THESIS OF THE DOCTORAL (PhD) DISSERTATION
SURFACE CHEMISTRY OF MOLYBDENA CONTAINING
CATALYSTS
Written by:
HEIDER NASZER
M.Sc. environmental engineer
Supervisor: Prof. Dr. ÁKOS RÉDEY
Doctoral School of Chemical Engineering and Material Sciences
INSTITUTIONAL DEPARTMENT OF ENVIRONMENTAL ENGINEERING
AND CHEMICAL TECHNOLOGY
FACULTY OF ENGINEERING
UNIVERSITY OF PANNONIA
Veszprém
2008
DOKTORI (PhD) ÉRTEKEZÉS
MOLIBDÉN TARTALMÚ KATALIZÁTOROK FELÜLETKÉMIAI
TULAJDONSÁGAINAK VIZSGÁLATA
Készítette: NASZER HEIDER
okleveles környezetmérnök
Témavezető:
Dr. RÉDEY ÁKOS
egyetemi tanár
Készült a Pannon Egyetem
Vegyészmérnöki Tudományok és Anyagtudományok Doktori Iskola Keretében
PANNON EGYETEM
MÉRNÖKI KAR
KÖRNYEZETMÉRNÖKI ÉS KÉMIAI TECHNOLÓGIA INTÉZETI TANSZÉK
Veszprém
2008
2
MOLIBDÉN TARTALMÚ KATALIZÁTOROK FELÜLETKÉMIAI
TULAJDONSÁGAINAK VIZSGÁLATA
Értekezés doktori (PhD) fokozat elnyerése érdekében
Írta:
NASZER HEIDER
Készült a Pannon Egyetem Vegyészmérnöki Tudományok és Anyagtudományok Doktori
Iskolájához tartozóan. Témavezető: Dr. RÉDEY ÁKOS Elfogadásra javaslom (igen / nem) ………………………. (aláírás) A jelölt a doktori szigorlaton…......... % -ot ért el. Az értekezést bírálóként elfogadásra javaslom: Bíráló neve: …........................…................. (igen /nem) ………………………. (aláírás) Bíráló neve: …........................…................. (igen /nem) ………………………. (aláírás) A jelölt az értekezés nyilvános vitáján…..........% - ot ért el. Veszprém, ………………………….
A Bíráló Bizottság elnöke A doktori (PhD) oklevél minősítése…................................. ………………………… Az EDT elnöke
3
CONTENTS
CHAPTER 1. INTRODUCTION AND LITERARY OVERVIEW 8
1.1. γ-Al2O3 and molybdena-alumina catalysts 8
1.1.1. The surface hydroxyls of γ-Al2O3 and molybdena-alumina 13
1.2. Current status of catalysts preparation 17
1.2.1. The impregnation method 17
1.2.2. Adsorption methods 18
1.3. Molybdena with CeO2 and SnO2 semiconductors 22
1.4. Direct conversion of methane under nonoxidative conditions 27
1.5. Characterization of the catalysts 31
1.5.1. Physical characterization 32
1.5.2. Surface characterization and IR spectroscopy 34
1.5.3. FTIR spectroscopic detection of CO adsorption on the catalysts 37
1.6. OBJECTIVES 40
CHAPTER 2. EXPERIMENTAL 41
2.1. Materials and catalyst preparation 41
2.2. Catalyst characterization methods and techniques 42
CHAPTER 3. RESULTS 44
3.1. Surface texturing 44
3.2. X-ray diffraction 48
3.2.1. XRD patterns of Mo/Al2O3, Ce-Mo/Al2O3 and Mo/CeO2 48
3.2.2. XRD patterns of Sn-Mo/Al2O3 and Mo/SnO2 51
3.3. Thermal analysis 56
3.3.1. TG and DTA of Mo/Al2O3 56
3.3.2. TG and DTA of Ce-Mo/Al2O3 and Mo/CeO2 58
3.3.3. TG and DTA of Sn-Mo/Al2O3 and Mo/SnO2 60
3.4. Electron Spin Resonance (ESR) measurements 62
4
3.5. In situ DRIFT spectroscopy measurements 65
3.5.1. DRIFT spectra of the calcined samples under vacuum 65
3.5.1.1. DRIFT spectra of γ-Al2O3 and CeO2 65
3.5.1.2. DRIFT spectra of Mo/Al2O3, Mo/CeO2 and Ce-Mo/Al2O3 66
3.5.1.3. DRIFT spectra of SnO2, Mo/SnO2 and Sn-Mo/Al2O3 68
3.5.2. CO chemisorption 70
3.5.2.1. CO chemisorption on Mo/Al2O3 70
3.5.2.2. CO chemisorption on CeO2, Mo/CeO2 and Ce-Mo/Al2O3 72
3.5.2.3. CO chemisorption on SnO2, Mo/SnO2 and Sn-Mo/Al2O3 74
3.5.3. In situ DRIFT results on methane transformation in absence of oxygen 76
3.5.3.1. Methane transformation on Mo/Al2O3 77
3.5.3.2. Methane transformation on Ce-Mo/Al2O3 78
3.5.3.3. Methane transformation on Mo/CeO2 79
3.5.3.4. Methane transformation on Sn-Mo/Al2O3 80
3.5.3.5. Methane transformation on Mo/SnO2 81
3.5.3.6. DRIFT spectra after methane reaction under vacuum 82
CHAPTER 4. DISCUSSION 84
4.1. Surface texturing 84
4.2. X-ray diffraction 84
4.3. Thermal analysis 86
4.4. Electron Spin Resonance (ESR) 87
4.5. In situ DRIFT spectroscopy 88
4.5.1. DRIFT spectra of the calcined samples under vacuum 88
4.5.2. CO chemisorption 89
4.5.3. In situ DRIFT studies on methane transformation in absence of oxygen 93
SUMMARY AND CONCLUSIONS 104
ACKNOWLEDGMENTS 107
REFERENCES 108
THESES 113
PUBLICATIONS 115
5
KIVONAT
Az impregnálással és mechanikus keveréssel készült Mo/Al2O3, Mo/CeO2, Mo/SnO2,
Ce-Mo/Al2O3 és Sn-Mo/Al2O3 katalizátorok struktúrájának, termikus stabilitásának, a hordozó
és a Mo közötti kölcsönhatásoknak, aktivitásuknak, a Mo ionok diszperziójának és a felületi
tulajdonságoknak vizsgálatára Induktív Csatolású Plazma (ICP), N2 adszorpció/deszorpció
(BET módszer), Termikus analízis (TG-DTA), X-ray diffrakció (XRD), Elektron Spin
Rezonancia (ESR) és Diffúz-reflexiós Fourier Transzformációs Infravörös (DRIFT)
módszereket alkalmaztam. A katalizátorok aktivitásának összehasonlítása CO adszorpció és a
CH4 konverziója alapján történt.
A BET, XRD és DRIFT eredmények összevetése alapján megállapítható, hogy a kalcinálás
hőmérséklete és időtartama, az oldat pH-ja és a hordozó izoelektromos pontja hatással vannak
a Mo ionok felületi diszperziójára. Másrészt, a cérium növelte a polimer Mo ionok felületi
koncentrációját (főleg hordozóként) cérium-molibdénátok formájában, illetve Mo-O-Ce
kötések megjelenésével (630, 875 cm-1), és újabb kettős O=Mo=O (995 és 1035 cm-1) kötések
jelentek meg, amelyek polimer MoO3 alakulatokra jellemzők.
A termikus analízis alapján megállapítható, hogy a Mo/Al2O3 termikusan a legstabilabb
(900°C-ig), ugyanakkor a Mo/CeO2, Mo/SnO2 és Ce-Mo/Al2O3 minták 700°C feletti
hőmérsékleten történő hőkezelése során morfológiai és kristályszerkezeti változások
következnek be, így lehetővé válik a fémionok diffúziója és kation csere játszódhat le köztük.
A szén monoxid 100°C-on történő kemiszorpciója során a karbonátok különböző formái és
fém-karbonil kapcsolódások jelentek meg, amelyek vákuumban stabilak maradtak. Viszont
ezek kialakulásához szükséges a katalizátorok redukált formája, továbbá a redukció során
kialakult koordinatívan telítetlen helyek (CUS), a kristályrácsban lévő oxigén ionok és a
felületi hidroxil csoportok jelentős szerepet játszanak. Másrészt, a Mo0 atomokat tartalmazó
alakulatok 700°C-on történő redukció után tűnnek fel, amelyekhez terminális, illetve
hídkötésű kapcsolódó szénmonoxid DRIFT sávjai jelentek meg (2025, 2002, 1994 cm-1).
A DRIFT vizsgálatok igazolták, hogy 800ºC-on redukált katalizátorokon a metán 700ºC-os
bontása során a felületi karbonátok különböző formái jelentek meg, illetve ezek bomlásából
származó CO és CO2 fejlődése megfigyelhető volt. Ez a rácsbeli oxigén reakcióképességének
és a hidrogén redukció során kialakult Lewis savas helyeknek köszönhető. Azonban, a
Mo/CeO2 és Mo/SnO2 katalizátorok jelentős aktivitást mutattak (főleg Mo/SnO2) a metán
reakcióban, ami feltehetően a nagymértékben diszpergált MoO3 klaszterek, illetve Ce+3/Ce+4
és Sn2+/Sn4+ redox párok együtthatásával magyarázható.
6
ABSTRACT
Mo/Al2O3, Mo/CeO2, Mo/SnO2, Ce-Mo/Al2O3 and Sn-Mo/Al2O3 catalysts were prepared by
impregnation and co-precipitation methods. The catalysts were characterized by Inductive
Coupled Plasma (ICP), Thermal analysis (TG and DTA), X-ray diffractometry, Electron Spin
Resonance (ESR), Diffuse reflectance Infrared (DRIFT) spectroscopy as well as N2
adsorption/desorption (BET method) techniques to examine their structural characteristics,
thermal stability, mutual interactions between Mo and the support, catalytic activity as well as
the dispersity of Mo ions and surface structure. The samples were compared to determine
what kinds of adspecies participate efficiently during CO adsorption and CH4 decomposition.
Combining the results obtained from BET, XRD and DRIFT investigations one may suggest
that beyond the Mo dispersion the formation of MoO3 clusters was found to respond to the
calcination temperature and time as well as to the solution pH and the isoelectric point of the
solid support. On the other hand, the introduction of ceria resulted in different molecular
formulae with Mo (particularly as a support). This led to the increase of polymerized surface
Mo species so as to forming Mo-O-Ce (bands at 630 and 875 cm-1) linkages, besides the
formation of coupled O=Mo=O bonds at 995 and 1035 cm-1 indicative of polymeric MoO3.
From thermal analysis, it can be inferred that Mo/Al2O3 is the thermally most stable material
in the temperature range used in the experiment (up to 900°C). Whereas Mo/CeO2, Mo/SnO2
and Ce-Mo/Al2O3 samples undergo morphological and structural modifications above 700°C
resulting in lattice defects, which motivate the mobility of metal ions and thus enhance the
possibility of cation exchange between them.
Additionally, the formation of metal-carbonyl species and various types of carbonates through
CO chemisorption at 100°C needs reduced catalysts containing coordinatively unsaturated
sites (CUS), oxygen vacancies and hydroxyl groups. On the other hand, the bands protruding
at 2025, 2002 and 1994 cm-1 are very likely associated with the terminally and bridged CO σ-
bonded to metallic Mo(0) species appearing after reduction at 700°C. Thus, CO being provided
as weakly adsorbed metal-carbonyls migrating towards the oxides through interfacial sites to
form carbonates being stable even under vacuum at room temperature.
Methane is retained at 700ºC by the catalysts reduced at 800ºC and generates various
carbonate species which decompose to CO and CO2 implying the existence of reactive lattice
oxygen in addition to Lewis acid sites possessed by hydrogen reduction. However, the
Mo/CeO2 and Mo/SnO2 materials presented marked activity (especially Mo/SnO2) in CH4
decomposition. This activity is presumably related to highly dispersed MoO3 clusters besides
the Ce3+/Ce4+ and Sn2+/Sn4+ redox couples as further emphasized by means of DRIFT results.
7
1. INTRODUCTION AND LITERARY OVERVIEW
1.1. γ-Al2O3 and molybdena-alumina catalysts Molybdena containing catalysts have been the subject of numerous studies due to their
potential role in different important reactions such as hydrodesulphurization (HDS),
hydrogenation (HYD), oxidative dehydrogenation (ODH), hydrodeoxygenation (HDO)
hydrodenitrogenation, hydrometallization, isomerization, epoxidation, partial oxidation of
alkanes and alkene metathesis reactions [1-5]. Presently, there is a continuous interest in Mo-
containing catalysts directed toward understanding the catalytic properties, beside extensive
studies have been addressed to characterize the structure of the catalyst surface. The chemical
studies of molybdena-alumina catalysts include reduction or sulphidation. These two
treatments are generally needed to activate the catalyst for most of the catalytic reactions [6].
The sulphidation occurs readily above 300°C, and the extent of sulphidation increases with
temperature. However, the catalyst sulphur content is limited at a given temperature. The
major sulphiding reaction is through the exchange of oxygen associated with molybdenum for
sulphur. The molybdena phase of the sulphided catalyst was found stoichiometrically close to
MoS2 [3, 6-10]. Perhaps, the most detailed analysis of the reduction of molybdena-alumina
catalyst is the works of Hall, Massoth, Wang and Rédey [3-12]. In their work, the overall
average oxidation state was determined from the gas consumed on oxidation or reduction and
an attempt was made to separate this average into percentages of Mo6+, Mo5+, Mo4+ and Mo3+
on the basis of H+ retained and vacancies created upon reduction with H2. It was found that
the reducibility of the catalyst after 2 hours reduction at 500°C under 1 atm H2 increased from
approximately e/Mo = 0.5 to e/Mo = 2 (e/Mo: electrons gained per Mo atom) as the Mo
loading increased from 2% to 10%, while the extent of reduction of 25% Mo loading catalyst
reached e/Mo = 2.7 in just one hour reduction. It was also noted that the extent of reduction at
short times for the 10% and 25% catalysts were actually greater than for bulk MoO3 implying
the presence of a MoO3 phase on the catalysts but with a smaller particle size than the low
surface area of bulk MoO3. In view of all reduction studies, it has been generally supposed
that from the amounts of H2 consumed and water generated in the reduction step, an amount
of hydrogen retained on the reduced catalyst could be calculated. Two types of adsorbed
hydrogen were characterized: (1) reversibly adsorbed (HR) which is removable as H2 by
evacuation above 450°C, and (2) irreversibly retained (HI) which could not be removed by
evacuation. The latter could only be removed as H2O by either evacuation at higher
temperature or by reoxidation. In addition to the retained hydrogen, water was produced
8
during the reduction. The amount of water increased with increasing the extent of reduction.
The formation of water was attributed to the removal of oxygen from the molybdena and
consequently created anion vacancies on the surface. On reoxidation, two processes occurred:
the anion vacancies were filled and the HI was reacted to H2O. Given these surface processes,
the following material balance equations were written:
[H2] = [WR] + HR +HI (Eq-1)
[O] = [WR] + HI (Eq-2)
[Wo] = HI – WA (Eq-3)
Where WR and Wo are the quantities of water produced in the reduction and reoxidation
steps, respectively, and WA is water retained by alumina after the reoxidation step
(ideally WA =0) [3, 7, 12].
The redox study was extended by Hall and Lo Jacono who calculated HI per molybdenum
(HI/Mo) and vacancies per molybdenum (□/Mo) at various extents of reduction (Hc/Mo) all
measured as atom/atom [13]. The quantity of hydrogen (Hc) consumed during the reduction
process can be expressed by the equation:
Hc = WR + HR + HI (Eq-4)
At low extents of reduction (Hc/Mo ≤ 1), two HI were present per vacancy created. However,
as reduction was further increased, HI reached a limiting level of about HI/Mo = 0.45 while
□/Mo increased more rapidly. The formation of HI and vacancies was interpreted by assuming
that molybdena forms an epitaxial monolayer on Al2O3. Accordingly, HI was taken as a
measure of Mo5+ and each vacancy was equated to one Mo4+. The reduction process was
pictured as follows:
9
Fig. 1. The reduction process of molybdena according to Hall and Lo Jacono
Later, FTIR study of the surface hydroxyl groups of the same catalyst was conducted by
Millman [14]. These results led to a minor modification in the reduction process of Mo/Al2O3
(Fig. 2). The spectra from catalysts reduced with H2 were indistinguishable in the OH region
from the alumina support or from the unreduced catalysts. Indeed, all five alumina OH bands
were found and no new ones were detected. Thus, the intensity increased on reduction was
10
probably due to the recoveries of Al-OH rather than the formation of Mo-OH. Accordingly,
the authors presented a modified picture for the reduction process of Mo/Al2O3:
Fig. 2. The reduction process of molybdena according to Millman
The implication is, however, that reduction reverses, the process of monolayer formation as
evidenced by the increased intensities of the OH region.
Other studies have suggested that patches of molybdena may be present on the surface,
particularly at high Mo loading. Giordano et al. studied molybdena-alumina catalysts
containing up to 30% Mo by various chemical and physical techniques [15]. These included
11
differential thermal analysis (DTA), thermogravimetric analysis (TGA) infrared and diffuse
reflectance spectroscopy. The authors deduced that below 4% Mo loading the Mo ions are
highly dispersed in the form of MoO42-. At increasing Mo contents (up to 10-15%), a
progressive increase of structures with bridged oxygen atoms in mixed tetrahedral and
octahedral environment (schemes IV and V surrounding Mo6+) were found.
Scheme 1.
They obtained MoO3-rich catalysts at Mo loading higher than 15% prevailing octahedral
configuration attributed to species like schemes V and VI. It was clearly demonstrated by
these authors that the molybdena surface structure could be altered by the Mo concentration
and by the calcination procedures. The Mo concentration was found to play an important role
in the nature of the molybdena surface structure. They suggested that tetrahedral MoO42- is the
main species in the catalysts with less than 5% Mo. Octahedrally coordinated polymolybdates
were formed with more than 5% becoming the dominant species as the loading increased,
while crystalline Al2(MoO4)3 phase was observed in the catalysts calcined at elevated
temperatures (>600°C) and MoO3 clusters were formed when the loading reached 16% and
20%, respectively. The degree of the polymerization becomes higher and higher as the Mo
loading is further increased. After the surface has been saturated (high Mo content), the
formation of bulk MoO3 and Al2(MoO4)3 occurs at different extents depending on the
preparation and calcination conditions.
12
Another surface polymolybdate structure was proposed by Medema et al. to explain their
Raman results:
Scheme 2.
It was suggested that supported molybdena species are aligned as one-dimensional chains
which result in Mo-O-Mo bonds, yielding octahedral coordination for Mo6+ [16, 17].
Despite the variations made by various authors, a consistent pattern has emerged. At low Mo
loading, tetrahedral MoO42- species dominate the surface. Increasing loading is generally
accompanied by the formation of a polymolybdate phase. On the basis of these studies, it is
generally known that the surface of molybdena remains mainly containing monomeric Mo
species. Moreover, polymeric molybdate species and free MoO3 can be formed at high Mo
contents taking into account the calcination temperature and time as well as the solution pH,
the isoelectric point and surface area of the solid support [12-17].
1.1.1. The surface hydroxyls of γ-Al2O3 and molybdena-alumina γ-Al2O3 has a defect spinel structure, a high surface area, a certain degree of acidity, and
forms solid solutions with transition metal oxides such as NiO and CoO. Above 900°C, γ-
Al2O3 is transformed into α-Al2O3, which has hexagonal structure and smaller surface area.
Even at lower temperatures a slow phase transition occurs, which shortens the catalyst
lifetime. Therefore, the incorporation of small amounts (1-2%) of SiO2 or ZrO2 in γ-Al2O3
shifts the γ → α transformation to higher temperature and increases the stability of the catalyst
[1, 2]. The alumina surface hydroxyl groups have been extensively studied by infrared
spectroscopy. Thermal dehydroxylation studies of γ-Al2O3 by IR spectroscopy were first
executed by Peri who concluded that Al-OH groups close to each other could form water, and
desorbed from the surface [18, 19]. The spectra displayed five different OH frequencies at
300°C for Al-OH groups. Assuming that the (100) crystal face is dominantly exposed on
13
γ-Al2O3 surface, Peri proposed a model predicting the existence of five types of Al-OH
groups. The model attributes the different OH frequencies to differing numbers of
surrounding surface oxide sites. In other terms, the difference of Al-OH frequency depended
upon the number of the nearest neighbour oxygen anions surrounded a particular Al-OH
group. Peri surmised that the high wavenumber band at 3800 cm-1 was an OH group with four
neighbouring oxygen anions. Whereas the low wavenumber band at 3700 cm-1 was an OH
group with no oxygen neighbours, but four aluminium ions. The other three OH groups giving
rise to the intermediate stretching frequencies at 3730, 3745, 3780 cm-1 were supposed to
have one, two and three nearest neighbouring oxygen anions, respectively. As the temperature
increased, the lower frequency associated (H-bonded) hydroxyl groups, being closest
together, were removed first, leaving isolated Al-OH groups to be removed at higher
temperatures after heating to 700°C. Upon heating to 900°C no hydroxyl groups were
detected by IR.
Knözinger and Ratnasamy, assuming that all of crystallite faces of Al2O3 (100, 110 and 111)
have an equal chance of projection on γ-Al2O3 surface, proposed a model that assigned the
different OH frequencies of the five types of Al-OH groups to tetrahedrally and octahedrally
coordinated Al3+sites in terminal and bridged configuration [20, 21]. These assignments are
shown in Fig. 3.
Fig. 3. Assignments of OH groups on γ-Al2O3 surface according to Knözinger and Ratnasamy
14
Despite the studies mentioned above, none has attempted to correlate directly the extent of
dehydroxylation with the production of Lewis acid sites as a function of the treatment
temperature. Nevertheless, in the spirit of the work of Zaki and Knözinger who undertook a
low temperature IR study of the interaction between Al-OH groups and adsorbed CO, five
bands of different types of hydroxyl groups were assigned on Al2O3 dehydroxylated at 500°C,
two types of associated Al-OH were removed first, leaving three types of isolated Al-OH
groups to be removed at higher temperature. They observed two low temperature CO
adsorption sites at -193°C on Al2O3 and they attributed the first adsorption site (2140-2150
cm-1) to CO hydrogen-bonded to Al-OH groups and a physisorbed CO layer, while the second
CO adsorption site (2195-2213 and 2238 cm-1) was assigned to tetrahedrally and octahedrally
coordinated Al+3 sites [22, 23].
Later, John and Ballinger undertook a low temperature (-196°C) IR study of the interaction
between Al-OH groups and adsorbed CO, in order to examine the acid-base properties of the
hydroxyl groups on Al2O3 after dehydroxylation at elevated temperatures [24]. They found an
approximately direct correlation between the elimination of surface hydroxyl groups and the
increase in integrated absorbance of CO on Lewis acid sites of Al2O3 produced during the
dehydroxylation. Furthermore, they observed two Al3+ adsorption sites: the first develops
following mild dehydroxylation, and the second appears only after dehydroxylation at 600°C
and higher due to the presence of a mixture of Lewis acid sites on highly dehydroxylated
Al2O3 surface. From these studies, it has become apparent that the acidity of Al2O3 develops
when it is dehydroxylated, not because of the OH groups themselves. Consequently, the
removal of water and/or hydroxyl groups (surface ligands), coordinatively unsaturated (CUS),
anions (oxygen ions) and cations (exposed Al3+, anion vacancies) are created.
Numerous IR studies reported that both tetrahedrally and octahedrally coordinated Mo6+ are
present in impregnated Mo/Al2O3 catalysts, and the octahedral/tetrahedral ratio increased with
increasing Mo loading. Infrared spectra of molybdena-alumina have also shown that alumina
surface hydroxyls are eliminated by the addition of molybdena [25-29]. These results
provided strong evidence of bonding between molybdena and the alumina surface. The
formation of tetrahedrally coordinated Mo6+ on the surface of alumina was pictured as the
replacement of the terminal hydroxyl groups as shown in the equation below [12, 30]:
15
(Eq-5)
Rédey and Millman independently measured the loss of OH groups on alumina in the
formation of molybdena-alumina catalyst by determination of the total hydrogen contents by
exchange with D2 using the isotope dilution method [3, 14]. The number of hydroxyls
eliminated in the formation of 8% Mo/Al2O3 catalyst (∆OH/Mo) was determined to be about
1.7 0.6, which is in agreement with the epitaxial monolayer model (Eq-5). ±
Interestingly, the hydroxyls, which are replaced by molybdena when the catalyst is formed
(calcined), reappear when the catalyst is partly reduced in H2.
The fragments arising from the dissociative adsorption of water on the surface of metal oxides
give rise to hydroxyl groups that are potentially more or less active Brönsted acid sites. Such
surface hydroxyl groups can be detected, directly, recording the IR spectra of the oxide
catalyst powders in the region 3800-3000 cm-1, where the O-H stretching modes (νOH's)
occur [12-24]. Although the position and shape of the ν(OH) bands of such surface hydroxyl
groups is informative on their coordination, these data do not give straightforward information
on their Brönsted acidity. In fact, as for example, the position of the OH band over a basic
catalyst like MgO, of a weakly acidic catalyst as amorphous silica and of a strong Brönsted
acidic catalyst like silica-alumina is almost the same (3745 ± 3 cm-1). Moreover, very acidic
catalysts like, for example, sulphated zirconia and titania do not present any definite sharp
ν(OH) band, while others, like zeolite ZSM-5 and silica-alumina, show sharp ν(OH) bands.
These facts are due to the following main reasons:
1. The ν(OH) frequency depends not only on the O-H bond strength, but also on the nature of
the M-O(H) bond, i.e. the element(s) to which the OH is bonded.
2. In any case, even for OH's bonded to the same element, the function ν(OH) versus acidity is
not necessarily linear but can present a maximum.
3. The state of the OH groups on the surface also depends on the basic strength of the oxide
ions. In fact, in very covalent structures, like for silica-alumina and zeolites, where oxygens
are almost not basic, the acidic OH's are responsible for rather sharp and well-defined bands,
while when the nearest oxygens are more or less basic, the acidic OH's give rise to
H-bonding, with a shift down and a broadening of the ν(OH) band [25-31].
16
1.2. Current status of catalysts preparation Different methods have been described in the literature for the preparation of these catalysts.
Among these methods, the impregnation (IW) and the equilibrium adsorption (EA) methods
have been widely used [1, 2, 3].
1.2.1. The impregnation method On most of the previous studies, the molybdena-alumina catalysts were prepared by the
impregnation method, which is the most widely used preparation technique for supported
heterogeneous catalysts. This procedure involves making a solution of the active component
having a volume equal to the pore volume of the support material [3-15].
The solution is then impregnated into the support material and mixed evenly where it is
completely taken up. The solid is first dried at about 150°C and then calcined at elevated
temperatures, in general, between 400 to 600°C (Fig. 4). Consequently, the resulting
preparations have varied somewhat depending upon the initial molybdenum compounds used,
the initial pH value of the molybdenum solution, the pore structure of the support and the
preparation skill. One obvious disadvantage of this preparation method has been that
molybdena cannot be uniformly distributed over the support surface.
Knözinger et al. studied molybdena-alumina catalysts prepared by impregnation at pH = 6 and
11 and concluded from Raman results that the catalyst prepared at pH = 6 contained a large
amount of bulk MoO3 while that prepared at pH = 11 did not [32]. Moreover, the same results
showed that this technique does not lead to uniform coverage. Accordingly, it can be argued
that this technique is not well defined. For instance, the pH value of the solution is not usually
specified when using this technique in addition to the pre-treatment conditions including
calcination temperature, time and ambient atmosphere. For example, it has been reported by
Hercules and co-workers that increased calcination temperature and time favour the formation
of Al2(MoO4)3, which is believed to be a sub-surface species, while bulk MoO3 was observed
at low calcination temperature and short time [33-35].
In summary, the final catalyst made by the impregnation method may be affected by the
solution pH value, the mesh size of the support used, and other variables such as calcination
temperature and time. Therefore, it is difficult to reproduce the catalyst by this method from
one batch to another.
17
Fig. 4. Production of supported molybdena catalysts by impregnation method
1.2.2. Adsorption methods These methods for preparation of molybdena-alumina catalysts were first used by Sonnemans
and Mars in both liquid and gas phases [36]. In the preparation from liquid phase a fresh
solution of 1% ammonium paramolybdate (pH 1-9) was flowed through an alumina bed for a
period of 4 hours (until the concentration of the effluent solution was found the same as that
of the entering reagent). The final Mo content was found to depend on the solution pH value.
It reached 21% in MoO3 when prepared at pH = 1, but decreased as the pH was increased. At
pH = 9, the catalyst contained only 2% Mo. They also reported that the concentration of the
solution had an effect on the amount adsorbed (the concentration range studied was in the
range between 0.2 and 1%). They concluded that Mo/Al2O3 with monolayer coverage could
be achieved by either gas phase or liquid phase adsorption method. The authors suggested that
the pH effect on the molybdenum content resulted from a change in the mean size of the
polymolybdate ions, which is a function of the pH.
In contrast, Iannibello and co-authors who showed the same pH dependency interpreted the
pH effect as due to a higher fraction of protonated alumina hydroxyls at lower pH leading to a
higher adsorption of anions [37]. They observed that the pH decreased rapidly at first and then
increased slowly to a steady value. The authors attributed the fast initial pH decrease to
18
exchange of ammonium ions with protons of alumina and the slow subsequent increase to
exchange of molybdate ions with surface hydroxyl groups as represented by Eq-6 as [MoO4]2-
was adsorbed, the equilibrium (Eq-7) would shift toward the right:
Al-OH + [MoO4]2- ⇔ Al-O-MoO3- + OH- (Eq-6)
[Mo7O24]6- + 4H2O ⇔ 7[MoO4]2- + 8H+ (Eq-7)
Actually, at pH values above 7 or 8, Mo4+ occurred as the tetrahedral monomeric molybdate
anions, but polymerization occurred at concentration in excess of 10-4 M at somewhat lower
pH values. The two major equilibriums [3, 12, 37], which may occur in the solution, may be
written as follows (Eqs. 7 and 8):
8MoO42- + 12H+ ⇔ Mo8O26
4- + 6H2O (Eq-8)
The major molybdate species at concentration above 10-3 M that are present at various pH
ranges can be roughly illustrated by the scheme below [11, 12]:
Scheme 3.
However, the well-understood principle of colloid chemistry can provide an explanation for
the variations in Mo loading with pH and the reversibility of the adsorption process. The
surface of a solid oxide in an aqueous solution is generally electrically charged. This charge
may be attributed to: (a) dissociation of surface hydroxyl groups or (b) adsorption of protons
formed by hydrolysis of H2O. These two mechanisms can qualitatively explain the pH
dependence of surface charge and the existence of a pH resulting in zero net charge, called the
isoelectric point of the solid (IEPS) or zero point of charge (ZPC) [38-40].
19
Based on the concept of IEPS, the surface chemistry of an oxide with an aqueous solution can
conveniently be expressed as written by Parfitt [41]:
M-OH2+ M-OH ⇔ ⇔ M-O– + H+ (Eq-9)
Decreasing pH IEPS Increasing pH ⇐ ⇒
The equilibrium indicates that at IEPS of the oxide, the net surface charge is zero, although
this does not necessarily mean that all the surface hydroxyls are in the form of M-OH because
they vary in acidity depending on their coordination. The net zero surface charge simply
means the concentrations of M-OH2+ and M-O– are equal at IEPS. When the pH is adjusted
lower than the IEPS, the concentration of M-OH2+ will be higher than M-O–, consequently the
surface will carry a net positive charge:
++ −⇔+− surfsurf MOHaqHMOH 2)( (Eq-10)
The situation is reversed when the solution pH is adjusted higher than the IEPS, and the oxide
surface will carry a net negative charge:
)(
2
aqHMOMOHor
OHMOOHMOH
surfsurf
surfsurf
+−
−−
+−⇔−
+−⇔+−
(Eq-11)
The same factors play an important role in the adsorption of metal oxyanions on the oxide
support. Consequently, the fact that the charge on the surface can be adjusted by varying the
solution pH. Thus, the amount adsorbed can be varied by changing the zeta potential of the
surface. Accordingly, the IEPS has a great influence on the band position of the adsorbed
species. The IEPS of several oxide supports is shown in Table 1.
20
Table 1. Comparison between IEPS of several oxide supports
Oxide supports Isoelectric point (IEPS)
SiO2 1-2
γ-Al2O3 6-8
TiO2 4.7-5 (Rutile), 5.7-6.2 (Anatase)
MgO 12.1-12.7
ZrO2 6.6-7.1
CeO2 (cerianite) 6.7-6.8
SnO2 (cassiterite) 4.2
J. Sarrin and co-workers compared the reducibility and activity of two series of molybdena-
alumina catalysts prepared by impregnation (IW) and adsorption methods (EA) [42]. Their
results showed that for both series of catalysts, the reducibility of the molybdenum species
increases as the Mo loading increases, in agreement with the literature [12-15]. For the
catalysts with similar loading, the (IW) series showed a higher surface coverage of the Mo
phase and higher degree of reduction than for the corresponding (EA) preparations. The
reducibility data were consistent with the catalytic results and oxygen chemisorption results.
The (IW) preparations (with similar Mo loading) were more active in the isomerization of
1-butene and chemisorbed larger amount of oxygen than their (EA) counterparts. The
differences in reducibility can be ascribed to a nonuniform repartition of the molybdenum
species between the external and internal surfaces of alumina for (IW) preparations, which
may contain a greater fraction of easily reducible polymeric Mo species than their more
uniform (EA) counterparts. Another possible explanation may stem from the decoration effect
of the Mo species by Al3+ ions. The latter may arise from the dissolution of alumina, which is
favoured on the (EA) series due to the long contact time between the Mo solution and
alumina.
Other studies indicate that the preparation method does not influence the molecular structure
of the Mo species present on Al2O3. Thus, for a given Mo loading, the nature of the Mo
species is independent of the preparation method [43-47]. However, it is not clear, how a
given preparation method may induce the Mo speciation.
21
1.3. Molybdena with CeO2 and SnO2 semiconductors Promoters are the subject of great interest in catalyst research due to their remarkable
influence on the activity, selectivity and stability of industrial catalysts. It is sometimes
difficult to define precisely the function of the promoters that has not been elucidated [1, 2].
Ni and Co promoters in Mo/Al2O3 catalysts are well-known for their success in the
hydrodesulphurization (HDS) of petroleum feedstock and coal liquefaction products [48]. The
promoting role of both promoters was found to increase the Mo dispersion and reduction, in
addition to the increase in H2 mobility, an intercalation effect with MoS2, a decrease in
deactivation and an increase in surface segregation of mixed sulphide phases. For instance,
when Co was introduced into Mo/Al2O3, various effects occurred. Free MoO3, as well as
Al2(MoO4)3, was converted into CoMoO4, and the Co addition resulted in a decrease of the
isolated Mo tetrahedral concentration and favoured the formation of the polymeric form.
Moreover, Topsøe et al. reported direct evidence of Co and Mo existing in the active form as
a Co-Mo-S surface phase [49, 50].
The purpose of doping a semi conductive carrier in order to enhance the catalytic activity of
supported metal catalysts has recently been applied in developing “three-way” catalysts [51].
The conductivity of semiconductors is generally low but can be considerably changed by
either incorporating with other oxides or upon pre-treatments. Their crystal lattices tend to
release or take up oxygen. Therefore, alloying the metals with semiconductors can increase or
decrease the activity. This effect has some industrial relevance since can both accelerate
desired reactions and suppress undesired reactions [52, 53]. For instance, the addition of Sn to
Pd gives selective catalysts for the removal of acetylene from ethylene streams [54].
In the case of n-type semiconductors (e.g. ZnO, TiO2, CeO2, SnO2), a pre-treatment in a
reducing atmosphere generates electron-donor levels (oxygen vacancies VO, metal under a
lower oxidation state), which increases the free electron concentration [55-57].
The presence of electron-donor levels gives rise to electronic transitions, which may occur in
the infrared region:
M(n−1)+ → Mn+ + e− (Eq-12)
VO → VO+ + e− (Eq-13)
VO+ → VO
2+ + e− (Eq-14) For instance, on heating or by reaction with reducing gases such as H2, CO for n-type
semiconductors such as CeO2, the release of oxygen can be described as below:
22
CeO2 ↔ CeO2-x + ½ xO2(g) (Eq-15)
Oox ↔ Vo + ½ O2(g) (Eq-16)
Ce4+ + Vo ↔ Ce3+ + Vo+ (Eq-17)
Where VO, VO
+ and VO2+ are the neutral and ionized oxygen vacancies, respectively, OO
x is
the lattice oxygen [2].
On the other hand, the adsorption of oxygen on a nonstoichiometric oxide, containing oxygen
vacancies VO, generates lattice oxygen OOx. At the same time, metal ions are oxidized at the
surface and the conductivity is lowered for n-type semiconductors because the oxygen acts as
an electron acceptor:
½ O2(g) + Vo ↔ Oox (Eq-18)
Ce3+ + O2 ↔ Ce4+ + O2– (Eq-19)
The adsorption of oxygen results eventually in complete coverage of the surface by O– or O2-
and the heat of the adsorption remains practically constant while the surface becomes
saturated with oxygen and negatively polarized.
Considering the adsorption of hydrogen on n-type semiconductors, it has been shown that H2
mainly undergoes heterolytic dissociation [2, 58-60]:
M2+ + O2- + H2 → M+ H + OH– (Eq-20)
On heating, the hydroxyl ions are decomposed to water and anionic defects, and a
corresponding number of metal cations are reduced to atoms. In this strong chemisorption, a
free electron or positive hole from the lattice is involved in the chemisorptive bonding. This
changes the electrical charge of the adsorption center, which can then transfer its charge to the
adsorbed molecule. Thus, chemisorbed hydrogen acts as an electron donor and increases the
conductivity of n-type semiconductors. Furthermore, the change in the electrical charge
density on the surface can hinder the further adsorption of the same gas. A decrease in the
heat of adsorption with increasing degree of coverage is then observed, and hence a deviation
from Langmuir adsorption isotherm occurs [2, 61-65].
However, when a metal is applied to an appropriate n-type semiconductor, its electron density
increases [1, 2]. The general behaviour of some nonstoichiometric semiconductor oxides is
summarized in Table 2.
23
Table 2. Behaviour of nonstoichiometric semiconductor oxides
n-type p-type
For instance TiO2, SnO2, CeO2 NiO, CoO, FeO
Type of conductivity Electrons Positive holes
Addition of M21+ O oxides Lowers conductivity Increases conductivity
Addition of M23+O3 oxides Increases conductivity Lowers conductivity
Adsorption of O2, N2O Lowers conductivity Increases conductivity
Adsorption of H2, CO Increases conductivity Lowers conductivity The optical absorption of semiconducting oxides arises from five different phenomena: (i)
intrinsic absorption, corresponding to transitions between (full) valence bands and (empty)
conduction bands, which occur often in the UV–visible range and sometimes in the near-
infrared (NIR) (for narrow gap semiconductors); (ii) transitions between valence bands, called
intervalence transitions, only observed in p-type materials, which may appear in the NIR; (iii)
free carrier absorption, arising from transitions within one band; (iv) transitions of an
electron to or from a localized state; (v) lattice vibrational absorption. Their mid-infrared
examination offers special difficulties due to mainly the transition types (iii) and (iv), which
involve the absorbance due to free carriers and electron- or hole-donors, whose concentration
depends on the semiconduction type, the surrounding atmosphere and the temperature [62].
These difficulties are seriously enhanced when the sample under study is a metal supported on
an n-type semiconducting support, the reduction of the support is then greatly favoured by the
metal, e.g. through activation in vacuum and spill over of hydrogen or CO. For example, this
has been observed in the case of metals supported on ZnO and ceria [63-65].
Lanthanide ions of variable valence particularly Ce3+/4+ usually lead to nonstoichiometric
CeO2-x. It has been reported earlier that the latter aspect and the defect structure on ceria was
due to oxygen vacancies accompanied by triply and/or quadruply Ce interstitial to maintain
the electrical neutrality. However, lately, oxygen vacancies have been finally affirmed as the
prevailing defects neglecting the negligible effect of Ce interstitials in such a fluorite-
structured oxide system [66-72]. The availability of these defect sites on the surface is
probably related to their high bulk concentration. The oxygen anions (O2-) on ceria surface
may be one, two or three coordinated to cerium cations. From what has been conferred for
CeO2 in addition to its role as either oxygen storage and release or thermal stabilizer. It has
been used either as a promoter or as a support for metal catalysts in many applications since
24
ceria has a beneficial effect for CO oxidation and NOx reduction under both stoichiometry
and excess oxygen beside for CO/NO reaction, CO and CO2 hydrogenation. Since the
catalytic oxidation of CO has acquired tremendous attraction lately particularly in connection
with world-wide endeavors to curb the detrimental impacts of automotive emissions on the
atmosphere. Although the detailed mechanism of the reactions mentioned above is still
unknown, the researchers clearly assigned the promotional effect of the catalysts to the role of
ceria in creating Ce3+/4+ redox couple [73-75]. However, due to the limited supply of precious
metals and some impractical properties, an attention has been given to transition metals and
their oxides as catalysts supported on CeO2 or doped with CeO2, since the ability of ceria to
donate oxygen to supported metals is also a key feature in other catalytic reactions like for
example, catalytic combustion and water gas shift reaction [76-78].
M. Mokhtar investigated the influence of ceria on Mo/TiO2 and found that the presence of
ceria leads to increase the concentration of polymerized surface Mo oxide species, and rather
initiated the formation of MoO3 over-layers. Additionally, the involvement of ceria, on the
other hand, retarded the strong association rendered between Mo and Ti and thus stimulated
the formation of discrete amounts of the corresponding oxides. More specifically, ceria was
found to work as a mediator between Mo and Ti [79, 80].
Lucia and co-workers found that the presence of cerium in the Mo-Sn system increases the
rate of ethanol dehydrogenation as well as the selectivity to acetic acid and acetaldehyde. In
addition, it caused changes in the distribution of Mo species and in the textural properties, but
mostly increasing the basicity of the catalyst [81].
Stannic oxide thin films are attractive for many applications due to their unique physical
properties such as high electrical conductivity, high transparency in the visible part of
spectrum, and high reflectivity in the IR region. In particular, tin oxide films are stable at high
temperatures, have excellent resistance to strong acids and bases at room temperature, are
resistant to mechanical wear, and have very good adhesion to many substrates [82-85]. Thus,
transparent and electrically conductive stannic oxide films are widely used for a variety of
applications. Briefly, these applications include: as electrodes in electroluminescent displays,
imaging devices, protective coatings, antireflection coatings, gas and chemical sensors,
transducers applications based on transparent conductors and other optoelectronic devices.
Furthermore, tin oxide films are more stable than the other transparent conducting oxide
(TCO) films such as zinc oxide (ZnO). Moreover, they have a lower material cost. Recently,
the synthesis of ultra fine tin oxide particles is of great technological and scientific interest
25
owing to their superior physical and chemical properties and their use as either catalysts for
the oxidation of organic compounds or gas sensors [83-86].
As the electrical conductivity of SnO2 derived from the variable valence on the Sn atomic
center is very sensitive to oxidative and reducing atmospheres, tin oxides as gas sensors
detecting a trace amount of the gases have been applied to processes in chemical, heretical
and fermentation industries to control the amount of the harmful wastes discharged from the
plants, the explosion of the combustible gases and incomplete combustion, exhaust gases
from automobiles [87-90]. However, molybdenum–tin thin films seems to be promising gas
sensors. It has recently been stated that addition of MoO3 to SnO2 increases the sensor
response to CO and NO2 [91-94].
All the above properties have led to intense research of SnO2 coatings over the past few
decades. Currently, numerous techniques exist for the preparation of tin oxide films such as
chemical vapor deposition, spray pyrolysis, sputtering, and sol–gel deposition [95-98].
Nevertheless, the ability of SnO2 to generate defects has been only recently shown to induce
interesting performances for supported Pd catalysts, e.g. in deNOx reactions [99]. Conversely,
tin dioxide has received limited attention in the catalysis field and the use of Mo-Sn oxides in
selective oxidation appears to be unique industrial application [100-109].
On the other hand, Mo/SnO2 catalysts have been used for selective oxidation reactions due to
their high activity. Niwa et al. [100] studying the methanol oxidation with several supported
molybdenum catalysts found the following sequence of activity:
Mo/SnO2 > Mo/Fe2O3 > Mo/ZrO2 > Mo/TiO2 > Mo/Al2O3
Goncalves et al. [101] and Medeiros et al. [102] have shown that acetic acid can be obtained
from ethanol oxidation in only one-step with high yield when Mo/SnO2 catalysts prepared by
precipitation procedure are used.
Recently, Liu et al. [103] have shown that Mo/SnO2 catalysts are very active for the oxidation
of dimethyl ether although they are more selective for formaldehyde than Mo/Al2O3 catalyst.
V. Lochar claimed that the activity of MoO3/SnO2 catalyst for methanol oxidation could be
associated with its Brönsted and Lewis acidity as the result of the catalyst reduction [104].
Other catalytic activity results suggest the existence of synergy between the apparently pure
phases of MoO3 and SnO2. Therefore, MoO3 in close contact with SnO2 has shown to be
much more active and selective than the individual pure phases. The high dispersion of
26
molybdenum species on the highly reducible SnO2 support was suggested to be responsible
for the exceptional activity of these catalysts [105-109].
However, the information available in the literature on the interpretation of infrared and
Raman spectra of Mo/Sn and Mo/Ce compounds prepared on different surfaces is rather
limited. In assigning vibrational spectra, some DFT (Discrete Fourier Transform) calculations
or some vibrational spectroscopic data relating to Mo/Sn and Mo/Ce systems can be relied on
[65, 80, 88, 91, 97, 104-109]. On the other hand, few papers can be found in the literature on
IR and TG studies related to either Mo/Sn or Mo/Ce system in contrast to publications
relating to noble metals with ceria and tin.
Anyhow, a lot of debates concerning the role played by ceria and tin oxides necessitate further
studies in order to explore the influence of CeO2 and SnO2 on the structure and surface
characteristics of molybdena for better understanding the nature, structure and the physico-
chemical properties of these oxides, since the nature of the interactions between metal oxides
and supports are often attributed to the complexity of these systems and differences in the
preparation and experimental conditions adopted.
1.4. Direct conversion of methane under nonoxidative conditions An important task confronting catalytic chemists is how to realize direct conversion of
methane to versatile fuels and valuable chemicals by building up the desired C–C (or C–O)
bond. Thermodynamic constraints on the reactions in which all four C–H bonds of CH4 are
totally destroyed, such as CH4 reforming into synthesis gas or CH4 decomposition into carbon
and hydrogen, are much easier to overcome than the reactions in which only one or two of the
C–H bonds are broken under either oxidative or nonoxidative conditions [110-117]. Direct
conversion of CH4 with the assistance of oxidants is thermodynamically more favourable than
that under nonoxidative conditions. Therefore, the direct conversion of CH4 under the aid of
oxidants has received much more attention than that under nonoxidative conditions, especially
when considering the production of fuels and valuable chemicals from CH4 [118-123].
With the urge to quest for renewable energy and cleaner fuels, it is recognized that hydrogen
energy will inevitably replace fossil fuel energy in the near future due to the fact that the
burning of hydrogen is pollution free. However, it is a practical way to produce H2 from CH4
due to its high H/C atomic ratio and great abundance in reserves. Therefore, the direct
conversion of CH4 under nonoxidative conditions into H2 and/or H2 accompanied with basic
chemicals is closely related to the effective utilization of CH4-containing resources and thus to
sustainable progress and development of the living conditions of humankind [124-127].
27
The direct conversion of CH4 under nonoxidative conditions is thermodynamically
unfavorable. Nevertheless, as an alternative approach, it has still attracted the attention of
many researchers. In heterogeneous catalysis, various metals have been discovered that can
chemisorb CH4 at moderate temperatures and that can decompose CH4 to C and H2 at higher
temperatures [128-137].
Amariglio and co-workers reported a “two-step” process on Pt, Ru, and Co in isothermal
experiments [128-130]. In a series of publications, the authors suggested that C–C bonding
could take place between H-deficient and CHx formed during the first step of methane
chemisorption, while H2 saturated the alkane precursors in the second step and removed them
from the surface. In view of the fact that hydrogenation at a temperature lower than that of
CH4 chemisorption is favorable for lessening hydrogenolysis. The authors reported a
nonoxidative conversion of methane to higher hydrocarbons through a dual temperature two-
step reaction on Pt/SiO2 and Ru/SiO2 catalysts. Indeed, when chemisorption of methane was
set at a fixed temperature (usually lower than 320°C), the selectivity to heavier alkanes
increased with the lowering of hydrogenation temperature on both catalysts. On the other
hand, when the hydrogenation temperature was less than 120°C, hydrogenolysis was
negligible, and thus the variations of the products can only be attributed to the changes
affected by the adlayer formed during the chemisorption of methane at a certain temperature.
It was discovered that the products of C2+ hydrocarbons at every hydrogenation temperature
displayed a maximum versus the methane chemisorption temperature on both catalysts. In the
case of the Pt/SiO2 catalyst, mainly C2H6 and n-C5H12 were produced during the first minute
of the reaction. This illustrates that C–C bonds could form during CH4 adsorption, and the
authors assumed a surface intermediate of C5 precursor bonded on dispersed and coordinately
unsaturated Pt atoms.
Van Santon et al. suggested that CH4 first dissociated on a precious metal to form carbide and
H2. Then, the carbide was hydrogenated by H2 to produce higher hydrocarbons. C–C bonds
were supposed to be created during the hydrogenation step. Since the reactivity of the CHx
surface intermediates formed from CO and CH4 was quite similar. The authors suggested that
the chain-growth probability would depend on the metal–carbon bond strength and that the
mechanism of C–C bond formation in the two-step route should be related to that occurring in
the Fisher–Tropsch reaction. They also demonstrated that the homologation of olefins (C2H4,
C3H6, etc.) with methane could occur over Ru/SiO2 and Co/SiO2 catalysts [131, 132].
28
The two-step route is also feasible over a number of oxide- or zeolite-supported transition
metal and bimetal catalysts. Solymosi and Cserenyi illustrated that over a Cu-promoted
Rh/SiO2 catalyst, the enhanced formation of C2H6 and higher hydrocarbons could be observed
in the two-step process [133, 134].
Guczi et al. reported that the chemisorption of CH4 at 250°C and the subsequent
hydrogenation of the CHx species at 250°C over Co–Pt/NaY and Co–Pt/Al2O3 performed the
best of all the catalysts tested. The chemisorbed CHx species had the highest concentration,
and all CHx species were hydrogenated in the second step, giving a selectivity of C2+ close to
84%. They found that there was a correlation between the hydrogen content of the surface
CHx species (the optimum value of x being around 2) and the chain length of the
hydrocarbons produced in the hydrogenation step in their mechanistic study of the two-step
process [135]. Later, they reported that the two-step process could be simplified into a one-
step process with a C2+ hydrocarbons production higher than that obtained in the two-step
process over Co–Pt/NaY bimetallic catalyst. These results could be obtained if the CH4 was
pulsed with H2/He mixture at 250°C [136].
Bradford reported the results of the isothermal, nonoxidative, two-step conversion of CH4 to
C2+ hydrocarbons over supported and unsupported Pt and Ru catalysts at moderate
temperatures and elevated pressures. It was shown that an increase in reaction pressure
increased the branching and molecular weight distribution of the product [137].
Several researchers suggested the preparation of a multifunctional catalyst to avoid the use of
a two-step process. Furthermore, it has been reported that dehydrogenative coupling of CH4
without any oxidant could be carried out over Pt–SO4/ZrO2 catalysts. A steady conversion of
0.2% (the equilibrium conversion of CH4 into C2H6 and H2 is estimated to be 0.6%) was
observed after the catalyst was reduced in H2 at 500°C [138, 139].
On the other hand, in order to overcome the thermodynamic limit and to enhance the
reactivity for obtaining high yields in direct conversion of CH4 under nonoxidative conditions,
plasma excitation has also been attempted. The product distribution is dependent on the
method by which plasma excitation is produced. For example, in pulsed corona discharges at
atmospheric pressure, C2 hydrocarbons (mainly C2H2) were obtained with a high selectivity of
around 70 to 90%. In microwave plasmas, the product distribution shifted from C2H6 to C2H4
and finally to C2H2 with an increase in power density. By introducing a proper catalyst into
the microwave plasma reactor, CH4 could be converted to higher hydrocarbons at atmospheric
pressure. In addition, with a CH4 and H2 mixture as the feed gas, the selectivity to C2H2 was
88% and that to C2H4 was 6% at a CH4 conversion of 76% [140, 141]. Here, again, the main
29
drawback is the low energy efficiency to drive this thermodynamically unfavorable reaction.
Thermodynamically, the transformation of CH4 under nonoxidative conditions is more
favorable to aromatics than to olefins. The direct conversion of CH4 to aromatics was tested
on several catalysts in either a pulse or a flow reactor. Wang et al. reported on the
dehydroaromatization of methane (MDA) for the formation of aromatics (mainly C6H6) and
H2 under a nonoxidizing condition in a continuous flow reactor on Mo/HZSM-5 catalysts
[142]. More detailed studies on the reaction revealed that the channel structure and acidity of
the HZSM-5 zeolite, as well as the valence and location of the Mo species, are crucial factors
for the catalytic performance of the Mo/HZSM-5 catalysts. In addition, W/HZSM-5 and
Re/HZSM-5 are also reported to be active elements for MDA [142-144].
Solymosi and co-workers [147-152] and Lunsford and co-workers [153-155] characterized
the Mo/HZSM-5 catalyst by means of XPS and found that during the initial induction period,
the original Mo6+ ions in the zeolite were reduced by CH4 to Mo2C, accompanied by the
depositing of carbonaceous cokes. They suggested that Mo2C provides active sites for C2H4
formation from CH4, while the acidic sites catalyze the subsequent conversion to C6H6. The
Mo2C species probably are highly dispersed on the outer surface, and some of them reside in
the channels of the zeolite. Meanwhile, the spectra of Mo/HZSM-5 samples reacted with CH4
at 700°C for one and 24 hrs were basically identical to the Mo2C reference spectrum, except
for a partial contribution from the Mo oxide, given rise by MoOxCy. These authors claimed
that the Mo oxide species dispersed in the HZSM-5 framework might migrate onto the
external surface of the HZSM-5, be converted by CH4 to Mo2C, and disperse on the support
surface. Therefore, the carbonaceous deposits created in MDA are in various forms and play
different roles. First, Mo2C and/or MoOxCy, which are possibly active species for CH4
activation, are formed during the induction period. Second, the formation of the active
intermediates, the CHx species, follows the activation of CH4 on Mo2C and/or MoOxCy. The
last one to be formed is coke leading to the deactivation of the catalyst. It is understandable
that there are some similarities between the carbonaceous species formed in MDA and those
formed in the first step of the two-step process, since both reactions are carried out under
nonoxidative conditions, and Mo2C shows some precious metal-like properties.
In spite of the fact that the reaction is thermodynamically unfavorable under pressurized
conditions and that 10% CO2 added to the feed totally suppresses the activity of the 2 wt%
Mo/HZSM-5 catalyst. Ichikawa and co-workers found that an increase in CH4 pressure and
the addition of small amounts of CO and CO2 (less than 3%) to the CH4 feed enhanced the
catalyst stability in the reaction [156, 157]. By increasing the CH4 pressure, the formation
30
rates of C6H6 and hydrocarbons could be moderately increased. This kind of pressure
relationship may be related to a sufficient supply of H2 from CH4 and a suitable concentration
of surface carbon species CHx for the formation of aromatic products. By using a CO and CH4
mixture as the feed to conduct the reaction, the authors suggested that CO dissociated on the
Mo sites to form the active carbon species CHx. The dissociated oxygen species [O] from CO
might react with the surface inert carbon species to regenerate CO, resulting in the
suppression of coke formation on the catalyst. These results imply that although the Brönsted
acid sites are necessary, excess Brönsted acid sites are detrimental for the reaction, since
severe coke formation will occur on them [157-159].
Considerable efforts have been devoted to developing active and selective catalysts and
understanding the bifunctionality of Mo/HZSM-5 catalysts and the nature of carbonaceous
deposits formed during the reaction. However, neither new active and selective catalysts nor a
thorough understanding of the mechanism of the reaction has been achieved.
However, despite all substantial research efforts into nonoxidative two-step or one-step CH4
homologation, its low efficiency is the main problem to further developing it as a commercial
process. In any case, these studies enhanced our knowledge in direct conversion of CH4 under
nonoxidative conditions, particularly methane dehydroaromatization, and stimulated chemists
to explore new methane conversion processes.
1.5. Characterization of the catalysts Both of the physical and the chemical properties of a catalyst must be known if relationships
between the structure and activity, selectivity, and lifetime are to be revealed. There are many
techniques commercially available for the analysis of catalysts opening up new possibilities
for fundamental catalyst research. In this section, I will encounter some methods for
characterizing catalysts and not discuss their capabilities and limitations that have been
described in the literature [1, 2].
31
1.5.1. Physical characterization The significance of the physical surface is based on the fact that the bonding forces of the
atoms in the outer most atomic layers are not saturated. Thus, these atoms experience
structural reorganization (relaxation, reconstruction) and a high chemical reactivity. On the
other hand, the thermodynamic properties of these layers are different from the bulk.
The distribution of pores across the inner and outer surfaces is an important property of the
catalyst. The texture generally refers to the pore structure of the particles including the pore
size distribution, pore volume and pore shape which are determined by gas adsorption
(usually N2, BET method) at relatively low pressures (low values of p/po = pressure/saturation
pressure) for microporous materials [1, 2].
Although the specific surface area is one of the most important parameters of catalysts and
can be determined by the multipoint BET method. However, there is no direct relationship
between catalyst activity and the surface area. Such predictions can only be made by
chemisorption of appropriate gases such as H2, O2, CO, NO at room temperature or above,
respectively [2-4].
However, additional physical techniques can be very useful such as X-ray diffraction (XRD)
that is very essential to determine the crystalline structure and crystallite sizes of the catalysts
despite its limitation to detect particles smaller than about 2 nm. Therefore, the missing of any
discernable diffraction lines does not necessarily prove that the phase in question is absent [1].
Additionally, thermal analysis is an indispensable tool to follow completely the thermal
behaviour of the catalysts, the changes in the composition of the catalysts should be followed
as a function of the temperature, as well. The use of thermogravimetric/differential
thermogravimetric analysis (TGA/DTA) combined with different types of spectroscopic
techniques is essential to reveal the thermal stability of the catalysts. The thermal analysis
combined with mass spectrometry (TG-MS) of different mixed oxide systems can provide
better understanding of the removal and formation of different components in the catalysts.
Thus, knowing the thermal history of the catalysts, tailoring the thermal properties of prepared
catalysts can be facilitated [83, 160-165].
On the other hand, electron paramagnetic resonance (EPR) or electron spin resonance (ESR)
is a technique for studying chemical species that have one or more unpaired electrons, such as
organic and inorganic free radicals or inorganic complexes possessing a transition metal ion.
The basic physical concepts of EPR are analogous to those of nuclear magnetic resonance
(NMR), but it is electron spins that are excited instead of spins of atomic nuclei.
32
Because most stable molecules have all their electrons paired, the EPR technique is less
widely used than NMR. However, this limitation to paramagnetic species also means that the
EPR technique is one of great specificity, since ordinary chemical solvents and matrices do
not give rise to EPR spectra [166-168].
The EPR technique depends on the fact that certain atomic systems have a permanent
magnetic moment. The energy levels of the magnetic system are influenced by the
surrounding atoms and by external magnetic fields. Transitions among the levels can be
detected by monitoring the power absorbed from an alternating magnetic field, just as
ordinary atomic transitions are detected by absorption of light. Comparing the observed
transitions with model calculations then lets us deduce some features of the environment
around the moment [169, 170].
There are a number of ways for condensed matter to retain some magnetic moments, the most
important of which involve certain unusual molecules, transition-group atoms, or particular
point defects in solids. Molecular NO and NO2 both have an odd number of electrons and
hence a permanent magnetic moment. Similarly, many large molecules can exist with an odd
number of electrons. Completing this group, the ground state of O2 happens to be a partially
filled shell with corresponding moment. Transition-group atoms have incomplete 3d, 4d, 5d,
4f or 5f shells. Bonding of these atoms often involves higher-energy p or s electrons, leaving
the unpaired d or f electrons relatively undisturbed. When this occurs, the atom or ion retains
nearly the full atomic moment. Finally, certain defects such as vacancies or foreign atoms in a
crystal may gain or lose an electron relative to the chemically bonded host, thereby producing
a localized moment. The Hamiltonian for the ion is then the sum of several terms [171]:
(Eq-21)
The first term H0 is the usual free-atom Hamiltonian, except for two parts, which are written
explicitly. The spin-orbit coupling is the second term, while the (AI S) term describes the
"hyperfine" coupling of the electronic spin to the nuclear spin I. We use Hcf to represent the
electrical interaction of the paramagnetic species with the neighbouring atoms, including
effects due to bonding. In the simplest approximation, the interaction can be thought of as due
to point charges at the surrounding host sites, hence the common name "crystal field". By
relating Hcf to the observed EPR spectrum we hope to learn something about the surroundings
33
of the ion. The last term is the Zeeman interaction. In principle, we also should include the
nuclear Zeeman, but in such cases, it is too weak to be of concern [172].
When the hyperfine interaction is present, we must solve a slightly more complicated
problem, since the electronic energy will depend on the orientation of the nucleus as well as
the applied field. For instance, for 3d ions, the spin-orbit interaction is weak, but the crystal
field strength can be comparable to the electron-electron interaction contained in H0. We must
usually, therefore, include Hcf from the beginning and then treat the spin-orbit, hyperfine and
Zeeman terms as perturbations [170-173].
This technique is very sensitive, the detection of very small quantities of paramagnetic
substances (10-6 g) being possible, and it can even distinguish between isolated and
clusterized paramagnetic centers (agglomeration of paramagnetic centers in certain portions).
Furthermore, according to existing literature the purity of the sample can be measured by
assigning the corresponding signals to different ionic impurities in the solid sample [174].
1.5.2. Surface characterization and IR spectroscopy Generally, the features of a surface determine the properties of a material and therefore must
be characterized:
• Topology, morphology
• Elemental composition
• Chemical bonding of elements
• Structure (geometric and electronic)
Since high information content, high spatial resolution and high absolute and relative
detection power are required, that can only be met by physical techniques based on the
interaction of photons, electrons, ions and electrical fields with the material investigated [2].
Complete understanding of catalytic reactions mechanisms, including the nature of adsorbed
intermediates is highly desirable. However, as such should reasonably be expected to provide
major assistance in reaching the goals of better catalysts and improved catalytic processes
from a better fundamental understanding of catalyst surface chemistry. This is an area in
which infrared (IR) spectroscopy undoubtedly makes further major contributions. A variety of
IR techniques can be used to obtain information on the surface chemistry of different solids.
Special meaning have investigations carried out under the reaction conditions. This includes
spectral characteristics of reaction components, surface changes due to temperature treatment
and many others. In principle, all forms of IR spectroscopy, including transmission-
34
absorption, diffuse reflectance (DRIFT), ATR (attenuates total reflection) and photo acoustic
spectroscopy (PAS), are suitable for in situ measurements. The principal information obtained
with all these techniques is equivalent and local availability and experimental necessities such
as the sample particle size and the molecular extinction coefficient of the sample may
dominate personal choices. For most practical experimental reasons the vast majority of
experiments are currently performed in the transmission-absorption and the diffuse
reflectance mode. This is more related to the design of cells to be used as reactor than to the
principal problems of the other techniques. The IR cell in which the catalyst sample is pre-
treated and subsequently studied is extremely important in surface studies [175, 176].
The cell is normally chosen to suit the purposes of a particular study. Some features are
usually of overriding importance in a given application. Various complex schemes have been
designed to seal the reactor cell with IR-transparent windows so that the IR cell can be
operated at elevated temperatures and pressures. The development of high temperature and
pressure IR cells has permitted the observation of adsorbates under reaction conditions. These
cells may serve as a differential reactor for steady-state reaction, temperature-programmed
reaction or desorption, and unsteady-state reaction studies. Therefore, the IR cell suitable for
investigations of catalyzed reactions must fulfil two requirements: (a) it must allow the
recording of IR spectra under in situ reaction conditions, and (b) its volume and construction
must assure good mixing the gases inside, and the feasible space velocities must allow
flexible variations of the conversion exposing the catalyst to the reactants and products that
can be analysed precisely at the exit of the reactor cell [176].
However, the use of the infrared spectroscopy in heterogeneous catalysis can be classified: (1)
the determination of catalyst bulk structures, (2) identification of adsorbed species and surface
active sites, (3) characterization of surface hydroxyl groups (including Brönsted sites) and (4)
examining the catalyst surface structures (Fig. 5).
35
Fig. 5. Applications of IR spectroscopy in catalysis and surface science
From the discussions above and studies of others, IR can provide information concerning the
metal-oxygen vibrations in the region below 1000 cm-1 [177]. Therefore, the metal-oxygen
vibrational frequencies can be roughly divided into five characteristic ranges as shown in the
table below:
Table 3. IR characteristics of metal-oxygen bonds
Vibrational mode Wavenumber (cm-1)
Symmetric and asymmetric stretches of M=O 900-1100
Asymmetric stretches of M-O-M 700-900
Symmetric stretches of M-O-M 500-700
Bending vibrations of M=O 310-400
Deformation vibrations of M-O-M ~ 200
36
Diffuse reflectance spectroscopy (DRIFT) The optical phenomenon known as diffuse reflectance is commonly used in the UV–Vis, NIR,
and MIR regions to obtain molecular spectroscopic information [175-177]. When it is applied
in MIR area with a Fourier transform it is known as diffuse reflectance infrared Fourier
transform spectroscopy (DRIFTS). It is usually used to obtain spectra of powders with
minimum sample preparation. The collection and analysis of surface-reflected
electromagnetic radiation as a function of frequency or wavelength obtain a reflectance
spectrum. Two different types of reflection can occur: regular or specular reflection usually
associated with reflection from smooth, polished surfaces like mirrors, and diffuse reflection
associated with reflection from so-called mat or dull surfaces textured-like powders. In diffuse
reflectance spectroscopy, electromagnetic radiation reflected from dull surfaces is collected
and analysed. If a sample to be analysed is not shiny, and whatever reason is not amenable to
conventional transmission spectroscopy, diffuse reflectance spectroscopy is a logical
alternative. The advent of FTIR spectrometers has led to the widespread application of DRIFT
becoming a valuable technique since hardly any sample preparation is necessary. This implies
that the DRIFT technique is potentially of great value for in situ studies of catalytic systems.
One of the interesting advantages of DRIFT is the prevention of typical transmission
problems at high wavenumbers (due to scattering) and at low wavenumbers (due to strong
absorption of catalyst carriers). Another advantage is the high sensitivity of DRIFT.
Furthermore, the catalyst powder need not be compressed to obtain high quality spectra. This
improves the reproducibility and the activation of the catalyst. Nevertheless, there is a limited
number of commercially available, heatable and evacuable DRIFT cells in combination with a
specially designed optical system that is suitable for the in situ activation and pre-treatment of
catalyst samples at high temperatures and pressures [176-178].
1.5.3. FTIR Spectroscopic detection of CO adsorption on the catalysts As already cited, coordinatively unsaturated cations exposed on the surface of ionic oxides
give rise to surface Lewis acid sites. Consequently, basic molecules can interact with these
sites by forming a new coordination bond, so completing or increasing the overall
coordination at the surface cation. The stronger are the polarizing power of the Lewis acidic
cation (charge to ionic radius ratio) and the basic strength of the adsorbate, the stronger is the
Lewis interaction. Upon this interaction, electrons flow from the basic molecules towards the
catalyst surface. These electronic perturbations as well as the molecular symmetry lowering
37
arising from this contact are the causes of a vibrational perturbation of the adsorbate. In most
cases, the vibrational perturbation only consists in shifts of the vibrational frequencies, the
more pronounced, the stronger is the interaction, i.e. the greater is the Lewis strength of the
surface site. Accordingly, the shift of the position of some very sensitive bands of the
adsorbate upon adsorption can be taken as a measure of the Lewis acid strength of the surface
sites and the decrease in frequency has been associated with increasing acid strength [175].
As the type of probe molecule chosen will influence the obtained characteristics of the probed
solid and, hence, will also affect the structure–activity relationship derived, the choice of the
appropriate molecule is very important. On the use of carbon and nitrogen monoxides as
probes for the surface cationic centers it is evident that the carbon and nitrogen monoxides are
very weak bases and largely used for the surface characterization of cationic centers on metal
oxide surfaces [176-179].
CO implies a triple bond between C and O, according to the literature, the stretching
frequency for the free molecule in the CO gas is measured at 2143 cm-1 (see the roto-
vibrational absorption band in Fig. 6). In principle, the electrons are distributed symmetrically
between C and O atoms, so that the lower positive charge of the C nucleus with respect to O
implies the formation of a dipole with the negative charge at the C atom in spite of the lower
electronegativity of C with respect to the O nucleus. For this reason, the CO molecule tends to
interact through the C end with cationic centers. This interaction is rather weak, usually
completely reversible by outgassing at room temperature and should be studied at room or
lower temperature (e.g. at liquid nitrogen, 77 K) [177].
According to theoretical calculations, the metal-CO interaction is a simple polarization, with
no formation of a true coordinative σ-bond with the cationic center. This interaction tends to
increase the CO bond order, so that the CO stretching frequency tends to increase upon it.
Accordingly, the experimental measure of the CO stretching frequency for CO interacting
with surface cations can be taken as a measure of the polarizing power of the cation or, in
other terms, of its Lewis acidity. However, when the cation or the metal atom contains,
besides empty orbitals, also full or partly filled d-type orbitals, they can interact with the
empty π*- type orbitals of CO, via a π-type electron backdonation from the metal to CO. This
implies that these antibonding orbitals become partly filled so that the bond order and the CO
stretching frequency are decreased by this last interaction [178-181]. In this case, the
interaction can become very strong and very stable metal-carbonyl complexes (carbonates)
can be formed. The experimental CO stretching frequency in this case is a complex function
of the electron accepting power of the cation (Lewis acidity) and of its π-type electron
38
donating power. Accordingly, the CO stretching frequency of CO adsorbed on several metal
cations is very informative on the oxidation state of the adsorbing ion. The associated process
can be easily followed by IR spectroscopy [179-182].
In fact, acidic OH groups form very weak hydrogen bonded adducts with both CO and H2
probes and the shift of the stretching frequency induced by the perturbation is proportional to
the charge present on the hydrogen (and hence indirectly upon its Brönsted acidity). On the
contrary, basic OH groups do not show any tendency to interact with CO and H2 [183-187].
Yet, a few have been published on IR study of CO adsorption on Mo-containing catalysts in
which the Mo loading was mostly 8-10 %. For instance, Zs. Németh undertook a low
temperature IR study at -196°C of the interaction between Mo/Al2O3 (8 %) and CO within the
extent of reduction from 500°C up to 900°C. A correlation was found between the extent of
reduction and the increase of adsorbed CO species. Furthermore, the five bands detected
(1991, 2025, 2052, 2159, 2205 cm-1) were assigned to molybdenum having different valence
states as the result of the higher reduction without the detection of adsorbed carbonate species.
Moreover, two bands at 1991 and 2025 cm-1 assigned to metallic Mo0 were observable after
reduction at 700°C [188, 189].
Fig. 6. FTIR roto-vibrational spectra of gaseous CO (left) and NO (right)
39
1.6. OBJECTIVES
Knowing that a lot of researchers prepared and studied Mo/Al2O3 catalysts with low Mo
loadings (8-10 wt%), not taking into account the isoelectric point of the solid support and the
calcination temperature and time, since as expected free MoO3 clusters could not be produced
at Mo loading lower than 15 wt% on Al2O3 depending on the calcination temperature and
time. For this reason, my first objective was to prepare MoO3-rich catalysts and thus how
affects the final catalyst.
The second objective was to explore the influence of ceria and tin (as either promoters or
supports) on molybdena, the changes with adsorption of molybdates in comparison with the
Mo/Al2O3 catalyst since IR and thermal analysis studies regarding these systems are rather
limited. So that the knowledge gained from the catalyst formation should be helpful in the
characterization of the resulting catalyst.
A complementary approach, to help reducing such uncertainties in composition and bonding,
is to study these catalysts, which are chosen to mimick the real catalysts as closely as possible
in terms of overall chemical composition. Therefore, the third objective was to study these
model catalysts by various characterization techniques. Each characterization technique has
its limitations, thus it is dangerous to rely on one method. Whenever, various techniques were
applied, not only to obtain supplemental information about the catalysts but also to serve
mutual checks. These included the multipoint BET analysis method, X-ray diffraction (XRD),
Thermal analysis (TG-DTA), Electron Spin Resonance (ESR) and Diffuse reflectance Fourier
Transform Infrared (DRIFT) spectroscopy.
Additionally, their reduction characteristics with H2 and their activity towards CO adsorption
and CH4 decomposition were also aimed to provide insight into their surface characteristics.
The overall objective was to seek correlations and differences between these catalysts to
collate the results providing better understanding and contributing to the identification of the
physicochemical characteristics of Mo-containing catalysts that could influence their catalytic
behaviour in many reactions.
40
2. EXPERIMENTAL
2.1. Materials and catalyst preparation The γ-Al2O3 support (Ketjen CK300), with BET surface area of 200 m2/g, mean pore diameter
determined by the BJH method of 6.37 nm and pore volume of 0.51 cm3/g, was used for
catalyst preparation. Additionally, the CeO2 formed by the thermal decomposition of
Ce(NO3)3·6H2O (Lobo Feinchemie) in air at 600°C for 6 hrs was found to have a surface area
of 80 m2/g with a mean pore diameter of 12.8 nm and pore volume of 0.016 cm3/g. The SnO2
(purity of 99%, Vega Refinery, Romania) with a surface area of 9 m2/g was found to have a
mean pore diameter of 23.54 nm and pore volume of 0.026 cm3/g.
The Mo/Al2O3, Mo/CeO2 and Mo/SnO2 catalysts were prepared by impregnation of the
supports (γ-Al2O3, CeO2, SnO2) with an aqueous ammonium heptamolybdate (NH4)6
Mo7O24.4H2O (Merck) solution at pH = 2, the pH was adjusted by adding 96 v/v% HNO3.
The concentration of heptamolybdate was that required to not only obtain surface
concentrations close to the theoretical monolayer coverage but also go beyond a compact
single lamella of molybdenum oxide structure [12-15]. The contact time for a given
concentration with the solid (γ-Al2O3, CeO2, SnO2) was extended for 72 hrs and the
temperature of the rotary shaker was maintained at 25°C.
In the second method, the CeO2 and SnO2 oxides were mechanically mixed with 20 wt%
Mo/Al2O3 (impregnated and dried) by adding 10 v/v% HNO3 to prepare CeO2-20 wt%
Mo/Al2O3 and SnO2-20 wt% Mo/Al2O3 samples (containing 5 wt% of CeO2 and 5 wt% of
SnO2, designated as Ce-Mo/Al2O3 and Sn-Mo/Al2O3).
For loadings corresponding to more than 15 wt% Mo on alumina, both Al2(MoO4)3 and bulk
MoO3 are reported with their formation being favoured by both increased calcination
temperature (up to 700°C) and increased calcination time [12-15, 32-35]. So that all samples
were dried in air at 150°C for 6 hrs followed by calcination in air at 600°C for 12 hrs.
41
2.2. Catalyst characterization methods and techniques a) Inductive Coupled Plasma–Atomic Absorption Spectroscopy (ICP-AAS) was
performed in order to verify the amounts of Mo loading. Therefore, the molybdenum loading
was 20 wt% in Mo/Al2O3, 12 wt% in Mo/CeO2 and 4 wt% in Mo/SnO2 samples.
b) Total surface area, SBET, was calculated using the multipoint BET analysis method.
The assessment of mesoporosity was estimated from the adsorption branches via BJH method.
The nitrogen adsorption/desorption isotherms were measured at -196°C using a conventional
volumetric apparatus.
c) The crystalline structure and particle size were determined by X-ray diffraction
(XRD) experiments performed on a Philips PW3710 diffractometer run with Ni-filtered
copper (CuKα) radiation at 50 kV and 10 mA with a scanning speed of 2θ = 2.5°/min and
X-ray wavelength of λ = 0.154056 nm. The XRD patterns were obtained at room temperature
and crystalline phases were identified by comparison with PDF (Powder Diffraction Files)
standards from ICDD.
d) TG and DTA analyses were conducted for 50 mg of uncalcined samples using
automatically recorded Labsys-TG (France) units. The rate of heating was 10°C/min under a
flow of dry air (Messer, Hungary) from room temperature up to 900°C.
e) The Electronic Paramagnetic Resonance (EPR) measurements were carried out with a
Bruker ELEXSYS 500 spectrometer operating at the X-band microwave frequency of 9.29
GHz and magnetic field modulation of 100 kHz, with ER041 XG microwave bridge (a
microwave power of 1 mW and a modulation amplitude of 10 G, recording time of 1 s).
The EPR spectra were recorded and analyzed (using the Win-EPR data acquisition software)
for the calcined samples before and after CO adsorption at room temperature and at -196°C
using a conventional liquid nitrogen flow system.
f) Diffuse Reflectance Fourier Transform Infrared (DRIFT) spectroscopic studies were
performed by Bruker Tensor 27 spectrometer with dry air cooled DTGS detector. In addition,
the sample compartment is always flushed with dry air to avoid detecting of environmental
gases such as CO2 and H2O. In all cases, in situ DRIFT spectra were recorded using
42
Thermo-Nicolet DRIFT cell by co-addition of 126 scans at 4 cm-1 resolution. The cell
equipped with Thermo-Nicolet chamber specially designed for high temperature/vacuum
studies (up to 900°C) was fitted with KBr disks for low temperature and with CaF2 disks for
high temperature measurements. The operating temperature with deviation of ± 3°C was
controlled by using Thermo-Nicolet proportional temperature controller connected to the cell.
In situ DRIFT spectra expressed as absorbance unit versus frequency were collected after
taking the background spectrum for each experiment and can be evaluated accordingly.
Spectral manipulations such as baseline correction, smoothing, normalization, band
component analysis and deconvolution were performed using “OPUS” and “GRAMS”
software packages. In order to enhance the apparent resolution of a spectrum, or to decrease
the line width as well as spectral ranges comprising broad and overlapping lines can thus be
separated into sharp single lines using Fourier Self-Deconvolution function.
Hence, the DRIFT reactor cell was attached to a glass gas handling/ultrahigh vacuum system
and can be evacuated at up to 1.33x10-8 bar. Thus, with the aid of this system the following
gases were used for treatments and reactions:
• Reduction in hydrogen (Messer, purity of 99.995%)
• CO (Messer, purity of 99.99%)
• O2 and N2 (Messer, purity of 99.6%)
• CH4 (Fluka, purity of 99.99%)
• He (99.99% purity)
43
3. RESULTS 3.1. Surface texturing The Mo surface density is expressed as the number of Mo atoms per nanometer square of
surface area (Mo atoms/nm2). It was obtained by the equation [65]:
Mo surface density = (MoO3% x 6.02 x 1023) / (surface area x 144 x 1018) (Eq-22) Where the unit of the surface area is m2/g and 144 is the molar weight of MoO3. The values of
surface areas, cumulative pore volume, Vp (cm3g−1) and the average mesopore diameter for
the calcined materials were estimated and cited in Table 4, whereas Figs. 7, 8, 9 and 10 show
the cumulative mesopore volume and differential pore distribution curves versus pore
diameter plot of the samples calcined at 600°C.
The Mo/Al2O3 catalyst has an average mesopore diameter around 6.2 nm and a continuous
distribution of larger pores that extend well into the macroporous region. As this sample
contains high Mo amount, its texture may be related to Al2O3 because the pore size
distribution curve obtained for Al2O3 (not shown) is very similar. Accordingly, it may be
inferred that the large pores present in the Mo/Al2O3 sample may be related to the interstices
resulting from the packing of the Al2O3 particles, which behave as pores in the N2 adsorption
measurements. The Ce-Mo/Al2O3 catalyst has an average mesopore diameter around 6.9 nm
and being these mesopores of larger diameter in contrast with Mo/Al2O3. More specifically,
upon addition of cerium the average mesopore size increased in Ce-Mo/Al2O3 sample. This
increase in the average mesopore size is responsible for the significant reduction of the
specific surface area of this sample (Table 4).
This finding might indicate the enforced location of Ce species in the pores of this sample
leading to an effective pore widening. Increasing the pore diameter of the former sample
when compared with the others is indicative to the role played by cerium-molybdate and
CeO2 and to their partial deposition inside the micropores creating their own pore system and
thus enhancing the pore diameter. On the other hand, these species might be deposited on
internal surfaces blocking some pores in Ce-Mo/Al2O3 and thus, reducing the surface area
comparatively. Nevertheless, some Mo nanoparticles presumably entrapped in the micropores
of ceria oxide that can also cause expansion in unit cell and pore diameter.
The Sn-Mo/Al2O3 catalyst has an average mesopore diameter around 5.85 nm. Decreasing the
pore radius of this sample when compared with the rest of samples indicates the probability of
the presence of Sn species as separate phases probably as oxides. On the other hand, some Sn
44
particles might be entrapped in internal surfaces blocking some pores in Mo/Al2O3 and thus,
reducing the surface area (34 m2/g) comparatively.
For Mo/SnO2 catalyst, the cumulative mesopore volume distribution is bimodal (Fig. 10),
with an average mesopore diameter around 22.43 nm. On the other hand, the pore diameter
size decreased slightly for Mo/CeO2 and Mo/SnO2 samples in comparison with CeO2 and
SnO2. This may propose that the diffusion of MoO3 did not proceed into the CeO2 and SnO2
bulk during the preparation.
Table 4. Textural characteristics derived from BET method for the materials calcined at 600°C
Catalyst
Average mesopore
diameter (nm)
Total mesopore volume (cm3/g)
BET Surface
area (m2/g)
Mo density (atom/nm2)
Atom ratio (Mo/M)
γ-Al2O3 6.37 0.51 200 - - 20 wt%
Mo/Al2O36.20 0.225 117.6 7 Mo/Al=0.134
5 wt% Ce-20 wt% Mo/Al2O3
6.9 0.072 29.6 28 Mo/Ce=6.67
CeO2 12.8 0.016 80 - - 12 wt%
Mo/CeO212.5 0.0053 16.2 31 Mo/Ce=0.24
SnO2 23.54 0.026 9 - - 5 wt% Sn-20
wt% Mo/Al2O35.85 0.073 34 24.6 Mo/Sn=6
4 wt% Mo/SnO2 22.43 0.023 3.86 43 Mo/Sn=0.066
45
Fig. 7. Cumulative mesopore volume distribution vs. pore diameter of Mo/Al2O3,
Mo/CeO2 and Ce-Mo/Al2O3 samples
Fig. 8. Differential pore distribution vs. pore diameter of Mo/Al2O3, Mo/CeO2 and
Ce-Mo/Al2O3 samples
46
Fig. 9. Cumulative mesopore volume distribution vs. pore diameter of Mo/SnO2 and Sn-Mo/Al2O3 samples
Fig. 10. Differential pore distribution vs. pore diameter of Mo/SnO2 and Sn-Mo/Al2O3
samples
47
3.2. X-ray diffraction Figs. 11 and 13 show the diffractograms of the samples calcined at 600°C. The values of the
average crystallite size ( d ) of the crystalline phases were calculated by using the
Debye–Scherrer equation [1]:
θβλ
coskd = (Eq-23)
Where λ = 0.154056 nm, is the X-ray wavelength, k the particle shape factor (0.9 for cubic
particles), β the full width of the peak considered at half maximum (in rad), and θ is the half
of the diffraction angle.
3.2.1. XRD patterns of Mo/Al2O3, Ce-Mo/Al2O3 and Mo/CeO2 It can be seen, the XRD patterns of Mo/Al2O3 exhibited diffraction lines and d-spacing values
corresponding to Al2O3 (tetrahedral and octahedral), orthorhombic Al2(MoO4)3 and octahedral
MoO3 crystallites (Fig. 11). Moreover, the profile of XRD peaks of Mo crystallites can be
attributed to different types of particles: (i) huge particles, responsible for the fine part of the
peak with high intensity (Al2(MoO4)3), and (ii) smaller free MoO3 crystallites giving a low
intensity of the peak (broadening/low intensity peaks) from the XRD patterns of crystalline
species. This can be related to a less crystalline phase and well-dispersed MoO3 and/or MoO3
particles are under limit to be detected by XRD technique (< 2 nm). Indeed, the calculated
average crystallite size of MoO3 was 1.7-3.4 nm in Mo/Al2O3 sample (Table 5).
Consequently, the XRD crystallite size calculation takes into account essentially the biggest
crystallites. However, taking into account that the ionic radius of Mo6+ (0.62 Å) is larger than
that of Al3+ (0.50 Å), therefore, the Mo6+ ions do not incorporate into the lattice of Al2O3 to
replace Al3+ ions suggesting that some molybdena supported on the Al2O3 may spread over
the support. Nevertheless, Mo can diffuse into defect sites of alumina, and/or interact with the
surface hydroxyl groups of alumina becoming difficult reducible and strongly bound to form
Mo-O-Al bonds in the form of Al2(MoO4)3 [12-17, 42-47].
On the other hand, the diffractograms of Ce-Mo/Al2O3 and Mo/CeO2 exhibited diffraction
lines and d-spacing values at 3.7, 3.1, 2.7, 1.9 and 1.6 Å that correspond to (110), (111),
(200), (220) and (311) phases, respectively, of the CeO2 cubic structure [53-65]. Furthermore,
the XRD patterns of Ce-Mo/Al2O3 display a noticeable decrease of the line intensities and
crystallite size of MoO3 (1.4-2.7 nm) and Al2(MoO4)3 (2.8-5.6 nm) phases in the latter
material which was perceived comparatively to Mo/Al2O3 patterns upon the addition of ceria.
48
The diffraction patterns of Mo/CeO2 material exhibited well-ordered MoO3 crystallites at d-
spacing of 6.25, 3.9, 3.4, 2.83 and 1.84 Å. Thus, the increase in crystallinity of the former
sample could be caused by the presence of more cerium-molybdate in contrast with Ce-
Mo/Al2O3 that became predominant on the small particles, whereas different surfaces such as
(110) and (111) CeO2 phases were exposed on large CeO2 particles (Table 5).
Generally, this sample revealed the presence of new lines with d-spacing values pertaining to
various phases including CeO2 and cerium-molybdate (three crystalline structures) those most
probably exposed to the external surface in contrast with Ce-Mo/Al2O3 sample in which only
one crystalline structure of cerium-molybdate (Ce2Mo3O12) was found (Fig. 11).
The value of the average particle size of MoO3 crystallites in Mo/CeO2 was between 1.8-4.6
nm, which was higher (within the experimental error) than that of 1.4-2.7 nm in
Ce-Mo/Al2O3. On the contrary, the average particle size of CeO2 crystallites was lower in
Mo/CeO2 (Table 5). This may demonstrate that incorporating Mo in ceria, in Mo/CeO2,
effectively prevents particles agglomeration allowing the material to maintain its dispersion.
The presence of the residual peaks of (110) and (111) phase of CeO2, which is the dominant
phase forming cubic ceria, indicates that ceria structure is still intact. Ceria crystallizes in a
cubic fluorite structure where each cerium cation is coordinated by eight equivalents nearest
neighbour oxygen anions at the corner of the cube. The models for clean (111) and (110)
surfaces are shown in Fig. 12, respectively, and are fully relaxed under the restriction of fixed
cell parameters and fixed geometry of the lower layers. The CeO2 (111) surface relaxation is
quite small, with the Ce sub-lattice remaining unperturbed and with the [O] sub-lattice
undergoing small changes in inter-layer distances of around 0.03 Å. The displacements show
a slight contraction of the first few layers for the (111) surface. The relaxation of the (110)
surface exhibits a reverse behaviour compared to that of the (111) surface. Specifically, the
oxygen sub-lattice remains essentially unaffected, while the cerium sub-lattice undergoes
larger changes in the inter-planar distances of ±0.23 Å [66-72].
The flat ideal (110) surface becomes slightly rumpled upon relaxation, with Ce atoms shifted
towards the center of the slab by 0.13 Å relative to the surface O atoms, this equals 2.4% of
the ceria bulk lattice constant. In addition, the Ce–O bond length at the surface contracts by
0.04–2.324 Å compared to the bond length in the bulk by 2.36 Å [73-79].
49
Fig. 11. XRD patterns of Mo/Al2O3, Ce-Mo/Al2O3 and Mo/CeO2 calcined at 600°C
Table 5. Crystallite sizes derived from XRD data of Mo/Al2O3, Ce-Mo/Al2O3 and Mo/CeO2 materials calcined at 600ºC
The average crystallite sizes (nm) Catalyst CeO2 MoO3 Al2(MoO4)3
Mo/Al2O3 - 1.7-3.4 3.3-6.9 Ce-Mo/Al2O3 17-20 1.4-2.7 2.8-5.6
Mo/CeO2 7-9 1.8-4.6 -
Fig. 12. Slab models for ceria surfaces (111 and 110), where the yellow and red spheres
represent Ce and O atoms.
50
3.2.2. XRD patterns of Sn-Mo/Al2O3 and Mo/SnO2 Fig. 13 displays the X-ray diffraction patterns of Sn-Mo/Al2O3 and Mo/SnO2 catalysts
calcined at 600°C. The diffractogram of Sn-Mo/Al2O3 exhibits reflection peaks and d-spacing
values corresponding to Al2O3, orthorhombic Al2(MoO4)3, tetragonal SnO2 (cassiterite) but
does not show reflections related to crystalline MoO3. On the contrary, phase transformations
either MoO3 into monoclinic MoO2 or SnO2 into SnO were observed.
On the other hand, the XRD patterns of Mo/SnO2 show reflection peaks and d-spacing values
attributed only to two phases in the form of MoO3 (crystallite size between 7-10 nm) and
SnO2 (crystallite size between 3.2-5.8 nm) (Table 6). However, assuming for simplicity a
completely ionic structure, Mo6+ ions, with atomic radius of 0.62 Å can migrate to the surface
and from the surface to the sub-layers. The oxygen vacancies of the surface and of the sub-
layers can easily reduce Mo6+ to Mo5+ or Mo4+ with atomic radius of 0.63 and 0.65 Å,
respectively. As Sn4+ has atomic radius of 0.71 Å, Mo6+, Mo5+ and Mo4+ can easily occupy
the tin lattice sites giving rise to a solid solution. Despite this fact, no mixed oxide phase was
observed between Mo and Sn ions in both samples even after the calcination at 600°C. Since
the Mo/SnO2 catalyst obtained by impregnation only showed the presence of the cassiterite
phase and MoO3, while the Sn-Mo/Al2O3 catalyst obtained by co-precipitation showed the
presence of MoO2, SnO2 and SnO without reflection peaks related to MoO3.
Tin oxide SnO2, in its pure form, has the tetragonal crystalline structure at room temperature
and normal pressure [84-101]. SnO2 is a more densely packed crystal where each tin atom is
surrounded by a slightly distorted oxygen octahedron while in SnO the tin atoms sit on the
vertices of pyramids with an oxygen square basis (Fig. 14).
51
Fig. 13. XRD patterns of Sn-Mo/Al2O3 and Mo/SnO2 calcined at 600°C
Table 6. Crystallite sizes derived from XRD data of Sn-Mo/Al2O3 and Mo/SnO2 materials
calcined at 600ºC
The average crystallite sizes (nm) Catalyst MoO3 MoO2 SnO2 SnO Al2(MoO4)3Sn-Mo/Al2O3 - 1.7-2.3 1.7-4.5 1.8-2.4 2.8-6.2
Mo/SnO2 7-10 - 3.2-5.8 - -
52
Fig. 14 shows atomic configurations of SnO2 and SnO. (a) is the structure of SnO2 unit cell
where the dashed lines link the O atoms forming an octahedron surrounding a tin atom. (b) is
the structure of SnO unit cell where the dashed lines link the O atoms and the tin atom
forming square-based pyramids. (c) shows the projection of the SnO2 unit cell onto a (100)
plane showing the traces of the alternate O and Sn atomic planes and (d) shows the projection
of SnO unit cell onto a (010) plane showing the traces of the O and Sn atomic planes.
Fig. 14. Atomic configurations of SnO2 and SnO: (a) Structure of SnO2 unit cell. (b) Structure of SnO unit cell. (c) Projection of the SnO2 unit cell onto a (100) plane showing the traces of
the alternate O and Sn atomic planes. (d) Projection of SnO unit cell onto a (010) plane showing the traces of the O and Sn atomic planes.
53
On the other hand, the coordination of the metal atom in MoO3 can be best considered as that
of a markedly distorted octahedron, although it can be easily deduced from the MoO3
tetrahedron as a basic unit. However, the clean (100) MoO3 has a layer structure in which
each layer is built up of MoO6 octahedrons at two levels, connected along the z-axis by
common edges and corners, so as to forming zigzag rows and along x-axis by common
corners only. Moreover, each layer exhibits, in the direction of z-axis, oxygen atoms, which
are common for three different octahedrons. Each octahedron also shares, along x-axis, two
oxygen atoms with two neighbouring octahedrons. Besides, for each MoO6 octahedron there
is only one oxygen atom which is doubly bound (Mo=O) to the molybdenum atom. It
occupies different positions along the y-axis. For each MoO6 octahedron at the higher level
this oxygen atom points up. On the contrary, for each MoO6 octahedron at the lower level this
oxygen atom points down [194-197].
The structure of the MoO6 octahedron is shown in Fig. 15. In addition, the idealized
arrangement of molybdenum and oxygen atoms on the (100) plane of MoO3 is shown in
Fig. 16. The clean (100) plane presents coordinately unsaturated Mo6+ ions with one bridging
O2- ion missing from their coordination sphere. Thus, it acquires the formal uncompensated
charge (+1). This plane also contains unsaturated bridging O2- ions with one Mo6+ missing
and a formal uncompensated charge of (-1). According to this approach, two types of surface
oxygen atoms can be distinguished: tightly bound inactive lattice oxygen atoms and weakly
bound active lattice oxygen atoms that may play an important role in MoO3 activity.
54
Fig. 15. The MoO6 octahedron structure. Mo (shaded circles), O (open circles).
Fig. 16. Space fill idealised (100) MoO3 face. The uncompensated charges are also shown.
Mo (shaded circles), O (open circles).
55
3.3. Thermal Analysis The thermogravimetric (mass loss, TG), derivative thermogravimetric (rate of mass loss,
DTG) and differential thermal analysis (DTA) curves of 50 mg of uncalcined samples, over
the range from room temperature up to 900°C in the flow of an atmosphere of dry air, are
shown in Figs. 18, 19, 20, 21 and 22.
3.3.1. TG and DTA of Mo/Al2O3 The TG-DTG and DTA curves recorded in argon flow for pure (NH4)6Mo7O24.4H2O (AHMT)
at a scanning speed of 5°C/min up to 600°C are shown in Fig. 17. The TG-DTG exhibit three
decomposition peaks with the maximum located at 127, 220 and 307°C respectively. The TG
shows that AHMT loses weight with heating over three steps.
The first endothermic peak at 100-127°C was followed by 7.3% mass loss corresponding to
the loss of three H2O molecules and two molecules of NH3, which is consistent with the
theoretical value (according to the literature, the theoretical weight loss, assuming release of
three water molecules and two NH3 molecules is 7.1%) [160–162]. The second endothermic
peak at 200-225°C which was accompanied by a mass loss of 4.4% is probably related to the
elimination of (NH4)2O (i.e. 2NH3 + H2O) which is in agreement with the theoretical value
(4.2%). The third endothermic peak takes place within 267-327°C was associated with 7.4%
weight loss (theoretical value 7.1%) and corresponds to the evolution of two NH3 molecules
together with three molecules of H2O. Moreover, the exothermic peak of maxima at 327°C
confirms the crystallization of new phases (probably MoO3). Therefore, the overall mass loss
was 19.1%. These three stages can be explained to proceed according to the following
equations:
(NH4)6Mo7O24.4H2O (NH⎯⎯⎯⎯⎯⎯ °− C127100 →
→
→
4)4Mo7O23.2H2O + 3H2O + 2NH3 (Eq-24)
(NH4)4Mo7O23.2H2O (NH⎯⎯⎯⎯⎯⎯ °− C2252004)2Mo7O22.2H2O + H2O + 2NH3 (Eq-25)
(NH4)2Mo7O22.2H2O ⎯⎯⎯⎯⎯⎯ °− C327267 7MoO3 + 3H2O + 2NH3 (Eq-26)
56
Fig. 17. TG-DTG and DTA curves for AHMT in argon atmosphere
The TG-DTG and DTA curves recorded in dry air flow for Mo/Al2O3 exhibit a large mass loss
(37.8%) up to 300°C with an endothermic peak of maxima at 83°C in the DTA thermogram
(Fig. 18). This is corresponding to the loss of structural and intercalated water and to the
decomposition of (NH4)6 Mo7O24.4H2O [160, 161].
The DTA thermogram of this sample displays a very big exothermic peak of a maximum at
around 300°C that could be due to the process of the formation of new phases probably to the
crystallization of MoO3 species [162-165].
However, the three decomposition steps of (NH4)6Mo7O24.4H2O are missed on the TG-DTA
curves of Mo/Al2O3 due to several factors such as: first, Mo/Al2O3 is impregnated and dried,
second, the heating atmosphere (dry air) and heating rate (10°C/min) are different from those
(argon flow, 5°C/min) applied for thermal analysis of pure (AHMT).
57
Fig. 18. TG-DTG and DTA curves of Mo/Al2O3
3.3.2. TG and DTA of Ce-Mo/Al2O3 and Mo/CeO2 TG-DTG and DTA curves recorded for Ce-Mo/Al2O3 (Fig. 19) indicate that the weight loss
(20.5%) under 300°C is less than that of Mo/Al2O3. This mass loss is due to the removal of
water and to the decomposition of Ce(NO3)3 and (NH4)6 Mo7O24.4H2O producing the
corresponding oxides according to the equation:
(NH4)6 Mo7O24.4H2O + Ce(NO3)3 + 2O2 → 7MoO3 + CeO2 + 3NO2 + 6NH3 + 12H2O
(Eq-27)
This indicates that the weight loss is highly affected by the mode of the preparation that in
turn affects the interaction mode of CeO2–MoO3 compounds. The latter affects by its turn the
thermal stabilities of the produced compounds. In the meanwhile, the position of the
endothermic peak shifts to a higher value with a maximum at 101°C, whereas the big
exothermic peak remained with the maximum at around 300°C comparatively.
58
Of particular interest, the TG and DTA of Ce-Mo/Al2O3 display a large mass loss (23%)
extended between 700 and 900°C with an endothermic peak of maxima at 758°C indicating
that the material undergoes morphological and structural modifications. This may be due to
the high oxygen storage and release of ceria resulting in lattice defects and thus enhancing the
mobility of Mo and Ce ions. Thus, the crystalline structure is progressively modified
comparatively favoring the possible mutual diffusion of the Mo and Ce ions. The last mass
loss in this material may characterize the sublimation of some molybdena produced [65, 80].
The thermal decomposition course presented for Mo/CeO2 is shown to be of multistep and
represented by four endothermic peaks of maxima at 61, 98, 173 and 737°C in the DTA curve
(Fig. 20) besides a big exothermic peak of a maximum at 308°C. As can be seen, this material
contributes a total weight loss around 34.7%. On the other hand, it is worth mentioning that
the decomposition temperature of either Mo or Ce precursor salts in the synthesized materials
are not similar revealing that the interactions between them are varied according to altering
the preparation methods and their content, which indeed affect the mobility of Mo species.
Fig. 19. TG-DTG and DTA curves of Ce-Mo/Al2O3
59
Fig. 20. TG-DTG and DTA curves of Mo/CeO2
3.3.3. TG and DTA of Sn-Mo/Al2O3 and Mo/SnO2 On the TG curve of Sn-Mo/Al2O3 a large weight loss step is marked (38.3%) quite similar to
that of Mo/Al2O3 (37.8%) comparatively implying that the addition of SnO2 did not affect the
thermal behaviour of Mo/Al2O3. This step with 38.3% mass loss of the original mass
corresponds to dehydration and decomposition of ammonium heptamolybdate (Fig. 21).
The DTA curve of Sn-Mo/Al2O3 shows three endothermic peaks at 69, 103 and 211°C
indicating the presence of weakly bounded, strongly bonded and structural water species.
The TG curve of Mo/SnO2 is definitely different (Fig. 22). Two mass loss steps are observed.
The rate in the first stage of its mass loss between 70 and 300°C is relatively low (11.4%),
while the mass loss in the second stage is higher (41.7%). The overall weight loss measured
up to 900°C is 53.1%. The DTA curve of Mo/SnO2 shows first an endothermic dehydration
occurring between 70 and 200°C (temperature peaks marked at 71 and 197°C). It is followed
by the exothermic effect starting from 300°C (peaks marked at 327 and 396°C) indicating the
process of crystallization in this sample (probably formation of molybdenum oxide).
60
Nevertheless, the DTA peaks centred at 797 and 846°C explicitly are two strongly
endothermic heat effects of the main degradation presumably due to the release of some
volatile tin and molybdena species and/or due to surface and structure modifications of the
contact between phases leading to lattice defects, and thus enhancing the mobility of the two
ions and mutual interaction between them. These changes indicate that Mo can migrate, some
Mo and Sn species can be destroyed, while others are created or increased. It is, however,
hard to estimate if new segregated phases are formed upon thermal treatment above 750°C
that can be followed more certainly by TG-MS studies [83].
Fig. 21. TG-DTG and DTA curves of Sn-Mo/Al2O3
Fig. 22. TG-DTG and DTA curves of Mo/SnO2
61
3.4. Electron Spin Resonance (ESR) measurements The EPR spectra recorded at room and liquid nitrogen temperature for Mo/Al2O3, Mo/CeO2
and Mo/SnO2 samples calcined at 600°C before and after interaction with CO are shown in
Figs. 23, 24, 25 and 26.
The EPR spectra can best be analysed using a spin-Hamiltonian of the form: H = βSgB + SAI,
where the symbols have their usual meaning (Eq-21). It contains the electronic Zeeman term
(β is the Bohr magneton, S = 1/2 the electron spin, g is the g-tensor and B is the applied field),
perturbed by the hyper-fine coupling term between the unpaired electron and the nuclear spin
of molybdenum (I), whereas (A) being the hyperfine structure tensor [166-169].
The EPR spectra of calcined Mo/Al2O3 show a weak hyperfine structure with six lines in both
parallel (g║) and perpendicular (g┴) bands (Fig. 23). The intense central line arises even from 96Mo isotope which has the nuclear spin I = 0, while the lower intensity lines correspond to
the hyperfine structure from the 95Mo and 97Mo isotopes which have the nuclear spin I= 5/2.
However, the nuclear magnetic moments of 95Mo and 97Mo being close, so the isotope
splitting is not resolved in the EPR spectra.
The EPR lines of Mo/Al2O3 between 3400-3600G becoming more prominent at both room
temperature and -196°C after CO reaction due to the presence of Mo5+ paramagnetic centers
(Mo6+ diamagnetic) and hence there is an increase of concentration of paramagnetic centers in
this sample. Accordingly, Mo6+ ions were reduced by CO to Mo5+. It should be mentioned
that Al2O3 does not give rise to EPR lines. Thus, the spectra obtained for Mo/Al2O3 may be
considered as the hyperfine structure typical for isolated Mo5+ ions and its widening is due to
the contribution of the clustered Mo5+ ions coupled by the dipolar interactions between Mo5+
paramagnetic ions (between the electron spin and its surrounding nuclei) leading to the line
broadening [170-174].
The EPR spectra of Mo/CeO2 show (Fig. 24) a complex EPR signal (due to the phase
containing a mixture of oxides) generated by the two Mo5+ and Ce3+ paramagnetic centers
present in the sample after CO interaction. Thus, both the support and the active phase are
reduced by CO. More specifically, the band at 3370G is assigned to the Ce3+ ions, while the
bands between 3400-3600G are assigned to Mo5+ ions (both Mo6+ and Ce4+ are diamagnetic).
On the other hand, the EPR spectra of Mo/SnO2 show the presence of Mo5+ ions after CO
interaction, the EPR spectrum of Mo/SnO2 did not show any resonance signals of O- and O2-
paramagnetic species neither before nor after CO contact (Fig. 25). Meanwhile, Vo oxygen
vacancy was observed in powdered Mo/SnO2 sample at both temperatures (r.t. and –196°C)
62
after CO interaction. Accordingly, it is attributed to an oxygen defect, the single ionised
oxygen vacancy Vo, produced by CO interaction:
CO + Oo ↔ Vo + CO2 (Eq-28)
Vo ↔Vo+ + e– (Eq-29)
Where Oo is the lattice oxygen, V0 is a neutral oxygen vacancy and e– an electron in the
conduction band. It is worth mentioning that the EPR spectrum of calcined Mo/SnO2 did not
show any resonance signals due to the fact that both Mo6+ and Sn4+ are diamagnetic.
Fig. 23. The EPR spectra of Mo/Al2O3 recorded at –196°C before and after CO contact
Fig. 24. The EPR spectra of Mo/CeO2 recorded at –196°C before and after CO contact
63
Fig. 25. The EPR spectra of Mo/SnO2 recorded at –196°C before and after CO contact
Fig. 26. The EPR spectra of the samples recorded at room temperature after CO contact
64
3.5. In situ DRIFT spectroscopy measurements 3.5.1. DRIFT spectra of the calcined samples under vacuum All DRIFT spectra were recorded at room temperature for samples calcined at 600°C after
evacuation at 1.33x10-8 bar. The IR assignments of metal-oxygen bonds are shown in Table 7.
3.5.1.1. DRIFT spectra of γ-Al2O3 and CeO2 The DRIFT spectra of the γ-Al2O3 and CeO2 supports are depicted in Fig. 27. The spectrum of
γ-Al2O3 displays bands at 416, 485, 690 and 744 cm-1 are tentatively assigned to terminal
νs(Al-O) (416 cm-1), νas(Al-O) (485 cm-1) and bridging νs(Al-O-Al) (690 cm-1), νas(Al-O-Al)
(744 cm-1) vibration modes [12-17]. On the other hand, the bands at 1017 and 1161 cm-1 can
be assigned to the in plane and the out of plane deformation modes of OH groups being in the
bulk of Al2O3. The band at 1633 cm-1 is related to the deformation vibration of water (H-OH).
The spectrum of CeO2 exhibits bands at 415 cm-1 νs(Ce-O) and 480 cm-1 νas(Ce-O) due to
symmetric and asymmetric stretching modes of Ce-O terminal bonds, while the bands at 687
and 740 cm-1 are assigned to νs(Ce-O-Ce) and νas(Ce-O-Ce) bridging bonds [65-73].
Additionally, two bands at 1016, 1153 and 1633 cm-1 can be attributed to the deformation
modes of OH groups and water as mentioned above. Moreover, the broad band in 3200-3800
cm-1 region in the two spectra is assigned to isolated and associated hydroxyl groups bound to
tetrahedrally and octahedrally coordinated Al3+ sites and to CeO2 surface [18-31, 74-80].
Fig. 27. In situ DRIFT spectra of γ-Al2O3 and CeO2 under vacuum
65
3.5.1.2. DRIFT spectra of Mo/Al2O3, Mo/CeO2 and Ce-Mo/Al2O3 The spectrum of Mo/Al2O3 shows bands at 415 cm-1 and 490 cm-1 corresponding to bridging
stretching modes of νs(Mo-O-Al) and νas(Mo-O-Al) indicating the strong association of Mo
species (MoO42-) with Al2O3 support (Fig. 28) that was confirmed by XRD. These two bands
may have stemmed from overlapping with Al-O vibrations since they are broader than 416
and 485 cm-1 bands corresponding to (Al-O) vibrations. On the other hand, the band at 995
cm-1 that corresponds to νas(Mo=O) terminal stretching in bulk MoO3 indicates strong features
for microcrystalline MoO3 species [12-15]. One may notice that the oxygen anion of the
terminal Mo=O bond when exposed to the ambient atmosphere can easily interact with
moisture, resulting in the formation of hydrated surface species.
On the other hand, in order to help reducing uncertainties in composition and bonding of
Mo/Al2O3 the spectrum of 8 wt% Mo/Al2O3 (prepared by others at the Institute of
Environmental Engineering with 196 m2/g surface area) was recorded at the same conditions
to compare with that of 20 wt% Mo/Al2O3. As can be seen (Fig. 29), the bands at 423 and 512
cm-1 can be assigned to νs(Mo-O-Al) and νas(Mo-O-Al) vibration modes, whereas the band at
957 cm-1 is likely corresponding (terminal νasMo=O stretches) either to isolated tetrahedral
MoO42- species or bonded to alumina surface (Eq-5). However, the absence of this band in the
spectrum of 20 wt% Mo/Al2O3 confirms the evidence for the appearance of polymeric MoO42-
that is further emphasized by the band appearing at 749 cm-1 assigned to νs(Mo-O-Mo), which
shifts to higher frequency (757 cm-1) in Ce-Mo/Al2O3 spectrum [12-17].
The spectra of Mo/CeO2 and Ce-Mo/Al2O3 exhibit similar characteristic IR bands and can be
characterized by new bands at 630, 757, 875, and 1035 cm-1 with the absence of the bands at
415 and 490 cm-1 if compared with that of Mo/Al2O3 and CeO2 (Figs. 27 and 28).
However, the fact that the addition of ceria contributes to the formation of new phases so as to
forming Mo-O-Ce linkages that were represented by the bands at 630 cm-1 νs(Mo-O-Ce) and
875 cm-1 νas(Mo-O-Ce), while the bands separated by 40 cm-1 with different relative
intensities at 995 and 1035 cm-1 are associated with the formation of coupled νs(O=Mo=O)
bonds indicative of the presence of polymeric MoO3 implying that IR is more confined to too
small crystallites than XRD did.
Additionally, the proposed interaction between Mo and ceria is further emphasized by the
absence of the vibration modes of Ce-O at 415 and 687 cm-1 observed in CeO2 spectrum.
66
Fig. 28. In situ DRIFT spectra of 20% Mo/Al2O3, Ce-Mo/Al2O3 and Mo/CeO2 under vacuum
Fig. 29. In situ DRIFT spectra of 8% Mo/Al2O3 under vacuum
67
3.5.1.3. DRIFT spectra of SnO2, Mo/SnO2 and Sn-Mo/Al2O3 The DRIFT spectra recorded of SnO2, Mo/SnO2 and Sn-Mo/Al2O3 samples are depicted in
Fig. 30. The spectra display bands at 1634, 740, 690, 630, 590 and 480 cm-1.
The bands at 480 νs(Sn-O) and 590 cm-1 νas(Sn-O) are assigned to symmetric and asymmetric
stretching modes of (Sn-O) terminal bonds. The bands centred at 630 and 690 cm-1
correspond to νs(O-Sn-O) and νas(O-Sn-O) vibrations. Moreover, the peak at 740 cm-1 is
assigned to bridged species of νas(Sn-O-Sn), while the discrete peak at 1031 cm-1 is related to
the lattice of different oxygen-bridged species of tin with the shoulder at 1123 cm-1 to
terminal γ(Sn-OH), besides the band at 1634 cm-1 observed in all samples due to the
deformation vibration mode of the H-O-H bond of water.
The spectrum of Mo/SnO2 exhibits additional band at 995 cm-1 attributed to νas(Mo=O)
terminal stretching in bulk MoO3 indicating strong features for free MoO3 clusters.
On the other hand, in the spectrum of Sn-Mo/Al2O3 the band at 761 cm-1 is related to
νs(Mo-O-Mo) associated with polymolybdates but it is noticeable to observe that the
vibrational peak at 995 cm-1 is absent in this spectrum in agreement with XRD results since
no MoO3 reflection peaks were detected by XRD.
Furthermore, no bands arising from Mo–O–Sn vibrations were observed (in line with XRD
results) in both Mo/SnO2 and Sn-Mo/Al2O3 samples that undoubtedly appear at 860-870 cm-1.
However, it is important to stress that the Mo–O–Mo unit is associated with polymolybdates,
while the Mo–O–Sn is related to isolated tetrahedral Mo species. Therefore, FTIR results
allow these two kinds of coordinations at hydrated conditions to be distinguished [91-98].
68
Fig. 30. In situ DRIFT spectra of SnO2, Mo/SnO2 and Sn-Mo/Al2O3 under vacuum
Table 7. Assignments of IR characteristics of metal-oxygen bonds
Assignment Wavenumber (cm-1) Reference Terminal νsCe-O, νasCe-O Bridging νsCe-O-Ce, νasCe-O-Ce
415, 480 690, 687, 740 65-80, 180-187
Terminal νsAl-O, νasAl-O Bridging νsAl-O-Al, νasAl-O-Al
416, 485 690, 744 12-31
In plane and out of plane deformation modes of OH groups
1123, 1153,1161 1016, 1017 82, 175-177
Deformation vibration of water 1633, 1634, 1637 25-31, 82 νsMo-O-Al, νasMo-O-Al stretching 415, 423, 490, 512 12-17, 32-37 Terminal νasMo=O 957, 995 12-15, 42-47 Coupled O=Mo=O 995 and 1035 12-15, 34, 65, 80 νsMo-O-Mo 749, 757, 761 12-17 νsMo-O-Ce, νasMo-O-Ce 630, 875 65, 79-81 Terminal νsSn-O, νasSn-O, νsO-Sn-O, νasO-Sn-O Bridging νsSn-O-Sn νasSn-O-Sn
480, 590 630, 690
740, 1031 82, 90-99
OH groups to tetrahedrally and octahedrally coordinated Al3+ sites
3200-3800 18-31
OH groups of SnO2 and CeO2 3200-3800 90-99, 180-187
69
3.5.2. CO chemisorption
Hence, a new sample (50 mg) was used for each experiment and the pre-treatment for all
samples as follows:
1. The sample was mounted in the cell fitted with CaF2 disks (which are IR transparent up to
1000 cm−1) followed by evacuation at room temperature and at 1.33x10-8 bar for 30 min.
2. The sample was reduced at different temperature in hydrogen flow (60 cm3/min, 1 bar,
10°C/min of heating rate) up to 800°C.
3. Then the system was cooled back to room temperature, flushed with N2 and then evacuated
at room temperature and 1.33x10-8 bar for 30 min.
4. Hence, the CO adsorption was carried out at 4x10-2 bar and in temperature range 20-100°C
for one hour. All the DRIFT spectra (MIR) were preserved following evacuation of the cell at
room temperature and at about 1.33x10-8 bar for 30 min.
On the other hand, since no appreciable CO adsorption bands appeared below 100°C only the
spectra of CO adsorption at 100°C are subtracted and deconvoluted. The IR assignments of
metal-CO bonds are shown in Table 8.
3.5.2.1. CO chemisorption on Mo/Al2O3 The spectra were presented after subtracting the spectra of the gas phase and the sample prior
to the adsorption. The spectra of CO adsorbed at 100°C on 20 wt% Mo/Al2O3 reduced at
600°C, 700°C and 800°C are depicted in Fig. 31.
Four different υ(CO) bands can be observed in the spectrum of Mo/Al2O3 reduced at 600°C.
The two bands at νs1388 and νas1497 cm-1 are originated from monodentate carbonate from
CO adsorbed on molybdena, while the other two bands are due to CO species adsorbed on
metal cations. Thus, the CO adsorption band at 2048 cm-1 can be assigned as CO bonded to
Mo4+ and/or Mo3+ ions. Furthermore, the band at 2194 cm-1 is assigned to σ-bonded CO
adsorbed on octahedrally coordinated Al3+ sites which are present on the surface more than
tetrahedral one, having stronger Lewis acidity, since the lower wavenumber has been
associated with increasing acid strength [20-28]. However, this band became more prominent
and shifted to 2197 cm-1 after reduction of Mo/Al2O3 at 800°C due to the increase in the CO
stretching frequency above of the free CO gas molecule frequency (2143 cm-1) with strong
Lewis acid character.
The spectrum of CO adsorbed on Mo/Al2O3 reduced at 700°C exhibits new bands: (i) two
bands at νs1351 and νas1578 cm-1 due to symmetric and asymmetric vibrations of formate
70
species that is further confirmed by the broad band at about 2685 cm-1 attributed to ν(CH)
stretching, (ii) two bands at νs1485 and νas1550 cm-1 are assigned to vibrations of
carboxylates, (iii) two bands at νs1385 and νas1447 cm-1 can be assigned to vibrations of
monodentate species. In addition, the bands appearing at νas1415, νs1715 and νs1850 cm-1
indicate the characteristics of free carbonate species on the surface. However, the band at
2048 cm-1 appears with a shoulder at 1994 cm-1 that can be assigned to bridged Mo0-CO
indicating the presence of small amounts of metallic Mo0. These two bands shifted to 2050
and 2002 cm-1 on the catalyst reduced at 800°C.
It can be seen that further reduction up to 800°C enhanced the adsorption of CO to form
bridged carbonates (νas1230 and νs1765 cm-1) and bicarbonates (νs1447 and νas1600 cm-1).
The spectrum also reveals bands located at νas1281, νs1653 cm-1 belonging to bidentates,
besides the bands at νs1385, νas1560 cm-1 assigned to formate. On the other hand, the bands
protruding at 2025 and 2050 cm-1 are created upon the interaction between Mo ions and
adsorbed CO molecule. More specifically, the band observed at 2025 cm-1 with a shoulder at
2002 cm-1 are very likely associated with the terminally configured CO σ-bonded to metallic
Mo(0) species. Whereas the band at 2048 cm-1 that was first seen after reduction at 600°C
shifted to 2050 cm-1 due to the presence of lower molybdenum valence states as the result of
the higher reduction at 800°C when the average molybdenum oxidation number was
estimated to be 1,6 [3-9]. This band can be assigned to CO π-bonding to Mo2+ and/or Mo1+ in
harmony with some reported results in the literature [12-15].
Another interesting point of CO adsorption behaviour on Mo/Al2O3 surface reduced at 800°C
is that the band at 2197 cm-1 is more prominent and broader than the band at 2194 cm-1. This
band is probably formed by overlapping the band corresponding to octahedral alumina sites
with a band generated by interaction of CO with molybdena hydroxyls (Mo-OH bonded).
Consequently, the lower charge and higher reduction of Mo lead to a weaker σ-bond and
enhance a π-type electron backdonation to form Mo-CO complexes (carbonates). This
contributes to a CO stretching frequency below the CO gas frequency (2143 cm-1).
71
Fig. 31. In situ DRIFT spectra of CO adsorbed on Mo/Al2O3 reduced at different temperature 3.5.2.2. CO chemisorption on CeO2, Mo/CeO2 and Ce-Mo/Al2O3 reduced at 800°C
On the basis of the results that can be found extensively in the literature regarding the
reduction of Mo, Ce and Sn [3-12, 101-106, 181-186] and taking into account the XRD and
TG-DTA results and those obtained upon CO adsorption on 20 wt% Mo/Al2O3 reduced at
different temperatures. It was decided to perform CO adsorption and CH4 reaction on the
catalysts only reduced at 800°C.
In situ DRIFT spectra of CO adsorption on Ce-Mo/Al2O3, Mo/CeO2 and CeO2 catalysts
reduced at 800°C are depicted in Fig. 32. The DRIFT spectra exhibit reactivity patterns
initiating from adsorbed CO and various types of carbonate on reduced catalysts.
The spectrum of Ce-Mo/Al2O3 exhibits an indicative band at 2198 cm-1 of CO adsorbed on
coordinatively unsaturated Al3+ sites (Aloct3+–CO), whereas a band at 2170 cm-1 is attributed
to CO linearly bonded to Ce4+ cations that is observed in all Ce-containing catalysts. On the
basis of spectral investigations of various carbonate compounds, which can be found
elsewhere [175-178], one can suppose that the surface of Ce-Mo/Al2O3 is covered with
various types of carbonate: monodentate (νs1420 and νas1540 cm-1), bidentate carbonate
(νas1320 and νs1680 cm-1) and bicarbonate (νs1460, νas1630 cm-1 and 1220 cm-1 γOH) with the
absence of Ce3+–CO band (2150 cm-1) in this spectrum.
72
The DRIFT spectra recorded on reduced CeO2 and Mo/CeO2 following CO adsorption show a
band at 2150 cm-1 assigned to CO linearly bonded to Ce3+ together with the appearance of
Ce4+-CO besides the band at 2025 cm-1 due to CO adsorbed to metallic Mo0 as mentioned
previously. Nevertheless, the absence of other Mo-CO bands may imply that Mo is almost
totally reduced and/or Mo-CO bonds were very weak and disappeared under vacuum or
formed carbonates via π-type electron backdonation [176-179].
One can notice that the reduced surface of CeO2 and Mo/CeO2 after CO chemisorption
exposes clearly monodentate carbonate (νs1430 and νas1570 cm-1), bidentate (νas1340, νs1680
and νas1320, νs1690 cm-1), bicarbonate (1220, 1230 cm-1 γOH, νs1460, νs1490 and
νas1630 cm-1), and bridged carbonate (νas1285 and νs1750 cm-1).
Of particular interest, the CO adsorption on Mo/CeO2 shifts the corresponding peak positions
of carbonates to higher wavenumbers implying the increase in CO surface coverage and
different distribution of electrons comparatively.
Fig. 32. In situ DRIFT spectra of CO adsorbed on CeO2, Mo/CeO2 and Ce-Mo/Al2O3 catalysts reduced at 800°C
73
3.5.2.3. CO chemisorption on SnO2, Mo/SnO2 and Sn-Mo/Al2O3 reduced at 800°C
The spectra of CO adsorbed on the samples are depicted in Fig. 33. The spectrum of SnO2
exhibits bands assigned to CO-Sn4+ (2237 cm-1) and CO-Sn2+ (2145 cm-1) entities that also
appear in the spectrum of Mo/SnO2, besides the bands at νas1257 and νs1597 cm-1 belonging
to bidentate species, while the bands at νs1387 and νas1480 cm-1 are likely arising from
monodentate species. The spectrum of Mo/SnO2 displays band positions at νas1247 and
νs1775 cm-1 with 1094 cm-1 γ(COO)– due to the formation of bridged carbonate, whereas the
bands at νs1445 and νas1585 cm-1 with 1197 cm-1 γ(OH) can be assigned to bicarbonate
species. In addition, the peaks appearing at νas1507 and νs1889 cm-1 indicate the presence of
free carbonates on the surface. A band located at 1994 cm-1 is related to bridged CO adsorbed
on metallic Mo(0) species (Mo(0)-CO) that also observed in the spectrum of Sn-Mo/Al2O3,
while the band with low intensity at 2089 cm-1 can be assigned to Mo4+CO and/or Mo3+CO.
On the other hand, the spectrum of Sn-Mo/Al2O3 shows bands at νs1432 and νas1623 cm-1
with 3634 cm-1 ν(OH) and 1227 cm-1 γ(OH) ascribed to bicarbonate species. Furthermore, the
peak located initially at 2057 cm-1 can be assigned to Mo2+CO and/or Mo1+CO sites.
Additionally, the new band at 2210 cm-1 is ascribed to Aloct3+–CO sites whereas the band
corresponding to CO-Sn4+ shifts to 2244 cm-1. However, it is worth mentioning that CO-Sn2+
band (2145 cm-1) is absent in Sn-Mo/Al2O3 spectrum. One may propose that Sn2+ ions (SnO)
are located in the internal surface of Mo/Al2O3 particles as separate phase and/or the CO-Sn2+
bond was weak so as to forming CO complexes (carbonates).
74
Fig. 33. In situ DRIFT spectra of CO adsorbed on SnO2, Mo/SnO2 and Sn-Mo/Al2O3
catalysts reduced at 800°C
Table 8. Assignments of IR bands observed upon CO chemisorption on metal cations
Assignment Wavenumber (cm-1) Reference Mo0 1994, 2002, 2025 188, 189
Mo3+/4+ 2048, 2089 12-15, 45, 117, 188, 189 Mo1+/2+ 2050, 2057 12-15, 45, 117, 188, 189 Ce3+/4+ 2150, 2170 80, 180-187
+3octAl 2194, 2197, 2198, 2210 20-28
Sn2+/4+ 2145, 2237, 2244 91-99
75
3.5.3. In situ DRIFT results on methane transformation in absence of oxygen The procedure of nonoxidative CH4 reaction on 50 mg of the catalysts was as follows:
1. The DRIFT reactor cell was attached to the glass gas handling/vacuum system.
2. Reducing in H2 flow (60 cm3/min, 1 bar, 10°C/min of heating rate) up to 800°C.
3. The system was cooled back to room temperature and evacuated at about 1.33x10-8
bar for 30 min.
4. Recording DRIFT spectra
5. Once, introduction of CH4 (110 cm3) into the reaction space (the whole volume
including the DRIFT reactor cell and gas-circulation system was 387.3 cm3) and the
reaction was carried out for 1 hr at 700°C and 4x10-2 bar.
6. CH4 gas circulation with the aid of a glass gas-circulation pump (built in the system
with max. speed of 200 cm3/min). Helium was added to improve the efficiency of the
gas circulation.
7. Recording DRIFT spectra every 20 min.
The interaction of methane with the surface of the catalysts reduced in hydrogen flow at
800°C has been studied for 1 hr in order to reveal surface intermediate species that allow a
better interpretation of the reaction pathway. The same amount of CH4 was introduced into
the reaction space during the experiments. Due to the limitations posed by CaF2 disks, only
bands appearing at wave number > 1000cm-1 can be detected in this experiment.
The samples were compared to determine precisely what kinds of adspecies participate
efficiently to carbon storage during CH4 transformation. Three series of temperature
programmed and time-dependent DRIFT spectra have been recorded under reaction
conditions on each of the following samples: 20 wt% Mo/Al2O3, 12 wt% Mo/CeO2, 5 wt%
Ce-20 wt% Mo/Al2O3, 4 wt% Mo/SnO2, 5 wt% Sn-20 wt% Mo/Al2O3.
76
3.5.3.1. Methane transformation on Mo/Al2O3 reduced at 800°C Fig. 34 shows the time-dependent DRIFT spectra of Mo/Al2O3 surface species formed after
exposure to methane. At 3013 cm-1 a discrete peak is observed due to the CH4 gas phase. The
bands at 3113 and 2920 cm-1 are assigned to νasCH and νsCH stretching, while the bands at
1355 and 1302 cm-1 attributed to asymmetric and symmetric deformation vibration of (CH)
(δasCH and δsCH) in CH4 gas. Additionally, the spectra showed two bands centred at νs1385
and νas1570 cm-1 due to symmetric and asymmetric vibration of surface formate that was
further emphasized by the presence of the band at 2730 cm-1 assigned to ν(CH) stretching.
On the other hand, the spectra exhibit bands at 2110 and 2190 cm-1 that are likely attributed to
roto-vibrational frequencies of CO gas phase resulting from decomposition of formate species
and/or from a reforming reaction between the carbonaceous deposits generated by CH4
decomposition and oxygen vacancies on the surface. Furthermore, the band at 1265 cm-1 is
likely assigned to twisting vibration mode of δe(-CH3 methyl), besides the band at 1113 cm-1
pertaining to γ(OH). More specifically, the intensity of the peaks assigned to CH4 gas phase
decreased slightly as the reaction proceeds implying that methane conversion is low.
Fig. 34. In situ DRIFT spectra of CH4 reaction on Mo/Al2O3 reduced at 800°C
77
3.5.3.2. Methane transformation on Ce-Mo/Al2O3 reduced at 800°C Fig. 35 shows the DRIFT spectra of reduced Ce-Mo/Al2O3 during the interaction with
methane, respectively. In addition to the bands corresponding to CH4 gas phase (3013 cm-1)
and bands assigned to different vibrational modes of (CH) (3113, 2927, 1350 and 1302 cm-1)
as mentioned above, the spectra also show the presence of a small amount of CO2 (band at
2310 cm-1) in the first 20 min of CH4 reaction. In contrast, after 40 min of the reaction the
spectra exhibit two bands at νs1511 and νas1655 cm-1 originating from symmetric and
asymmetric vibration modes of carbonate species besides the bands at 2310 and 2355 cm-1
belonging to CO2 gas phase with higher intensity after 60 min of the reaction.
These carbonates appear in the form of bicarbonates after 60 min, which is confirmed by the
presence of the peaks at 1240 γ(OH) and 3737 cm-1 ν(OH), whereas the bands at 2110 and
2180 cm-1 are very likely due to CO gas phase, while the band at 1095 cm-1 liberated from
deformation vibration of γ(COO)¯ being adsorbed on the surface. However, the thermal
stability of bicarbonates is reportedly lower than that of other types of carbonates, their
decomposition produces CO2 [178-184].
Fig. 35. In situ DRIFT spectra of CH4 reaction on Ce-Mo/Al2O3 reduced at 800°C
78
3.5.3.3. Methane transformation on Mo/CeO2 reduced at 800°C The time-dependent DRIFT spectra of CH4 decomposition over reduced Mo/CeO2 are
depicted in Fig. 36. The spectra exhibit discrete bands at 3113, 3013, 2920, 1355 and 1302
cm-1 corresponding to different vibration modes of CH4 gas phase as mentioned previously.
The infrared spectrum after 40 min of reaction exhibits peaks at νs1480 and νas1553 cm−1,
with a band at 2737 cm−1 ν(CH) due to the formate species, respectively. While those
assigned to bicarbonates are observed at νs1520, νas1657 cm-1 that was further emphasized by
OH peaks at 3737 cm-1 ν(OH) and 1120 cm-1 γ(OH). The carbonate bands are essentially in
the same positions up to the end of the experiment. The spectra exhibit IR features near 2115
and 2192 cm−1 (CO gas phase), besides the bands at 2312 and 2353 cm-1 (CO2 gas phase)
indicating the presence of CO and CO2 during CH4 decomposition. These bands become more
prominent after 60 min implying the high conversion of CH4 that was confirmed by the
gradual decrease of the band intensity corresponding to the CH4 gas phase (3013 cm-1).
However, it is possible that formate species are adsorbed on Mo while bicarbonates on Ce but
the beneficial effects might have occurred either because the cooperation between the partial
CeOx and Mo generated sites with higher activity, and/or because the oxidative properties of
CeO2 increased the dissociation of CH4.
Fig. 36. In situ DRIFT spectra of CH4 reaction on Mo/CeO2 reduced at 800°C
79
3.5.3.4. Methane transformation on Sn-Mo/Al2O3 reduced at 800°C The time-dependent DRIFT spectra of Sn-Mo/Al2O3 during CH4 reaction are shown in Fig.
37. Several bands are due to fundamentals, overtones vibrations of CH4 gas phase (3113,
3013, 2930, 1350 and 1303 cm-1) as mentioned in the previous sections. It can be seen that
CO and CO2 (2100, 2180 and 2360 cm-1 bands) appear after 40 min. These bands are more
prominent after 60 min of CH4 reaction. Nevertheless, after 60 min the spectra present bands
at νs1370, νas1587 and 2860 cm-1 ν(CH) indicative of formate species, while the band at 1060
cm-1 is liberated from deformation vibration of γ(COO)¯, whereas the bands at νs1470, νas1620
and 3673 cm-1 ν(OH) are practically arising from bicarbonate species. However, the presence
of the former species can be evaluated in analogy to what has been observed for CH4
decomposition on Mo/Al2O3 and Ce-Mo/Al2O3. It is worth pointing out that the intensity of
the peaks corresponding to CH4 gas phase decreased slightly till the end of the experiment
implying that the CH4 conversion is low.
Fig. 37. In situ DRIFT spectra of CH4 reaction on Sn-Mo/Al2O3 reduced at 800°C
80
3.5.3.5. Methane transformation on Mo/SnO2 reduced at 800°C Fig. 38 shows the time-dependent DRIFT spectra of surface species formed upon contact of
Mo/SnO2 with CH4. In addition to the bands belonging to CH4 gas phase (3113, 3013, 2940,
1350 and 1303 cm-1) the DRIFT spectra exhibit a discrete band at 1680 cm-1 with the band at
1190 cm-1 γ(H-CO) are typical for ν(C=O) stretching in coordinatively bonded formaldehyde
on the Lewis acid sites. Accordingly, this suggests that the first detectable adsorbed species to
be adsorbed formaldehyde intermediate accompanied with the presence of a large amount of
CO2 (2310 and 2360 cm-1). However, formaldehyde completely disappeared after 40 min of
CH4 reaction while the intensity of CO2 band gradually increased till the end of the reaction.
The spectra also show the liberation of small amounts of CO (2110 and 2187 cm-1) after 40
min. Furthermore, the bands located at νs1540, νas1620 and 3651 cm-1 ν(OH) are associated
with IR characteristic features of adsorbed bicarbonate species.
Since as can be seen, the intensity of CH4 gas phase peaks gradually decreased even after 40
min and completely disappeared by the end of the experiment demonstrating that the CH4
conversion is very high and almost complete CH4 oxidation was achieved.
Fig. 38. In situ DRIFT spectra of CH4 reaction on Mo/SnO2 reduced at 800°C
81
3.5.3.6. DRIFT spectra after methane reaction under vacuum As can be seen from the spectra obtained after the cell evacuation for 20 min at room
temperature (r.t.) and 1.33x10-8 bar, some formate and carbonate species are still adsorbed on
the surfaces (Figs. 39 and 40). Therefore, the spectrum of Mo/Al2O3 displays bands at νs1455
and 2857 cm-1 ν(CH) assigned to formates, while the band at 1626 cm-1 is due to the
deformation vibration of water (H-OH) with the band appearing at 1116 cm-1 is assigned to
γ(OH) being in the bulk of the catalyst, which shifted to higher wavenumber (1125 cm-1) in
the spectra of Ce-Mo/Al2O3 and Mo/CeO2 (Fig. 39).
In addition, the spectrum of Ce-Mo/Al2O3 exhibits new bands at νs1437 cm-1 assigned to
carbonate, whereas the band at 2171 cm-1 is very likely assigned to Ce4+-CO that also
appeared in the spectrum of Mo/CeO2. In contrast, the spectrum of Mo/CeO2 also reveals
bands located at 1437, 1650, 1734, 1839, 2873 and 3747 cm-1 are closer to the bands positions
for formates (νs1437, νas1650, and ν(CH) 2873 cm-1) and bridged carbonates (νs1734 and
νs1839 cm-1) being adsorbed on different sites (Mo and Ce).
The spectrum of Sn-Mo/Al2O3 exhibits bands at νs1453, νas1587 and 2789 cm-1 ν(CH)
indicative of formate species whereas the spectrum of Mo/SnO2 shows bands at νas1267 and
νs1878 cm-1 assigned to bridged carbonates. Moreover, the band observed at 1053 cm-1 is due
to the deformation vibration of γ(COO)-, while the band at 1127 cm-1 is assigned to γ(OH),
beside the band at 1633 cm-1 observed in the two spectrums is due to the deformation
vibration of water (H-OH) (Fig. 40).
Of particular interest, the broad hydroxyl band reappears in all spectra with lower intensity
comparing with the spectra of the calcined catalysts besides new OH groups (1116, 1125,
1127 and 3747 cm-1) are still present even after evacuation, that may be explained by the
presence of strongly bound hydroxyl groups encapsulated in the oxide bulk (Figs. 39 and 40).
82
Fig. 39. In situ DRIFT spectra of Mo/Al2O3, Mo/CeO2 and Ce-Mo/Al2O3
after CH4 reaction under vacuum
Fig. 40. In situ DRIFT spectra of Mo/SnO2 and Sn-Mo/Al2O3
after CH4 reaction under vacuum
83
4. DISCUSSION 4.1. Surface texturing Inspecting of the data compiled in the Table 4 reveals the mesoporosity nature of these
materials. In addition, the computed values of mean pore radius of the different samples,
which are comparable to each other show the mesoporosity and decreasing the total pore
volume with increasing the pore diameter of all samples comparatively (Figs. 7 and 9) may
indicate that their pores are narrow as well as they are deep [70-72].
Therefore, it is worth mentioning that the significant decrease of the surface area of the
samples is rather due to blocking pores by the oxides added (MoO3, SnO2, CeO2). On the
other hand, this may also demonstrate that molybdena is well-dispersed on the supports.
4.2. X-ray diffraction The presence of Al2O3 and the Mo dispersion obtained on Al2O3 may suppress stronger
interaction between Mo and Ce and may contribute to particles agglomeration not allowing
the material to maintain its dispersion. This was confirmed by the presence of only one
crystalline structure of cerium-molybdate (Ce2Mo3O12) in addition to the decrease in the
intensity of XRD lines assigned to Al2(MoO4)3 and MoO3 in the Ce-Mo/Al2O3 sample, while
three crystalline structures of cerium-molybdate were identified in Mo/CeO2 sample (Fig. 11).
The substitution of some Ce4+ ions by Mo6+ favours the formation of defects in the ceria–
molybdena lattice that induce a distortion of the oxygen sub-lattice. This is because the ionic
radius of Mo6+ (0.62 Å) is too small compared to that of Ce4+ (1.01 Å) to accommodate all the
oxygen. This distortion increases with the Mo content and is responsible for the progressive
change of the symmetry of solid solutions. It is not possible to identify clearly the
composition at which phase separation occurs, as this depends on several factors including
sample preparation and treatment, although generally ceria-rich composition crystallizes
easily with a cubic symmetry while intermediate and molybdena-rich compositions prefer an
octahedral phase. From this picture, it is expected that solid solutions having the highest
concentration of Mo will show the best redox behaviour. However, this is not completely true,
since a high MoO3 content will reduce the quantity of active redox elements. Therefore, a
detailed balance between structural defects and Ce or Mo content must be reached for
optimum dispersions that in turn affect the structure of these systems [65, 76-79].
Combining the results obtained from XRD and BET investigations, it can be anticipated that
the introduction of cerium ion promotes aggregation of the Mo particles mainly when using as
84
a support with high Mo loading as emphasized by diminishing in the surface area and total
pore volume besides the formation of different molecular formulae between Mo and Ce.
Thus, the interaction between molybdena and ceria probably due to charge effects and may be
based on the fact that CeO2 possesses a strong basic property. This indeed affects the particle
size of CeO2 and MoO3 crystallites. More specifically, in principle decreasing the particle size
of the two interacting ions helps their diffusion and thus their mutual interaction that is highly
affected by the mode of the preparation that in turn affects the thermal stability and
crystallinity of the produced compounds.
In conformity, IR and XRD studies by M. M. Mohamed et al. have emphasized that the CeO2
substrate promoted the aggregation of Mo ions up to 8 wt% loading when impregnation
method was adopted for preparation [80].
The use of SnO2 as promoter in Mo/Al2O3 leads to surface structure definitely different from
that of Mo/SnO2 (Fig. 13). We may tentatively attribute the discrepancy between the two
solids to sintering through condensation of the external, octahedral molybdenum layers
between different particles during the calcination of Sn-Mo/Al2O3 sample. While the
superficial molybdate ions could inhibit sintering of the SnO2 support during the calcination
of Mo/SnO2. However, stannous and stannic oxide coexist frequently either due to an oxygen
loss associated with the reduction of SnO2, or to the oxidation of SnO. Two main mechanisms
are responsible for sensing: (1) The bulk diffusion of oxygen from outside into the oxide,
compensating an original deficiency of oxygen, which is typical of most oxides. (2) The low
temperature chemisorptions of environmental gases on the surface of multiple grains,
charging the surface state and charge distribution inside the grain. The oxygen characteristics
at the grain surfaces of the porous SnO2 materials are strongly dependent on the surrounding
gas atmosphere [90-99].
Hence, it can be anticipated that no linkages were observed between Mo and Sn ions that
presumably occur at temperature higher than 700°C. Additionally, the high Mo loading leads
to the decrease of SnO2 crystal size (the decreased peaks intensity assigned to SnO2 in the two
catalysts) probably due to polarization effects and the presence of molybdate species strongly
affects the growth of SnO2 crystals in line with some results in the literature indicating that
Mo disturbs the SnO2 crystallization mainly at higher Mo loading [100-107].
In conformity, F. Goncalves and co-workers showed that the presence of the cassiterite phase,
the crystal size and surface area strongly depends on the molybdenum loading. It was shown
that the presence of molybdenum oxides inhibits crystal growth of the SnO2 support, leading
to material with a high specific area and high dispersion of molybdenum and tin at high
85
molybdenum loading when co-precipitation and impregnation methods were adopted. They
showed a pronounced influence of the preparation method on the crystallinity and catalytic
properties of Mo-Sn system [101].
N. G. Valente and co-workers have shown that no phase transition takes place at temperatures
up to 700°C of MoO3 and SnO2 mixed mechanically. They claimed that there are synergetic
cooperations between SnO2 and MoO3 to explain the catalytic behaviour in the oxidation of
methanol. So that MoO3 crystals would be oriented on the surface of SnO2 creating
catalytically active sites that would not be present in the isolated phases [105].
Anyhow, the fact that doping of transition metal oxides to SnO2 dramatically influences the
defect and the sintering behaviour of tin oxide. Nevertheless, MoO3 with SnO2 exhibits many
interesting features. This is the reason for the numerous studies of their physical and chemical
properties [101-109].
4.3. Thermal Analysis Several authors have studied the mechanisms of the thermal decomposition of ammonium
heptamolybdate (AHMT). The decomposition process took place via the formation of
different intermediate compounds, which decompose readily yielding solid MoO3 (Eqs. 24-
26). However, similar results were obtained by A. Said and S. A. Halawy [160] who found the
three mass losses corresponding to the thermal decomposition of pure AHMT in nitrogen flow
that were slightly different (7.2, 4.3 and 6.8%) from the results obtained in the present work,
respectively. Details of the mechanism of the thermal decomposition of ammonium
heptamolybdate yielding MoO3 have been given elsewhere [80, 160-162].
Some authors studied SnO2 for its capacity to dissolve cations such as Sb and Mo. Okamoto et
al. suggested [84] through quantitative XRD and catalytic study that SnO2 could dissolve
Mo6+ in spite of charge unbalance leading to high catalytic activity of the Mo/SnO2 system in
the oxidative dehydrogenation of sec-butyric alcohol. While other results did not show
significant dissolution of Mo6+ in the SnO2 structure even at high Mo concentrations after
either calcination or reduction up to 700°C [102-108].
Anyhow, it is worth pointing out that the dissolution of some molybdena in SnO2 was
observed by the end of this experiment and H2SO4 was used to clean the sample cup because
of the adhesive form of the sample. This confirms the hypothesis of significant molybdenum
dissolution in tin oxide above 750°C.
86
However, probable transformation of the Mo-Sn oxide system needs further studies as well as
the formation of new phases (e.g. formation of a new oxide phase associating Mo with Sn)
during the thermal treatment of this system above 700°C.
Comparing the TG-DTA results, it can be anticipated that Mo/Al2O3 is the thermally most
stable material in the temperature range used in the experiment (900°C). Whereas
Ce-Mo/Al2O3, Mo/CeO2 and Mo/SnO2 samples undergo morphological and textural
modifications throughout the thermal behaviour above 700°C resulting in lattice defects
(Schottky and Frenkel defects) which motivate the mobility of Mo, Ce and Sn ions and thus
enhance the possibility of interaction between them (Figs. 18-22).
However, the fact that the five samples showed varying net mass loss implying that they have
not comparable molecular formulae. Accordingly, although the type of active sites is more or
less similar in all samples, different reaction pathways may result in different activities.
4.4. Electron Spin Resonance (ESR) The EPR spectra of the samples showed the presence of Mo5+ and Ce3+ paramagnetic centers
after CO interaction at both temperatures (r.t. and -196°C). One may notice that the two types
of EPR spectra (after CO interaction at both r.t. and -196°C) are quite similar for all samples
but more prominent and in some cases (e.g. for Mo/Al2O3) broader at -196°C (Figs. 23-26).
According to the literature, the increase of the line width is due to the clusterisation of the
paramagnetic centers in the sample. On the other hand, if any oxygen inside the sample tube,
it will condense at the liquid nitrogen temperature. Therefore, the condensed oxygen adsorbs
on the surface of the solid sample. Consequently, the interaction between the oxygen dipoles
adsorbed on the surface may lead to the broadening of the EPR lines [166-174].
In conformity, O. Cozar et al. and others found the six lines of Mo in their ESR study of
molybdenum-lead-phosphate glasses. Furthermore, the existence of three types of symmetry
sites for Mo5+ ions has been shown. The dipole–dipole and super exchange coupled Mo5+ ions
appeared and their number increased with the Mo content in samples heat-treated in nitrogen
atmosphere [166-169].
B. Kamp et al. studied by ESR spectroscopy the oxygen defects on SnO2 in oxygen
atmosphere at elevated temperatures and pressures and it was found that oxygen defects on
SnO2 occur above 700ºC [170, 171].
F. Morazzoni et al. in either SnO2 or Pt/SnO2 samples observed no paramagnetic species
corresponding to SnO2 after treatment with CO/Ar. The EPR spectra showed only a
symmetrical resonance line that was attributed to an oxygen defect (Eqs. 28-29) [172].
87
4.5. In situ DRIFT spectroscopy 4.5.1. DRIFT spectra of the calcined samples under vacuum Although the band positions of such surface hydroxyl groups can be distinguishable by
recording and deconvoluting IR spectra in 3000-3800 cm-1 region upon adsorption of
molecules such as pyridine or upon dehydroxylation at elevated temperatures. Details of OH
yield on alumina and ceria have been given elsewhere [18-31, 180-187].
For instance, concerning OH groups on CeO2, M. I. Zaki and co-workers identified and
differentiated acidic and basic OH’s at 3583, 3621, 3652 and 3684 cm-1 involving in pyridine
adsorption on CeO2 [180].
The comparison between the spectra of 8 wt% Mo/Al2O3 and 20 wt% Mo/Al2O3 (Figs. 28 and
29) permits to infer that free MoO3 clusters and polymeric molybdate species can only be
formed at higher Mo loading in line with some results in the literature [12-17, 32-35].
As the result of the proposed interaction between molybdena and ceria on the basis of the
strong basicity possessed by CeO2, this interaction is further emphasized by forming
Mo-O-Ce linkages that were represented by the bands at 630 cm-1 νs(Mo-O-Ce) and 875 cm-1
νas(Mo-O-Ce) in addition to the bands at 995 and 1035 cm-1 associated with the formation of
coupled νs(O=Mo=O) bonds (Fig. 28). Accordingly, ceria can attribute to the increase of
polymerized surface Mo species. Nevertheless, the expected competition between Ce and Mo
with alumina hydroxyl groups may suppress stronger interactions between them (decreased
intensity of OH region in Ce-Mo/Al2O3 spectrum).
Furthermore, no bands arising from Mo–O–Sn vibrations were observed in both spectra of
Mo/SnO2 and Sn-Mo/Al2O3 samples (in line with XRD results). This means that the
molybdenum oxide may spread over the tin oxide readily after calcination at 600°C (Fig. 30).
According to literature sources tin can form –Sn(OH)2–O–Sn(OH)2–O– type polymer chains
in which the OH groups remain stable even at elevated temperatures [82-90]. The
concentration of surface hydroxide of tin oxide may be consistent with the formation of solid
acidity, because its concentration was found to be high up to 700°C of the calcination
temperature of tin oxide. Several bands in 3300-3800 cm-1 region due to fundamentals of OH
yield on SnO2 have been given elsewhere [91-99].
88
4.5.2. CO chemisorption It is reasonable to suggest that the CO adsorption on Mo/Al2O3 reduced at different
temperature occurs now mainly on small crystallites of MoO3. However, the presence of
formate and carboxylate indicates that the adsorbed CO on Mo reacts with the geminal
hydroxyls on the surface to form these species implying the high reactivity of the lattice
oxygen such as Mo=O and Mo–O–Mo. This means that CO adsorption involves the lattice
oxygen of Mo activated at different sites of the surface throughout the reduction pretreatment.
Nevertheless, the formation of formate species may proceed through a surface reaction
between the weakly adsorbed CO to Mo species and a reactive H atom adsorbed on the
surface upon H2 reduction via (LH) mechanism leading to the formation (with the
participation of lattice oxygen of MoO3 and/or hydroxyls) of various carbonate species which
become more notable on the surface of Mo/Al2O3 reduced at 800°C (Fig. 31). Accordingly,
this supports the following reaction steps:
Scheme 4.
Hence, one may suggest that the presence of MoO3 clusters makes the surface eligible to
liberate various carbonates upon CO adsorption at 100°C. It can be anticipated that there is an
approximately linear correlation between the increase of the extent of reduction and the
increasing integrated absorbance of CO adsorbed on 20 wt% Mo/Al2O3 catalyst (Fig. 31).
On the other hand, the reduction of the catalysts to produce lower oxidation states of Mo and
more Ce3+ species activated the surface to be eligible for CO chemisorption as metal-CO as
well as carbonate species. Thus, the chemisorption of CO involves oxygen in the catalysts
such oxygen could be present in Ce–O–Mo and/or Ce–O–Ce associates. The noticed gain in
intensity for the bands in conjunction with the carbonate species is formerly observed. These
species are relatively stable under vacuum at room temperature (Fig. 32).
It has been revealed that the hydrogen reduction of Ce and Mo at 800°C (when Ce and Mo are
almost totally reduced) improves the surface reactivity as well as the stability of carbonates,
89
and leads to the presence of small amounts of metallic Mo species after reduction at 700°C.
One may suggest that the intervention of the cerium couple (Ce3+/Ce4+) was appreciable in
Mo/CeO2 following the reduction process and CO adsorption probably due to forming
stabilized species of ceria oxide and molybdena-cerium [54, 65, 80].
However, the absence of Ce3+–CO band (2150 cm-1) in Ce-Mo/Al2O3 spectrum (Fig. 32)
indicates that Ce ions are preferentially located in the internal surface of Mo/Al2O3 particles,
and mostly in compensating positions substituting Mo ions in Ce-Mo/Al2O3, whereas for
Mo/CeO2 the majority of Ce ions are present on the external surface as cerium-molybdate and
CeO2 moieties in line with BET and XRD results.
On the other hand, CO-Sn2+ band (2145 cm-1) is also absent in Sn-Mo/Al2O3 spectrum. One
may propose that Sn2+ ions are located in the internal surface of Mo/Al2O3 and/or the CO-Sn2+
bond was very weak so as to forming CO complexes like carbonates (Fig. 33). Consequently,
CO being provided as weakly adsorbed metal-carbonyls migrating towards the oxides through
interfacial sites to form carbonates via a π-type electron backdonation [175-179].
Some infrared spectroscopy (IR) investigations of CO adsorption on ceria have found that CO
not only adsorbed weakly in vertical orientation as on other oxides, but also interacts strongly
with the CeO2 surface, forming carbonate and/or inorganic carboxylate like complexes even at
room temperature [180-183]. The formation of various carbonate species like complexes of
CO on ceria suggests that CO is oxidized by the lattice oxygen of ceria, or in other words, the
CeO2 surface is partially reduced by CO.
Studies by Liu et al. on the adsorption of NO2 and SO2 on zirconium doped ceria found that
the oxygen sites of the Ce(111) surface are able to react with NO2 and SO2 to form surface
sulphates and nitrates at r.t. [184]. This suggests that there is the possibility of forming surface
carbonates after adsorbing CO.
On the other hand, other experimental studies have found no evidence for chemisorption of
CO on ceria or they have found that chemisorption is only observed after surface
pre-treatment. For example, through photoemission studies of CO adsorption on ultra-thin
ordered CeO2 (111) and CeO2 (110) surfaces [185-187].
Berner et al. found that CO exposure does not affect unreduced CeO2 surface [185].
Z. Yang et al. found that the adsorption of CO exists different features on (111) and (110)
surfaces of CeO2. While only weak adsorptions were found on (111) surface, both weak and
strong adsorptions exist on (110) surface [186]. Electrostatic interactions were involved in the
weak interactions, while covalent bonding was developed in the strong adsorptions.
90
The geometries of the strong interaction modes for CO on Ce(110) surface all involved the
formation of carbonates on the surface. Their calculated results were in agreement with the
experimental results. These confirm experimental observations that CO not only adsorbed
weakly in a linear form (vertical orientation), but also interacts strongly with the CeO2 surface
forming considerable amounts of carbonate species even at room temperature. In their study,
only the stochiometric non-defective (110) and (111) surfaces of ceria have been examined.
Therefore, they concluded that the adsorptions of CO on the (111) and (110) surfaces have
distinctly different properties. On the (111) surface, only weak adsorption (physisorption)
modes exist, whereas, for the (110) surface, not only weak adsorption modes exist but also a
few strong adsorption modes exist that can be classified as chemisorption. In all cases, the
strong adsorption modes involved the formation of carbonate species on the surface.
All these conflicting results in the literature on whether or not surface complexes exist suggest
that the surface properties and the adsorption behaviour of CO on the ceria surfaces are
strongly related to the nature of the surfaces used in the experiments. Therefore, some
experiments have simultaneously observed both weak and strong interactions for the
adsorption of CO on cerium oxide, while others found only weak interactions. Why both
strong and weak interactions of CO on ceria occur in some cases, while only weak
interactions are observed in others remains unclear. Although some theoretical investigations
on stoichiometric bulk ceria and its surfaces have been published, there have been few
theoretical investigations on the adsorption of CO on ceria and no first-principles studies.
Thus, the mechanisms and atomic level understanding of the interaction of CO on ceria
surfaces have not yet been adequately addressed theoretically.
Considering the present results and taking into account some results concerning CO
adsorption on Mo, Ce and Al oxides. It can be anticipated that the interaction between Mo and
supports has an effect on the adsorption properties of CO, which may indicate the different
electronic effects in the catalysts. Within this context, intimate coupling of Mo and Ce ions of
different oxidation states has great facilities for electron exchange interactions. Thus, the
electron-mobile environment necessitated by redox reactions is established that has a great
share in enhancing the CO adsorption. On the other hand, the geometric effects also play an
important role for the Mo–support interactions, the variance of the reducibility may be caused
by the different locations of Mo on the supports or the different orientations of Mo
combination to the supports. The special Ce–Mo–Al interaction may be originated from both
the electronic effect that affects the CO adsorption and the geometric effect that may
contribute to the reducibility variance of MoO3.
91
Furthermore, different methods for sample preparation and pre-treatment will result in
surfaces with different orientations, different oxidation states (e.g. unreduced and reduced),
and varying degrees of surface contamination. Additionally, defects such as step edges,
vacancies and partial surface reconstructions allow the formation of surface carbonates upon
CO adsorption [176-187].
Accordingly, the active sites for CO adsorption in the five samples are not similar and the
relative intensities of the same band positions are different indicating a various distribution of
electrons. Moreover, the pronounced increase in absorbance can be easily interpreted as due
to a decreased concentration of free electrons and of electrons trapped in oxygen vacancies
upon CO adsorption. For instance, the creation of Sn2+ donor levels and oxygen vacancies
upon reduction and CO adsorption as well as the variations of point defects and carbonates
may be described as follows:
Upon H2 reduction:
SnO2 + H2 ↔ SnO + H2O (Eq-30)
SnO2 ↔ SnO2−x + xO (Eq-31)
O2-(lattice) +H2 ↔ VO
2- + H2O (Eq-32)
Sn4+ + VO2- ↔ Sn2+ + VO (Eq-33)
Upon CO adsorption:
CO(ads) + O2-
(lattice) ↔ (COO)– + VO (Eq-34)
The increase in electron concentration arises from:
VO ↔ VO+ + e− (Eq-35)
VO+ ↔ VO
2+ + e− (Eq-36)
Sn2+ ↔ Sn4+ + 2e− (Eq-37)
Where VO, VO+ and VO
2+ are the neutral and ionized oxygen vacancies, respectively,
according to the Kröger–Vink notation [177-179].
92
These results are in agreement with the literature. For instance, Popescu et al. and others
found that CO adsorption on SnO2 even at room temperature gives rise to CO–Sn2+ and CO–
Sn4+ end-on species and to various carbonate entities [99, 106-109].
Anyhow, the formation of carbonate species throughout the contact with CO implies the
existence of reactive oxygen on the surfaces after the reduction in H2 flow at 800°C, which
creates point defects (native and foreign atoms) that can act as both donors and acceptors.
The interaction of CO gas with these surfaces led to different changes in the lattice oxygen
contents on such surface, this by its turn changes the amount of adsorbed carbonate species.
Some discrepancies concerning CO adsorption on the five samples are practically due to
several factors such as the nature of the sample, the stoichiometry of oxides, the presence of
lattice defects, the size and shape of the particles and hydroxyl groups concentration.
However, the dissolution of Mo was also observed in Mo/SnO2 sample reduced at 800ºC by
the end of this experiment (here, again H2SO4 was used for cleaning).
4.5.3. In situ DRIFT studies on methane transformation in absence of oxygen It is generally accepted that methane is mainly activated on metallic surfaces. The electron
donation from the HOMO of CH4 to the lowest unfilled molecular orbitals of metal surface
should dominate dissociative CH4 adsorption. However, the fact that the carburization of Mo
species by CH4 is thermodynamically possible, therefore CH4 can interact with Mo vacancies
according to:
Mo[ ]Mo + CH4 ↔ Mo[C]Mo + 2H2 (Eq-38)
In this state, carbon can be extracted from the site, producing CO and regenerating the
vacancy, as shown:
(Eq-39)
H)x4(CHMoeCHMoCH x_
44 −+⎯⎯ →⎯+⎯⎯ →⎯ + (Eq-40) (Eq-41)
Mo[C]Mo· · ·OCO ↔ Mo[ ]Mo + 2CO (Eq-42)
93
Where (♦) and (■) represent adsorption sites on Mo oxide. Thus, the formation of CO from
CH4 would not be a simple dual site reaction involving adsorbed carbon and lattice oxygen,
but involves a solid-state reaction. This mechanism would be consistent with the high
activation energy of CH4 conversion on the supported catalyst.
Accordingly, the reduction of Mo by H2 at high temperature provides reactive H atoms able to
form OH groups and oxygen vacancies on the surface, particularly those present in the bulk,
are the driving force for CH4 activation. Furthermore, hydroxyls are also formed during the
activation step due to the reaction of the lattice oxygen with hydrogen as dihydrogen arising
from methane dehydrogenation or as water arising from the initial decomposition of methane.
It comes therefore that these rather basic OH groups are required for formate formation
supporting the following reaction steps:
H2(g) + 2Mo → 2Mo-H (Eq-43)
Mo-H + O2– → Mo + OH– + e– (Eq-44)
CO + OH– ↔ HCOO – (Eq-45)
Based on these results, it can be inferred that the loss of OH groups during CH4 conversion
does not necessarily implicate these species in the reaction to produce CO as the generation of
formate species from CO would also account for the consumption of hydroxyls (Fig. 34).
Lin et al. carried out a comparative FTIR study on the interaction of CH4 with silica, alumina,
and HZSM-5. The results demonstrated that OH groups played a very important role in CH4
adsorption. When an interaction between the OH groups and CH4 took place, the band shift of
the OH groups varied and the strength of the interaction decreased in accordance with the
order of their acidities (Si–OH–Al > Al–OH > Si-OH). The authors considered the possibility
that CH4 is activated by interacting with a proton leading to a heterolytic cleavage of a C–H
bond of CH4 [110].
Recent studies using supported and unsupported Mo compounds indicate that interactions
with methane at temperature around 700°C lead to the formation of Mo2C, which is
considered as the active site for the formation of CH2 and CH3 fragments. These carbides may
be destroyed by reaction with air or CO2 [111-113].
However, it is well-known that alumina support has the lowest oxygen mobility and non-
reducible alumina support is unable to store carbonaceous adspecies, the present results show
that molybdenum oxide can activate methane and oxidize it into surface formates and carbon
94
monoxide but methane conversion is low. Consequently, this CH4 partial oxidation involves
the interaction of methane with lattice oxygen anions and surface OH groups created by the
previous reduction step and/or upon the initial decomposition of CH4.
Iglesia and co-workers [114] claimed that selective silanation of external acid sites on
HZSM-5 by using large organosilane molecules could decrease the content of acid sites as
well as the number of MoOx species retained on the external surface, which were regarded as
key factors for coke formation during MDA. On samples prepared using silica-modified
HZSM-5, acid sites, MoOx precursors, and active MoCx species formed during the CH4
reaction at 677°C were found to predominately reside within the zeolite channels, where
spatial constraints could inhibit the bimolecular chain-growth pathways. Consequently, the
selectivity of hydrocarbons on a 4% Mo/silica-modified HZSM-5 increased by about 30% in
comparison with that on a 4% Mo/HZSM-5.
On ceria containing catalysts, to explain the fractional presence of CO2 and CO a mechanism
can be proposed (Figs. 35 and 36), as well as taking into account the carbon exchange
between CH4 and the surface. Surface reactions describe the proposed mechanism:
(Eq-46)
Meanwhile, the CO2 could adsorb on the carbon filled-site from the reaction above, extract
the carbon to produce CO, then the carbon is removed from the active surface replenishing a
vacancy according to the reverse Boudouard reaction:
(Eq-47)
Moreover, ceria could simultaneously undergo a redox reaction in the presence of CO2:
COCeO2COOCe 2232 +↔+ (Eq-48)
(Eq-49)
Where (■) represents adsorption sites on ceria, on the other hand, the CO could also reduce
the ceria by reversing reaction, and further undergo CO disproportionation as the consequence
of the occurrence of the Boudouard equilibrium:
(Eq-50)
95
However, such higher methane conversions are not necessarily related to the larger amounts
of coke and faster deactivation of the catalyst, because deactivation also critically depends on
the rate with which the coke can be removed by CO2 under reaction conditions when CO2 can
react with the CHx species as well as with CH4:
CO2 + CHx → 2CO + (x/2) H2 (Eq-51)
Therefore, in comparison with the CH4 reaction on Mo/Al2O3 the presence of CeO2
contributes to the rapid activation of CH4, thereby accelerating the carbon gasification
reaction to produce CO2. Thus, this supports the conclusion that CO2 is readily dissociated to
CO and adsorbed oxygen over reduced CeO2 by filling up the oxygen vacancies in Ce3+
species. The occurrence of Ce4+/Ce3+ redox couple generates oxygen vacancies and releases
free electrons. Free electrons transfer readily from Ce3+ to π* orbital of CO2 to activate CO2.
The increased CO2 then decomposes to CO and active surface oxygen, reacts with the CHx
species and enhances the catalytic activity of CH4 decomposition since with the aid of
Ce4+/Ce3+ redox couple, CO2 is more readily activated to release more surface oxygen, and the
rate of carbon elimination has been accelerated.
Analogous observations were made by Darujati et al. who found that Ce promotion
dramatically improves the stability of the Mo2C/γ-Al2O3 catalysts. They claimed that Ce acted
to increase the oxidation resistance of Mo2C and avoid coking by CO2 activation via the redox
reaction (Eq-48), thereby helping to prevent oxidation of Mo2C by CO2. From their study the
addition of ceria promoter to Mo2C/γ-Al2O3 catalyst appears to alter the dry methane
reforming (DMR) mechanism proposed earlier for bulk Mo2C catalysts by enhancing
relatively strong CO2 adsorption and the role of ceria was found to influence the redox
reactions on the surface as well as the activity and stability of the catalyst [115, 116].
On the other hand, although the presence of CO and CHX as well as carbonates over Rh/Al2O3
and Rh/Mo/Al2O3 catalysts has been observed by Anderson et al. during IR study on CH4
decomposition at 400°C, which also facilitated carbide formation [117].
R. Wang et al. have reported that an interaction between Rh and CeO2 was induced by high
temperature reduction, which resulted in the creation of oxygen vacancies in ceria. They
concluded that the CO2 activation in CH4/CO2 reforming should be mainly favoured by
availability of Ce4+/Ce3+ redox couple in Rh–CeO2/Al2O3 catalyst, which was rather slow
process on Rh/Al2O3 catalyst. The Ce3+ species readily promoted CO2 dissociation into CO
and surface oxygen. The higher catalytic activity and coke resistance of Rh–CeO2/Al2O3 were
96
associated with the presence of the two-redox couples favoring the activation of both CH4 and
CO2. Their catalytic test results showed that the promotional effect of CeO2 on CO2
conversion was much higher than that on CH4 conversion [118].
Given that oxygen mobility is high in ceria, allowing substantial reduction of the bulk, and
given that CO2 gas is present under reaction conditions, one need not invoke surface mobility
of adsorbed CO or CO2 in order to explain the apparently high coverage of carbonate.
Formation of CO2 takes place in two distinct moments: a fraction of the CO is rapidly
oxidized to CO2 and the CO2 desorbs readily from the catalysts, while the remaining fraction
of CO/CO2 is slightly adsorbed and accumulated on the catalyst forming various carbonate
species (Eqs. 47-51).
When ceria and molybdena have been reduced at high temperature (>700°C), oxygen
vacancies, particularly those present in the bulk, seem to be the driving force for CH4
activation. Therefore, when CeO2 is exposed to H2 or CO, oxygen vacancies VO2- with two
electrons trapped have been created. Such a vacancy is a neutral entity with respect to the
surface lattice of ceria. It can easily lose an electron by spontaneous ionization becoming
singly positively charged with respect to the solid as follows [118-121]:
VO 2OOHH 222
)lattice(−+→+− (Eq-52)
−+ +−→− eOO VV 22 (Eq-53)
2CeO2 + H2 ↔ Ce2O3 + H2O (Eq-54)
Reduction of CeO2 with hydrogen is generally thought to occur via a stepwise mechanism,
first reduction of the outer most layers of Ce4+ (surface reduction), then reduction of the inner
Ce4+ layers (bulk reduction) at higher temperatures. A few mechanisms have been put forward
to account for this behaviour that comprises sequential steps of: (i) dissociation of
chemisorbed hydrogen with formation of OH groups, (ii) formation of anionic vacancies with
desorption of water by recombination of H and OH (with concomitant reduction of Ce4+ to
Ce3+) and (iii) diffusion of surface anionic vacancies into the bulk. This picture is consistent
with results obtained by Trovarelli and others upon the temperature programmed reduction
with hydrogen of high surface area CeO2 when the TPR profile showed two well-defined
peaks centred at approximately 600 and 800°C [119, 120]. Hence, the ability of ceria to be
97
easily reduced to nonstoichiometric oxides is related to the properties of fluorite structured-
mixed valence oxides to deviate from stoichiometry (Fig. 12).
When considering the hydrogen conversion, the dissociation is faster and occurs at much
lower temperature (200°C) than for the alkanes (> 600°C). Hence, the cleavage of the H-H
bond is likely fast and is probably not a rate-limiting step such as the cleavage of C-H bond of
the alkanes. It may also appear on different active sites. Therefore, the conversion of
hydrogen is able to maintain its level with time, despite the production of water.
As it was pointed out previously, CH4 transformation over Mo/CeO2 can be explained by
invoking a redox mechanism with a simple redox route for CH4 oxidation, which utilizes
oxygen activated from the support in a typical reduction/oxidation mechanism (Mars Van
Krevelen type) in which the catalyst undergoes a partial reduction by methane [124-127].
Oxygen storage is therefore important because it provides an alternative route for the
oxidation of CH4. An alternative redox route involves oxygen from the support, which reacts
with methane to form adsorbed CO2 in the form of carbonates. Decomposition of carbonates
is then stimulated which provides also reoxidation of the support:
22Ox22 xHxCOeVCeOMoCeOMoCH4 ++−+++−↔−+ −− (Eq-55)
)xy(COCeOVCeOCOeVCeOCO y22Ox222Ox22 ⟨+↔++↔−+++ −−−−
−−++ (Eq-56)
However, Ce4+/Ce3+ redox couple facilitates the elimination of CHx species by partial
oxidation, resulting in higher methane conversion and lower amounts of coke. This partial
oxidation of CHx species over Mo/CeO2 will continuously result in the creation of oxygen
vacancies and Ce4+/Ce3+ redox couple. Thus, CH4 decomposition acts as the supplier of a
hydrogen pool, while Ce4+/Ce3+ redox couple promotes CO2 activation by accepting electrons
and replenishing the oxygen vacancies. This demonstrates that the reoxidized ceria can be
reduced again by methane, regenerating the oxygen vacancies and releasing free electrons, so
the creation of oxygen vacancies of ceria is a reversible process in the reaction atmosphere.
Accordingly, the presence of ceria in the catalysts as either promoter or support leads to
significantly higher methane conversion especially on Mo/CeO2 by decreasing in C storage
capacity and therefore an increase in the CO2 release upon the decomposition step. This effect
is likely to favour carbon trapping via carbonates as shown previously by DRIFT
spectroscopy. However, the mean CO concentration was not affected by the ceria or tin since
it is essentially controlled by the enthalpy of desorption from the metal phase [117].
98
Regarding the intimate atomic mechanism involved in oxidation of carbon, several authors
pointed out the importance of redox properties of the catalyst. That is, the effectiveness of the
catalyst can be related to its ability to deliver oxygen from the lattice to carbon reactant in a
wide temperature range. Recently, it has been reported that the use of supports based on CeO2
confers interesting properties to CH4 decomposition catalysts due to high availability of
surface oxygen and high surface reducibility. Nevertheless, analysis discrepancies in the
outcome of the results from different laboratories derive from synthesis and treatment
procedures [116-121, 145, 146].
For instance, Craciun et al. studied the CH4 + H2O reaction over Rh, Pt and Pd supported on
ceria catalysts [145]. In their study, they proposed a mechanism, which involved a surface
reaction of the adsorbed oxygen on the ceria with the dissociated methane on the surface.
Their study led to the conclusion that oxygen transfer from the ceria to the noble metals was
the rate determining step, where the participation of lattice oxygen and catalyst reducibility
have shown to improve overall performances.
Similar results over Pt/CeO2 were reported by Otsuka et al. who found that the oxidation of
CH4 by CeO2 was thermodynamically available at above 600°C. The reduction degree of
CeO2 was significantly improved from 3.5% to 17.1% in the presence of Pt after the reaction
with CH4 [146].
Concerning the CH4 dissociation on Mo/SnO2, the generation of formaldehyde intermediate
upon CH4 reaction may indicate that the Mo/SnO2 has a high concentration of Lewis acid sites
(Fig. 38). However, since the free CH2O was not detected in the gas phase that undoubtedly
appears at around 1730 cm-1 [175-179], the absence of formaldehyde may be due to its low
concentration and/or due to surface reactions by extracting lattice oxygen to form additionally
CO and CO2. These steps are represented by the following reaction steps:
CH4 + 2MO → CH2O + H2O + 2M (Eq-57)
CH2O + MO → CO + H2O + M (Eq-58)
CH2O + 2MO → CO2 + H2O + 2M (Eq-59)
Where MO represents metal oxide surface site. Moreover, CO and H2 can also be produced
via the pyrolytic decomposition of CH2O:
CH2O → CO + H2 (Eq-60)
99
Meanwhile, exclusion of a direct pathway from CH4 to CH2O allows CO and CO2 to be
treated as a single product (CO)x in the analysis of the kinetics of CH2O formation and
consumption. Since the interconversion of CO and CO2 was not investigated, no attempt was
made to include the effects of this process in the modelling of CH2O formation and
consumption on Mo/SnO2. The following mechanistic reaction pathway can be proposed:
2,32
41
4 COCOOMHCHOlatticeOxCHnabstractioHCH ⎯⎯⎯ →⎯ −⎯⎯⎯⎯ →⎯−⎯⎯⎯⎯ →⎯
Scheme 5.
Accordingly, it is argued that step 1, hydrogen abstraction, is stated to be favoured by lattice
oxygen possessing more negative charge. The general trends observed with formaldehyde
selectivity were explained on the basis of electrophilicity of adsorbed oxygen enhancing the
rate of step 2 with respect to step 3. Consequently, it can be stated that acid-base bifunctional
catalysts would be more effective for formaldehyde production [100-106].
Niwa and Igarashi correlated the acidity and reducibility to the catalytic behaviour of
Mo/SnO2 system in the oxidative dehydrogenation of methanol converted into formaldehyde
selectively. They found that the generation of acid sites is also strongly affected by the
calcination temperature of tin oxide that affects by its turn the formation of acid sites on the
loaded molybdenum oxide [100].
Smith and Ozkan [190, 191] studied the partial oxidation of methane to formaldehyde over
MoO3 samples exposing different relative amounts of (010) and (100) plane areas. Their
experimental characterization studies suggest that the Mo=O sites residing preferentially on
the side planes could be promoting to the formation of formaldehyde, while the bridging sites
Mo–O–Mo mainly on the basal plane were more likely to lead to complete oxidation of CH4.
S. Chempath and A. T. Bell studied the partial oxidation of CH4 at 700ºC on Mo/SiO2
catalyst. They found that CH2O is the only initial product. As the CH4 conversion increased,
the CH2O selectivity decreased and the selectivity for CO and CO2 increased [192].
Finally, the present results permit to infer that Mo/SnO2 showed the highest CH4 conversion
among the catalysts studied leading almost to complete CH4 oxidation (Fig. 38). On the other
hand, the results also demonstrate the activity of the molybdena and tin lattice oxygen and its
participation in the reaction under study conditions.
100
One may suggest that the high catalytic activity of the Mo/SnO2 system towards CH4 might be
associated with the dissolution of Mo ions in the SnO2 crystals, since this dissolution was
observed at the end of the experiment and H2SO4 was used to clean the sample cup due to the
adhesive form of the sample.
The activation of methane is believed to be the key step in the conversion of methane.
According to Lunsford, methane activation occurs homolytically via the abstraction of a
hydrogen atom by oxygen anions present on oxide surface [153-155]:
[ ] ⋅+−→+−− 3CHOHCatCHOCat 4 (Eq-61)
This first hypothesis has been substantiated by the presence of methyl radicals on the surface.
The methyl radicals so formed on the surface could react with the catalyst and produce
methoxide ions, which on a subsequent reaction with water yield methanol. Further oxidation
of methanol or dehydrogenation of methoxide ion leads to the formation of formaldehyde.
However, the second hypothesis has been proposed by Sokolovskii and co-workers [193].
They suggested that methane activation proceeds heterolytically via the participation of acid-
base centers giving place to a proton detachment and the formation of a metal–methyl
compound, where the methyl results in being negatively charged:
−++−−+ +→+ 22
3422 OHMeCHCHOMe (Eq-62)
The coordinately unsaturated metal and ion paired with a strong nucleophile (O2-) ion may act
as an active center. Further oxidation of these surface methyl anions would lead to methyl
radicals, which then dimerize:
−+⋅+− ++→ eMeCHMeCH 2
32
3 (Eq-63) At the same time, the authors mentioned the possibility that some methyl radicals could
escape to the basal plane and dimerize subsequently in ethane or, in presence of additional
electron–hole pair excitations, to produce formaldehyde. In addition, the barrier to methyl
radical diffusion over O2- sites is approximately the same as the desorption energy (0.4 eV).
Unless CH2O is desorbed, additional electron–hole pair excitations should lead to further
dehydrogenation and surface reduction with the formation of CO and CO2.
101
Moreover, the results of B. Irigoyen et al. [194] obtained from their theoretical and
computational study of the CH4–MoO3 chemical interaction suggest that while the hydrogen
abstraction requires less energy for each consecutive step the overall process remains
endothermic. The different sequences and sites for hydrogen abstraction from methane,
analyzed over different layers exposing molybdenum and oxygen atoms, allowed them to
conclude that the heterolytic H-abstraction is an energetically more favorable process in
comparison to the homolytic one. Their results indicate that despite an important energy
barrier being necessary for the first C–H bond activation, the overall oxidation process is
kinetically more favoured in the heterolytic mechanism.
Other studies suggest that the activation of methane occurs on super acid catalysts as well as
on organometallic complexes at low temperatures via heterolytic cleavage of the C-H bond of
CH4 [195-197]. In addition, the controlled activation of the C–H bond of methane and the
formation of the C–C bond have been extremely important and common topics in transition
metals and methane chemistry, as well as in homogeneous and heterogeneous catalysis. In
fact, transition metal centers play crucial roles in the recent development of promising
catalytic systems for C–H activation reactions in heterogeneous catalysis, and considerable
mechanistic insight has been gained [123-157, 190-197].
However, the five samples were compared to determine precisely what kinds of adspecies
participate efficiently to C storage upon CH4 transformation. This carbon storage on the
reduced Mo/Al2O3 catalyst occurs through the formation of formate species, which in turn can
either desorb into the CO gas phase or be stored on the surface that can only be formed thanks
to oxygen provided by surface lattice oxygen reacting with carbonaceous species provided by
the methane decomposition. It can therefore be deduced that the formation of formates needs
reduced molybdena containing oxygen vacancies and contributes directly to an efficient C
storage, though it requires the reaction of the surface oxygen with hydrogen atoms arising
either from hydrogen reduction (Mo–H) or from the initial methane decomposition (Eqs. 43-
45). Indeed, no straightforward relationship exists between this capacity and the hydrogen
yield since the former depends on many factors such as the basicity of the catalyst (e.g. to
accommodate carbonate species), the availability of lattice oxygen and/or defects for ensuring
the formation and spill over of the various carbon-containing adspecies. In turn, the hydrogen
yield is essentially controlled by the methane decomposition over the metal phase.
However, CH4 decomposition is associated with the hydroxyl consumption at high
temperatures (disappearance of OH region) in accordance with some results reported in the
literature indicating the involvement of OH groups in CH4 activation [110-118].
102
The reappearance of OH region (with lower intensity) after CH4 reaction (Figs. 39 and 40)
confirms that some hydrogen provided by the methane dissociation is still adsorbed on the
metal oxide surfaces supporting the mechanism of CH4 dissociation mentioned previously.
Furthermore, the presence of CO and CO2 implying the existence of reactive lattice oxygen
that may be due to the weakening of the covalence of the metal-oxygen bonds and/or the
enhancement of the mobility of lattice oxygen sites in addition to Lewis acid sites induced by
hydrogen reduction (Mo with lower oxidation states, Ce4+/3+ and Sn4+/2+ redox couples).
The fact that the coke provoked by carbon deposition has shown to intervene effectively in
different amounts on all the catalysts during CH4 decomposition because of the black colour
of the catalysts and this can greatly reduce the IR signal intensities.
Nevertheless, in any case, as evaluated from the spectra using Kubelka-Munk function to
obtain quantitative analysis [175-179], there is an approximately linear correlation between
the amount of converted methane and the sum of the amount of CO and CO2 gases plus the
amount of carbon stored on the catalyst, thus confirming a closed carbon mass balance within
the experimental error.
Finally, in contrast with the considerable amount of papers published about the activation of
methane on pure transition metals, the theoretical studies of this reaction on metal oxides have
not yet succeeded in giving a sufficiently clear and complete explanation of this process and
the reaction pathway is still up for debate.
103
SUMMARY AND CONCLUSIONS The Mo/Al2O3, Mo/CeO2, Mo/SnO2, Ce-Mo/Al2O3 and Sn-Mo/Al2O3 samples prepared by
impregnation and co-precipitation methods showed either different Mo dispersity or catalytic
activity towards CO adsorption and CH4 transformation. These catalysts with high Mo
loadings were prepared at low pH taking into account the calcination temperature (600°C) and
time as well as the solution pH, the isoelectric point and surface area of the solid support.
Two types of surface molybdena species have been identified. These are the surface bound
MoO42- polymeric species due to the concomitant strong interaction revealed between Mo and
support and free MoO3 crystallites that validated by XRD and DRIFT results. Accordingly,
high Mo loadings (>15 wt% on Al2O3) are necessary to obtain considerable amounts of free
MoO3 favorable for keeping the active metal in a higher dispersion state. Free MoO3
crystallites are more easily reducible than MoO42- molybdates with tetrahedral configuration
strongly bound to Al2O3 (Mo-O-Al bonds) in the form of Al2(MoO4)3.
The Mo/Al2O3 material had the highest specific surface area (117.6 m2/g) and rather presented
mesopores of regular distribution. However, this material showed the highest thermal stability
while the Mo/CeO2, Mo/SnO2, Ce-Mo/Al2O3 samples undergo structural modifications above
700°C. This may be due to high oxygen storage and release of ceria and tin resulting in lattice
defects and thus enhancing the mobility of metal ions and mutual interactions between them.
One may indicate that the introduction of cerium promotes aggregation of the Mo particles
(particularly as a support) probably due to charge effects and/or to the strong basic property
possessed by CeO2 as further emphasized by means of DRIFT (Mo–O–Ce linkages) and XRD
(decreased crystallites size of MoO3 species and different molecular formulae between Mo
and Ce). This led to the increase of polymerized surface Mo species besides the formation of
coupled O=Mo=O bonds (bands at 995 and 1035 cm-1) indicative of polymeric MoO3.
In the meanwhile, doping Mo/Al2O3 with SnO2 leads to surface structure definitely different
from that of Mo/SnO2. In the case of Sn-Mo/Al2O3, MoO3 crystals completely disappeared
and transformed into MoO2 with the presence of SnO, whereas Mo/SnO2 was formed only by
the two phases of MoO3 and SnO2 with high crystallinity but in both catalysts no linkages
were observed between Mo and Sn ions after calcination at 600°C and molybdate species
strongly affect the growth of SnO2 crystals. The major point that should be outlined is that the
characteristic of Mo-Sn system is the result of the preparation method employed.
104
On the other hand, the thermal behaviour of Mo/SnO2 showed the dissolution of molybdena
ions in the SnO2 crystals above 750°C. Thereby its high activity revealed in CH4 conversion
might be associated with the former. It is reasonable to note that the marked dissolution was
observed in all thermal treatments of Mo/SnO2 above 750°C (TG-DTA, CO adsorption and
CH4 reaction after H2 reduction at 800°C) with somewhat different extents.
One may notice that the reduction of the catalysts improves the surface reactivity leading to
the presence of small amounts of metallic Mo after reduction at 700°C. Moreover, further
reduction up to 800°C enhanced the adsorption of CO so as to forming various types of
carbonates, therefore as the reduction continues more coordinatively unsaturated (CUS) sites
will be produced and thus resulting in more CO adsorption sites.
CO chemisorption at 100°C on Mo/Al2O3 reduced at different temperature mainly occurred
on small MoO3 crystallites with only one band appearing at 2197 cm-1 corresponding to
octahedral alumina sites (Aloct+3–CO) implying that Al2(MoO4)3 is difficult reducible.
Furthermore, CO chemisorption at 100°C on Mo/Al2O3 leads to the formation of formates,
carboxylates and carbonates. On the other hand, CO chemisorption on Mo/Al2O3, Mo/CeO2,
Mo/SnO2, Ce-Mo/Al2O3 and Sn-Mo/Al2O3 catalysts reduced at 800°C involves oxygen in the
catalysts such oxygen could be present in Mo=O, Mo–O–Mo, Ce–O–Mo, Ce–O–Ce and
Sn–O–Sn associates to form carbonates. It is reasonable to suggest that the formation of
carbonate species involves reduced catalysts with hydroxyl groups and oxygen vacancies
implying the existence of reactive lattice oxygen that may be due to the weakening of the
covalence of the metal-oxygen lattice bonds and/or the enhancement of the mobility of lattice
oxygen sites. Within this context, intimate coupling of Mo with Ce and Sn ions of different
oxidation states has great facilities for electron exchange interactions. Thus, the electron-
mobile environment necessitated by redox reactions is established that has a great share in
enhancing the CO adsorption and therefore, leading to apparently high coverage of carbonate
species on the catalysts reduced at 800°C.
Concerning the CH4 decomposition on Mo/Al2O3 the results showed that molybdena oxide
could activate methane and oxidize it into surface formates and carbon monoxide but methane
conversion is low. However, for CH4 decomposition on Mo/CeO2 and Ce-Mo/Al2O3 the
results permit to infer that the beneficial effects occurred either because the cooperation
between Ce and Mo interfacial active sites generated with higher activity or because the
oxidative properties of CeO2 increased the dissociation of CH4 resulting in the liberation of
the apparently high coverage of carbonates besides CO and CO2, and as a result the methane
conversion increased mainly on Mo/CeO2.
105
For CH4 decomposition on Sn-Mo/Al2O3 the CH4 conversion is low with the formation of
carbonate species while Mo/SnO2 revealed the highest CH4 activity (high conversion into
CO2) and selectivity leading to the formation of formaldehyde intermediate and to almost total
CH4 oxidation too. This relevant activity and selectivity is presumably due to the dissolution
of molybdena ions in the SnO2 crystals resulting in more active sites.
Comparatively, the presence of CeO2 and SnO2 contributes to the rapid activation of CH4,
thereby accelerating the carbon gasification reaction to produce CO2 decreasing the coke
could be provoked by carbon deposition that has shown to intervene effectively with
somewhat different rates throughout CH4 decomposition. Accordingly, the following order of
the catalysts can be affirmable in accordance with the increase of CH4 conversion:
Mo/SnO2 > Mo/CeO2 > Ce-Mo/Al2O3 > Sn-Mo/Al2O3 > Mo/Al2O3
Finally, the present results suggest that Mo/CeO2 and Mo/SnO2 reduced at 800°C have the
most likely active species for CO adsorption and CH4 decomposition (especially Mo/SnO2)
due to highly dispersed MoO3 species besides Ce3+/Ce4+ and Sn2+/Sn4+ redox couples that
have high capacity towards oxygen. These species were responsible for their catalytic activity
revealed in CH4 oxidation.
Indeed, these results in correlation with the literature suggested that the higher dispersion of
MoO3 on a highly reducible support leads to more active and selective catalysts by optimizing
the interaction with the support throughout the preparation procedure.
Furthermore, it is inferred that there is an approximately linear correlation either between the
increase of the extent of reduction up to 800°C and the increasing integrated absorbance of CO
adsorbed on the catalysts or between the amount of converted methane and the sum of the
amount of CO and CO2 gases plus the amount of carbon stored on the catalysts.
106
ACKNOWLEDGMENTS I wish to express my greatest thanks to my supervisor Professor Ákos Rédey for enabling this
research to be feasible and for his friendly guidance and his incredible assistance.
I would like to express my sincere appreciation to the following persons whose support made
this research work possible:
Dr. József Kovács and Dr. Tibor Egyházy for their extraordinary help in making some
measurements.
Prof. Monica Caldararu from the Institute of Physical Chemistry “ilie murgulescu” of the
Romanian Academy of Sciences and Dr. Roman Dula from the Institute of Catalysis and
Surface Chemistry in Krakow for their knowledge and cooperation, guidance and help in the
Electron Spin Resonance (ESR) measurements.
Prof. János Kristóf for a very good course in infrared spectroscopy.
Prof. Dénes Kalló, Prof. Pál Tétényi and Dr. Jenő Hancsók for their advices in preparing the
manuscript.
Dr. Tatiana Yuzhakova, Pál Bui and all my colleagues at the Institutional Department of
Environmental Engineering and Chemical Technology for their friendship and encouragement
received from them throughout my research work.
107
REFERENCES
1. Szabó Z., Kalló D.: Kontakt katalízis, Akadémiai Könyvkiadó, Budapest 1966. 2. H. Jens: Industrial Catalysis, a practical approach, Wiley-VCH, Weinheim 1999. 3. Rédey Á.: Akadémiai Doktori, Veszprémi Egyetem, 1994. 4. Á. Rédey, W. K. Hall: J. Catal., 108, 185 (1987). 5. Á. Rédey, J. Goldwasser, W. K. Hall: J. Catal., 113, 82 (1988). 6. F.E. Massoth: J. Catal., 36, 164 (1975). 7. F.E. Massoth: J. Catal., 30, 204 (1973). 8. F.E. Massoth: Adv. Catal., 27, 265 (1978). 9. W. K. Hall and F. E. Massoth: J. Catal.: 34, 41 (1974). 10. J. Valyon, W. K. Hall: J. Catal., 84, 216 (1983). 11. L. Wang, W. K. Hall: J. Catal., 77, 232 (1982). 12. L. Wang: PhD dissertation, University of Wisconsin-Milwaukee, 1982. 13. W.K. Hall and M. LoJacono: Proc. VIth intl. Cong. Catal., 1, 246 (1977). 14. W. S. Millman and W. K. Hall: J. Catal., 60, 404 (1979). 15. N. Giordano, J.C.J. Bart, A. Vaghi, A. Castellan, G. Martinotti: J. Catal., 36, 81 (1975) 16. J. Medema: J. Catal., 37, 91 (1975). 17. J. Medema, C. Van Stam, V. H. J. De Beer, A. J. A. Konings, D. C. Koningsberger: J.
Catal. 53, 386 (1978). 18. J. P. Peri: J. Phys. Chem., 86, 1615 (1982). 19. J. P. Peri: J. ACS Symp. Ser. no., 288, 422 (1985). 20. H. Knözinger, P. Ratnasamy: Catal. Rev. Sci. Eng., 17 (1), 31-70 (1978). 21. P. Ratnasamy, H. Knözinger: J. Catal., 54, 155 (1978). 22. M.I. Zaki, H. Knözinger: J. Spectrochim. Acta., 43A, 1455 (1987). 23. M.I. Zaki, H. Knözinger: J. Mater. Chem. Phys., 17, 201 (1987). 24. H. Ballinger, J.T. Yates: J. Langmuir, 7 (12), 3041 (1991). 25. J. Goldwasser, S. M. Fang, M. Houalla, W. K. Hall: J. Catal., 115, 34 (1989). 26. R. Nakamura, D. Pioch, R.G. Bowman, R.L.Jr. Burwell: J. Catal., 93, 388 (1985). 27. E.A. Paukshtis, E.N. Yurchenko: J. Russ. Chem. Rev., 52, 242 (1983). 28. T. P. Beebe, P. Gelin, J. T. Yates: J. of Surf. Sci., 148, 526 (1984). 29. J. Chung, J. P. Zhang, R. L. Burwell: J. of Catal., 116, 506 (1989). 30. M. Dufaux, M. Che, and C. Naccache: J. Chim. Phys., 67, 527 (1970). 31. M. Digne, P. Sautet, P. Raybaud, P. Euzen, H. Toulhoat: J. of Catal., 211, 1-5 (2002). 32. H. Knözinger, H. P. Boehm: J. Cat. Sci. and Techn., Vol. 4, 39 (1983). 33. D. M. Hercules, M. Yamada, J. Yasumaru, M. Houalla: J. Phys. Chem., 95, 7037
(1991). 34. D. M. Hercules, V. Sazo, L. Gonzalez, J. Goldwasser, M. Houalla: Surf. Interface
Anal. 23, 367 (1995). 35. D. M. Hercules, J. Goldwasser, C. Scott, M. J. Perez Zurita, M. Houalla: Catal. Lett.,
32, 273 (1995). 36. J. Sonnemans, P. Mars: J. Catal. 31, 209 (1973). 37. A. Iannibello, F. Trifiro: Z. Anorg: Allg. Chem., 413, 293 (1975). 38. S. Ardizzone and S. Trasatti: Adv. Coll. Inter. Sci., 64, 173 (1996). 39. L. K. Chau and, M. D. Porter: J. Coll. Inter. Sci., 145, 283 (1995). 40. E. McCafferty, J. P. Wightman: J. Coll. Inter. Sci., 194, 344–355 (1997). 41. G. D. Parfitt and D.W. Fuerstenau: J. Coll. Sci. 48, 415 (1976). 42. J. Sarrın, O. Noguera, H. Royo, M. J. Perez Zurita, C. Scott, M.R. Goldwasser, J.
Goldwasser, M. Houalla: J. Mol. Catal. A: Chemical 144, 441–450 (1999). 43. J. Engelhardt, J. Goldwasser, W. K. Hall: J. Catal., 76, 48 (1982).
108
44. J. Engelhardt, J. Goldwasser, W. K. Hall: J. Mol. Catal., 15, 173 (1982). 45. N. Koizumi, G. Bian, K. Murai, T. Ozaki, M. Yamada: J. Mol. Catal. A: Chemical,
207, 171–180 (2004). 46. E. C. DeCanio, D. A. Storm: J. of Catal., 130, 653 (1991). 47. I. Oliveros, M. J. Perez Zurita, C. Scott, M. R. Goldwasser, J. Goldwasser, S. Rondon,
M. Houalla, D.M. Hercules: J. Catal., 171, 485 (1997). 48. P. Tétényi, T. I. Korányi, M. Dobrovolszky, T. Koltai, K. Matusek, Z. Paál: Fuel Proc.
Tech. 61, 55-71 (1999). 49. N. Y. Topsøe, H. Topsøe: J. Catal., 84, 386 (1983). 50. N. Y. Topsøe, H. Topsøe: J. Catal., 139, 631 (1993). 51. A. J. Brungs, A. P. E. York, J. B. Claridge, C. Marquez-Alvarez, M. L. H. Green:
Catal. Lett., 70, 117 (2000). 52. B. Claridge, A.P.E. York, C. Marquez-Alvarez, A.J. Brungs, J. Sloan, S.C. Tsang,
M.L.H. Green: J. Catal. 180, 85 (1998). 53. A. P. E. York, J. B. Claridge, A. J. Tsang, M. L. H. Green: Chem. Comm., 39, (1997). 54. S. Yokota, K. Okumura, M. Niwa: Catal. Lett., 84, 131 (2002). 55. E. McCafferty, P. M. Natishan, G. K Hubler: Corros. Sci., 35, 239 (1993). 56. E. McCafferty: Corros. Sci., 37, 481 (1995). 57. M. Mantel, J. P.Wightman: Surf. Inter. Anal., 21, 595 (1994). 58. G. Ramis, G. Busca: J. Mol. Struct., 193, 93 (1989). 59. G. Ramis, G. Busca, M. Turco, E. Kotur, R. J.Willey: J. Catal., 157, 523 (1995). 60. G. Ramis, G. Busca, C. Cristiani, L. Lietti, P. Forzatti: Langmuir, 8, 1744 (1992). 61. C. U. I. Odenbrand, J.G.M. Brandin, G. Busca: J. Catal., 135, 505 (1992). 62. M. Trombetta, G. Busca, S. A. Rossini, V. Piccoli: J. Catal., 168, 349 (1997). 63. H. C. Yao: J. Catal., 70, 440 (1981). 64. M. C. Kung, H. H. Kung: J. Catal. Rev.-Sci. Eng., 27 (3), 425 (1985). 65. L. Dong and Y. Chen: Farady Trans., 92, 22 (1996). 66. I. E. Wachs, A. M. Turek: J. Phys. Chem., 96, 5000-5007 (1992). 67. E. Odier, Y. Schuurman, K. Barral, C. Mirodatos: Surf. Sci. Catal., 147, 79 (2004). 68. E. Odier, Y. Schuurman, H. Zanthoff, C. Millet, C. Mirodatos, Stud. Surf. Sci. Catal.,
133, 327 (2001). 69. K. M. S. Khalil: J. Catal., 178, 198 (1998). 70. K. M. S. Khalil, L. A. Elkabee, B. Murphy: Micro. Meso. Mater., 78, 83 (2005). 71. K. M. S. Khalil, L. A. Elkabee, B. Murphy: J. Coll. Interf. Sci., 287, 534 (2005). 72. K. M. S. Khalil: J. Coll. Interf. Sci., 307, 172–180 (2007). 73. F. Boccuzzi, A. Chiorino, S. Tsubota, M. Haruta: J. Phys. Chem., 100, 3625 (1996). 74. M. Manzoli, F. Boccuzzi, A. Chiorino, F. Vindigni, W. Deng, M. F. Stephanopoulos:
J. Catal., 245, 308–315 (2007). 75. F. Menegazzo, M. Manzoli, A. Chiorino, F. Boccuzzi, T. Tabakova, M. Signoretto, F.
Pinna, N. Pernicone: J. Catal., 237, 431 (2006). 76. T. Tabakova, F. Boccuzzi, M. Manzoli, D. Andreeva: Appl. Catal. A, 252, 385 (2003). 77. F. Boccuzzi, A. Chiorino, M. Manzoli, D. Andreeva, T. Tabakova: J. Catal., 188, 176
(1999). 78. M. Manzoli, A. Chiorino, F. Boccuzzi: Appl. Catal. B, 52, 259 (2004). 79. M. M. Mohamed: Appl. Catal. A, 267,135 (2004). 80. M. M. Mohamed, S. M. A. Katib: Appl. Catal. A, 287, 236–243 (2005). 81. G. A. Lucia, F. M. Goncalves, P. R.S. Medeiros: Appl. Catal. A: General 208,
265–270 (2001). 82. J. Kristóf, H. Nasser, R. L. Frost, A. De Battisti, Á. Rédey: Appl. Surf. Sci., 242,
13–20 (2005).
109
83. J. Kristóf, N. Heider, R. L. Frost and V. Vágvölgyi: J. Therm. Anal. Calorim. Vol. 78, 687–695 (2004).
84. Y. Okamoto, K. E. Oh-Hiraki and S. Teranishi: J. Catal. 71, 99 (1981). 85. G. Kozma, Á. Kukovecz, Z. Kónya: J. Mol. Structure 834, 430–434 (2007). 86. J. Madarász, P. Bombicz, M. Okuyac, S. Kanekoc, Gy. Pokol: J. Anal. Appl. Pyr. 72,
209–214 (2004). 87. R. S. Patil, M. D. Uplane, P. S. Patil: Appl. Surf. Sci. 252, 8050–8056 (2006). 88. S. Mihaiu, G. Postole, M. Carata, M. Caldararu, D. Crisan, N. Dragan, M. Zaharescu:
J. Eur Ceramic Society 24, 963–967 (2004). 89. K. V. R. Chary, K. R. Reddy, G. Kishan, J. W. Niemantsverdriet, Gerhard Mestl: J.
Catal., 226, 283–291 (2004). 90. J. M. Herrmann, F. Villain, L. G. Appel: Appl. Catal. A: General 240, 177–182
(2003). 91. E. Comini, M. Ferroni, V. Guidi, G. Martinelli, G. Sberveglieri: Sens. and Acts. B 95,
157–161 (2003). 92. L. B. Gulina, V. P. Tolstoy: Thin Solid Films 440, 74–77, (2003). 93. Y. Maehara, T. Kimura, H. Kawanowa, Y. Gotoh: J. Crys. Growth 275, 1619–1624
(2005). 94. T. Serin, N. Serin, S. Karadeniz, H. Sarı, N. Tugluoglu, O. Pakma: J. Non-Crys. Solids
352, 209–215 (2006). 95. M. Ivanovskaya, E. Lutynskaya, P. Bogdanov. Sens. and Acts. B 48, 387–391 (1998). 96. A. Chiorino, G. Ghiotti, F. Prinetto, M. C. Carotta, D. Gnani, G. Martinelli: Sens. and
Acts. B 58, 338–349 (1999). 97. A. Chiorino, G. Ghiotti, M. C. Carotta, G. Martinelli: Sens. and Acts. B 47, 205–212
(1998). 98. E. Heracleous, M. Machli, A. A. Lemonidou, I. A. Vasalos: J. Mol. Catal. A:
Chemical 232, 29–39 (2005). 99. D. Amalric-Popescu, F. Bozon-Verduraz: Catal. Today 70, 139–154 (2001).
100. M. Niwa, J. Igarashi: Catal. Today 52, 71-81 (1999). 101. F. Goncalves, P. R. S. Medeiros, J. G. Eon, L. G. Appel: Appl. Catal. A: General 193,
195–202 (2000). 102. P. R. S. Medeiros, L. G. Appel: Appl. Catal. A: General 231, 125–130 (2002). 103. H. Liu, P. Cheung, E. Iglesia: J. Catal. 217, 222-229 (2003). 104. V. Lochar: Appl. Catal. A: General 309, 33–36 (2006). 105. N. G. Valente, L. E. Cadtis, O. F. Gorriz, L. A. Arrtia, J. B. Rivarola: Appl. Catal. A:
General 153, 119-132 (1997). 106. M. Daturi and L. G. Appel: J. Catal., 209, 427–432 (2002). 107. R. Rai, T. D. Senguttuvan, S. T. Lakshmikumar: Comp. Mater. Sci., 37, 15–19 (2006). 108. C. da Silva, L. G. Appel, S. Braun: Appl. Catal. A: General 314, 172–178 (2006). 109. J. Lercher, E. Heracleous, A. Lemonidou: J. Appl. Catal., 264, 73-80 (2004). 110. L. Lin, Y. Xu: Appl. Catal. A, 188, 53 (1999). 111. Y. Yang, W. Li, H. Xu: React. Kinet. Catal. Lett., 77, 155 (2002). 112. T. Armaroli, D. Minoux, S. Gautier, P. Euzen: Appl. Catal. A, 251 241–253 (2003). 113. D. C. LaMont, W. J. Thomson: Chem. Eng. Sci., 60, 3553 – 3559 (2005). 114. E. Iglesia, W. Ding, G. D. Meitzner: J. Catal., 206, 14 (2002). 115. A. R. S. Darujati, D. C. LaMont, W. J. Thomson: Appl. Catal., 253, 397 (2003). 116. A. R. S. Darujati, D. C. LaMont, A. J. Gilligan, A. S. Chellappa, W. J. Thomson:
Appl. Catal., 255, 239 (2003). 117. J. A. Anderson, Z. Wang, C. H. Rochester: J. Catal. 184, 213-223 (1999). 118. R. Wang, H. Xu, X. Liu, Q. Ge, W. Li: Appl. Catal. A, General 305, 204–210 (2006).
110
119. A. Trovarelli: Catal. Rev. Sci. Eng., 38, 439 (1996). 120. A. Trovarelli, M. Boaro, E. Rocchini, C. de Leitenburg, G. Dolcetti: J. Alloys Compd.,
374, 387 (2004). 121. M. Boaro, F. Giordano, S. Recchia, V. Dal Santo, M. Giona, A. Trovarelli: Appl.
Catal. B: Environ., 52, 225 (2004). 122. H. Jenő: Alternatív motorhajtóanyagok, Veszprémi Egyetemi Kiadó, Veszprém, 2004. 123. G. A. Olah, J. A. Olah: J. Am. Chem. Soc., 93, 1256 (1971). 124. T. Bunluesin, H. Cordatos, R.J. Gorte: J. Catal., 157, 222 (1995). 125. T. Bunluesin, E. S. Putna, R. J. Gorte: Catal. Lett., 41, 1 (1996). 126. T. Bunluesin, G. W. Graham, R. J. Gorte: Appl. Catal. B, 14, 105 (1997). 127. S. Hilaire, X. Wang, T. Luo, R. J. Gorte, J. Wagner: Appl. Catal. A, 215, 271 (2001). 128. H. Amariglio, P. Pareja, A. Amariglio: Catal. Today, 25, 113 (1995). 129. H. Amariglio, M. Belgued, P. Pareja, A. Amariglio: Catal. Lett., 31, 19 (1995). 130. H. Amariglio, A. Amariglio, M. Belgued, P. Pareja: J. Catal., 177, 113 (1998). 131. R. A. van Santon, T. Koerts, M. T. A. G. Deelen: J. Catal., 138, 101 (1992). 132. R. A. van Santon, T. Koerts: J. Mol. Catal., 74, 185 (1992). 133. F. Solymosi, A. Erdohelyi, J. Cserenyi, A. Felvegi: J. Catal., 147, 272 (1994). 134. F. Solymosi, J. Cserenyi: Catal. Lett., 34, 343 (1995). 135. L. Guczi, K. V. Sarma, L. Borko: J. Catal., 167, 495 (1997). 136. L. Guczi, L. Borko: Catal. Today, 64, 91 (2001). 137. M. C. J. Bradford: Catal. Lett., 66, 113 (2000). 138. G. Jacobs, L. Williams, U. Graham, G. A. Thomas, D. E. Sparks, B. H. Davis: Appl.
Catal. A: General, 252, 107–118 (2003). 139. H. Liu, T. Li, B. Tian, Y. Xu: Appl. Catal. A, 213, 103 (2001). 140. D. Ma, D. Wang, L. Su, Y. Shu, Y. Xu, X. Bao: J. Catal., 208, 260 (2002). 141. Y. Xu, X. Bao, L. Lin: J. Catal., 216, 386–395 (2003). 142. L. Wang, R. Ohnishi, M. Ichikawa: J. Catal., 190, 276 (2000). 143. C. Müller, C. Freysoldt, M. Baudin, K. Hermansson: Chem. Phys., 318, 180 (2005). 144. B. Herschend, M. Baudin, K. Hermansson: Chem. Phys., 328, 345–353(2006). 145. R. Craciun, W. Daniell, H. Knözinger: Appl. Catal. A, 230, 153 (2002). 146. K. Otsuka, Ye Wang, E. Sunada, I. Yamanaka: J. Catal., 175, 15 (1998). 147. F. Solymosi, A. Erdohelyi, A. Szoke: Catal. Lett., 32, 43 (1995). 148. A. Szoke, F. Solymosi: Appl. Catal. A, 142, 361 (1996). 149. J. Kiss, F. Solymosi: J. Catal., 179, 277 (1998). 150. F. Solymosi, L. Bugyi, A. Oszko: Catal. Lett., 57, 103 (1999). 151. F. Solymosi, and M. Pasztor: J. Catal. 104, 312 (1987). 152. A. Berko, F. Solymosi: J. Catal., 183, 387 (1999). 153. J. H. Lunsford: Stud. Surf. Sci. Catal., 81, 1 (1993). 154. J. H. Lunsford, D. Wang, M. P. Rosynek: Top. Catal., 3, 289 (1996). 155. J. H. Lunsford, D. Wang, M. P. Rosynek: J. Catal., 169, 347 (1997). 156. M. Ichikawa, Y. Shu, R. Ohnishi: J. Catal., 206, 134 (2002). 157. M. Ichikawa, R. Onishi, S. Liu, Q. Dong, L. Wang: J. Catal., 182, 92 (1999). 158. P. Mériaudeau, L. V. Tiep, V. T. T. Ha, C. Naccache, G. Szabo: J. Mol. Catal. A, 144,
469 (1999). 159. P. Mériaudeau, V. T. T. Ha, L. V. Tiep: Catal. Lett., 64, 49 (2000). 160. A. A. Said and S. A. Halawy: J. Therm. Anal. Cal., Vol. 41, 1075-1090 (1994). 161. L. B. Sharma and S. Batra: J. Therm. Anal. Cal., Vol. 34, 1273-1281 (1988). 162. N. R. E. Radwan, M. Mokhtar and G. A. El-Shobaky: J. Therm. Anal. Cal., Vol. 71,
977–986 (2003). 163. Z. Sarbak and W. K. Józwiak: J. Therm. Anal. Cal., 87, 229 (2007).
111
164. R. Lyszczek: J. Therm. Anal. Cal., 90, 533 (2007). 165. M. Blachnio, P. Staszczuk, G. Grodzicka, L. Lin and Y. X. Zhu: J. Therm. Anal. Cal.,
88, 601 (2007). 166. O. Cozar, D.A. Magdas, I. Ardelean: J. Non-Crys. Solids 354, 1032–1035 (2008). 167. L. Bih, A. Nadiri, H. Khemakhem, B. Elouadi: J. Mol. Struct. 872, 1–9 (2008). 168. D. Zayachuk, Ye. Polyhach, E. Slynko, O. Khandozhko, C. Rudowicz: Spectr. Acta
Part A 60, 1247–1256 (2004). 169. G. Mitrikas, G. Kordas, G. Fanourakis, E. Simoen: Nuc. Inst. and Meth. in Phy.
Research B 186, 181–185 (2002). 170. B. Kamp, R. Merkle, J. Maier: Sens. and Acts. B 77, 534-542 (2001). 171. B. Kamp, R. Merkl, R. Lauck, J. Maier: J. Solid State Chem. 178, 3027–3039 (2005). 172. F. Morazzoni, C. Canevali, N. Chiodini, C. Mari, R. Ruffo, R. Scotti, L. Armelao,
E. Tondello, L. Depero, E. Bontempi: Mats. Sci. and Eng. C 15, 167–169 (2001). 173. A. Martinez-Arias, J. Soria, J. C. Conesa, X. L. Seoane, A. Arcoya: J. Chem. Soc.
Faraday Trans. 91, 1679 (1995). 174. M. Figaj, K. D. Becker: Solid State Ionics, 141, 507-512 (2001). 175. L. Vien, D. Colthup, N. B. Fateley, W. G. and Grasselli, J. G., “The Handbook of
Infrared and Raman Characteristic Frequencies of Organic Molecules.” Academic Press, San Diego, 1991.
176. J. Ryczkowski: Catal. Today, 68, 263–381 (2001). 177. Guido Busca: Catal. Today, 41, 191-206 (1998). 178. A. Zecchina, C. Lamberti, S. Bordiga: Catal. Today, 41, 169-177 (1998). 179. A. Zecchina, D. Scarano, S. Bordiga, G. Ricchiardi, G. Spoto, F. Geobaldo: Catal.
Today, 27, 403 (1996). 180. M. I. Zaki, M. A. Hasan, F. A. Al-Sagheer, L. Pasupulety: Coll. and Surf. A:
Phys. and Eng. Aspects, 190, 261–274 (2001). 181. H. Zhu, Z. Qin, W. Shan, W. Shen, J. Wang: J. Catal., 225, 267–277 (2004). 182. H. Zhu, Z. Qin, W. Shan, W. Shen, J. Wang: J. Catal., 233, 41–50 (2005). 183. B. Herschend, M. Baudin, K. Hermansson: J. Chem. Phys., 120, 4939 (2004). 184. H. Liu, L. Su, H. Wang, W. Shen, X. Bao, Y. Xu: Appl. Catal. A, 236, 263 (2002). 185. U. Berner, K. Schierbaum, G. Jones, P. Wincott, S. Haq, G. Thornton: Surf. Sci., 467,
201 (2000). 186. Z. Yang, T. K. Woo, K. Hermansson: Chem. Phys. Lett., 396, 384–392 (2004). 187. El. Aneggi, M. Boaro, C. de Leitenburg, G. Dolcetti, A. Trovarelli:
J. Alloys and Compounds, 408–412, 1096–1102 (2006). 188. Németh T. Zs.: Doktori értekezés, Veszprémi Egyetem, 1996. 189. N. Heider, R. Ákos, Y. Tatiana, N. T. Zsolt and O. Tamás: Reac. Kinet. And
Catal. Letts., Vol. 92 No. 2, 329-335, (2007). 190. M. R. Smith, U.S. Ozkan: J. Catal., 141, 124 (1993). 191. M. R. Smith, L. Zhang, S. A. Driscoll, and U.S. Ozkan: Catal. Lett., 19, l-15 (1993). 192. S. Chempath, A. T. Bell: J. Catal., 247, 119–126 (2007). 193. V. D. Sokolosvskii, G. M. Aliev, O. V. Buyesvskaya, A. A. Davydov:
Catal. Today, 4, 293 (1989). 194. B. Irigoyen, N. Castellani, A. Juan: J. Mol. Catal. A, 129, 297–310 (1998). 195. I. A. Fisher, A. T. Bell: J. Catal., 172, 222 (1997). 196. A. Fritz, V. Pitchon: Appl. Catal. B, 13, 1 (1997). 197. X. Zhang, D. He, Q. Zhang, Q. Ye, B. Xu, Q. Zhu: Appl. Catal. A: General, 249,
107–117 (2003).
112
THESES The main results of the dissertation are summarized in the following thesis points:
1. Taking into account the calcination temperature and time as well as the solution pH,
the isoelectric point and the surface area of the solid support high Mo loadings (>15
wt% on Al2O3) are necessary to obtain considerable amounts of free MoO3 crystallites
favorable for keeping the active metal in a higher dispersion state. Thus, MoO3
clusters are more easily reducible than MoO42- molybdates with tetrahedral
configuration strongly bound to Al2O3 (Mo-O-Al bonds) in the form of Al2(MoO4)3
and the higher activity of Mo/Al2O3 may be associated with the former.
2. For high Mo loadings obtained two types of molybdena species were the predominant
surface species. These are the surface bound MoO42- polymeric species due to the
concomitant strong interaction revealed between Mo and support and free MoO3
crystallites that validated by XRD and DRIFT results.
3. The Mo/Al2O3 material showed the highest thermal stability up to 900°C while the
Mo/CeO2, Mo/SnO2, Ce-Mo/Al2O3 samples undergo morphological and structural
modifications above 700°C resulting in lattice defects, thus enhancing the mobility of
metal ions and the possibility of interactions between them.
4. The introduction of cerium promotes aggregation of the Mo particles (particularly as a
support) probably due to charge effects and/or to the strong basic property possessed
by CeO2 so as to forming different molecular formulae. This led to the increase of
polymerized surface Mo species besides the formation of coupled O=Mo=O bonds
indicative of polymeric MoO3 as further emphasized by means of DRIFT and XRD.
5. Concerning the use of SnO2, the major point that should be outlined:
a) The characteristic of Mo-Sn system is the result of the preparation method adopted.
However, doping Mo/Al2O3 with SnO2 leads to surface structure definitely different
from that of Mo/SnO2. Accordingly, when using SnO2 as promoter the MoO3 crystals
completely disappeared and transformed into MoO2 with the presence of SnO whereas
Mo/SnO2 was formed only by MoO3 and SnO2 oxides. Meanwhile, in both cases, no
linkages were observed between Mo and Sn ions after calcination at 600°C and
molybdate species strongly affect the growth of SnO2 crystals.
b) The thermal behaviour of Mo/SnO2 showed the dissolution of molybdena ions in
the SnO2 crystals above 750°C with somewhat different extents, thereby resulting in
more active sites and thus leading to a high catalytic activity of Mo/SnO2 catalyst.
113
6. The hydrogen reduction of the catalysts improves the surface reactivity generating
oxygen vacancies and coordinatively unsaturated sites (CUS) leading to the presence
of small amounts of metallic Mo after reduction at 700°C. Moreover, further reduction
up to 800°C enhanced their activity towards CO adsorption and CH4 dissociation.
7. The peculiarities of in situ DRIFT studies of CO adsorption and CH4 transformation
on the catalysts have been achieved under reaction conditions. Thus, the noticed gain
in the intensity for the bands in conjunction with various types of carbonate species
has been observed upon CO adsorption at 100°C and CH4 decomposition at 700°C.
This involves reduced catalysts containing coordinatively unsaturated sites (CUS) with
hydroxyl groups and oxygen vacancies so as to forming various carbonate species
implying the existence of the reactive lattice oxygen in the catalysts such oxygen
could be present in Mo=O, Mo–O–Mo, Ce–O–Mo, Ce–O–Ce and Sn–O–Sn entities.
8. The Mo/CeO2 and Mo/SnO2 catalysts reduced at 800°C have the most likely active
species for CO adsorption and CH4 dissociation. The highly dispersed MoO3 species
besides Ce3+/Ce4+ and Sn2+/Sn4+ redox couples that have high capacity towards oxygen
were responsible for the high catalytic activity revealed by Mo/SnO2 and Mo/CeO2.
Thus, the CH4 conversion increased in accordance with the following order of the
catalysts:
Mo/SnO2 > Mo/CeO2 > Ce-Mo/Al2O3 > Sn-Mo/Al2O3 > Mo/Al2O3
114
PUBLICATIONS
1. H. Nasser, Á. Rédey, T. Yuzhakova, J. Kovács: Thermal stability and surface
structure of Mo/CeO2 and Ce-doped Mo/Al2O3 catalysts, Journal of Thermal Analysis
and Calorimetry 2008, accepted for publication.
2. H. Nasser, Á. Rédey, T. Yuzhakova, J. Kovács: In situ DRIFT study of nonoxidative
methane reaction on Mo/SnO2 catalyst, Reaction Kinetics and Catalysis letters 2008,
accepted for publication.
3. H. Nasser, Á. Rédey, T. Yuzhakova, Zs. N. Tóth and T. Ollár: FTIR study of CO
adsorption on molybdena-alumina catalysts for surface characterization, Reaction
Kinetics and Catalysis letters, Vol. 92 No. 2, 329-335, 2007.
4. Nasser H., Rédey Á., Yuzhakova T.: Structure and Thermal Stability of Ceria-doped
Mo/Al2O3 Catalysts, Environmental Engineering and Management Journal, Vol. 5 No.
3, 425-432, 2006.
5. Nasser H., Rédey Á., Yuzhakova T: Structure and Surface Chemistry of Ceria-doped
Mo/Al2O3 Catalysts, MicroCAD International Conference Proceedings pp.
65-71, 2007.
6. Nasser H., Rédey Á., Yuzhakova T.: Környezetvédelmi felhasználású cérium tartalmú
Mo/Al2O3 katalizátorok struktúrája és termikus stabilitása, Országos
Környezetvédelmi Konferencia kiadványa pp. 285-293, 2006.
7. Nasser H., Kristóf J., Rédey Á., R. L. Frost, A. De Battisti: IrO2/SnO2 katalizátorok
képződési mechanizmusának vizsgálata és felületkémiai jellemzése, XIX Országos
Környezetvédelmi Konferencia kiadványa pp. 239-246, 2005.
8. J. Kristóf, H. Nasser, E. Horváth, R. L. Frost, A. De Battisti, Á. Rédey: Investigation
of SnO2 thin film evolution by thermoanalytical and spectroscopic methods, Applied
Surface Science, 242, 13–20, 2005.
9. J. Kristóf, H. Nasser, E. Horváth, R. L. Frost and V. Vágvölgyi: Investigation of
IrO2/SnO2 thin film evolution by thermoanalytical and spectroscopic methods, Journal
of Thermal Analysis and Calorimetry, Vol. 78, 687–695, 2004.
10. Yuzhakova T., Rédey Á., Caldararu M., Auroux A., Carata M., Postole G., Hornoiu
C., Nasser H.: Study of Pt/Sn-Al Catalyst for Environmental Application,
Environmental Engineering and Management Journal, Vol. 5 No. 4, 559-568, 2006.
115
11. Yuzhakova T., Rédey Á., Nasser H., Caldararu M., Strukova L., Gaál Z., Fazakas J.:
The Effect of Pretreatment on the Catalytic Properties of Supported Tin Catalysts,
Conference Proceedings, Paper Number 133, Brisbane, Chemeca 2005.
12. Yuzhakova T., Rédey Á., Holenda B., Domokos E., Nasser H., Caldararu M., Gaál Z.,
Fazakas J.: Surface Chemistry Studies on Molybdena-Alumina Catalysts, Conference
Proceedings, Paper Number 134, Brisbane, Chemeca 2005.
PRESENTATIONS
1. Nasser H., Rédey Á., Yuzhakova T., Caldararu M.: In-situ DRIFT Studies of Mo/SnO2
and Sn-doped Mo/Al2O3 Catalysts, 8th International Conference of the Romanian
Catalysis Society, Bucharest, Romania, June 21-23, 2007.
2. Nasser H., Rédey Á., Yuzhakova T., Caldararu M.: In-situ DRIFT Studies of Mo/SnO2
and Sn-doped Mo/Al2O3 Catalysts, Past and Present in DeNOx Catalysis, DeNOxCAT
2007, Uzlina, Romania, June 17-20, 2007.
3. Nasser H., Rédey Á., Yuzhakova T.: Structure and Surface Chemistry of Ceria-doped
Mo/Al2O3 Catalysts, MicroCAD International Scientific Conference, Miskolc, March
22-23, 2007.
4. Nasser H., Rédey Á., Yuzhakova T.: Structure and Thermal Stability of Ceria-doped
Mo/Al2O3 Catalysts, ICEEM03/EEMJ Conference, Iasi, Romania, September 21-24,
2006.
5. H. Nasser, J. Kristóf, R. L. Frost, A. De Battisti, Á. Rédey: Investigation of SnO2 thin
film evolution by thermoanalytical and spectroscopic methods, 10th EuCheMS
Conference on Chemistry and the Environment, Rimini: September 4-7, 2005.
6. Rédey Á., Yuzhakova T., Nasser H.: Environmental Impact Assessment Theory and
Practice, Key note lecture, ICEEM03/EEMJ Conference, Iasi, Romania, September
21-24, 2006.
7. Rédey Á., Kováts N., Yuzhakova T., Nasser H.: Environmental Impact Assessment,
Theory and Practice, Key note lecture, 1st European Chemistry Congress, Budapest,
Hungary, August 27-31, 2006.
8. Yuzhakova T., Rédey Á., Auroux A., Caldararu M., Carata M., Postole G., Hornoiu
C., Nasser H., Popescu I., Sandulescu I., Strukova L., Fazakas J.: Effect of Pt Loading
on the Surface and Catalytic Behaviour of Tin Containing Catalysts, 1st European
Chemistry Congress Budapest, Hungary, August 27-31, 2006.
116
9. Yuzhakova T., Rédey Á., Holenda B., Domokos E., Nasser H., Caldararu M., Gaál Z.,
Fazakas J.: Surface Chemistry Studies on Molybdena-Alumina Catalysts, Chemeca
2005, Brisbane, Australia, September 25-28, 2005.
10. Yuzhakova T., Rédey Á., Nasser H., Caldararu M., Strukova L., Gaál Z., Fazakas J.:
The Effect of Pretreatment on the Catalytic Properties of Supported Tin Catalysts,
Chemeca 2005, Brisbane, Australia, September 25-28, 2005.
11. Yuzhakova T., Rédey Á., Caldararu M., Auroux A., Scurtu M., Postole G., Nasser H.,
Fazakas J.: Adsorption Capacity of Tin Oxide Supported Catalysts to Capture Air
Pollutants, 2nd International Conference on Thermal Engines and Environmental
Engineering, Galati, Romania, June 7-9, 2007.
12. Yuzhakova T., Caldararu M., Hornoiu C., Auroux A., Rédey Á., Nasser H.:
Correlation Between Physico-Chemical Characteristics and Electrical Capacitance of
Tin Oxide Supported on Alumina Catalysts, 8th International Conference of the
Romanian Catalysis Society, Bucharest, Romania, June 21-23, 2007.
AWARDS: 1. Ceepus mobility grant to the University of Czestochowa, Poland 2005.
2. Ceepus mobility grant to the University of Poznan, Poland 2006.
3. Ceepus mobility grant to the Technical University of Cluj, Romania 2007.
4. Inclusion in Who’s Who in the World, 2009 Edition.
117