the catalytic activity. This agree with the

95
 Results and discussion

results obtained in the catalytic tests, where the samples with P5+ as central atom
(cat.Cs2H and Cs3) showed to be more active than samples containing Si4+ (cat.K3H
and K4).

On the opposite, the substitution of H+ with alkaline cations (Cs+ and K+) caused a
modification on the secondary structure of the POM and an increase of the surface
area, due to a growth of the dimensions. However, this substitution produces a
reduction of the acid strength, because of the introduction of alkaline cations, having a
lower electronegativity than the protons. Moreover, it increases the electrons density
of the external shell of POM and consequently an higher deprotonation energy has to
be reached to catalyze the dehydration reaction: this produces a loss of the activity.
These remarks could explain the differences observed in the experiments performed
with the four POM samples: catalyst less substituted (cat. Cs2H and K3H) are more
active than their corresponding totally substituted (cat. Cs3 and K4). Furthermore,
catalysts which are nominally completely substituted still show acid characteristics, as
demonstrated in the previous experiments. Scientific literature confirms these
properties taking into account the method of preparation of the materials [182, 183]:
actually, the POM synthetized as precipitate is a mixture containing some traces of the
original non-substituted POM (H3PW12O40 and H4SiW12O40 respectively) in amount
depending on the condition of the preparation (reactant amount, calcination
temperature, precipitation method and activation treatment) and this allows cat. Cs2H
and K3H to exhibit acid properties.

The nature of the products obtained in the reactivity tests of 1-butanol agrees with the
studies reported in scientific literature concerning the dehydration of ethanol
catalyzed by POM [184]: the products are respectively its olefin (ethylene), aldehyde
(acetaldehyde) and ether (diethyl ether). This is due to the mechanism of alcohols
dehydration, which expects two competitive routes:

1. Intramolecular route: it leads to the formation of the olefin, through

decomposition of the alcohol adsorbed on the catalyst surface; it is the favorite
route, especially at high temperature;

96
 Results and discussion

2. Intermolecular route: it leads to the formation of the ether, through the reaction
between a physisorbed molecule of alcohol and a chemisorbed one; this
pathway is preferred at low temperature.

In addition to these principal routes, other collateral reaction could occur at high
temperature, generating aldehyde and coke.

97
 Results and discussion

4.2.2 Reactivity tests with VOPO4·2H2O catalysts

Vanadyl phosphate dihydrated (VOPO4∙2H2O) is a V5+ phase commonly used for the
so-called VPD route, which is one of the methods of preparation utilized for the
synthesis of the precursor VOHPO4∙0,5H2O [220]. This preparation is a “two-step”
route, since in the first step the formation of the phase VOPO4∙2H2O (VPD) occurs,
while in the second one the VPD is reduced by the addition of an alcohol or a mixture
of alcohols to finally obtain the precursor VOHPO4∙0,5H2O [29, 221, 222, 223].
Since the trasformation of the precursor to the active phase is topotactic in the
traditional synthesis in the organic medium [29], also in the VPD route the
dehydration to VHP takes place retaining its microstucture and for this reason the
control of the preparation of VOPO4∙2H2O is essential to fine-tune the morphology of
the active phase [224, 193].
Vanadyl phosphate dihydrated has a layered structure composed of repeated
octahedral [VO6] and tetrahedral [PO4]3- units arranged in the same tetragonal crystal
group of the  and VOPO4 phase, with tha addition of two crystallization water
molecules [193, 225, 226]. One of the charachteritics of VPD is that in general all of
the VOPO4 phases could convert more or less easily to VPD after an hydration
treatment; the VOPO4 phase that more exhibits this attitude is VOPO4 [227].
Concerning the dehydration of 1-butanol, VPD was suggested as possible catalyst for
the reaction since V/P/O based catalysts are known to have acid properties [90, 91,
92].
First of all we carried out a blank experiment in order to study the reactivity of 1-
butanol in gas phase as function of temperature in absence of the inert filler and
without any catalyst. The blank tests were carried either in inert or oxidizing
environment respectively, feeding 1% butanol in He or in air. In both cases the
conversion of 1-butanol was negligible, since the average value obtained was 5%.

After the blank tests the reactivity of VOPO4∙2H2O was investigated feeding 1% 1-
butanol in He with residence time W/F=1.33 g·s·ml-1: Figure 4.27 reports the results
obtained while varying the reaction temperature.

98
 Results and discussion

VPD 1% 1-butanol in He
80 100

90
70
80
60
Selectivities (%)

Conversion (%)
70
50
60

40 50

40
30
30
20
20
10
10

0 0
180 200 220 240 260 280 300 320
Temperature (°C)

Figure 4.27: Catalytic behaviour of VPD catalyst in 1-butanol dehydration as a function of

temperature. Conditions: feed 1 mol% 1-butanol in He, W/F 1.33 g·s· mL-1. Symbols: 1-butanol
conversion (). Selectivity to 1-butene (), 2-butenes (), n-butyl ether, () butyraldehyde (), lack
to carbon balance (+).

Comparing the results with the performances obtained with Keggin polyoxometalates,
we can observed that VOPO4∙2H2O is less active than POM catalysts, since the total
conversion of 1-butanol could not be reached even at 300°C. Moreover, the lack to
carbon balance indicated that the adsorbtion of 1-butanol over the catalyst surface
occurred at low temperature (30% at 200-220°C); also at high temperature the balance
was not total, since probably some coke was formed, as demonstrated by the carbon
missing between 20 and 25% at 280 and 300°C. Although the performance of VPD
catalyst in the dehydration of 1-butanol was at least worse than that one obtained with
POM, we decided to investigate the performances of the catalyst in a different
environment, in order to test if VOPO4∙2H2O could compensate for its lower activity
with an higher operational flexibility.
Since the aim of our study is the synthesis of MA starting from 1-butanol, we have the
keep in mind the possibility to carry out the process in the “one-reactor”
configuration, arranged with two separated catalytic beds (the first is the acid catalyst

99
 Results and discussion

for 1-butanol dehydration and the second is the redox one, for the oxidation of
butenes to MA). Since in this configuration 1-butanol has to be fed together with air,
the acid catalyst has to be able to perform the dehydration reaction in oxidizing
condition, that mean in presence of air. For this reason, we investigated the reactivity
of VOPO4∙2H2O feeding 1% 1-butanol in air with residence time W/F=1.33 g·s·ml-1:
Figure 4.28 reports the results obtained while varying the reaction temperature.

VPD 1% 1-butanol in air

90 100

80 90

70 80
Selectivities (%)

Conversion (%)
70
60
60
50
50
40
40
30
30
20 20
10 10

0 0
260 280 300 320 340 360 380 400

Temperature (°C)

Figure 4.28: Catalytic behavior of VPD catalyst in 1-butanol dehydration as a function of temperature.
Conditions: feed 1 mol% 1-butanol, 20% O2, remaining: N2; W/F 1.33 g·s· mL-1. Symbols: 1-butanol
conversion (). Selectivity to 1-butene (), 2-butenes (), CO (), CO2 (), furan () and “lights”
().

VOPO4∙2H2O catalyst showed a different behavior in presence of air: total conversion

was reached in the entire range of temperature (280-380 °C) and an high selectivity to
1-butene and 2-butenes was shown, accompanied by a limited formation of oxidative
products (COx). Even if on increasing the temperature the selectivity to 2-butenes
becomes lower, the VPD catalysts gives surprising performances at 300°C: in fact, the
total conversion of 1-butanol was accompanied by a total selectivity to butenes;
moreover, contrary to what shown with the POM catalysts studied before, with VPD

100
 Results and discussion

the carbon balance was close to 100%. For these reasons, VPD catalyst could be
considered as a interesting candidate for the 1-butanol dehydration in presence of air
in the “one-reactor” configuration.
In order to deeply investigate the behavior of VPD in an inert environment, we carried
out two reactivity tests as function of time-on stream at two temperatures, 260 °C and
320°C. The results are shown in Figure 4.29 and Figure 4.30.

VPD 1% 1-butanol in He at 260 °C

100 100

90 90

80 80

Conversion (%)
Selectivity (%)

70 70

60 60

50 50

40 40

30 30

20 20

10 10

0 0
0 1 2 3 4 5 6 7
Time (h)

Figure 4.29: Catalytic behavior of VPD catalyst in 1-butanol dehydration as a function of time-on
stream. Conditions: feed 1 mol% 1-butanol, 20% O2, remaining: N2; W/F 1.33 g·s· mL-1. Temperature:
260°C. Symbols: 1-butanol conversion (). Selectivity to 1-butene (), 2-butenes (), n-butyl ether,
() butyraldehyde () and furan (); carbon missing (+).

101
 Results and discussion

VPD 1% 1-butanol in He at 320 °C

80 100

90
70
80
60

Conversion (%)
Selectivity (%)

70
50
60

40 50

40
30
30
20
20
10
10

0 0
0 1 2 3 4 5 6 7 8

Time (h)

Figure 4.30: Catalytic behavior of VPD catalyst in 1-butanol dehydration as a function of time-on
stream. Conditions: feed 1 mol% 1-butanol in He; W/F 1.33 g·s· mL-1. Temperature: 320°C. Symbols:
1-butanol conversion (). Selectivity to 1-butene (), 2-butenes (), n-butyl ether, ().
butyraldehyde () and furan (); carbon missing (+).

The reactivity tests reported in Figure 4.29 and Figure 4.30 have shown that in both
cases the conversion of 1-butanol progressively decreases during the reaction time: in
fact, while in the first hour of reaction the catalyst gives a conversion between 80 and
90%, after seven hours the conversion dropped down to 50-55%. This means that a
process of deactivation was occurring and led to a worsening of the catalytic
performances. Despite the VPD catalyst have exhibited a good behavior in presence
of air, this behavior was not followed in He, since the material deactivated very fast
after few hours of reaction, at both high and low temperature: for this reason our
interest to VPD was focused on its utilization in oxidizing environment.
The study on VOPO4∙2H2O catalyst in the dehydration of 1-butanol was continued
testing its behavior in air as function of time-on stream at 260°C. The graph of the
catalytic test is reported in Figure 4.31.

102
 Results and discussion

VPD 1% 1-butanol in air at 260°C

100 100

90 90

80 80

70 70

Conversion (%)
Selectivity (%)

60 60

50 50

40 40

30 30

20 20

10 10

0 0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Time (h)

Figure 4.31: Catalytic behavior of VPD catalyst in 1-butanol dehydration as a function of time-on
stream. Conditions: feed 1 mol% 1-butanol in He2; W/F 1.33 g·s· mL-1. Temperature: 260°C. Symbols:
1-butanol conversion (). Selectivity to CO2 (), acetic acid (), 1-butene (), 2-butenes (), n-
butyl ether, () butyraldehyde () and furan ().

The experiments have shown that VPD catalyst exhibited an interesting catalytic
performance also during time:

 at 260°C in air the conversion was almost total (always higher than 96%);
 despite the oxidizing condition, the catalyst gave essentially only the product
derived from the dehydration of 1-butanol, i. e. 1-butene and 2-butenes; in
fact, the carbon oxides and other oxidative products, such as butyraldehyde
and furan, were produced with selectivity lower than 3%.
 the catalyst kept stable performances while increasing the time of reaction; in
particular, on the contrary of its behavior revealed in He, the deactivation did
not occur (the value of conversion did not decrease) and the selectivity to
butenes did not drop.

In conclusion, the reactivity tests have demonstrated that the VOPO4∙2H2O is a

promising catalyst for the dehydration of 1-butanol to 1-butene and 2-butenes in

103
 Results and discussion

presence of air: in fact, not only the system is highly active and selective to butenes
but also it gives the products with stable perfomance during time. For this reason, it
could be considered a good candidate for the first bed in the “one-reactor”
configuration for the synthesis of maleic anhydride starting from 1-butanol.

4.2.3 Characterization of VOPO4·2H2O catalysts

VOPO4∙2H2O catalyst was characterized by means of Raman spectroscopy and XRD:

Figure 4.32 shows the most significant spectra collected on the fresh catalyst.

1
.
= VOPO4 2 H2O
1= 1-VOPO4

a.u.

1
1

1

1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200
-1
Raman shift (cm )

Figure 4.32: Raman spectra of fresh VOPO4·2H2O catalyst.

The Raman analysis have confirmed that VOPO4·2H2O is the main phase, as
demonstrated by the bands detected at 1038, 985, 954, 543, 280 cm-1. Otherwise, also
some traces of αI-VOPO4 were detected, confirming that strong relationship between
VOPO4·2H2O and αI-VOPO4: in fact, VPD could be easily transformed to αI-
VOPO4, since the latter is the VOPO4 phase which is more quickly hydrated to VPP.
The Raman analysis were carried out also on the VPD samples after reaction with 1-
butanol either in air or in He: the spectra collected are reported respectively in Figure
4.33 and Figure 4.34.

104
 Results and discussion

.
1 = VOPO4 2 H2O
1= 1-VOPO4

a.u.

1 
1

1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200
-1
Raman shift (cm )

Figure 4.33: Raman spectra of spent VOPO4·2H2O catalyst after reactivity tests carried out feeding 1%
of 1-butanol in air.

1
.
= VOPO4 2 H2O
1= 1-VOPO4
= (VO)2P2O7
a.u.

1 
1

1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200
-1
Raman shift (cm )

Figure 4.34: Raman spectra of spent VOPO4·2H2O catalyst after reactivity tests carried out feeding 1%
of 1-butanol in He (different positions of analysis).

First of all, the spectra registered over the catalyst after reaction in air have shown that
no substantial modifications occurred on the catalyst surface, since VOPO4·2H2O

105
 Results and discussion

continued to be the main phase, with some traces of αI-VOPO4. On the other hand,
even if VOPO4·2H2O and αI-VOPO4 were still present also in the catalyst after
reaction in He, other compounds had formed, as revealed by the bands attributed to
(VO)2P2O7. The formation of VPP could be explained considering the different gas
phase conditions: in fact, in the case of the reaction conducted in air, the environment
was oxidizing and preserved the oxidized state of V, as demonstrate by the presence
of VOPO4 phases. On the opposite, when VPD was used feeding 1% of 1-butanol in
He, the alcohol produced reducing conditions, that led to the formation of a reduced
vanadium phase, in this case (VO)2P2O7.
In order to confirm differences observed by Raman spectroscopy, we analyzed the
VPD sample by means of X-ray diffraction. The diffractograms of the fresh and spent
(after reaction) catalyst are reported in Figure 4.35.

= VOPO4.2H2O
a.u. = αI-VOPO4
= VPP
= ω-VOPO4

After reaction
(1-butanol in He)

Fresh

10 20 30 40 50
Position ( 2Theta)

Figure 4.35: XRD diffractograms of fresh and spent (after reaction) VPD catalysts

XRD analysis confirmed what was observed in the Raman spectra: the fresh catalyst
was mainly composed by VOPO4·2H2O and αI-VOPO4, while in the sample after
reaction (in reducing condition) the (VO)2P2O7 phase was observed.

106
 Results and discussion

4.2.4 Reactivity tests with vanadyl pyrophosphate DuPont catalyst

One of the peculiarity of the vanadyl pyrophosphate catalyst is its multifunctionality:

in fact, the presence of strong Brønsted and Lewis acid sites [228, 229] and of redox
sites represented by vanadium atoms in different oxidation states [19] has made
possible the application of VPP in selective oxidations of several organic molecules,
for example the oxidation of methacrolein to methacrylic acid [230], the
ammoxidation of propane to acrylonitrile [231], the ammoxidation of picolines [232],
the furan synthesis from butadiene [233] and the selective oxidation of decaline and
tetrahydrophtalic anhydride [234]. For this reason vanadyl purophosphate furnished
by DuPont was suggested also for the dehydration of 1-butanol. First al all, we
investigated the reactivity of VPP in 1-butanol dehydration in inert environment, i.e.
in presence of He. The results are reported in Figure 4.36.

VPP DuPont 1% 1-butanol in He

100 100
90 90
80 80
Selectivities (%)

70 70

Conversion (%)
60 60
50 50
40 40
30 30
20 20
10 10
0 0
220 230 240 250 260 270 280 290 300 310 320 330 340 350
Temperature (°C)

Figure 4.36: Catalytic behaviour of DuPont VPP catalyst in 1-butanol dehydration as a function of
temperature. Conditions: feed 1 mol% 1-butanol in He, W/F 1.33 g·s· mL-1. Symbols: 1-butanol
conversion (). Selectivity to 1-butene (), 2-butenes () and butyraldehyde ().

107
 Results and discussion

The experiments have shown that the catalyst is active and it reaches the total
conversion in the entire range of temperature explored; moreover, it is selective to 1-
butene and 2-butenes, with minor amount of butyraldehyde. In particular, vanadyl
pyrophosphate is more active than VPD (Figure 4.27), as demonstrated by the total
conversion of 1-butanol obtained even at low temperature. This is due to its acidity
[90, 91], that is probably higher than that one shown by VPD: the acidity of VPP
favours the dehydration of 1-butanol, leading to better catalytic performances.

In order to investigate if these performances are stable during time on stream or if

deactivation occurs, the reactivity of the catalyst was examined at 260°C as function
of the reaction time: the results are reported in Figure 4.37.

VPP DuPont 1% 1-butanol in He at 260°C

100 100

90 90

80 80
Selectivities (%)

Conversion (%)
70 70

60 60

50 50

40 40

30 30

20 20

10 10

0 0
0 1 2 3 4 5 6 7 8 9 10
Reaction time (h)

Figure 4.37: Catalytic behaviour of DuPont VPP catalyst in 1-butanol dehydration as a function of
reaction time. Conditions: feed 1 mol% 1-butanol in He, W/F 1.33 g·s· mL-1. Temperature: 260°C.
Symbols: 1-butanol conversion (). Selectivity to 1-butene (), 2-butenes (), butyraldehyde ().

The experiments have confirmed that not only VPP is active but also that it does not
deactivate as fast as VPD in the same conditions: in fact, the value of conversion
decreased only from 99% to 95% after 9 hours of reactions. Despite this, still the

108
 Results and discussion

deactivation shown is non-negligible. This is an interesting result, since vanadyl

pyrophosphate catalyst exhibits acid and redox properties, it could be a promising
catalyst for the “one-pot” reaction to maleic anhydride from 1-butanol. For this reason
we decided to carried out the study of the catalytic behaviour in presence of air, as
function of the reaction time at 260°C: the results are reported in Figure 4.38.

VPP DuPont 1% 1-butanol in air

100 100

90 90

80 80
Selectivities (%)

Conversion (%)
70 70

60 60

50 50

40 40

30 30

20 20

10 10

0 0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Reaction time (h)

Figure 4.38: Catalytic behaviour of DuPont VPP catalyst in 1-butanol dehydration as a function of
reaction time. Conditions: feed 1 mol% 1-butanol in air, W/F 1.33 g·s· mL-1. Temperature: 260°C.
Symbols: 1-butanol conversion (). Selectivity to CO (), CO2 (), 1-butene (), 2-butenes (),
acetic acid + acrylic acid (), maleic anhydride (MA) (), phtalic anhydride (PA) (), butyraldehyde
(), furan (+).

Figure 4.38 shows that the catalyst now has a stable catalytic behavior, with no
deactivation, as it was for the VPD catalyst under the same reaction conditions.
However, it was clearly less selective than the VPD, because of the greater formation
of by-products such as butyladehyde and other oxygenated compounds.

Concluding, the best performance, in terms of both catalyst stability and selectivity to
butenes, was shown by the VPD catalyst in oxidizing conditions.

109
 Results and discussion

4.2.5 1-butanol oxidehydration to MA with vanadyl pyrophosphate

DuPont catalyst

The reactivity tests carried out with DuPont vanadyl pyrophosphate in the dehydration
of 1-butanol has demonstrated that the material could be used to produced butenes in
presence air with interesting results. Since VPP is a multifunctional catalyst, we
decided to investigate if it could be utilized also for the “one-pot” configuration to
produce directly MA from 1-butanol, i. e. it could catalyze the dehydration of 1-
butanol and the selective oxidation of butenes to MA.

4.2.5.1 Thermal reactivity experiments: blank tests

First, we carried out experiments aimed at studying the reactivity of 1-butanol in the
gas phase, without any catalyst, both with and without oxygen. Figure 4.39 shows the
results obtained in the presence of oxygen; the empty space in the reactor was filled
in with inert material (steatite). The range of temperatures chosen for the experiments
was that typically used for butene oxidation to MA, that is, between 300°C and
400°C. Under aerobic conditions, with the feed composition used, 1-butanol already
converted completely at 370°C.

Figure 4.39 shows the main products obtained, i.e. butenes, butyraldehyde, CO and
CO2; the latter prevailed at 400°C, while olefins were the dominant products at the
intermediate temperature of 350°C. However, the overall selectivity to these
compounds was largely lower than 100% at 300°C, whereas it was close to 80% at
350, 370, and 400°C. In fact, many other by-products formed, especially at low
temperature (these compounds were not quantified individually): formaldehyde
(which showed an increasing yield with the rise in temperature), formic acid, acetic
acid, acrylic acid, tetrahydrofuran, dihydrofuran, 1-propanol, propionaldehyde,
dibutylether, and the butyl esters of formic acid, acetic acid, propionic acid, and
butanoic acid; other olefins were propylene, isobutene, and butadiene.

110
 Results and discussion

Figure 4.39: Catalytic behaviour of the “inert” (steatite) used in chemical-1-butanol (CB)
oxidehydration as a function of temperature. Conditions: feed 1 mol% 1-butanol, 20% O2, remaining:
N2. Symbols: 1-butanol conversion (). Selectivity to butenes (), butadiene (), acetic acid +
acrylic acid (), CO (), CO2 (), and butyraldehyde ().

The overall amount of these by-products was considerably lower at T  350°C than at
300°C; those which formed in greater amounts at both temperatures were
propionaldehyde, tetrahydrofuran, and butylformate. Moreover, 2-butenal, benzene,
and 2-ethylacrolein (compounds which, however, did not form at 300°C) were also
observed. At 400°C, the only by-products obtained (besides those shown in Figure
4.39) were: acrolein, furan, acetic acid, methylvinylketone, tetrahydrofuran, and
benzene. These results indicate that 1-butanol is extremely reactive and transforms
into a wide range of oxidised compounds, which are typically observed in gas-phase
radical oxidations; the presence of a solid in the reactor may facilitate the occurrence
of these reactions, especially those involving the oxidative transformations of both the
reactant and the products intermediately formed.
When the same experiments were carried out with an oxygen-free feed, but with
steatite always filling the reactor, 1-butanol conversion was 40% at 300°C, 88% at
350°C and 98% at 400°C; the prevailing products were butenes (overall selectivity
56%, 68% and 78% at 300, 350 and 400°C, respectively), whereas the selectivity to

111
 Results and discussion

butyraldehyde was always lower than 3%. It may be concluded that in the temperature
range examined, the thermal dehydrogenation of 1-butanol occurs only to a minor
extent; in the presence of oxygen, however, the formation of butyraldehyde becomes
much greater. On the other hand, the thermal dehydration of 1-butanol is very
common under all the conditions examined.
Because of the significant contribution of the “inert” material used to fill the void in
the reactor, we compared the catalytic behaviour obtained with steatite and that shown
either in the presence of corundum  another material typically used as an inert filler
for catalytic experiments in lab-scale reactors  or without any inert material at all. In
the latter case, the aim is that of highlighting the contribution of homogeneous
reactions, due to the fact that the glass walls of the reactor should not contribute to 1-
butanol conversion. Figure 4.40 compares the results obtained with steatite (data
taken from Figure 4.39) and corundum in an empty reactor, at 370°C.

Figure 4.40: Results of reactivity experiments as a function of the inert material filling the reactor
(with no catalyst). Conditions: feed 1 mol% 1-butanol, 20% O2, remaining: N2. Temperature 370°C.

It can be seen that, with the empty reactor, 1-butanol conversion was lower than that
recorded when the inert was filling the reactor. The nature of the products formed was
similar in the three cases, but with a different relative amount of the various

112
 Results and discussion

compounds. Based on these results, it was decided not to fill the empty space in the
reactor with any material during catalytic experiments; since the main role of the inert
material was that of inducing the heating of the gaseous flow before reaching the
catalytic bed, we modified the set-up of the reactor’s pre-treatment inlet feed, in order
to achieve an efficient gas feed heating even without the inert material.
The most important outcome of these experiments is that, in order to be selective to
MA, a catalyst for 1-butanol oxidative dehydration working in an oxidising
atmosphere has to be very efficient in 1-butanol dehydration, in order to further
enhance the rate of 1-butene formation with respect to dehydrogenation, thus limiting
the parallel formation of butyraldehyde; the latter is the precursor of several of the
many by-products identified, but cannot be transformed into MA. Furthermore,
butyraldehyde must also be very efficient in 1-butene oxidation into MA. This
suggests that vanadyl pyrophosphate (VPP)  the catalyst used industrially for n-
butane oxidation to MA  is the possible candidate catalyst for this reaction, since not
only is VPP selective in butene oxidation into MA [191, 192], but it is also
characterised by the acidic properties needed for alcohol dehydration [89, 235]. In
fact, 1-butanol has been previously used in some studies as a possible reactant for MA
synthesis [236, 237], for the purpose of demonstrating the polyfunctional
characteristics of VPP.

4.2.5.2 The effect of temperature

After the blank tests it was possible to choice the best inert filler and the best reactor
configuration, so the catalytic tests were carried out using corindone and not filling
the empty space of the reactor above the catalytic bed. Vanadyl pyrophospate catalyst
furnished by DuPont was suggested as the candidate to study the one-pot
oxidehydration of 1-butanol to MA. At first, the reactivity was investigated feeding
1% 1-butanol in air with residence time W/F=1.33 g·s·ml-1Figure 4.41 reports the
results obtained using the VPP catalyst, while varying the reaction temperature.

113
 Results and discussion

Figure 4.41: Catalytic behaviour of the VPP DuPont catalyst in chemical-1-butanol oxidehydration as
a function of temperature. Conditions: feed 1 mol% 1-butanol, 20% O2, remaining: N2; W/F 1.33 g·s·
mL-1. Symbols: 1-butanol conversion (). Selectivity to Maleic Anhydride MA (), 1-butene (), 2-
butenes (), acetic acid + acrylic acid (), CO (), CO2 (), Phthalic anhydride PA () and “lights”
().

Under the conditions chosen, the conversion of 1-butanol was completed over the
entire range of temperature investigated; products obtained were MA, light acids
(acrylic and acetic acid), carbon oxides, phthalic anhydride (PA), and minor amounts
of other oxygenated compounds: furan and formaldehyde. Butenes formed at 340°C
with a molar ratio between the two positional isomers that was different from the
equilibrium value; the latter is close to 3.7 at 330°C for 2-butenes/1-butene molar
ratio [217], whereas the experimental one was close to 2 at 300°C, and to 1 at 320 °C
and 340°C. The decrease in the molar ratio shown at increasing temperature and the
concomitant formation of MA and PA suggest that the isomerisation of 1-butene, the
primary product of 1-butanol dehydration, into 2-butenes is slower than the
consecutive oxidation occurring upon the olefins, and that 2-butenes are more quickly
oxidised than is 1-butene.
The formation of PA, with a maximum yield of 12% shown at 340°C, may occur via a
Diels-Alder reaction between the intermediately formed butadiene and MA. In the

114
 Results and discussion

case of n-butane oxidation, the formation of PA becomes important only under the
conditions of surface saturation due to the coverage of active sites by olefinic
intermediates, an event that  with the VPP catalyst  only occurs under alkane-rich
feed conditions [238, 239, 240]. Under these circumstances, the over-reduction of the
V sites leads to the prevailing occurrence of bimolecular reactions, such as the
cycloaddition of MA and butadiene, with a minor contribution of olefin oxidation to
produce more MA. However, it is worth remembering that, in the case of n-pentane
oxidation to MA and PA, the mechanism proposed did not involve any Diels-Alder
reaction between pentadiene and MA, but rather the oxidation of dialkylaromatics,
which are formed by olefin dimerisation, oxidative dehydrocyclisation. and
aromatisation [241, 242, 243]. Therefore, it cannot be ruled out that even in the case
of 1-butanol oxidation, the formation of PA may indeed occur by the oxidation of o-
xylene, the latter being formed by the oxidative dehydrocyclodimerisation of butenes.
Indeed, it is worth noting that during 1-butanol oxidation the butadiene selectivity
recorded was very low, i.e. less than 1%, over the entire temperature range examined
(for example, it was 0.2% at 340°C). Even though butadiene can be oxidised to MA,
if butadiene were the key reaction intermediate for MA and PA formation, it would
have formed in a greater amount, since the reactivity of butenes (which were formed
in large amounts, see Figure 4.41) over VPP catalyst is not much different from that
of butadiene (both are by far more reactive than n-butane) [27, 191, 192].
The highest selectivity to MA was observed at 340°C (39%); in the range 300-340°C,
the increase in MA selectivity occurred with a concomitant decrease in selectivity to
butenes, whereas at higher temperatures, the MA selectivity decline was due to the
prevailing formation of CO and CO2. Acetic and acrylic acids also formed in
relatively high amounts, with a maximum overall selectivity close to 20% at 320°C;
the two compounds formed in similar amounts. Butyraldehyde selectivity was very
low (e.g. 0.1% at 340°C).
These results indicate that at low temperatures the prevalent reactions were: i) the
transformation of 1-butanol into butenes, and ii) the oxidative cleavage of either
alcohol or olefins into acetic acid and acrylic acid. When the temperature was raised,

115
 Results and discussion

butenes were transformed into MA and PA, whereas the light acids were oxidised into
COx.

4.2.5.3 The effect of contact time

In order to investigate the scheme of reaction and identify the primary and the
secondary products, catalytic tests at different contact times were carried out. The
temperature chosen was 340°C, i.e. the condition where the DuPont catalyst showed
the highest MA yield. Figure 4.42 plots the effect of contact time on catalytic
performance, at 340°C.

Figure 4.42: Catalytic behaviour of the VPP DuPont catalyst in chemical-1-butanol oxidehydration as
a function of the W/F ratio. Conditions: feed 1 mol% 1-butanol, 20% O2, remaining: N2; temperature
340°C. Symbols: 1-butanol conversion (). Selectivity to Maleic Anhydride MA (), 1-butene (),
2-butenes (), acetic acid + acrylic acid (), CO (), CO2 (), Phthalic anhydride PA () and
“lights” ().

From these experiments, the kinetic relationship between butenes and both MA and
PA is apparent; in fact, the selectivity to both olefins declined when increasing the
contact time, while  at the same time  the selectivity to MA and PA increased. The

116
 Results and discussion

molar ratio between 2-butenes and 1-butene decreased from a value close to 2.2 (a
value which still is lower than the thermodynamic one), shown for a very short
contact time, to a value close to 1. This is a further proof that the isomerisation of 1-
butene to butenes is relatively slow at low temperatures, under conditions at which the
rate of the consecutive oxidation is negligible. Thereafter, when either the temperature
or the contact time is increased, the consecutive oxidehydrogenation into MA occurs
at a faster rate on 2-butenes than on 1-butene.
The selectivity to butadiene (not shown in the figure) also decreased, but remained
always lower than 1%, even for very short contact times. It can also be seen that the
conversion of 1-butanol was less than 100% (but close to 98%) only for extremely
short contact times. A surprising effect was that in the W/F range between 0.6 and 1.3
g·s·mL-1, the only consecutive reaction was the transformation of butenes into MA,
whereas the formation of all the other products, including COx and PA, did not show
any further increase. This may be due to the fact that under conditions in which the
surface coverage is small, because of the low amount of residual butenes, the catalyst
becomes very selective in the oxidation of butenes into MA. In other words, the
optimal conditions for the VPP catalyst should be those in which a high O2/(1-butanol
+ olefins) ratio makes it possible to maintain a clean and selective surface, with a
relatively abundant concentration of “free” oxidising sites. These hypotheses will be
confirmed by the experiments described below. Another somehow surprising result is
that the initial selectivity  that is, the selectivity extrapolated to nil conversion  was
almost zero for CO and CO2. This means that the parallel combustion of 1-butanol
gives a marginal contribution to the formation of these compounds, at least at 340°C,
whereas the major contribution derives from the consecutive overoxidation of
butenes.

4.2.5.4 The effect of oxygen partial pressure

Considering the results obtained in the catalytic experiments in function of contact

time, it could be hypothesized that the selective path to MA and PA, instead of the
unselective one to COx and acids, was function of the availability of V5+ sites on the

117
 Results and discussion

catalyst surface. This means that when the consecutive transformation of butenes to
MA is not kinetically favoured, a great amount of butenes keep adsorbed on the
catalyst surface, occupying the active sites and giving a lowering in MA yield; on the
other hand, when olefins are converted to the oxidized products rapidly, the
concentration of free active sites is higher, giving and enhancement of the catalytic
performances. In order to deeply investigate the role of oxidized vanadium sites
experiments varying the oxygen partial pressure were carried out. Figure 4.43 shows
the effect of oxygen molar fraction on catalytic behaviour.

Figure 4.43: Catalytic behaviour of the VPP DuPont catalyst in chemical-1-butanol oxidehydration as
a function of the O2 molar fraction. Conditions: feed 1 mol% 1-butanol; temperature 340°C, W/F 1.33
g s mL-1. Symbols: 1-butanol conversion (). Selectivity to Maleic Anhydride MA (), 1-butene (),
2-butenes (), acetic acid + acrylic acid (), CO (), CO2 (), Phthalic anhydride PA () and
“lights” ().

Clearly, an increase in the O2 content led to an increase in both MA and PA

selectivity, whereas there was no effect on light acid selectivity (this was confirmed
by repeated experiments carried out by increasing and decreasing the temperature; this
behaviour turned out to be fully reproducible). Selectivity to CO and CO2 increased
slightly in the range 0.04-to-0.10 O2 molar fraction, then decreased until the O2 molar

118
 Results and discussion

fraction was equal to 0.20 (at the latter conditions the maximum MA and PA
selectivity of 39 and 12%, respectively, was recorded); lastly, both increased again
when the O2 molar fraction was equal to 0.26. It is evident that both low (less than
0.10 molar fraction) and high (more than 0.20) oxygen partial pressure lead to poor
selectivity either because of the prevailing selectivity to olefins, or because of the
dominant CO and CO2 formation. Therefore, there is a range where the increased O2
partial pressure shows a strongly positive effect on both MA and PA yields, because
of the decreased selectivity to butenes (precursors for MA and PA formation) and COx
(formed by the consecutive overoxidation of both MA and PA). It is important to note
that such an effect is not observed in the reaction of MA synthesis by n-butane
oxidation: in this case, in fact, the interaction between the reactant and VPP surface is
very weak. Therefore, the saturation of the catalytically active surface is shown only
for very high n-butane partial pressure (under the so-called hydrocarbon-rich feed);
otherwise, being the concentration of adsorbed olefins very low, there is a large
availability of oxidising Vanadium sites. In fact, the best selectivity to MA from n-
butane was observed with the lowest O2/alkane molar ratio in the feed, when different
feed compositions were compared under differential conditions [244]. This could also
explain why, under normal conditions, PA forms in traces only during n-butane
oxidation. Instead, from 1-butene the effect of O2 partial pressure was similar to that
observed with 1-butanol, with a maximum MA selectivity obtained for a
O2/hydrocarbon molar feed ratio close to 30 [191].
In general, it may be concluded that for a consecutive reaction pattern leading in the
end to CO2, with MA and PA as intermediate products, the selectivity to these latter
compounds is a function of the fraction of oxidising sites on the catalyst surface
which is available under steady state conditions, the latter being, in turn, a function of
both the O2/reactant feed and the nature of the reactant [89, 191, 245].

4.2.5.5 The effect of 1-butanol concentration

As shown by the previous tests, the availability of V5+ sites favours the selective
oxidation of butenes and butadiene to MA and PA, so theoretically an increase in the

119
 Results and discussion

1-butanol partial pressure should produce a reduction on MA selectivity. VPP DuPont

catalyst was tested using 2% 1-butanol in function of the temperature, without varying
the other reaction conditions (see Figure 4.44).

Figure 4.44: Catalytic behaviour of the VPP DuPont catalyst in (chemical)-1-butanol oxidehydration
as a function of the temperature. Conditions: feed 2 mol% 1-butanol, 20% O2, remaining: N2; W/F 1.33
g s mL-1. Symbols: 1-butanol conversion (). Selectivity to Maleic Anhydride MA (), 1-butene (),
2-butenes (), acetic acid + acrylic acid (), CO (), CO2 (), Phthalic anhydride PA () and
“lights” ().

The results reported in Figure 4.44 have shown clearly the detrimental effect of the
increase of 1-butanol partial pressure, confirming the above-stated: in fact, a
comparison of tests carried out with 1-butanol-rich (2%) and 1-butanol-lean (1%) feed
shows that the selectivity to both MA and PA was much lower than that seen in
experiments carried out with 1% 1-butanol in the feed. In the intermediate
temperature range, the selectivity to CO and CO2 was much higher, while in the low T
range the selectivity to olefins was also higher when compared to leaner 1-butanol
conditions. These results indicate that, on the one hand, the dehydration of 1-butanol
is hindered, again because of the surface saturation which is obviously more
significant because of the higher gas-phase concentration of 1-butanol. On the other

120
 Results and discussion

hand, it seems that under conditions in which the availability of oxidising sites is
scarce, the adsorbed unsaturated intermediates are not oxidised into MA (or PA
either), but are instead preferentially transformed into CO and CO2. It is possible that
a direct attack of adsorbed olefins by gas-phase oxygen occurs: an event which has
been demonstrated to contribute to overoxidation under conditions of VPP surface
saturation [239, 240].

The data obtained highlight that butenes are the intermediates in 1-butanol
oxidehydration into MA. Moreover, the catalytic reaction is very selective in the first
step of alcohol dehydration into butenes (the initial selectivity to butenes is very high,
as shown in Figure 4.41), whereas the kinetically consecutive step of butene oxidation
into MA shows a limited selectivity only, because of the several side reactions.
Another important result is that most of the by-products observed in thermal
experiments (Figure 4.39 and Figure 4.40), are not formed during catalytic
experiments. This means that either they are burnt by the catalyst to COx, or that the
catalytic transformation of 1-butanol into butenes and of butenes into oxidised
compounds prevails over thermal reactions. Clearly, a contribution of the thermal, but
selective, dehydration of 1-butanol to butenes formation cannot be ruled out.

4.2.5.6 1-butene oxidation to MA

At this point it is clear that butenes are the primary products of the oxidehydration of
1-butanol, so lastly we checked the reactivity of the catalyst in 1-butene oxidation.
Results are reported in Figure 4.45.

121
 Results and discussion

Figure 4.45: Catalytic behaviour of the VPP DuPont catalyst in 1-butene oxidation as a function of
temperature. Conditions: feed 1 mol% 1-butene, 20% O2, remaining: N2; W/F 1.33 g s mL-1. Symbols:
1-butene conversion (). Selectivity to Maleic Anhydride MA (), 2-butenes (), acetic acid +
acrylic acid (), CO (), CO2 (), Phthalic anhydride PA () and “lights” ().

At low temperature the prevailing products were 2-butenes: their selectivity, however,
decreased when the temperature was raised. The highest yield to MA was 46%,
slightly higher than that shown from 1-butanol (39%); it is also worth noting that the
reactivity of 1-butene was slightly lower when compared to that of 1-butanol. These
data can be interpreted by taking into account the different degree of interaction of the
two molecules with the VPP surface. The other products obtained were the same as
those from 1-butanol: PA, CO, CO2, acetic acid and acrylic acid.

4.2.5.7 A DRIFTS study of the mechanism of 1-butanol

oxidehydration

In order to understand the interaction of 1-butanol with the catalyst surface and
investigate on the reaction mechanism, the alcohol was first adsorbed over VPP at
140°C; Figure 4.46 shows the spectra obtained after the alcohol adsorption using

122
 Results and discussion

either He or Air as the carrier (8 mL min-1): the experimental conditions are reported
in paragraph 3.1.4.

Figure 4.46: DRIFTS spectra for 1-butanol adsorption on VPP DuPont catalyst at 140°C. Carrier: He
or air, flow rate 8 mL/min.

In both cases 1-butanol interacted strongly with the OH groups present on the catalyst
surface as seen by the negative peaks at 3743 and 3663 cm-1. The positive peaks at
around 3250 and 2970-2880 cm-1 correspond to the OH and C-H stretching of
adsorbed butanol, respectively. In the case of the oxidising atmosphere, the peaks at
1727, 1627, and 1468 cm-1 indicate the presence of oxidised species even at this low
temperature; these bands may be attributed to either crotonic or maleic acid [246].
These experiments confirm the strong interaction of alcohol with the VPP surface,
thus supporting the hypothesis of the saturation of the surface.
When the adsorption was performed at 300°C, differences among spectra recorded
under different conditions were more evident (Figure 4.47).

123
 Results and discussion

Figure 4.47: DRIFTS spectra for 1-butanol adsorption on VPP DuPont catalyst at 300°C. Carrier: He
or air, flow rate 8 mL/min.

In the case of an inert atmosphere, the most important features were the negative
peaks at high frequency and broad bands at around 1630 cm-1 and 973 cm-1. The
former may correspond to a C=C stretching of an olefin (butene) adsorbed on the
surface, while the latter is usually attributed to a V=O stretching [247]. Therefore, in
this case (when no air is present) the catalyst is able to dehydrate the butanol
producing butenes, but this alters the catalyst surface. On the other hand, when air is
present, olefins are also formed (butenes and butadiene in this case) but they are
further oxidised as evidenced by both the sharp peak present at 1778 cm -1 and the
shoulder at 1850 cm-1, which have been attributed previously in literature to the O-
C=O stretching of adsorbed MA [248]. The bands at 1730 and 1616 cm-1 indicate
either the presence of species with a C=O and/or C=C bond, such as butyraldehyde, 2-
propanal, furan, and alkenals, or some adsorbed unreacted olefins. Since there were
no changes in the low frequency range, this confirms that oxygen is necessary to keep
the catalyst surface species unaltered. As regards the OH interaction, clearly in

124
 Results and discussion

oxidant atmosphere the bands were less intense and this agrees with the fact (observed
in reactivity tests) that air (O2) is needed in order to have a less adsorbed and thus
more reactive molecule, which is more quickly oxidised because of the greater
availability of V5+ sites.

Temperature programmed desorption (TPD) experiments were also carried out. In

this set of experiments, 1-butanol adsorption over VPP DuPont catalyst was
performed at 140°C and then stopped to monitor the reactive transformations of the
adsorbed species with temperature, either in He or in air. The evolution of the surface
modifications were followed by collecting DRIFTS spectra in He (Figure 4.48) and in
air (Figure 4.49).

Figure 4.48: DRIFTS spectra for 1-butanol adsorption on VPP catalyst at 140°C and its desorption
with temperature (TPD) in inert atmosphere (He).

125
 Results and discussion

Figure 4.49: DRIFTS spectra for 1-butanol adsorption on VPP catalyst at 140°C and its desorption
with temperature (TPD) in oxidising environment..

As observed from the spectra, in the case of inert atmosphere (Figure 4.48) the
adsorbed 1-butanol evolved mainly towards butadiene (broad band centred at 1602
cm-1) but then other olefins and carbonylic species also formed (C=C and C=O bands
around 1630-1724 cm-1); this occurred concomitantly with the increase of the band at
978 cm-1, thus confirming that vanadyl (V4+) surface species developed during the
transformation of 1-butanol under inert atmosphere. The OH interaction was so strong
that it was maintained throughout the whole temperature program (negative bands); in
fact, a slightly decreased intensity was seen only at T>320°C. On the other hand,
when air was present (Figure 4.49), oxidised carbonyl species were more readily
formed, corresponding to further oxidation compounds; particularly, the bands at
around 1780 and 1850 were assigned to adsorbed MA.
Therefore, it may be concluded that the catalyst under anaerobic conditions is able to
dehydrate 1-butanol and oxidise olefins until crotonic or maleic acid; however,
oxygen from the gas phase becomes necessary because it helps to keep the vanadium

126
 Results and discussion

centres effectively oxidised, while also making it possible for the intermediate to
dehydrate and desorb, in the end producing MA.
Lastly, operando-mode experiments were also performed by feeding 1-butanol over
the VPP catalyst at 360°C, with and without air: these results are presented in Figure
4.50 and Figure 4.51.

Time on stream
Figure 4.50: Operando-DRIFTS spectra collected during the feeding of 1-butanol at 360°C in
anaerobic atmosphere (He) over VPP DuPont catalyst, at increasing time-on stream of adsorption.
Time on stream

Figure 4.51: Operando-DRIFTS spectra collected during the feeding of 1-butanol at 360°C in air over
VPP DuPont catalyst, at increasing time-on stream of adsorption.

127
 Results and discussion

These tests have shown not only the bands previously attributed, but also their
gradually formation during time-on stream of adsorption, confirming the essential role
of oxygen for the success of the reaction. Moreover, the relative amounts of the
products desorbed from the catalyst surface were detected during the operando-tests:
results are reported in Figure 4.52.

Figure 4.52: Comparison of the relative amount of the main products detected during the feeding of
1-butanol at 360°C in anaerobic atmosphere (blue) and in presence of oxygen (red).

In line with the previous tests, they confirmed that MA could be formed only when air
was present; in addition, with oxygen in the feed, the relative amounts of butenes

128
 Results and discussion

were smaller (less accumulation), whereas those of other intermediates such as

crotonaldehyde and furan, were larger.

4.2.5.8 The reactivity of intermediates and the reaction mechanism:

the role of crotonaldehyde

At this point, the reactivity tests and the DRIFTS experiments have suggested that
crotonaldehyde may be a key intermediate in the mechanism of 1-butanol
oxidehydration to MA, in order to investigate the possible reaction mechanism.
Therefore, we carried out experiments feeding crotonaldehyde (as reactant) and air
over the VPP catalyst; we used a very diluted feed of the aldehyde, 0.2 mol% in air, in
order to simulate the low concentration of the adsorbed compound during 1-butanol
oxidation.
The results obtained are plotted in Figure 4.53; as a matter of fact, the yield to MA
was close to 50%, with by-products CO, CO2, acetic acid, acrylic acid, and furan, the
latter being found at 260 and 280°C only.

Figure 4.53: Catalytic behaviour of the VPP DuPont catalyst in crotonaldehyde oxidation in function
of temperature. Conditions: feed 0.3 mol% crotonaldehyde, 20% O 2, remaining: N2; W/F 1.33 g s mL-1.
Symbols: crotonaldehyde conversion (). Selectivity to maleic anhydride MA (), acetic acid +
acrylic acid (), CO (), and CO2 ().

129
 Results and discussion

The rapid decline in selectivity seen at 340°C  which, conversely, is the temperature
at which the best selectivity to MA from 1-butanol was seen  may be ascribed to an
enhanced combustion rate of crotonaldehyde. Indeed, once again the oxidation
properties of VPP may be strongly affected by the reactant-to-oxygen ratio in the
feed. With crotonaldehyde, which is expected to show a weaker interaction with the
catalyst surface compared to 1-butanol, the use of a low crotonaldehyde-to-oxygen
feed ratio may lead to a catalyst overoxidation at 340°C, which is responsible, in the
end, for the high selectivity to CO2. This hypothesis was confirmed by catalytic tests
carried out in order to show the effect of the inlet molar fraction of crotonaldehyde on
MA selectivity. Results are reported in Figure 4.54.

Figure 4.54: Catalytic behaviour of the VPP DuPont catalyst in crotonaldehyde oxidation in function
of inlet molar concentration of crotonaldehyde. Conditions: 20% O 2, remaining: N2; W/F 1.33 g s mL-1;
temperature 340°C. Symbols: crotonaldehyde conversion (). Selectivity to maleic anhydride MA
(), acetic acid + acrylic acid (), CO (), and CO2 ().

It can be seen that an increased feed ratio led to a remarkable increase in selectivity,
with a yield to MA shown at a high crotonaldehyde inlet molar fraction similar to that
achieved from 1-butanol under similar reaction conditions (i.e. at similar reactant-to-
oxygen feed ratio).

130
 Results and discussion

As for the hypothesis of crotonaldehyde as the key reaction intermediate in C 4

oxidation to MA, this possibility has already been described in literature. In fact, it
has been reported that crotonaldehyde can be oxidised to maleic acid by using a
catalyst based on V/Mo/O supported on pumice, with 46% yield [249]. Cicmanec et
al. [250] reported that, with a similar catalyst, crotonaldehyde follows two parallel
pathways of oxidation, where the trans form is oxidised to crotonic acid, while the
latter is then oxidised to MA, whereas the cis form is oxidised to furan; the latter may
then be oxidised to MA via 2(3H)-furanone and 2(5H)-furanone intermediates. The
mechanism involving crotonic acid as the intermediate in crotonaldehyde oxidation
was also formerly proposed by Church and Bitha [251], using various types of V
oxide catalysts. Xue and Schrader [252] suggested that in n-butane oxidation over
VPP the main route for MA formation occurs with crotonaldehyde as the key reaction
intermediate. However, it cannot be ruled out, as suggested by Wenig and Schrader
[248], that the formation of crotonaldehyde or crotonic acid is due to the oxidation of
butadiene and not of butene. Crotonaldehyde can also be oxidised to furan with high
selectivity by using polyoxometallate catalysts [253]. Indeed, furan is another
molecule which was proposed as an intermediate during C4 oxidation with VPP
catalyst. For example, Guliants et al. [236, 237] reported that the oxidation of 2-
butene occurs via the formation of furan, which is then oxidised to MA with the
generation of a furan endoperoxide produced by reaction with a V5+-OO species.
Through a dedicated experiment we also confirmed that furan can be oxidised with
59% yield to MA, by using 0.3% furan in air as inlet feed.

Another set of experiments was carried out using crotonaldehyde as the reactant,
while varying the contact time, in order to detect the possible formation of furan as an
intermediate during crotonaldehyde oxidation to MA (Figure 4.55).

131
 Results and discussion

Figure 4.55: Catalytic behaviour of the VPP DuPont catalyst in crotonaldehyde oxidation in function
of contact time W/F. Conditions: 0.3% crotonaldehyde, 20% O 2, remaining: N2; W/F 1.33 g s mL-1;
temperature 260°C. Symbols: crotonaldehyde conversion (). Selectivity to maleic anhydride MA
(), acetic acid + acrylic acid (), CO (), and CO2 ().

Our results demonstrated that, indeed, furan forms directly from the aldehyde.
However, it was also shown that the C balance was poor, especially at low values of
contact time.

We characterised the used catalyst after the experiment carried out at low contact time
and recorded the IR spectra by means of DRIFT spectroscopy. The analysis are
reported in Figure 4.56.

132
 Results and discussion

Figure 4.56: Top: Diffuse-Reflectance IR spectra of the fresh catalyst and of the spent catalyst after
reactivity experiments with crotonaldehyde at 0.18 s contact time (Figure 4.55). Bottom: difference
spectra.

As shown in Figure 4.56, the spectra highlighted the presence of several compounds
at the adsorbed state, such as MA, crotonic acid, maleic acid, butenes and butadiene.
The catalyst spent after reaction at longer contact time, conditions at which the C
balance was improved, showed the presence of minor amounts of these adsorbed
compounds. Therefore, at low contact time the catalyst retained adsorbed molecules
which are precursors for the formation of MA: crotonic acid and maleic acid. When
the contact time was increased, the selectivity to MA increased too; this was due to
both the oxidation of furan - the selectivity of which in fact decreased -, and of maleic
acid, which was retained on the catalyst surface at low contact time but desorbed as

133
 Results and discussion

MA when the contact time was increased, with concomitant improvement of the C
balance.
In the following table (Table 4.1) the assignments of the main bands detected are
reported.

Band, cm-1 Intensity (KM) Assigned to

1467 0,265 Crotonic acid

1601 0,515 Butadiene

1672 0,160 Butene

1725 0,322 Maleic acid

1780 0,893 Maleic anhydride

1853 0,247 Maleic anhydride

Table 4.1: Table with assignment of main bands detected in DRIFTS experiments.
Band attributions were given based on refs [246, 248].

All these experiments demonstrated that the mechanism for MA formation likely
occurs with crotonaldehyde as the key intermediate compound. The latter undergoes
two parallel reactions of transformation, one leading to furan, and the other to maleic
acid, possibly via crotonic acid. Both furan and maleic acid are intermediates for MA
formation, via oxidation or via dehydration, respectively. The relative importance of
the two parallel ways is likely a function of the degree of surface availability of
oxidising V sites, and thus of the degree of surface coverage by adsorbed species.
Another important difference is that furan desorbs easily from the VPP surface, as
demonstrated by DRIFTS-MS operando experiments, which showed no proof for the
retention of furan at the adsorbed state, while furan was detected in a relatively large
amount in the gas phase. Conversely, maleic acid was strongly interacting with the
catalyst surface; in order to desorb this compound it was necessary to use conditions
which fostered the development of a clean and oxidised surface.

134
 Results and discussion

Figure 4.57 summarises the proposed reaction network for the transformation of 1-
butanol to MA, as inferred by combining our DRIFTS and reactivity experimental
results.

Figure 4.57: Reaction scheme for 1-butanol oxidehydration to MA.

The mechanism involves the dehydration of 1-butanol to 1-butene, the isomerisation

to 2-butenes, and the oxidation of the latter to crotonaldehyde. Crotonaldehyde can
either cyclises to yield furan or be oxidised until maleic acid. Maleic acid desorbs in
the form of MA, whereas furan can desorb and be consecutively re-adsorbed and
further oxidised to MA. It is also important to note that the endoperoxide compound
formed by furan oxidation can rearrange either into 5-hydroxy-2(5H)-furanone [237],
or 4-oxo crotonic acid, which is also a precursor of maleic acid; this way, the two
parallel reactions would share the same intermediate, maleic acid, as also suggested
by DRIFTS experiments.

135
 Results and discussion

4.2.6 Bio-butanol to maleic anhydride

After having established the effect of the main reaction parameters on catalytic
behaviour using a chemically sourced 1-butanol (CB), we conducted experiments
using three different types of BB, whose main impurities are shown in Table 3.1,
compared with those of CB. It is important to note that all the samples had purity
degrees higher than 99.5%, with a residual water content that in all cases was less than
1%; the latter is expected to have a negligible effect on reactivity behaviour, since the
amount of water generated during the reaction was much greater than that present in
1-butanol. However, there were differences in the amount and type of impurities
present, and the following rank of impurity contents may be assumed (as from a semi-
quantitative assessment inferred from GC and GC-MS analysis): BB1 > BB2 > BB3.
It is also worth noting that the pH of the aqueous solution obtained by dissolving 1 ml
of BB in 14 ml of water was slightly different in the 3 samples (Table 3.1). Lastly, we
also searched for metallic impurities in BB samples by means of detailed ICP and
XRF analysis, but no traces of them were found.
The catalytic behaviour shown by sample BB1 (the one containing the greater amount
of impurities) is reported in Figure 4.58.

Figure 4.58: Catalytic behaviour of the VPP DuPont catalyst in oxidehydration of BB1.

136
 Results and discussion

Symbols: 1-butanol conversion (). Selectivity to Maleic Anhydride MA (), 1-butene (), 2-
butenes (), acetic acid + acrylic acid (), CO (), CO2 (), Phthalic anhydride PA () and “lights”
().

A comparison with the behaviour shown by the CB sample (Figure 4.41) highlights
the following differences:

 With BB1, the maximum yield to MA shown was only 31%, at 340°C (vs 39%
for CB, at 340°C). On the other hand, the selectivity to PA was higher than
that shown with CB; at 320°C, the PA yield was 25%, and 17% at 340°C; with
CB, the PA yield was no higher than 12%. Therefore, the best overall
selectivity to MA+PA, at total 1-butanol conversion, was 51% with CB, and
48% with BB1.
 With BB1, the selectivity to CO + CO2 was greater at low temperatures, but
similar (or even slightly lower) at high temperatures, compared to the
selectivity obtained with CB.
 The selectivity to light acids and butenes was also lower with BB1. Moreover,
traces of other by-products were found, such as lighter aldehydes, which
probably formed from small amounts of alcohols other than 1-butanol, which
were contained as impurities in BB.

In order to explain the drop in MA selectivity shown with bio-butanol, we monitored

the performances during the time-on-stream, comparing the effect on VPP DuPont
catalyst for chemical-butanol (CB) and for bio-butanol (BB1). Time-on-stream
experiments are reported in Figure 4.59 and Figure 4.60.

137
 Results and discussion

Figure 4.59: Catalytic behaviour of the VPP DuPont catalyst in chemical-1-butanol (CB)
oxidehydration. Symbols: 1-butanol conversion (). Selectivity to Maleic Anhydride (), 1-butene
(), 2-butenes (), acetic acid + acrylic acid (), CO (), CO2 (), Phthalic anhydride () and
“lights” (). Temperature 340°C.

Figure 4.60: Catalytic behaviour of the VPP DuPont catalyst in bio-1-butanol (BB1) oxidehydration.
Symbols: 1-butanol conversion (). Selectivity to Maleic Anhydride (), 1-butene (), 2-butenes

138
 Results and discussion

(), acetic acid + acrylic acid (), CO (), CO2 (), Phthalic anhydride () and “lights”
().Temperature 340°C.

These results demonstrated that the difference in the catalytic behaviour between the
two reactants was likely related to some impurities present in BB1, which
accumulated on the catalyst surface and modified the surface properties during the
first hours of reaction time. In fact, the two reactants gave a similar performance at the
beginning of the experiment time (reaction conditions having been fixed at 340°C);
however, in the case of the CB feed, the performance remained quite stable, whereas
in the case of BB1 the performance progressively changed, with a decline in
selectivity to MA and a corresponding increased selectivity to PA.
It may be concluded that the progressive change in the catalyst surface characteristics
had an effect mainly on the variation of the PA/MA yield ratio, but affected the
overall yield to the two compounds to a lower extent; it is worth reminding that PA
formed from a consecutive reaction on MA.
In regards to the nature of the impurities which might be responsible for the poisoning
effect, it should be noted that most of the impurities had characteristics not much
different from those of the reaction products obtained during 1-butanol
oxidehydration. The only exception was a carbazole compound, whose tentatively
identified formula was 5H-Naphtho[2,3-C]carbazole-5-methyl. Carbazoles are
antibiotics which can be released by microorganisms during the final period of the
fermentation process [254].

Lastly, the different reactivity behaviours of bio-butanols reported in Table 3.1 were
tested over VPP DuPont catalyst. The first sample was BB1pur: the bio-butanol was
purified to lower the impurity content, by means of adsorption over bleaching earth
enriched with activated charcoal (see Experimental). The catalytic results are reported
in Figure 4.61.

139
 Results and discussion

Figure 4.61: Catalytic behaviour of the VPP DuPont catalyst in BB1pur oxidehydration. Symbols: 1-
butanol conversion (). Selectivity to Maleic Anhydride (), 1-butene (), 2-butenes (), acetic acid
+ acrylic acid (), CO (), CO2 (), Phthalic anhydride () and “lights” ().

Comparing these results with the test carried out with BB1, we can observe that the
purification was only partially successful, since the maximum yield to MA increased
from 31% at 340°C to 33% at 360°C, keeping far from the 39% of yield obtained with
CB. For this reason the sample BB1 was purified again with a stronger treatment,
consisting in a double amount of bleaching earth used (0.06 g instead of 0.03) and a
longer time of stirring (24 h instead of 6 h): the bio-butanol obtained was named BB1-
superpur. In order to check if the second purification was successful the sample BB1-
superpur was tested with VPP DuPont in the same reaction condition: the results of
the catalytic test are reported in Figure 4.62.

140
 Results and discussion

Figure 4.62: Catalytic behaviour of the VPP DuPont catalyst in BB1-superpur oxidehydration.
Symbols: 1-butanol conversion (). Selectivity to Maleic Anhydride (), 1-butene (), 2-butenes
(), acetic acid + acrylic acid (), CO (), CO2 (), Phthalic anhydride () and “lights” ().

The catalytic test demonstrated that the removal of the impurities was efficient, since
at 360°C the same value of yield to MA obtained with CB was shown. This results
confirmed again that the worsening of the performances shown with bio-butanol was
due to the presence of some impurities which accumulates on the catalytic surface.
Moreover, with an efficient purification treatment they could be removed giving
results quite similar to that one obtained with CB.

We continued the reactivity experiments testing the behaviour of other two types of
bio-butanol, the BB2 and the BB3: the results are reported respectively in Figure 4.63
and Figure 4.64.

141
 Results and discussion

Figure 4.63: Catalytic behaviour of the VPP DuPont catalyst in BB2 oxidehydration. Symbols: 1-
butanol conversion (). Selectivity to Maleic Anhydride (), 1-butene (), 2-butenes (), acetic acid
+ acrylic acid (), CO (), CO2 (), Phthalic anhydride () and “lights” ().

Figure 4.64: Catalytic behaviour of the bulk VPP catalyst in BB3 oxidehydration.

142
 Results and discussion

Symbols: 1-butanol conversion (). Selectivity to Maleic Anhydride (), 1-butene (), 2-butenes
(), acetic acid + acrylic acid (), CO (), CO2 (), Phthalic anhydride () and “lights” ().

These samples exhibits different values of yields to MA, 32% for BB2 and 40% for
BB3, both at 360°C (see Table 3.1). These results confirmed again that the presence
of impurities greatly affect the catalytic response of bio-butanol over VPP DuPont
catalysts.

To summarize, Figure 4.65 compares the best yield to MA and the best MA+PA yield
(in all cases at total 1-butanol conversion, at 340-360°C).

Figure 4.65: Best MA and PA yields obtained with VPP DuPont catalyst feeding either chemical 1-
butanol (CB), or bio-1-butanol. Reaction conditions as in Figure 4.41.

It is apparent that samples BB1 and BB2 both led to lower yields if compared to CB,
but the BB3 sample (the purest amongst the BB samples) showed a catalytic
behaviour quite similar to CB. The purification of BB1 led to an increased yield to
MA, with a corresponding lower yield to PA; lastly, sample BB1-superpur, which had
undergone a prolonged treatment for the removal of heavier impurities, showed the
same behaviour as CB.

143
 Results and discussion

In conclusions the impurities contained in bio-butanol affected the overall process

selectivity; however, a purification treatment made it possible to obtain a yield to MA
which was similar to that achieved from chemically sourced 1-butanol.

4.2.7 VPP DuPont catalyst characterization

The catalyst used for the reactivity tests with 1-butanol and bio-butanol was an
industrial catalyst delivered by DuPont, composed by 90%wt. of vanadyl
pyrophosphate and 10%wt. of silica. The catalyst was characterized by Raman
spectroscopy, X-ray diffractometry and BET surface area analysis; the conditions are
reported in the Experimental section. The samples were analyzed before reactivity
tests (fresh) and after (spent).

Raman analysis

Raman spectroscopy is a very powerful analysis for VPO-based catalysts, because

these materials are well known to show typical and well-defined Raman bands. In
fact, it is possible to distinguish the main phase (VO)2P2O7, the precursor
VOHPO4·0,5 H2O and the several V3+ and V+5 phases, in particular αI, αII, β, δ, ε, γ,
ω-VOPO4, and hydrated phases VOPO4·nH2O. In order to verify which kinds of
phases are present on VPP DuPont catalysts, we collected a series of spectra changing
the focused position over the sample. The spectra of fresh catalyst are reported in
Figure 4.66. The same analysis was conducted over spent VPP DuPont catalyst:
examples of the Raman spectra are reported in Figure 4.67.

144
 Results and discussion

= (VO)2P2O7

a.u.

1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200
-1
Raman shift (cm )

Figure 4.66: Raman spectra of fresh VPP DuPont catalyst.

Figure 4.67: Raman spectra of spent VPP DuPont catalyst after reaction feeding 1%1-butanol in air.

145
 Results and discussion

As expected, the analysis on the fresh catalyst has confirmed that vanadyl
pyrophosphate is the main phase, since only in few points the presence of minor
amount of the oxidized and hydrated phase VOPO4·nH2O has been detected.

The Raman spectra of the spent sample have confirmed that VPP is still the main
phase, accompanied with the some traces of VOPO4·H2O: this means that the catalytic
tests did not substantially change the active phase of the catalyst.

X-ray diffraction analysis

XRD-diffraction analysis on fresh and spent VPP DuPont catalyst are reported in
Figure 4.68. Comparing the two diffractograms it is possible to note that the
diffraction patterns are the same, confirming the predominant presence of the vanadyl
pyrophosphate crystalline phase.

a.u •
• = (VO)2P2O7
•
•

After reaction
• •
• • • •

•
• • •
Fresh
•
•

10 20 30 40 50
Position ( 2Theta)
Figure 4.68: XRD diffractograms of fresh (red) and spent (blue) VPP DuPont catalysts.

146
 Results and discussion

BET surface area analysis

The surface area of the samples have been measured by BET method. For the fresh
catalyst this value is 40 m2/g, while for the spent one is 38 m2/g. For this reason we
can confirm that the catalytic tests have not altered the material in terms of surface
area.

147
 Conclusions

5 Conclusions

The synthesis of maleic anhydride (MA) has received for decades a special attention
due to its commercial viability and to its demand in the global market as an
intermediate chemical product. Its production has been commercialized mainly by
means of the partial oxidation of n-butane, a process catalyzed by a vanadium-
phosphorous-oxide (VPO) catalyst, typically vanadyl pyrophosphate (VO)2P2O7.
Otherwise, in recent years an increased interest to produce MA from renewable materials
has been observed. In this context my PhD thesis was focused on the improvement of the
selective oxidation of n-butane into maleic anhydride (MA), investigating the behavior
the industrially-produced vanadyl pyrophosphate catalyst (VPP), and on the study of a
new process for the synthesis of MA starting from a bio-based building block, 1-
butanol.

The first part financially and scientifically supported by Polynt SpA dealed with the
improvement of the synthesis of maleic anhydride starting from n-butane. The
reactivity tests were carried out in order to correlate how the catalytic behavior of
VPP (A catalyst) and Nb doped-VPP (X catalyst) could be affected by the reaction
conditions passing from plant-scale to the lab-scale. The A and X catalysts tested in
hydrocarbon-lean and rich conditions have shown that the differences between the
two samples could be observed in lab-scale only when an high concentration of n-
butane is fed. In fact, working at 1.7% n-butane in air (the concentration typically
used in fixed-bed reactors) the two catalysts did not differ and in laboratory scale they
could be considered equivalent, although their different composition. On the other
hand, working at 4% n-butane in air (the concentration typically used in fluidized-bed
reactors) the presence of Nb as dopant is needed to enhance the oxidative power of the
catalyst, becoming more reducing the environment. The improvement exhibited by
Nb-doped catalyst was explained by the formation of discete amount of δ-VOPO4,
which is known to be a selective V5+ phase. In-situ Raman analysis carried out on

148
 Conclusions

both A and X catalyst confirmed that Nb enhances the formation of δ-VOPO4, in fact
in the un-doped catalyst the phase was obtained only in traces and through the co-
feeding of water, which acted as promoter for the oxidation. In this context, the key
role of water in the modifications of the catalyst surface was investigated, being steam
an important co-product of the reaction. The co-feeding of steam in hydrocarbon-lean
condition over the A catalyst showed to have an improving effect on the
performances, since a significant increase in terms of selectivity to MA was
registered.
The study of vanadyl pyrophosphate as industrial catalyst for the selective oxidation
of n-butane to MA was continued focusing our attention on the catalyst ageing, in
terms of both performances and surface modification of the material. The gradual
deactivation of vanadyl pyrophosphate was monitored by reactivity tests and by
Raman spectroscopy: the analysis demonstrated that the ageing is a complex process
characterized by the formation of different V/P/O-based phases and by the retention
of adsorbed organic compounds that could alter the accessibility to the active sites.
More in details, the formation of the vanadyl metaphosphate VO(PO3)2, hydrated
phases (VOHPO4·nH2O), β-VOPO4 and coke deposits occurred. The presence of
vanadyl metaphosphate, characterized by a P/V ratio twice than the VPP one,
indicated that one of the effect of the ageing is the phosphorus surface enrichment and
it is one of the causes of the significant drop in activity registered. Together with
VO(PO3)2 phase, also the unselective and detrimental phase β-VOPO4 was
responsible of the worsening of the catalytic performances. Moreover, the ageing
proceeded through the alteration the desorption properties of the material, since the
water co-produced by the reaction led formation of hydrated phases (VOHPO4·nH2O)
and the organic compounds were strong adsorbed forming carbon deposits and
causing the blockage of the active sites.

The second part of my PhD thesis was focused on the study a new process for the
synthesis of MA starting from a bio-based building block, 1-butanol. This alcohol is
recently attracting attention, because it can be obtained from biomass and also used an
alternative fuel in place of bioethanol, since it may be also advantageously utilized to

149
 Conclusions

produce chemicals. The reaserch was supported by the FP7 programm of the EC
(EuroBioRef).

We examined the different possible configurations in order to carry out the reaction in
two steps or in one step: in particular we studied the dehydration of 1-butanol to
butenes using acid catalysts (Keggin polyoxomethalates and VOPO4·2H2O) and the
“one-pot” reaction (oxidehydration) using vanadyl pyrophosphate as multifunctional
catalyst.
The experiments carried out using Keggin polyoxomethalates supported on silica
(POM) demonstrated that these materials could be successfully utilized as catalyst for
the dehydration of 1-butanol to butenes: in fact, thanks to their acid properties, POM
gave butenes with high selectivity. The samples tested were Cs2HPW12O40,
Cs3PW12O40, K3HSiW12O40 and K4SiW12O40. The optimal range of temperature was
investigated in order to limitate the phenomena responsible of lack to carbon balance:
these materials were affected by the adsorption of the organic compound at low
temperature and by the formation of coke at high temperature. Moreover, the effect of
the chemical composition of the catalysts was outlined, in particular a key role ise
played by the nature of the central atom and by the subtitution of the protons with
alkaline cations. The Cs2HPW12O40 catalyst showed the best performance both in
terms of activity and selectivity, since it reached the total conversion at 150°C and the
highest values of selectivity to butenes. On the other hand, the K4SiW12O40 catalyst
was revealed to be the worst, the total conversion being obtained at 290°C with a
concomitant low selectivity to the dehydration products.
Concerning the dehydration of 1-butanol, the study was continued testing
VOPO4·2H2O (VPD), since V/P/O based catalysts are known to have acid properties.
VPD demonstrated to be a promising catalyst in oxidizing environment: in fact in
presence of air it exhibited high activity and selectivity with stable perfomance during
time and for this reason it was suggested as a good candidate for the dehydration of 1-
butanol in presence of air in the “one-reactor” configuration for the synthesis of
maleic anhydride from bio-butanol.

150
 Conclusions

Lastly, we investigated the feasibility of the selective oxidehydration of 1-butanol into

maleic anhydride. Experiments carried out with DuPont vanadyl pyrophosphate
demonstrated that the material could successfully catalyze the dehydration of 1-
butanol and the selective oxidation of butenes to MA, confirming its multifunctional
properties. By means of reactivity tests the optimal reaction conditions were found:
feeding 1%mol 1-butanol in air, the conversion of 1-butanol was completed over the
entire range of temperature and the highest selectivity to MA was observed at 340°C
(39%); the reaction also produced butenes, light acids (acrylic and acetic acid), carbon
oxides, phthalic anhydride (PA), and minor amounts of other oxygenated compounds
(furan and formaldehyde). At low temperatures the transformation of 1-butanol into
butenes and the oxidative cleavage of either alcohol or olefins into acetic acid and
acrylic acid were the prevalent reactions, while increasing the temperature, butenes
were transformed into MA and PA and the light acids were oxidised into COx.
The investigation on reaction scheme showed that 1-butene is the primary product of
the reaction and while its isomerisation to 2-butenes is relatively slow at low
temperatures, at high temperature or high the residence time the consecutive
oxidehydrogenation into MA occurs at a faster rate on 2-butenes than on 1-butene.
The detrimental effect of the increase of 1-butanol partial pressure and the positive
effect of the oxygen partial pressure showed that 1-butanol and butenes are strongly
adsorbed on the catalyst surface: this is way an high O2/(1-butanol + olefins) ratio is
needed in order to have a sufficient amount of “free” and oxidized active sites. The
strong interaction of alcohol with the VPP surface was demonstrated by DRIFT
experiments: in fact, oxygen from the gas phase is essential to keep the vanadium
centres oxidised, making possible the dehydration and the desorption of the
intermediates and at the end the production of MA.
The detection of crotonic acid in DRIFT experiments suggested that crotonaldehyde
may be a key intermediate in the mechanism of 1-butanol oxidehydration to MA. This
hypothesis was confirmed by catalytic tests and DRIFT experiments, since
crotonaldehyde undergoes two parallel reactions, one leading to furan, and the other
to maleic acid. Furan and maleic acid were found to be both intermediates for MA
formation, via oxidation or via dehydration, respectively. In conclusion, the

151
 Conclusions

mechanism of the oxidehydration of 1-butanol involves the dehydration of 1-butanol

to 1-butene, the isomerisation to 2-butenes, and the oxidation of the latter to
crotonaldehyde. Then, crotonaldehyde can either cyclises to yield furan or be oxidised
until maleic acid, which then could desorbs forming MA; on the other hand, furan can
desorb and be consecutively re-adsorbed and further oxidised to MA.

The study of the oxidehydration of 1-butanol to MA was concluded testing three

different types of bio-butanol, produced by fermentation of biomass. The first bio-
butanol (BB1), even if it reached the total conversion in the same range of
temperature investigated with the chemical 1-butanol (CB), gave an overall worsening
of the performances, being increased the selectivity to COx and lowered the selectivity
to MA and butenes. This behaviour was due to the accumulation on the catalyst
surface of some impurities present in bio-butanol, which led to the decline of the
performance. In particular, the responsible of the poisoning might be a carbazole
compound, an antibiotic produced that by microorganisms during the final period of
the fermentation process. With the aim of removing the impurities which negatively
affected the performances, bio-butanols were treated by means of adsorption over
bleaching earth enriched with activated charcoal: the catalytic test confirmed that the
removal of the impurities was efficient, since at 360°C the same value of yield to MA
obtained with CB was reproduced with BB1-superpur. In conclusion, the
oxidehydration of bio-butanol to maleic anhydride was successfully obtained,
demonstrating that also bio-butanol produced by fermentation could be easily utilize
to carry out the process, since a simple purification treatment made it possible to
remove the impurities present in it.

152
 Acknowledgements

6 Acknowledgments

The research leading to these results has received funding from the European Union
FP7 (FP7/2007-2013) under grant agreement n° 241718 EuroBioRef.

Heartfelt thanks to Polynt SpA for the financial and scientific support, in particular
thank to Carlotta, Silvia and Federico.

153
 Ringraziamenti

7 Ringraziamenti

Era l’ultima ora dell’ultimo giorno. Ricordo perfettamente quando suonò la

campanella e un senso di liberazione irruppe come non mai, alla sola idea che quei
banchi e quei libri sarebbero stati a breve soltanto un ricordo. Avevo deciso di
allontanarmi da quel mondo fatto di idee astratte e privo di concretezza; avevo deciso
che avrei preso una strada diversa da tutti, per poter toccare con mano quelle parole
che finora avevo solo potuto pronunciare o scrivere.
“Siamo schiavi delle leggi, per poter essere liberi”. Questa è stata ed è per me la
scienza, la chimica. In questo viaggio durato otto anni queste parole le ho sempre
portate con me, giorno dopo giorno, insieme alle odi dei poeti, alle invettive dei
grandi oratori, ai racconti degli storici e ai pensieri dei filosofi. La scienza mi ha
permesso di capire quanto l’uomo possa fare per il mondo, per gli altri, per se stesso.
Ne ho amato e ne amerò sempre la sua forza, la sua razionalità, il suo simbolismo e la
sua magia. Con la chimica possiamo capire il mondo e quello che ci circonda, ma
senza la comunicazione, la passione e la libertà anche la scienza nella sua concretezza
può rischiare di rimanere un contenitore sterile. Per questo mi piace pensare di essere
entrata a scuola da bambina, e di non esserne mai uscita. Mi piace pensare che i
grandi classici in fondo siano così reali e così attuali da poter essere d’aiuto nel
trasmettere tutto quello che la chimica fa per il progresso dell’uomo, nel suo
pragmatismo e nella sua etica.

In questo viaggio ho avuto la fortuna di incontrare tante persone straordinarie e di

aver imparato tantissime cose da ognuna di loro.

Ringrazio il prof. Fabrizio Cavani, per avermi seguito nel percorso del dottorato di
ricerca, per avermi dato supporto e fiducia.

Ringrazio Carlotta, Silvia e Federico, per avermi permesso di collaborare con loro,
dandomi l’opportunità di arricchire la ricerca scientifica.

154
 Ringraziamenti

Ringrazio tutto il gruppo di Catalisi, professori, ricercatori, dottorandi, borsisti,

assegnisti e tesisti. Grazie per essere un meraviglioso esempio di collaborazione su
ogni fronte, e per essere soprattutto degli amici prima di essere dei “colleghi”.

Grazie Ciccio per l’aiuto enorme che mi hai dato negli ultimi tempi e grazie Juliana
per il lavoro al DRIFT.

Grazie a tutti, Carlo, Mattia, Mugno, Taba, Max, Chiara, Diletta, Paris, Dodo, Grazia,
Sara, Alice, Paolo, Stefania, Claudia, Juliana, Cristian, Ciccio, Erica, Matte, Davide,
Alle, Fede, Beppe, Ale, Manuel, Dayanne, Matty, Ciaba, Silvia, Tommy,
Oriella…(sicuramente mi sarò dimenticata qualcuno).

Un ringraziamento speciale va ai miei tesisti, Giovanni, Mattia, Luca ed Alberto, per

essere stati la chiave di volta che mi ha permesso di capire quale fosse la strada da
prendere: grazie per essere stati una delle mie soddisfazioni più grandi in questi anni
di dottorato.

Infine grazie a Paolo, ai miei genitori, a mio fratello, alla mia famiglia e alle mie
amiche per essere il centro del mio mondo, per volermi bene, per avermi sopportato
pazientemente in questi anni, per essermi accanto e per condividere con me gioie e
dolori.

155
 Ringraziamenti

“Si quis in caelum ascendisset naturamque mundi et pulchritudinem siderum

perspexisset, insuavem illam admirationem ei fore; quae iucundissima fuisset, si
aliquem, cui narraret, habuisset”.

“Se un uomo salisse in cielo e contemplasse la natura dell'universo e la bellezza degli

astri, la meraviglia di tale visione non gli darebbe la gioia più intensa, come dovrebbe,
ma quasi un dispiacere, perché non avrebbe nessuno cui comunicarla”.

Cicerone “Laelius de amicitia” 88

Concipere animo non potes quantum momenti afferri mihi singulos dies videam.
“Mitte” inquis “et nobis ista quae tam efficacia expertus es.” Ego vero omnia in te
cupio transfundere, et in hoc aliquid gaudeo discere, ut doceam; nec me ulla res
delectabit, licet sit eximia et salutaris, quam mihi uni sciturus sum. Si cum hac
exceptione detur sapientia, ut illam inclusam teneam nec enuntiem, reiciam: nullius
boni sine socio iucunda possessio est.

Non puoi immaginare quali progressi io mi accorga di compiere giorno per giorno. Tu
mi dici: "Riferisci anche a me questo metodo che hai trovato così efficace."Certo,
desidero travasare in te tutto il mio sapere e sono lieto di imparare qualcosa appunto
per insegnarla. Di nessuna nozione potrei compiacermi, per quanto straordinaria e
vantaggiosa, se ne avessi conoscenza per me solo. Se mi fosse concessa la sapienza a
condizione di tenerla chiusa in me senza trasmetterla ad altri, rifiuterei: non dà gioia il
possesso di nessun bene, se non puoi dividerlo con altri.

Seneca “Epistulae morales ad Lucilium” (liber I, Epistola VI, 4-7)

156
 References

8 References

[1] MSDS of maleic anhydride

[2] Handbook of Chemistry & Physics, 95th edition, 2014-2015
[3] www.polynt.com Technical data sheet Maleic Anhydride Molten
[4] www.chemical.ihs.com
[5] www.mcgroup.co.uk
[6] Patent EP 1047687 (2006) Process for the production of gamma-butyrolactone
[7] Patent US 2772292 (1956) Hydrogenation of maleic anhydride
[8] Patent EP 0304696 (2006) Process for the preparation of 1,4-butanediol and/or tetrahydrofuran
[9] Patent US 5073650 (1991) Preparation of 1,4-butanediol
[10] www.thinking.nexant.com Report Abstract - Maleic Anhydride 07/08-8
[11] Ulmann’s Chemical Encyclopedia, Vol A, 16 (2011) 54
[12] K. Lohbeck,.H. Haferkorn, W. Fuhrmann, N. Fedtke, “Maleic and Fumaric Acids”, Ullmann's
Encyclopedia of Industrial Chemistry (2000)
[13] D. Lgs. n° 155 del 13/08/2010 “Attuazione della direttiva 2008/50/CE relativa alla qualità dell’aria
ambiente e per un’aria più pulita in Europa”
[14] F. Trifirò, R. K. Grasselli, Top. Catal., 57 (2014) 1188 - 1195
[15] T. A. Bither Jr, to E. I. Du Pont de Nemours and Company US Patent 4442226 (1984)
[16] F. Cavani, F. Trifirò, Chem. Tech., 24 (1994) 18
[17] P. Arpentier, F. Cavani, F. Trifirò, “The technology of catalytic oxidation”, Edition TECHNIP,
Paris (2001)
[18] G. Centi, F. Cavani, F. Trifirò, “Selective oxidation by heterogeneous catalysis”, Edition Kluwer
Academic Plenum Publishers (2001)
[19] F. Centi, F. Trifirò, J. R. Ebner and V. Franchetti, Chem. Rev., 88 (1988) 175
[20] F. Cavani, F. Trifirò, Appl. Catal. A, 88 (1992) 115
[21] R. M. Contractor, A. W. Sleight, Catal. Today, 3 (1988) 175
[22] N. F. Dummer, J. K. Bartley, G. J. Hutchings, Adv. Catal., 54 (2011) 189
[23] E. A. Lombardo, C. A. Sanchez, L. M.Cornaglia, Catal. Today, 15 (1992) 407
[24] G. Poli, I. Resta, O. Ruggeri, F. Trifirò, Appl. Catal. A Gen., 95 (1993) 117
[25] N. H. Batis, H. Batis, A. Ghorbel, J. C. Vedrine, J. C. Volta, J. Catal. 128 (1991) 248
[26] M. O’ Connor, F. Dason, B. K. Hodnett, Appl. Catal. A Gen., 64 (1990) 161

157
 Ringraziamenti

[27] F. Cavani, G. Centi, F. Trifirò, Appl. Catal., 9 (1984) 191-202

[28] H. S. Horowitz, C. M. Blackstone, A. W. Sleight, G. Teufer, Appl. Catal., 28 (1988) 1963
[29] J. W. Johnson, D. C. Johnston, A. J. Jacobson, J. F. Brody, J. Am. Chem. Soc., 106 (1984) 8123
[30] G. J. Hutchings, R. Higgins, Appl. Catal. A Gen., 154 (1997) 103
[31] S. Albonetti, F. Cavani, S. Ligi, F. Pierelli, F. Trifirò, F. Ghelfi, G. Mazzoni, Stud. Surf. Sci.Catal.,
143, (2002) 963
[32] E. C. Milberger, N. J. Bremer, D. N. Dria to Sohio (1982) US Patent 4350639
[33] M. E. Leonowicz, J. W. Johnson, J. F. Brody, H. F. Shannon, J. M. Newsam. J. Solid State Chem.,
56 (1985) 370
[34] J. W. Johnson, A. J. Jacobson, J. Angew. Chem., Int. Ed. Engl., 22 (1983) 412
[35] E. Bordes, P. Courtine, J. W. Johnson, J. Solid State Chem., 55 (1984) 270
[36] G. Busca, F. Cavani, G. Centi, F. Trifirò, J. Catal., 99 (1986) 400
[37] M. R. Thomson, J. R. Ebner in P.Ruyiz and P. Delmon “New development in selective oxidation
by heterogeneous catalysis”, Edition Elsevier, Amsterdam (1992) 353
[38] L. M. Cornaglia, C. Caspari, E. A. Lombardo, Appl. Catal., 15 (1991) 74
[39] R. M. Contractor, J. R. Ebner, M. J. Mummey in G. Centi and F. Trifirò “New development in
selective oxidations”, Edition Elsevier Science, Amsterdam (1990) 553
[40] S. Albonetti, F. Budi, F. Cavani, S. Ligi, G. Mazzoni, F. Pierelli, F. Trifirò, Stud. Surf. Sci.Catal.,
136 (2001) 141-146
[41] S. A. Linde, Yu E. Gorbunova, Dokl. Akad. Nauk SSSR, 245 (1979) 584
[42] P. T. Nguyen, R. D. Hoffman, A. W. Sleight, Mater. Res. Bull., 30 (1995) 1055
[43] Yu E. Gorbunova, S. A. Linde, Sov. Phys. Dokl., 24 (1979) 138
[44] T. Okamura, M. Misono, Catal. Today, 16 (1993) 61
[45] R. M. Contractor, Bergna, H. S. Horowitz, C. M. Blackstone, B. Malone, C. C. Torardi, B.
Griffiths, U. Chowdhry, A. W. Sleight, Catal. Today, 1 (1987) 47
[46] C. C. Torardi, J. C. Calabrese, Inorg. Chem., 23 (1984) 1308
[47] E. Bordes, Catal. Today, 1 (1987) 499
[48] E. Bordes, Catal. Today, 3 (1987) 163
[49] N. Duvauchelle, E. Kesteman, F. Oudet, E. Bordes, J. Solid State Chem., 137 (1998) 311-324
[50] L. O’ Mahony, J. Henry, D. Sutton, T. Curtin, B. K. Hodnett, Appl. Catal. A Gen., 253 (2003) 409
[51] M. Nakamura, K. Kawai, Y. Fujiwara, J. Catal., 34 (1974) 345
[52] T. Shimoda, T. Okuhara, M. Misono, Bull. Chem. Soc. Jpn., 58 (1985) 2163
[53] J. S. Buchanan, J. Apostolakis, S. Sundaresan, Appl. Catal., 19 (1985) 65
[54] B. K. Hodnett, B. Delmon, J. Catal., 88 (1984) 43
[55] A. Caldarelli, M. A. Bañares, C. Cortelli, S. Luciani, F. Cavani, Catal. Sci. Technol., 4 (2014) 419

158
 References

[56] P. T. Nguyen, A. W. Sleight, N. Roberts, W. W. Warren, J. Solid State Chem., 122 (1996) 259
[57] T. Okumara, K. Inumaru, M. Misono, ACS Symp. Ser., 523 (1993) 156
[58] K. Inumaru, T. Okumara, M. Misono, Chem. Lett., 21(10) (1992) 1955
[59] G. Mazzoni, F. Cavani, G. Stefani to Lonza (2005) US Patent 6956004
[60] Z. Shan, W. S. Frazee, M. J. Mummey, B. A. Horrel, to Hunstman WO 2012033635 Patent
[61] P. Ruiz, P. Bastians, L. Caussin, R. Reuse, L. Daza, D. Acosta, B. Delmon, Catal. Today, 16
(1993) 99
[62] F. Garbassi, J. C. J. Bart, F. Montino, G. Petrini, Appl. Catal., 16 (1985) 271
[63] F. Cavani, G. Centi, I. Manenti, F. Trifirò, Ind.. Eng. Chem. Prod. Res. Dev., 24 (1985) 221
[64] H. Morishige, J. Tamaki, N. Miura, N. Yamazoe, Chem. Lett., 19(9) (1990) 1513
[65] G. Centi, Catal. Today, 16 (1993) 5-26
[66] P. Mars, D.W. van Krevelen, Chem. Eng. Sci. Spec. Suppl., 3 (1954) 41.
[67] J.-C. Volta, Catal. Today, 32 (1996) 29
[68] Y. Schuurman, J. T. Gleaves, Ind. Eng. Chem. Res., 33 (1994) 2935
[69] G. J. Hutchings, A. Desmartin-Chomel, R. Olier, J.C. Volta, Nature, 368 (1994) 41
[70] M. Abon, K. Béré, A. Tuel, P. Delichere, J. Catal., 156 (1995) 28
[71] K. Ait-Lachgar, A. Tuel, M. Brun, J.M. Herrmann, J.M. Krafft, J.R. Martin, J.C. Volta, J. Catal.,
177 (1998) 224
[72] G. J. Hutchings, C.J. Kiely, M.T. Sananes-Schulz, A. Burrows, J.C. Volta, Catal. Today, 40
(1998) 273.
[73] U. Rodemerck, B. Kubias, H.W. Zanthoff, M. Baerns, Appl. Catal., 153 (1997) 203.
[74] U. Rodemerck, B. Kubias, H.W. Zanthoff, G.U. Wolf, M. Baerns, Appl. Catal., 153 (1997) 217
[75] E. Bordes, Catal. Today, 16 (1993) 27.
[76] E. Bordes, P. Courtine, J. Catal., 57 (1979) 236
[77] B. Jordan, C. Calvo, Can. J. Chem., 51 (1973) 2621
[78] D. Ballutaud, E. Bordes, P. Courtine, Mater. Res. Bull., 17 (1982) 519
[79] M. Tachez, F. Theobald, J. Bernard, A. W. Hewat, Rev. Chim. Miner., 19 (1982) 291
[80] E. Bordes, P. Courtine, J. Chem. Soc. Chem. Comm., 34 (1985) 294
[81] F. B. Ben Abdelouahab, R. Oliver, N. Guilhaume, F. Lefebvre, J.-C Volta, J. Catal., 134 (1992)
151
[82] F. Girgsdies, M. Schneider, A. Brückner, T. Ressler, R. Schlögl, Solid State Sciences, 11(2009)
1258-1264
[83] P. Amorós, G. S. Patience, J. R. Fernandez, M. J. Lorences, F. Díez, A. Vega, R. Cenni, Catal.
Today, 118 (2006) 32-38

159
 Ringraziamenti

[84] G. W. Coulston, S.R. Bare, H. Kung, K. Birkeland, G.K. Bethke, R. Harlow, N. Herron, P.L. Lee,
Science, 275 (1997) 191
[85] P. L. Gai, K. Kourtakis, D.R. Coulson, G.C. Sonnichsen, J. Phys. Chem. B, 101 (1997) 9916.
[86] R. Mallada, S. Sajip, C. J. Kiely, M. Menendez, J. Santamaria, J. Catal., 196 (2000) 1
[87] S. Mota, M. Abon, J.C. Volta, J.A. Dalmon, J. Catal., 193 (2000) 308.
[88] G. Busca, G. Ramis, V. Lorenzelli, J. Mol. Catal., 50 (1989) 231
[89] F. Cavani, F. Trifirò, Appl. Catal. A Gen., 157 (1997) 195
[90] L. M. Cornaglia, E. A. Lombardo, Stud. Surf. Sci. Catal., 90 (1994) 429
[91] G. Centi, G. Gollinelli, G. Busca, J. Phys. Chem., 94 (1990) 6813
[92] J. C. Vedrine, J. M. M. Millet, J.-C. Volta, Catal. Today, 32 (1996) 115-123
[93] D. J. Thompson, I. M. Ciobîcă, B. K. Hodnett, R. A. Van Santen, M. O. Fanning, Surf. Science,
547 (2003) 438-451
[94] Y. Schuurman, J. T. Gleaves, Catal. Today, 33 (1997) 25-37
[95] G. S. Patience, R. E. Bockrath, J. D. Sullivan, H. S. Horowitz, Ind. Eng. Chem. Res., 46 (2007)
4374-4381
[96] P. G. Pries de Oliveira, J. G. Eon, M. Chavant, A. S. Richè, V. Martin, S. Caldarelli, Catal.
Today, 57 (2000) 177-186
[97] V. V. Guliants, Catal. Today, 51 (1999) 255
[98] R. A. Overbeek, P.A. Warringa, M.J.D. Crombag, L.M. Visser, A.J. van Dillen, J.W. Geus, Appl.
Catal. A, 135 (1996) 209
[99] R. A. Overbeek, A.R.C.J. Pekelharing, A.J. van Dillen, J.W. Geus, Appl. Catal. A, 135 (1996) 231
[100] M. Ruitenbeek, R.A. Overbeek, A.J. van Dillen, D.C.Koningsberger, J.W. Geus, Recl. Trav.
Chim. Pays-Bas, 115 (1996) 519
[101] M. Nakamura, K. Kawai, Y. Fujiwara, J. Catal., 34 (1974) 345
[102] M. J. Ledoux, C. Crouzet, C. Pham-Huu, V. Turines, K. Kourtakis, P. Mills and J.J. Lerou, J.
Catal., 203 (2001) 495
[103] R. A. Overbeek, P.A. Warringa, M.J.D. Crombag, L.M. Visser, A.J. van Dillen and J.W. Geus,
Appl. Catal. A, 135 (1996) 209
[104] F. Cavani, D. De Santi, S. Luciani, A. Lofberg, E. Bordes-Richard, C. Cortelli, R. Leanza, Appl.
Catal. A, 376 (2010) 66–75
[105] M. Abon, J.M. Herrmann, J.C. Volta, Catal. Today, 71 (2001) 121
[106] L. Cornaglia, S. Irusta, E.A. Lombardo, M.C. Durupty, J.C. Volta, Catal. Today, 78 (2003) 291
[107] J. G. Hutchings, Appl. Catal., 72 (1991) 1
[108] G. J. Hutchings, R. Higgins, J. Catal., 162 (1996) 153

160
 References

[109] F. Cavani, C. Fumagalli, F. Ghelfi, C. Mazzoni, F. Pierelli to Lonza (2005) EP 1514598 A1

Patent
[110] D. Ye, A. Satsuma, T. Hattori, Y. Marakami, Appl. Catal., 69 (1991) L1
[111] D.-Q. Ye, M.-L. Fu, L. Hong, S.-Z. Rong, S. Cheng, X. Pang, Res. Chem. Intermed., 29 (2003)
271
[112] G. J. Hutchings, Catal. Today, 16 (1993) 139
[113] V. V. Guliants, J. B. Benziger, S. Sundaresan, I. E. Wachs, Stud. Surf. Sci. Catal., 130B (2000)
1721
[114] H. Liang, D.-Q. Ye, W.-M. Lin, J. Yun. Univ., 28 (2006) 68
[115] N. Ballarini, F. Cavani, C. Cortelli, S. Ligi, F. Pierelli, F. Trifirò, C. Fumagalli, G. Mazzoni, T.
Monti, Top. Catal., 38 (2006) 147
[116] I. Matsuura, T. Ishimura, S. Hayakawa, N. Kimura, Catal. Today, 28 (1996) 133-138
[117] V. V. Guliants, J. B. Benziger, S. Sundaresan, I. E. Wachs, A. M. Hirt, Catal. Lett., 62 (1999) 87-
91
[118] P. A. Agaskar, R. K. Grasselli, D. J. Buttrey, B. White, Stud. Surf. Sci. Catal., 110 (1997) 219-
226
[119] A. M. Duarte de Farias, W. De A. Gonzalez, P.G. Pries de Oliveira, J. G. Eon, J.M. Herrmann,
M. Aouine, S. Loridant, J.-C. Volta, J. Catal., 208 (2002) 238-246
[120] M. Eichelbaum, R. Stößer, A. Karpov, C.-K. Dobner, F. Rosowski, A. Trunschke, R. Schlögl,
Phys. Chem. Chem. Rev., 14 (2012) 1302-1312
[121] R. Fricke, H.-G. Jerchekewitz,, G. Lischke, G. Z. Ohlmann, Anorg. Allg. Chem., 23 (1979) 448
[122] F. Trifirò, Catal. Today, 16 (1993) 91
[123] B. K. Hodnett, B. Delmon, Appl. Catal., 15 (1985) 141
[124] L. Morselli, F. Trifirò, L. Urban, J. Catal., 75 (1982) 112
[125] M. Ai, J. Catal., 67 (1981) 110
[126] G. Centi, G. Fornasari and F. Trifirò, J. Catal., 89 (1984) 44
[127] S. Albonetti, F. Cavani and F. Trifirò, Catal. Rev. Sci. Eng., (1996) 413
[128] D. J. Thompson, I.M. Ciobîca, B.K. Hodnett, R.A. van Santen, M.O. Fanning, Surface Science,
547 (2003) 438
[129] D. J. Thompson, I.M. Ciobîca, B.K. Hodnett, R.A. van Santen, M.O. Fanning, Catal. Today, 91
(2004) 177
[130] G. Busca, E. Finocchio, G. Ramis, G. Ricchiaroli, Catal. Today, 32 (1996) 1330
[131] V. A. Zazhigalov, J. Haber, J. Stoch, V. M. Belousov, Appl. Catal., 96 (1993) 135
[132] M. R. Thompson, J. R. Ebner in P. Ruyiz, B. Delmon “New Developments in Selective Oxidation
by Heterogeneous Catalysis”, Edition Elsevier, Amsterdam, 1992, page 353

161
 Ringraziamenti

[133] P. L. Gai, K. Kourtakis, Science, 267 (1995) 661

[134] P. L. Gai, Topics Catal., 8 (1999) 97
[135] P. A. Agaskar, L. De Caul, R. K. Grasselli, Catal. Lett., 23 (1994) 339
[136] Z. Du, J. Ma, F. Wang, J. Liu, J. Xu, Green Chem., 13 (2011) 554-557
[137] H. Guo, G. Yin, J. Phys. Chem. C, 115 (2011) 17516-17522
[138] N. Alonso-Fagündez, M. López Granados, R. Mariscal, and M. Ojeda, ChemSusChem, 5 (2012)
1984-1990
[139] H.-D. Hahn, G. Dämbkes, N. Rupprich, H. Bahl, Ullmann’s Encyclopedia of Industrial
Chemistry, Vol. 6, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim (2012) pp. 417-430.
[140] T. Tsuchida, J. Kuboa, T. Yoshioka, S. Sakuma, T. Takeguchi, W. Ueda, J. Catal., 259 (2008)
183-189
[141] C. Yang, Z. Meng, J. Catal.,142 (1993) 37-44
[142] A. S. Ndou, N. Plint, N.J. Coville, Appl. Catal. A, 251 (2003) 337-345
[143] M. León, E. Díaz, A. Vega, S. Ordóñez, A. Auroux, Appl. Catal. B, 102 (2011) 590–599
[144] B. G. Harvey, H. A. Meylemans, J. Chem. Technol. Biotechnol., 86 (2011) 2
[145] N. Qureshi, T. C. Ezeji, Biofuels, Bioprod. Bioref., 2 (2008) 319
[146] M .R. Harper, K. M. V. Geem, S. P. Pyl, S. S. Merchant, G. B. Marin, W. H. Green, Combust.
Flame, 158 (2011) 2075
[147] R. Horyń, R. Klimkiewicz, Appl. Catal. A, 370 (2009) 72-77
[148] C. Cobzaru, S. Oprea, E. Dumitriu, V. Hulea, Appl. Catal. A, 351 (2008) 253-258
[149] A. Ungureanu, S. Royer, T. V. Hoang, D. Trong On, E. Dumitriu, S. Kaliaguine, Micropor.
Mesopor. Mater., 84 (2005) 283-296.
[150] H. Hattori, Appl. Catal. A, 222 (2001) 247-259
[151] J. Raskó, J. Kiss, Appl. Catal. A: Gen., 285 (2005) 252–260
[152] J. S. Kruger, R. Chakrabarti, R.J. Hermann, L.D. Schmidt, Appl. Catal. A, 411–412 (2012) 87–94
[153] A. A. Shoaibi, A.M. Dean, J. Fuel Cell Sci. Technol, 7 (2010) 041015-041018
[154] A. Iriondo, M.B. Guemez, J. Requies, V.L. Barrio, J.F. Cambra, P.L. Arias, J.L.G. Fierro, Stud.
Surf. Sci. Catal., 175 (2010) 453-456
[155] M. A. Parent, J.B. Moffat, Catal. Lett., 48 (1997) 135-143
[156] J. Macht, M.J. Janik, M. Neurock, E. Iglesia, J. Am. Chem. Soc., 130 (2008) 10369-10379
[157] K. Hashimoto, Y. Matsumura, M. Fukuchi, Y. Kera, Catal. Lett., 19 (1993) 375-381
[158] M. R. Mostafa, A.M. Youssef, Mater. Lett., 34 (1998) 405-410
[159] F. M. Bautista, B. Delmon, Appl. Catal. A, 130 (1995) 47-65
[160] V. Macho, M. Králik, E. Jurecekova, J. Hudec, L. Jurecek, Appl. Catal. A, 214 (2001) 251–257
[161] S. J. Joris, C.H. Amberg J. Fuel Cell Sci. Technol.,76 (1971) 3167-3171

162
 References

[162] G. Guiu, P. Grange, J. Catal., 168 (1997) 463-470

[163] A. Gil, M.A. Vicente, S.A. Korili, J. Catal., 229 (2005) 119-126
[164] S. Delsarte, M. Florea, F. Maugé, P. Grange, Catal. Today, 116 (2006) 216–225
[165] R. A. Hernandez, U. S. Ozkan, Ind. Eng. Chem. Res., 29 (1990) 1454
[166] U.S. Ozkan, G.L. Schrader, J. Catal., 95 (1985) 137
[167] J. McMurry, Chimica Organica, Ed. Piccin, 2009
[168] M. Perissinotto, M. Lenarda, L. Storaro, R. Ganzerla, J. Molec. Catal. A: Chem., 121 (1997) 103-
109
[169] G. Chen, S. Li, F. Jiao, Q. Yuan, Catal. Today, 125 (2007) 111-119
[170] P. Vazquez, L. Pizzio, C. Caceres, M. Blanco, H. Thomas, E. Alesso, L. Finkielsztein, B.
Lantano, G. Moltrasio, J. Aguirre, J. Molec. Catal. A: Chem., 161 (2000) 223-232
[171] M. J. Janik, J. Macht, E. Iglesia, M. Neurock, J. Phys. Chem., 113 (2009) 1872-1885
[172] P. Berteau, B. Delmon, J.-L. Dallons, A. Van Gysel, Appl. Catal., 70 (1991) 307-323
[173] M. A. Makarova, E. A. Paukshtis, J. M. Thomas, C. Williams, K.I. Zamaraev, Stud. Surf. Sci.
Catal., 75 (1993) 1711-1714
[174] L. Matachowski, A. Drelinkiewicz, E. Lalik, D. Mucha, B. Gil, Z. Brozek-Mucha, Z. Olojniczak,
Microporous Mesoporous Mater., 144 (2011) 46-56
[175] M. Ammam, J. Mater. Chem. A, 1 (2013) 6291-6312
[176] S. Lis, J. Alloys and Compounds, 300-301 (2000) 88-94
[177] X. Lopez, L. Vila-Nadal, X. Aparicio-Angles, J. M. Poblet, Phys. Procedia, 8 (2010) 94-103
[178] M. Sadakane, E. Steckhan, Chem. Rev., 98 (1998) 219-237
[179] I. A. Weinstock, Chem. Rev., 98 (1998) 113-170
[180] A. S. Dias, M. Pillinger, A. A. Valente, Appl. Catal. A: Gen., 285 (2005) 126-131
[181] V. Palermo, G. P. Romanelli, P. G. Vazquez, J. Molec. Catal. A: Chem., 373 (2013) 142-150
[182] L. Jing, J. Shi, F. Zhang, Y. Zhong, W. Zhu, Ind. Eng. Chem. Res., 52 (2013) 10095-10104
[183] P. Xiao, F. Dumur, M. A. Tehfe, B. Graff, J. P. Fouassier, D. Gigmes, J. Lalevee, Macromol.
Chem. Phys., 214 (2013) 1749-1755
[184] J. T. Rhule, C. L. Hill, D. A. Judd, R. F. Schinazi, Chem. Rev., 98 (1998) 327-357
[185] J. Berzelius, Ann. Phys. Chem., 82 (1826) 369
[186] M. T. Pope, “Heteropoly and Isopoly oxometalates”, Ed. Springer, 1983
[187] C. L. Hill, Chem. Rev., 98 (1998) 1-2
[188] L. Pauling, J. Am. Chem. Soc., 51 (1929) 2868
[189] J. F. Keggin, Nature, 131 (1933) 908
[190] A. Corma, A. Martinez, C. Martinez, J. Catal., 164 (1996) 422-432
[191] T. Okuhara, T. Nishimura, H. Watanabe, M. Misono, J. Mol. Catal., 74 (1992) 247256

163
 Ringraziamenti

[192] N. F. Dummer, W. Weng, C. Kiely, A.F. Carley, J.K. Bartley, C.J. Kiely, G.J. Hutchings, Appl.
Catal. A, 376 (2010) 47–55
[193] L. Griesel, J. K. Bartley, R. P. K. Wells, G. J. Hutchings, J. Mol. Cat. A: Chem., 220 (2004) 113
[194] G. Mazzoni, G. Stefani, F. Cavani, EP 804963 Patent (1997), assigned to Lonza group
[195] F. Cavani and J. H. Teles, ChemSusChem, 2 (2009) 508–534
[196] M. S. Haddad, R. A. Gustaferro, to Ineos (2013) US 20130102455 A1 Patent
[197] C. K. Dobner, S. Altwasser, W. Hagen, F. Rosowski, to BASF (2011) US 20110257414 A1
Patent
[198] A. Caldarelli, F. Cavani, F. Folco, S. Luciani, C. Cortelli, R. Leanza, Catal. Today, 157 (2010)
204-210
[199] H. Bluhm, M. Hävecker, E. Kleimenov, A. Knop-Gericke, A. Liskowski, R. Schlögl, D. S. Su,
Top. Catal., 23 (2003) 99-107
[200] M. Hävecker, A. Knop-Gericke, H. Bluhm, E. Kleimenov, R. W. Mayer, M. Fait, R. Schlögl,
Appl. Surf. Sci., 230 (2004) 272-282
[201] M. Hävecker, R.W. Mayer, A. Knop-Gericke, H. Bluhm, E. Kleimenov, A. Liskowski, D. S. Su,
R. Follath, F. G. Requejo, D. F. Ogletree, M. Salmeron, J. A. Lopez-Sanchez, J. K. Bartley, G. J.
Hutchings, R Schlögl, J. Phys. Chem. B, 107 (2003) 4587-4596
[202] H. W. Zanthoff, M. Sananes-Schultz, S. A. Buchholz, U. Rodemerck, B. Kubias, M. Baerns,
Appl. Catal. A, 172 (1998) 49–58
[203] E. W. Arnold, S. Sundaresan, Appl. Catal., 41 (1988) 225-239
[204] R. M. Contractor, H. S. Horowitz, G. M. Sisler, E. Bordes, Catal. Today, 37 (1997) 51-57
[205] E. Bordes, J. W. Johnson, A. Raminosova, P. Courtine, Mater. Sci. Monograph., 28B (1985) 887
[206] G. Stefani, P. Fontana to Lonza, US Patent 4181628 (1980)
[207] German Patent 2353136 (1973)
[208] K. R. Bakshi, US Patent 4123442 (1978)
[209] R. C. Edwards, P. H. Kilmer, C. A. Hudovich, P. L. Stauffenberg, Eur. Pat. Appl., EP123467
(1984)
[210] B. K. Hodnett, Catal. Rev.-Sci. Eng., 27 (1985) 373
[211] I. Mastuura, M. Yamazaki in G. Centi and F. Trifirò “New Developments in Selective
Oxidation”, Edition Elsevier, Amsterdam (1990) 563
[212] V. A. Zazhigalov, Teor. And Experim. Chem., 35 (1999) 265
[213] M. Pawlyta, J.-N. Rouzaud, S. Duber, Carbon, 84 (2015) 479-490
[214] A. C. Ferrari, J. Robertson, Phys. Rev. B, 61 (2000) 14095
[215] G. F. Froment, Stud. Surf. Sci. Catal., 6 (1980) 1
[216] M. Guisnet, P. Magnoux, Stud. Surf. Sci. Catal., 88 (1994) 53

164
 References

[217] D. M. Golden, K. W. Egger, S. W. Benson, J. Amer. Chem. Soc., 86 (1964) 5416

[218] L. Matachowski, A. Drelinkiewicz, E. Lalik, D. Mucha, B. Gil, Z. Brozek-Mucha, Z. Olojniczak,
Microporous Mesoporous Mater., 144 (2011) 46-56
[219] P. Berteau, M. Runet, B. Delmon, Acta Chim. Hung., 124 (1987) 25
[220] Y. H. Taufiq-Yap, M. H Looi, M. Z. Hussein, Z. Zainal, Asian Jour. Chem., 14 (2002) 1494-
1502
[221] I. J. Ellison, G. J. Hutchings, M. T. Sananes, J. C. Volta, J. Chem. Soc. Chem. Commun., 9
(1994) 1093
[222] Y. H. Taufiq-Yap, A. A. Rownaghi, M. Z. Hussein, R. Irmawati, Catal. Lett., 119 (2007) 64
[223] N. Yamamoto, N. Hiyoshi, T. Okuhara, Chem. Mater., 14 (2002) 3882
[224] N. Hiyoshi, N. Yamamoto, N. Ryumon, T. Okuhara, J. Catal., 221 (2004) 225-233
[225] G. Ladwig, Z. Anorg. Allg. Chem., 338 (1965) 266
[226] M. Trchová, P. Čapková, P. Matějka, K. Melánová, L. Beneš, E. Uhlıřová, J. Solid State Chem.,
148 (1999) 197
[227] C. J. Kiely, A. Burrows, S. Sajip, G. J. Hutchings, M. T. Sananes, A. Tuel, J. C. Volta, J. Catal.,
162 (1996) 31
[228] S. J. Puttock, C. H. Rochester, J. Chem. Soc. Faraday Trans. I, 82 (1986) 2773
[229] G. Busca, G. Centi, F. Trifirò, J. Am. Chem. Soc., 107 (1985) 7757
[230] M. Ai, J. Catal., 116 (1989) 23
[231] G. Centi, D. Pesheva, F. Trifirò, Appl. Catal., 33 (1987) 343
[232] A. Martin, B. Lücke, H. Seeboth, G. Ladwig, Appl. Catal., 49 (1989) 205
[233] G. Centi, F. Trifirò, J. Molec. Catal., 35 (1986) 255
[234] G. Centi, J. L. Nieto, F. Ungarelli, F. Trifirò, Catal. Letters, 4 (1990) 309
[235] M. Ai, S. Suzuki, Bull. Chem. Soc. Jpn., 47 (1974) 3074-3077
[236] V. V. Guliants, J.B. Benziger, S. Sundaresan, Stud. Surf. Sci. Catal., 101 (1996) 991-1000
[237] V. V. Guliants, S.A. Holmes, J. Mol. Catal. A, 175 (2001) 227-239
[238] F. Cavani, F. Trifirò, Catal. Today, 51 (1999) 561-580
[239] Ph. Arpentinier, F. Cavani, F. Trifirò, Catal. Today, 99 (2005) 15-22
[240] N. Ballarini, F. Cavani, C. Cortelli, F. Gasparini, A. Mignani, F. Pierelli, F. Trifirò, C. Fumagalli,
G. Mazzoni, Catal. Today, 99 (2005) 115-122
[241] F. Cavani, A. Colombo, F. Giuntoli, E. Gobbi, F. Trifirò, P. Vazquez, Catal. Today, 32 (1996)
125-132
[242] M. López Granados, J.L.G. Fierro, F. Cavani, A. Colombo, F. Giuntoli, F. Trifirò, Catal. Today,
40 (1998) 251-261

165
 Ringraziamenti

[243] G. Bignardi, F. Cavani, C. Cortelli, T. De Lucia, F. Pierelli, F. Trifirò, G. Mazzoni, C. Fumagalli,

T. Monti, J. Molec. Catal. A, 244 (2006) 244-251
[244] F. Cavani, G. Centi, A. Riva, F. Trifirò, Catal. Today, 1 (1987) 17-26
[245] F. Cavani, G. Centi, F. Trifirò, R.K. Grasselli, Catal. Today, 3 (1988) 185-198
[246] R. Wenig, G. Schrader, J. Phys. Chem., 91 (1987) 5674–5680
[247] B. Sakakini, J. Catal., 189 (2000) 253–262
[248] R. W. Wenig, G. L. Schrader, J. Phys. Chem., 91 (1987) 1911-1918
[249] S. K. Bhattacharyya, A. K. Kar, J. Catal., 1 (1962) 293-300
[250] P. Cicmanec, K. Syslova, J. Tichy, Topics Catal, 45 (2007) 229-232
[251] J.M. Church, P. Bitha, Ind. & Eng. Chem. Prod. Res. & Dev., 2 (1963) 61-66
[252] Z.-Y. Xue, G.L. Schrader, J. Catal., 184 (1999) 87-104
[253] M. Ai, T.-h. Tsai, A. Ozaki, Bull. Chem. Soc. Jpn., 53 (1980) 2647-2650
[254] P. Ruanpanun, Z. Teklemariam Dame, H. Laatsch, S. Lumyong, FEMS Microbiol Lett, 322
(2011) 77–81

166