Oxidative dehydrogenation of ethane in short contact time reactors

Chapter 1 Introduction 2 
 
where reaction tubes are exposed to flames. The highest wall operating temperature depends 
on the material of the coils, and can be as high as 1200°C. Modern radiant coils are built in 
nickel alloys and are quite expensive. Indeed, the capital cost of a furnace can be as high as 
20% of the total cost of an ethylene plant, and a third of it is for radiant coils. Tube service 
life is further reduced by carburization of the inside tube furnace, leading to approximately 
five-year service life in the hottest section. 
Coke deposition is a serious issue inside the radiant tubes, due to the long residence 
time of the hydrocarbons in the high temperature zone. Heat transfer efficiency is reduced by 
coke deposition, since coke acts as a thermal insulator, requiring increasing tube wall 
temperature for a given furnace loading and periodic shut down for its removal. Decoking is 
carried out burning out the coke with a mixture of steam and air, trying to prevent localized 
overheating and coil burnout (Kirk-Othmer, 1978). 
Another issue becoming increasingly serious is NO
x
 formation in the furnace. The 
high temperatures of the flames, higher when coke deposition reduces the heat transfer 
efficiency, lead to the formation of large amounts of thermal NO
x
, which are not in agreement 
with the increasingly strict air regulations. 
Ethylene is favorably produced using ethane as feedstock and, at difference from light 
naphtas or other feed fuels, ethane yields to higher ethylene selectivity. Typical yields from an 
ethylene plant that uses ethane as feedstock range from 50 to 60%, with selectivities to 
ethylene per atom of carbon up to 85% (Song et al., 1992). 
Nevertheless, among the hydrocarbon feedstocks, ethane requires the highest 
temperature and the longest residence time to obtain a commercially acceptable conversion, 
since it is the most refractory alkane after methane. In addition, it yields a harder, denser coke 
than liquid fuels.  
 
1.2 Background on ethane oxidative dehydrogenation 
The oxidative dehydrogenation of ethane (ethane ODH) has been investigated in the 
last 20 years as an alternative to dehydrogenation of ethane for ethylene production (Kolts et 
al., 1992; Cavani and Trifirò, 1995). The addition of oxygen to the feed makes indeed the 
process exothermic, and allows to conduct an autothermal process, as shown in the reactions 
1.1 and 1.2. 
C
2
H
6
 → C
2
H
4
 + H
2
   ∆H
r
° = + 136.3 kJ/mole 1.1 
C
2
H
6
 + ½ O
2
 → C
2
H
4
 + H
2
O  ∆H
r
° = - 105.5 kJ/mol e 1.2 
Chapter 1 Introduction 3 
 
The need for oxygen separation is outscored by the advantages in terms of materials, 
since the process no longer requires fired reactors. The heat is produced in situ, and there are 
no heat resistances. Moreover, the low temperature of the reactor (< 1000ºC), together with 
the fact that the process can be carried out in the absence of N
2
 in the feed, eliminates the 
formation of thermal NO
x
. Finally, coke formation is eliminated by the presence of the 
catalyst, at a few millisecond residence time (Torniainen et al., 1994). 
Nevertheless, the main difficulty in the ODH of ethane for ethylene production is that 
ethylene, even though is kinetically favored due to the presence of oxygen is not a 
thermodynamically stable product in the range of temperatures, which is required to attain 
relevant ethane conversions. Stable products are shown in Fig. 1.2. Between 800 and 1050°C 
mainly the formation of CO, H
2
 and CH
4
 occurs. Such product distribution is sensitive to the 
temperature but also to the ratio R = C
2
H
6
/O
2
 to such an extent that can be described by the 
equation 1.3. All of oxygen goes to CO, and the remaining ethane not reacted to CO leads to 
the formation of CH
4
 and H
2
. 
R C
2
H
6
 + O
2
 → 2 CO + (2R-2) CH
4
 + (4-R) H
2
   1.3 
Only for R = ∞ (without O
2
) the formation of a significant fraction of ethylene is 
predicted by thermodynamic equilibrium. 
It follows that for ethylene production two approaches can be pursued. 
1. Ethylene can be further kinetically favored by choice of an appropriate 
catalyst, for instance in packed bed with a highly selective catalyst. 
2. Being ethylene a significant intermediate, it can be recovered controlling the 
residence time and quenching the reaction before product degradation towards 
thermodynamic equilibrium, for instance in short contact time reactors. 
 
1.2.1 ODH in conventional reactors 
The first approach was followed in the last 20 years in conventional fixed bed reactors. 
The catalytic systems mainly studied can be divided into two groups. The first group consists 
of catalysts based on transition metal oxides, for low temperature applications (400 to 500°C). 
Such catalysts are active for the ODH of light hydrocarbons, mainly propane and butane 
through a redox mechanism, due to the high mobility of the oxygen of the lattice (Driscoll et 
al., 1985). The second group, for high temperature applications (600 to 700°C), is instead 
made of catalysts specific for the oxidative coupling of methane but effective also for ethane 
Chapter 1 Introduction 4 
 
ODH. These catalysts are characterized by basic sites and comprise ions or metal oxides of 
the group IA and IIA (Cavani and Trifirò, 1995). 
In our research group we deeply investigated both types of catalysts, and in particular 
the second one. From experimentation on basic oxides (promoted and non-promoted MgO, 
Sm
2
O
3
 and La
2
O
3
) used in the ODH reaction, as reported in a recent work conducted in our 
laboratories (Ciambelli et al., 2000; Ruoppolo, 2000), it was found that for temperatures 
above 700°C the selectivity of the process towards ethylene formation seems to be 
substantially independent on the composition of the catalyst, and is determined by gas-phase 
reactions. 
Thus, none of the catalysts investigated for oxidative dehydrogenation of ethane in 
conventional reactors seemed to be significant in large-scale applications, due to the 
limitations in yield and selectivity occurring by decreasing the dilution and increasing the 
ethane to oxygen ratio in the passage from lab scale to industrial scale (Cavani and Trifirò, 
1998). 
 
1.2.2 ODH in short contact time reactors 
In the second approach the development of new catalytic systems was joined to novel 
reactor configurations, namely short contact time reactors. 
The typical contact times in this kind of reactors are of the order of the milliseconds, 
and can be considered short with respect to the order of the residence times of comparable 
industrial processes (for instance in steam cracking the residence time is ~1 s). 
The major advantage of the use of short contact time reactors is that the residence time 
may be closely tuned on the characteristic time of the reactions to the desired products, while 
the reactions to undesired but thermodynamically favored products do not significantly occur 
due to the short time scale. In conclusion in a short contact time reactor high yields to non-
equilibrium products may be achieved. 
In addition, since a short contact time reactor is operated at high space velocity, it 
works under quasi-adiabatic conditions, due to the high rate of heat production (high flow 
rate) and the low heat dispersion (little reactor dimensions). The particular temperature profile 
(with temperature gradients up to 10
6
 K/cm) developing in such kind of reactors may also be 
significant in determining product distribution. 
In comparison with fixed beds, a number of issues may be overcome in short contact 
time reactors. At high flow rates, the radial and axial dispersion is lower (high number of 
Peclet), hence the radial temperature gradient is reduced and a uniform temperature may be 
Chapter 1 Introduction 5 
 
attained on the section; the residence time distribution is narrower; and mass transfer to the 
catalyst is enhanced. Above all, fixed beds are not adequate to be operated at short contact 
times mainly due to the unsustainable pressure drop that would be generate by high space 
velocities. Thus, the reactors suitable for short contact time applications are mainly structured 
reactors, gauzes or monoliths.  
What is also remarkable from an industrial standpoint is that high space velocity 
means high productivity (10
3
 times a conventional reactor) or smaller reactor dimensions. 
In the industrial field, the concept of short contact time reactor was first applied in 
reactors of oxidation of NH
3
 (reaction 1.4) and of ammoxidation of CH
4
 to HCN (reaction 
1.5). The reactors consist of noble metal gauzes and are operated at extremely short contact 
times (~10
-5
 s), since NO and HCN are not thermodynamically stable intermediates towards 
the formation of N
2
 and CO. 
NH
3
 + 45  O
2
 → NO + 23  H
2
O  ∆H
r
° = - 120.3 kJ/mol e 1.4 
NH
3
 + CH
4
 + 23 O
2
 → HCN + 3 H
2
O ∆H
r
° = - 313.8 kJ/mol e 1.5 
The Schmidt’s group at the University of Minnesota introduced the concept of short 
contact time first for syngas and then for ethylene production. Since 1993, they have 
demonstrated that olefins can be efficiently produced by the auto-thermal oxidative 
dehydrogenation of ethane over Pt containing structured reactors at short contact times 
(milliseconds) and temperatures in the range of 900–1000°C (Huff and Schimdt, 1993). Under 
such reaction conditions, it was shown that feeding a mixture C
2
H
6
/O
2
 in a molar ratio slightly 
lower than 2 (namely the stoichiometric value for ethane ODH) ethylene yield is high enough 
to be compared with the existing cracking processes. 
The choice of platinum-based catalysts for this process resulted by the investigations 
of Schmidt et al. (Schmidt et al., 1994) who compared the activity of different noble metals in 
short contact time reactors for partial oxidation processes, reporting Pt as the best catalyst for 
ethane ODH and Rh the most suitable in partial oxidation to syngas. The other noble metals 
tested for CH
4
 partial oxidation did not show promising performances or enough stability. Pd, 
for instance, quickly deactivated due to coke formation (Torniainen et al., 1994). Moreover, 
doping Pt with tin or copper resulted in improved performance for ethane ODH, with a 
significant increase in ethylene yield (Yokohama et al., 1996), which was even greater as a 
result of H
2
 addition (Bodke et al., 2000). 
Nature of these results remains disputed. Until recently, it has been reported that 
processes occurring in this kind of reactor are purely heterogeneous (Huff and Schmidt, 1993; 
Chapter 1 Introduction 6 
 
Huff and Schmidt, 1996; Bodke et al., 2000) and can be regulated by the appropriate choice of 
the active component (Schmidt and Huff, 1994), support morphology (Bodke et al., 1998), 
and reaction mixture composition (Huff and Schmidt, 1993). This seems to be true only at 
moderate temperatures, while above 800°C, it has been shown, both experimentally (Lødeng 
et al., 1999; Beretta et al., 2000; Beretta et al., 2001a and b; Mulla et al., 2001; Huff et al., 
2000) and theoretically (Huff et al., 2000; Beretta et al., 2001b; Zerkle et al., 2000), that the 
kinetics of homogeneous reaction paths is not negligible, even for millisecond residence time 
scales. 
The alternative possibility to a purely heterogeneous mechanism of ethylene 
production is a hetero-homogeneous reaction path. As proposed in recent works (Lødeng et 
al., 1999; Mulla et al., 2001), the catalyst could be assumed to act mostly as an igniting agent, 
active for total oxidation reactions, while ethylene-forming reactions mainly occur in the gas 
phase. In the same way, Bodke at al. (2000) sketched a short contact time reactor as a 2 zones 
reactor where in the first zone oxygen depletion occurs through C
2
H
6
 oxidation to CO
x
 and 
H
2
O with consequent temperature increase, while in the second zone, where O
2
 is absent, the 
endothermic dehydrogenation of ethane to ethylene and H
2
 takes place. Moreover, the authors 
also reported that the use of hydrogen co-feeding allows a minor ethane consumption in the 
first zone of the reactor, with H
2
 being oxidized easier than ethane over the Pt-Sn surface, and 
consequently C
2
H
6
 dehydrogenates in the second zone with very high selectivity (>80%) and 
yield (about 60%) to ethylene. Detailed heterogeneous and homogeneous chemical kinetic 
mechanisms employed in the two-dimensional computational fluid dynamics model 
developed by Zerkle et al. (2000) confirmed the main scheme of Bodke et al. and evidenced 
that the proportion between the contribution of heterogeneous and homogenous reactions to 
the ethane overall consumption is strongly dependent on C
2
H
6
/O
2
 feed ratio. More 
specifically, they reported that the formation of ethylene must be mainly attributed to 
heterogeneous dehydrogenation paths, especially when H
2
 is added to the feed. 
The use of materials other than noble metals has been considered by some authors 
(Mulla et al., 2002; Flick and Huff, 1999; Beretta and Forzatti, 2001). Flick and Huff 
proposed the use of Cr
2
O
3
 based foam monoliths in ethane ODH (1999), showing that they 
exhibit very promising performances, even better than Pt-based monoliths, but with a strongly 
limited catalyst life-time, likely due to the well-known deactivation processes occurring on 
transition metal oxides in the catalytic combustion of hydrocarbons (i.e. under operating 
conditions characterized by high temperatures, oxidizing atmosphere and the presence of 
water and CO
2
). 
Chapter 1 Introduction 7 
 
BaMnAl
11
O
19
 hexa-aluminate, characterized by a high thermal stability, was 
investigated by Beretta and Forzatti under autothermal conditions (2001). In comparison with 
Pt/α-Al
2
O
3
 catalyst, hexa-aluminates showed better performance, even though the higher 
preheat temperature required caused coke deposition. However, the authors did conclude that 
the presence of any catalyst could be useful to make the process autothermal, but reduces the 
maximum yield attainable in the homogenous phase by only heating the C
2
H
6
/O
2
 mixture. In 
particular, they clearly evidenced that Pt catalysts under controlled isothermal conditions, 
namely in the absence of any homogeneous contribution, do not exhibit any intrinsic activity 
in ethane ODH to ethylene, and in short contact time reactors are hence responsible of the 
higher amount of ethane oxidized to CO
x
 compared to a pure homogeneous process (Beretta 
et al., 2001; Beretta and Forzatti, 2001b). 
The morphology of the catalyst used for such processes has also been under 
investigation. Foams have been thoroughly studied since 1992 by Hickman and Schmidt, who 
found clear evidence that extruded honeycombed monoliths performed worse than foams in 
partial oxidation of methane over Pt based monoliths. Since then, foams have been mainly 
investigated for ODH reactions (Huff and Schmidt, 1993; Bodke et al., 2000; Flick and Huff, 
1998). Sadykov et al. (2000) used honeycombed monoliths in oxidative dehydrogenation of 
propane to reduce secondary catalytic reactions of propylene, through the straight channels of 
the monolithic support. Only limited interest has been devoted to noble metal gauzes (Lødeng 
et al., 1999), because of the large Pt consumption, due to volatilization at the high temperature 
of operation of gauze reactors. FeCr alloy elements packed in a quartz tube were used as 
alternative to foams, because of the high void fraction and large thermal conductivity (Beretta 
et al., 2001a and b). However, there is no assurance that foams are more useful than 
honeycomb-type monolith catalysts in ethane ODH, as the conclusions of the work of 
Hickman and Schmidt (1992) can be strictly applied only to the partial oxidation of CH
4
, in 
which the role of the homogeneous reaction paths is much less relevant, if not negligible, than 
in the case of ethane. ODH is different from partial oxidation to syngas, because the latter is a 
process that is widely accepted to occur to a larger extent on the catalytic surface, and is then 
more influenced by the larger geometric surface of foam monoliths. As a consequence, we 
think that honeycomb monoliths for the application of the short contact time concept in ethane 
ODH could be preferable, as their straight channels could better preserve ethylene from 
further reacting in secondary paths. 
Chapter 1 Introduction 8 
 
1.3 Understanding the oxidative dehydrogenation of ethane in short 
contact time reactors 
Initially, the process was claimed to occur by a pure heterogeneous reaction 
path, based on the experimental observation of large differences in performance 
among different noble metals (Torniainen et al., 1994). 
Huff and Schmidt (1996) proposed a heterogeneous model to describe ethane 
dehydrogenation on a Pt catalyst in 23 steps. Ethylene formation was assumed to occur on the 
surface following ethane dehydrogenation to ethyl and further dehydrogenation to ethylene. 
CO
x
 and H
2
O formation was accounted for through the decomposition of ethylene to C and H 
and following oxidation of these species. All the chemistry was assumed to occur on the 
catalyst surface, and the gas phase reactions played no role. 
The validity of the results of this simulation may be diminished by the consideration 
that at the high temperature predicted in the reactor (~1000ºC) significant homogeneous 
reactions may occur. 
The experiments and the simulations carried out by Beretta et al. (2000, 2001a and b) 
suggested that at low temperature on a Pt catalyst only CO
x
 and H
2
O are formed. At higher 
temperature, the ethylene yield observed at the exit of the Pt reactor can be entirely attributed 
to the gas phase reactions. The same authors claimed that the catalyst could be detrimental to 
the process, reducing the maximum attainable ethylene yield due to the formation of oxidized 
species (Beretta et al., 2001b). Even with homogeneous reactions, the catalyst is extremely 
important in igniting and sustaining the gas phase reactions (Mulla et al., 2001) and making 
the process feasible at short contact time. 
Henning and Schmidt (2002) performed a thorough investigation of the role of the 
homogeneous reactions in the ODH process. Downstream sampling of a Pt catalyst showed 
that most of the ethylene is formed in the gas phase. The catalyst is effective in oxidizing a 
part of ethane to CO
x
 and H
2
O, developing the heat necessary to sustain the gas phase 
reactions. At the exit of the catalyst, a fraction of O
2
 (∼70%) and a smaller fraction of ethane 
(∼40%) were consumed, while ethylene was mainly formed downstream through fast 
exothermic reactions sped up by the residual oxygen content of the stream, which also 
produces more heat. The addition of H
2
 to the mixture reduced the amount of ethane 
consumed and ethylene formed at the exit of the monolith. Once the O
2
 was consumed, slower 
endothermic reactions took place, leading to the formation of more dehydrogenation products. 
In these experiments, the back heat shield was removed, therefore significantly increasing the 
Chapter 1 Introduction 9 
 
heat losses due to radiation from the back face of the catalytic monolith. Henning showed that 
with addition of a back heat shield, the reactor is more adiabatic, and only 2 mm of catalyst 
are required to sustain the gas phase reactions. Varying the length of a Pt-spheres bed between 
2 and 10 mm, no major differences were found in conversion of the reactants and product 
distribution. 
This result further emphasizes that the role of the catalyst is critical in the first 
millimeters of the reactor in order to ignite and sustain the gas phase reactions, but, once those 
are ignited, the process occurs mainly in the gas phase. 
Based on the experimental evidence of the role of the gas phase reactions, 
Huff et al. proposed a kinetic model for ethane ODH (2000), assuming that the role 
of the catalyst was only to oxidize a fraction of the fuel, a so called sacrificial fuel, 
to CO
x
 and H
2
O. The heat released on the surface drove then the homogenous 
formation of ethylene. The authors proposed a non-adiabatic PF model, assuming 
that the rates of mass transfer between the gas phase and the catalyst surface are 
rapid compared with the corresponding rates of reaction in each respective phase. In 
addition, in the model the heat losses were accounted for as 25% of the heat 
produced on the surface. The unknown kinetic parameters of the surface model 
relative to ethane decomposition and dissociative adsorption and desorption of O
2
 
and H
2
 were adjusted to fit the experimental data. 
The model was based on the assumption of dissociative adsorption of ethane 
to C and H, and following oxidation of C and H. The rates of hydrogen and oxygen 
adsorption and desorption, together with the dissociative adsorption of ethane were 
considered as adjustable parameters. In addition, it was assumed that oxygen could 
adsorb non-competitively on the catalytic surface, while all the other species undergo 
competitive adsorption. This assumption was made in order to reduce the computing 
time of the model. Nevertheless, Bourane and Bianchi (2001) reported experimental 
evidence of non-competitive O
2
 adsorption in a study of CO oxidation on supported 
Pt particles. 
It must be highlighted that the Huff model, even though in good agreement 
with the experimental results relative to conditions under which the data fitting was 
performed, seems to fail in the case of H
2
 addition. 
More recently, Zerkle et al. (2000) proposed a detailed model for ethane 
oxidation on a Pt catalyst. The ethane surface chemistry was based on the model of 
Wolf et al. (1999) for non-oxidative methane conversion on Pt, and combined with 
Chapter 1 Introduction 10 
 
oxidative steps. This led to a heterogeneous mechanism consisting of 19 surface 
species and 82 elementary reactions. The heterogeneous mechanism was 
implemented together with a detailed homogenous mechanism in a 2-D model, since 
the authors claimed that mass and heat transfer are extremely significant factors in 
this kind of reactor. In particular, the mass transfer of the reactants from the bulk of 
the gas phase to the surface of the catalyst and the transfer of the heat produced on 
the surface to the gas phase were considered crucial points. 
The model predicted the percentage of formation of ethylene on the catalyst surface to 
be dependent on the composition of the feed, the ethylene being formed mainly 
heterogeneously when H
2
 was added to the reacting mixture. These findings of all surface 
chemistry are not in agreement with the experimental results reported by Henning and 
Schmidt (2002) and by Beretta et al. (2001b). 
In addition, the model proposed by Zerkle predicts the catalytic oxidation of ethane 
only at high temperature, while experimental observations and the results reported by Beretta 
showed evidence that ethane is oxidized on Pt at temperatures as low as 200ºC, and at even 
lower temperatures with H
2
 addition. 
In conclusion we can say that on the basis of recent experimental results, it is generally 
recognized that ethylene formation occurs in the gas phase. Less clear is what the role of the 
catalyst is, whether it can affect product distribution or is important only as ignitor of the 
reacting mixture. 
 
1.4 Aim of the work 
From the preceding sections results that in the ODH process in short contact time 
reactors a number of questions remains unresolved. In particular, the main issues that will be 
addressed in this work concern the possibility of using a catalyst other than platinum without 
affecting the performance but improving the reliability of the system from the standpoint of 
costs and thermal stability, the study of such a catalyst, focusing on the effect of the active 
phase and of the morphology of the catalyst support, the optimization of the process acting on 
the main parameters of operation of the reactor and finally the implementation of a model of 
the system to provide a deeper understanding of the phenomena occurring in it. 
In the previous sections, we have seen from a number of studies reported in literature 
(Huff and Schmidt, 1993; Flick and Huff, 1999; Beretta and Forzatti, 2001) that until now 
only noble metal-based catalysts turned out to be suitable for the ODH process at short 
Chapter 1 Introduction 11 
 
contact times. From other investigations (Beretta et al., 2001a and b) we have also seen that it 
seems quite established that ethylene formation occurs in the gas phase. Thus, a catalyst with 
good oxidation properties might be suitable to replace Pt, and can be a reliable alternative if 
such catalyst shows also high thermal stability and lower cost than Pt. In this work we 
propose the investigation of a non-noble metal-based catalysts for the oxidative 
dehydrogenation of ethane in short contact time reactors. In this field, we have recently 
developed novel structured catalytic systems, active in hydrocarbons deep oxidation processes 
and alternative to noble metals (Cimino et al., 2000). Specifically, the activity of the LaMnO
3
-
based monolithic catalyst has been proven to be stable in the catalytic combustion of methane 
under 100 hours operation in ignited conditions at 1000°C (Cimino et al., 2001b). To test the 
applicability of such a catalyst in ethane ODH at short contact time, we intend to perform 
reaction tests in comparison with Pt with the aim of evaluating the applicability and reliability 
of a significantly cheaper catalyst (compared to Pt-based ones) in the development of a novel 
and more efficient process of olefin production. 
The investigation of a novel catalyst, such as the LaMnO
3
 perovskite, can also be 
considered as a tool to assess the ultimate role of the catalyst, since changing the active phase 
undoubtedly affects the performance of the system. The direct comparison of a LaMnO
3
-
based catalyst with a Pt-based one and with a blank reactor can help us to deepen the 
understanding on the subject. In particular, the assumption that the catalyst is not important in 
determining the performance of the ODH reaction in a short contact time reactor (Beretta et 
al., 2001b) requires more proof. We believe that it is true that any catalyst able to ignite and 
sustain the gas phase reactions can be in theory employed in the system, but looking at 
ethylene selectivity, it is better that more CO
2
 than CO is formed on the catalyst surface. As 
intuitively understandable and shown in the stoichiometry of the reactions 1.3 and 1.4, less 
ethane is consumed and more heat is produced to react a given amount of oxygen when CO
2
 
is formed. 
72 C
2
H
6
 + O
2
 → 74 CO
2
 + 76 H
2
O ∆H
r
° = - 408.2 kJ/mol 1.3 
52 C
2
H
6
 + O
2
 → 54 CO + 56 H
2
O ∆H
r
° = - 345.1 kJ/mol 1.4 
Platinum is a highly active catalyst, very aggressive towards the oxidation of 
hydrocarbons. When O
2
 is the limiting reactant, a larger amount of ethane is rather oxidized to 
CO than CO
2
. Catalysts other than noble metals, less active but more selective towards the 
oxidation of CO to CO
2
 might be more effective in our process. Also in the gas phase, under 
conditions of short contact time reactors, larger amounts of CO are produced. A catalyst with 
Chapter 1 Introduction 12 
 
a marked activity towards CO
2
 formation may be highly beneficial to the process and perform 
better than Pt or no catalyst at all. 
Perovskite-based catalysts show quite peculiar catalytic properties both in being good 
catalysts of CO oxidation to CO
2
 as reported by Cimino et al. (2002), Ciambelli et al. (2001 
and 2002) and in being able to produce ethylene in mild oxidative conditions (Dai et al., 2000, 
Lee et al., 2001) and we reckon they can be suitable to the process. 
Nevertheless, our goal is to provide a deeper understanding of the properties required 
to be an optimal catalyst in the ODH process. Thus, we intend to extend the investigation to 
different active phases, having in mind the catalytic properties but also the properties of 
thermal stability. In particular, perovskites are extremely stable and we think that their lack of 
activity may be compensated by the dispersion of a noble metal, which is stabilized in the 
oxide matrix and used in lower amount since finely dispersed. Also other oxides, stable at 
high temperature, such as CeO
2
, may be suitable for the applications of ODH. 
The investigation of the catalyst may be pursued also focusing on the effect of other 
parameters of the catalytic reactor affecting the performance of the system. In particular, 
structured catalysts are often coated with a washcoat layer for a number of reasons. The 
washcoat increases the surface area, helps in preventing sintering of the active phase, 
improves the adhesion of the catalyst to the support and definitely changes the catalytic 
system. 
Another aspect that requires careful attention is the morphology of the support. Many 
supports are available on the market, but we will limit our investigation to ceramic supports, 
especially to foam and honeycomb morphology. Foam and honeycomb monoliths are 
different in the degree of randomness and vorticity, affecting dispersion both in radial and 
axial direction and consequently heat and mass transfer. They are different also in the ratio 
between homogeneous volume and geometric surface, which can influence the extent of 
reaction in the homogeneous phase and on the catalytic surface. A low homogeneous volume 
and a large geometric surface can yield to the quenching of the gas phase reactions and then to 
mainly oxidation products. Nevertheless, a large geometric surface per unit of volume is 
important to hold the heat produced on the catalyst, above all in the lab scale reactor.  
After considering the main issues regarding the catalyst and the catalytic reactor, this 
research will face the optimization of the process, and the effect of the main parameters of 
operation. A good regulation of these parameters would allow tuning the process towards the 
desired products and conversions.  
Chapter 1 Introduction 13 
 
The C
2
H
6
/O
2
 ratio affects both the selectivity of the reaction and its global 
exothermicity. Changing the fuel to oxygen ratio the product distribution is moved from 
olefins to syngas to deep oxidation products. 
The addition of sacrificial fuels to the reactor can be important to get additional hints 
about the role of the catalyst. The effect of H
2
 addition has already been studied (Bodke et al., 
2000) on Pt and Pt/Sn, finding deep differences between the two catalysts. The investigation 
of the effect of H
2
 addition will be extended to other active phases in the belief that it can be 
helpful in elucidating the role of the catalyst. The large market value of H
2
 suggests its 
recover rather than recycle to the reactor. We propose a more reasonable alternative such as 
appears the recycle of CO, which has no direct end uses.  
The investigation of other operation parameters, such as preheating and N
2
 dilution, 
will be also addressed in this work. Preheat is an interesting option since it can be easily 
achieved by recovering heat from the product stream. The interest in dilution stems instead 
from the possible choice between feeding pure oxygen (air separation beforehand) or air 
(nitrogen separation from the products afterwards) to the reactor. 
The optimization of the process requires further investigation also on the transient 
behavior of the system, the thermal stability of the catalyst and coke formation. Following the 
transient behavior may give a deeper understanding of the process and is also extremely 
important in the scale-up of the reactor. On a large scale, local hotspots connected to the 
ignition of a portion of the catalyst may yield to serious cracks in the material. In prevision of 
a possible scale-up, the catalyst is required also to be stable under a large number of hours of 
operation to avoid frequent shutdowns for its substitution. For this reason, also coke 
deposition is a serious issue, for instance in conventional steam cracking plants large amounts 
of carbon coke are formed because of the long residence time of the fuel at high temperature. 
In short contact time reactors gases are fed relatively cold and instantly heated up to the 
reaction temperature and also the residence time at high temperature (~1000°C) is relatively 
short (~5ms). Being the conditions of coke deposition different from those studied for steam 
cracking, also this point needs dedicated study. 
Finally, this work will also be devoted to investigate the main parameters dominating 
the whole system. Short contact time reactors for ODH of ethane involve a large number of 
phenomena, which over-impose each other, and understanding of the subject is still poor. To 
this purpose, the tools offered by experimental investigation are not accurate since do not 
allow to separate the different effects. With the help of numerical simulations, we intend to 
study ethylene formation under oxidative conditions at short contact times. 
Chapter 1 Introduction 14 
 
The models proposed until now are useful tools to fit the experimental data in a 
limited range of conditions, but with hundreds of surface and homogeneous reactions, the 
physical meaning of the events occurring in the reactor may be easily lost. 
In the present study, a new approach will be used to model such complicated system. 
First of all, homogeneous and heterogeneous reaction rates will be compared with respect to 
the transport phenomena. Nevertheless, it is difficult under the conditions of reaction to obtain 
surface kinetic data due to the over-imposition of the homogeneous and heterogeneous 
reactions. Thus we preferred to conduct a comparison between the intrinsic gas-phase kinetics 
and the transport fluxes, to provide insights into the real reaction path of ethane towards the 
production of ethylene. The catalytic reactions are controlled by mass transport to the surface. 
Thus, above a certain temperature, mass fluxes to the surface are too slow and the catalytic 
reactions become negligible. 
We aim at developing a mathematical model of a monolithic reactor in which 
homogeneous reactions are coupled to the heterogeneous catalytic combustion of ethane. The 
kinetic model we think to use for the homogeneous reactions is the mechanism proposed by 
Marinov et al. (1996) or by Mims-Dean (1994), for fuel-rich conditions. Instead, on the 
surface we assumed that only the total oxidation of ethane occurs. 
Only one channel of the monolithic reactor will be modeled by means of a two-
dimensional model, where mass and energy balance equations will be coupled with the 
Navier-Stokes equations, with the only simplification of the boundary-layer assumption. The 
system of partial differential equations will be solved by means of the module CRESLAF of 
the CHEMKIN3.6.2 package (Kee et al., 1990). 
This model will be considered a useful tool to determine the optimal conditions of 
operation and the properties required in an optimal catalyst. 
 
 
Chapter 1 Introduction 15 
 
 
Tube Furnace
Air H
2
  CH
4
C
2
H
6 
H
2
O 
H
2
O  CO
2
C
2
H
6
Heavies
Separations 
C
2
H
4
 
 
Fig. 1.1. Schematic of a steam-cracking furnace. A C
2
H
6
/H
2
O mixture is fed to radiant coils in 
a furnace, where the coils are exposed to open flames to achieve a fast heat transfer to the 
reactants. Product stream is then separated to recover ethylene.