3.

Investigaciones cinéticas de laODH de

etano sobre catalizadores de cloruro
alcalino soportados
The oxidative dehydrogenation of ethane is an energetically efficient process to selec-
tively produce ethene. Supported alkali chloride catalysts are a promising class of cat-
alysts, as they show a surprisingly high ethene selectivity. In the previous chapter, the
mechanism of ethane ODH over this class of catalysts has been elucidated, proposing a
novel Mars-van-Krevelen type mechanism, involving two interfaces being responsible for
ethane and oxygen activation. Also, elementary steps for the whole reaction have been
suggested. This chapter focuses on kinetic investigations of the ethane ODH and its side
reactions over supported alkali chloride catalysts. By combining steady state and tran-
sient experiments, kinetic information of the elementary steps could be obtained, finally
compiling kinetic data into an overall mathematical model being able to describe ethane
ODH and its side reactions. Furthermore, it could be shown that the ODH reaction is
not hindered by water and carbon dioxide, both forming as by-products.
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 69

3.1. Introduction

Oxidative dehydrogenation of ethane is a mildly exothermic reaction converting ethane

into ethene, common building block and intermediate for the production of many prod-
ucts. Even offering many advantages over steam cracking, the industrially state-of-the-
art-process, ethane ODH has not yet been exploited for large scale production of ethene.
ODH has however the potential to become economically feasible if outstandingly high
olefin selectivities can be achieved. Main side reactions are the total oxidation reactions
of ethane and ethene -one parallel and one consecutive side reaction. In a kinetic point
of view, high selectivities can be achieved if the ODH is much faster than both side
reactions. Those side reactions play a different role on various catalysts.
This work focuses on the kinetic description of ODH and its side reactions over sup-
ported alkali chloride catalysts, a novel, highly selective catalyst class. Chapter 2 reports
the ethane ODH mechanism over this novel type of catalyst, assuming that oxygen can
permeate through the alkali chloride overlayer, being activated at the interface between
support and melt. The active oxygen-chloride species, forming in low concentrations
and being ideally selective towards ethene, diffuses through the salt layer and activates
ethane at the interface between melt and gas phase. Objective of this work is a detailed
kinetic investigation of the ODH and a quantitative description of the side reactions
over supported alkali chloride catalysts. The elementary steps suggested are taken into
consideration for this model. To achieve this objective, kinetic information of single
elementary steps is determined via transient experiments in combination with steady
state experiments. The findings are used to build a model of the overall reaction. Be-
sides ODH, also the parallel and consecutive side reaction (total oxidation of ethane
and ethane) are investigated and regarded in the model. Detailed studies are performed
with a reference catalyst LiKCl/MgO + Dy2O3. Finally, a kinetic model will be derived
in order to describe all molecular processes occurring during ODH on supported alkali
chloride catalysts. Therefore, a rate equation for the ODH reaction on basis on the
mechanistic information (oxygen and ethane activation step) will be derived and used
in the mathematical fitting procedure. For the side reactions (combustion of ethane and
ethene towards carbon oxides), power-law models will be used.

3.2. Experimental

3.2.1. Catalyst preparation

The catalyst used for this study was prepared by wet impregnation as described in
Chapter 2.2 (entry 5 in Table 4).
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 79

3.2.2. Catalytic tests

Catalytic tests were performed in a plug flow reactor system as described in Chapter 2.
Steady state kinetic experiments were performed at atmospheric pressure, using vary-
ing partial pressures of ethane (99.5%) or ethene (99.5%) and oxygen (10.1% in He)
in balance He (99.99%). Catalysts had been on stream for 2 days before the kinetic
measurements to avoid initial overactivity. WHSV was varied between 0.8 h−1 and 2.0
h−1 in order to keep conversion levels low. Kinetic measurements were recorded at tem-
peratures between 500◦C and 650◦C, using different combinations of hydrocarbon and
oxygen partial pressures (minimum 20 mbar, maximum 100 mbar). For the experiments
to investigate product inhibition, part of the balance helium was substituted by carbon
dioxide or steam. In this case, the WSHV was 0.8 h−1 and reactions were performed in
a temperature window between 450◦C and 650◦C. Step experiments were performed by
heating the catalyst in helium, an oxygen loading period (10% O2 in He, 45 min), an
inert purge period with variable time (40 ml/min He) an a reaction in ethane (10% C2H6
in He, 45 min) at various temperatures. Gas analysis was performed via MS analysis.

3.2.3. Mathematical modeling

The overall kinetic model was implemented in MATLAB using material balances of a
plug flow reactor. The differential equations that describe the changing gas phase com-
positions down the length of the reactor are solved using the differential equation solver
odt23t built in MATLAB. Various parameters (activation energies and pre-exponential
factors of the different reactions) are optimized to fit the experimental data with the
nonlinear optimization function nlinfit. The responses being used in the parameter esti-
mation algorithm are the outlet streams of the products ethene and CO2 in the case of
ethane and oxygen in the feed, the outlet streams of the products CO2 and CO in the
case of ethene and oxygen in the feed .

3.3. Results

3.3.1. Kinetics of ODH and its side reactions on various

catalysts

In oder to determine the influence of different variables, a reference catalyst system

LiKCl/MgO+Dy2 O3 has been studied in detail as it is the reference catalyst used in the
mechanistic studies as well. Kinetic measurements have been performed at differential
conversion levels, varying the temperature and the reactant partial pressures. In order
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 71

to study the ethene total oxidation as well, the same measurements have been performed
with ethene as hydrocarbon in the feed. Table 7 compiles the temperature dependent
reaction orders regarding both reactants for the case of ethane.

Table 7.: Temperature dependent reaction orders regarding ethane and oxygen of
ethane ODH and total oxidation of ethane towards CO2 (LiKCl/MgO +
Dy2O3)

T / ◦C Ethane formation CO2 formation

RO ethane RO oxygen RO ethane RO oxygen
500 0.44 — — —
550 0.39 0.12 — —
600 0.54 0.14 0.97 0.35
625 0.56 0.15 0.77 0.43
650 0.59 0.21 0.63 0.46

For the catalysts supported on LiKCl/MgO + Dy2O3, the reactions order in oxygen is
around 0.1 up to 0.2 and the one for ethane around 0.5. Those values are similar to the
ones obtained for a similar catalyst.[30] It is obvious that the reaction orders increase
slightly with temperature for both reactants. For total oxidation towards CO2 , reaction
orders regarding ethane are higher than the reaction orders regarding oxygen. Both
reaction orders are below 1 for all temperatures.
Steady state operation only allows a determination of apparent reaction orders and
activation energies. Reaction orders for both reactants are notably lower than one.
As the activation of oxygen is rate determining (see Chapter 2) and thus slow,
increasing of the partial pressure of oxygen in the gas phase does not lead to a big
increase of the overall rate. The partial pressure of oxygen in the gas phase directly
only influences step 1 (molecular absorption of molecular oxygen in the liquid salt
melt, see Figure 22). Increasing the concentration of solved molecular oxygen in
the melt has apparently only a low effect on the overall reactivity. As the oxygen
activation step is slower than the hydrocarbon activation step, an increase of the
ethane partial pressure in the gas phase does not correlate linearly with the overall
ODH reactivity. Again, the oxygen activation step is slow, thus the concentration
of active intermediate species mainly governs activity. The reaction order of ethane
is lower than one, as it is present in excess. Reaction orders of both reactants
are between zero and one and change with temperature, which can be explained
by the fact that the kinetics of the elementary steps of ODH obviously change with
temperature. Thus, reaction orders of single steps have to be investigated independently.

Table 8 compiles the temperature dependent reaction orders for the case of ethene.
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 72

Table 8.: Temperature dependent reaction orders regarding ethene and oxygen of total
oxidation of ethene towards CO and CO2 (LiKCl/MgO + Dy2O3)

T / ◦C Ethane formation CO2 formation

RO ethane RO oxygen RO ethane RO oxygen
500 — — — —
550 — — — —
600 -0.26 0.28 — —
625 0.75 0.39 1.59 0.38
650 0.66 0.42 0.67 0.21

Reaction orders regarding ethane are higher than the reaction orders regarding oxygen
for all temperatures.
Complete kinetic data can be found in the supplementary information in Tables 21, 22,
23 and 24.

3.3.2. Influence of carbon dioxide and water on ODH

performance

To investigate to influence of the by-product carbon dioxide on the reactivity of the cat-
alyst for ethane ODH, experiments with a co-feed of 10% CO2 in total were performed.
Results for ethene productivity can be found in Figure 25.
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 73

ln r(C2H4) / ln ( mol s gCat. )

-1
0.36788
-1

0.13534

0.04979

0.01832
0.0011 0.0012 0.0013
-1 -1
T /K

Figure 25.: Influence of CO2 in the feed on reactivity (LiKCl/MgO +Dy2O3, ptotal = 1
bar, pEthane = pO2 = 70 mbar, ❑: without CO2-cofeed; ◦: with CO2-cofeed
(pCO2 = 100 mbar) )

The presence of CO2 does not decrease the ODH activity of the catalyst, thus the
reaction order regarding carbon dioxide in the ODH reaction is zero.
The formation of carbonates and their role as active species, suggested for other similar
catalysts, can thus be excluded. For example the reaction order of CO2 for a LiDyMg
mixed oxide catalyst is -0.5, suggesting that two active Li+-sites are blocked by one
CO2-molecule.[19] For supported eutectic alkali chlorides, it can be speculated that
CO2 does not show an affinity towards the chloride melt, and it can neither adsorb on
the suggested active site OCl − nor block it. The adsorption of CO2 on Li+ or K+ on
the surface of the melt is not likely due to the dynamically rearranging surface, but
can, however, not be excluded even for this supported chloride catalyst. As the cations
are not the active centers of a supported alkali chloride catalyst, the non-existing
blocking effect of CO 2 can be explained. Another hypothesis for the extraordinarily
high selectivities of supported alkali chloides is the temporary absorption of CO2 in the
liquid melt, what would enhance the olefin selectivity for a limited time. By co-feeding
CO2 for a longer time, this effect can be excluded. Thus, supported molten alkali
chlorides do not seem store carbon dioxide. For the kinetic model, the reaction is thus
zero order in carbon dioxide. Thus, CO2 does not have to be regarded in the derivation
of the rate equation of ethane ODH.

Additionally, the influence of water, product of ODH and side reactions, has been inves-
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 74

tigated. Measurements have been performed at different temperatures, as summarized

in Figure 26.

0.4
ln r(C2H4) / ln ( mol s gCat. )
-1
0.2
0.0
-0.2
-1

-0.4
-0.6
-0.8
-1.0
-1.2
-1.4
-1.6
-1.8
2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4
ln pH2O / ln (mbar)

Figure 26.: Effect of water on ODH reaction: □: 550◦C; •: 600◦C; K: 625◦C; V:

650◦C;

Apparently, water does not influence the reactivity of the ODH reaction thus a reaction
order of zero as well. Even water might be stored in low concentrations in the chloride
melt (see Chapter 2), it does definitely not affect the activity in ethane ODH. A blocking
of the active centers by water can thus be excluded. Furthermore, it can be excluded
that OCl −, the catalytically active intermediate, decomposes in the presence of water.
ODH reactivities thus only depend on the partial pressures of the reactants ethane and
oxygen.

3.3.3. Ethane ODH and its side reactions

In order to investigate whether COx is formed directly from ethane or via ethene, steady
state experiments with different space velocities and conversion levels have been per-
formed. Figure 27 presents the results of this study.
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 75

40

Y/%
20

0
20 30 40 50
X(C2H6) / %

Figure 27.: Steady state experiment with Li − K − Cl/MgO + Dy2O3 with variable
WHSV (T = 625◦C, p total = 1 bar, p Ethane = pO2 = 70 mbar, WHSV
variable; □: Ethene; •: CO2

Ethene yields increase linearly with conversion up to an ethane conversion of approx-

imately 40%, thus the olefin selectivity is constant up to this conversion. At higher
conversions, the linear trend slightly changes towards less olefin selectivity. Whereas
the CO2 curve also has a constant slope up to 40%, the yields at conversion levels over
40% increase less. Thus, CO x is mainly formed from ethane directly. However, at higher
ethene partial pressures in the gas phase, part of the formed ethene is converted to COx
in a consecutive side reaction. CO x formation pathways are further illustrated in Figure
28, where COx formation rates are presented, using both ethane and ethene as reactants
together with oxygen.

55 55
50 50
45 45
40 40
-1 -1

-1 -1
r(COx) / mol min Cat.

r(COx) / mol min Cat.

g

35 35
30 30
25 25
20 20
15 15
10 10
5 5
0 0
550 575 600 625 650 550 575 600 625 650
T / °C T / °C

(a) (b)

Figure 28.: COx formation rates from ethane (a) and ethene (b) (black: CO2, white:
CO; WHSV = variable, pO2 = pEthane = 80 mbar, ptotal=1 bar )
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 76

Considerable COx formation starts at 600◦C. CO 2formation rates are in general higher
for ethane as precursor. While CO is formed only at 650◦C using ethane as pre-cursor,
it can already be detected at 625◦C for the case of ethene. Whereas the dominant total
oxidation product is CO2 for the case of ethane, CO is formed in more considerable
amounts for the case of ethene as pre-cursor. For most of the temperatures studied,
more by-products form in the case of ethane as pre-cursor. Only for the temperature
of 625◦C, CO starts forming for ethene as pre-cursor, the total amount of formed by-
products is higher for the case of ethene in this case. Figure 29 presents the influence of
ethene selectivitiy on the oxygen partial pressure. It can be observed that more oxygen
in the feed leads to decreased selectivities. However, at 600◦C ethene selectivities are
still at very high levels.

95

94
S(C2H4) / %

93

92
20 30 40 50 60 70 80
p(O2) / mbar

Figure 29.: Influence of oxygen/ethane ratio on ethene selectivity (WHSV variable,

T=600◦C, pEthane = 40 mbar, ptotal =1 bar

It can be seen that the CO x formation rate using ethene as reactant is slightly higher
as the respective one with ethane as reactant. This explains why the olefin selectivity
decreases slightly with increasing ethene partial pressures in the gas phase. Blank tests
with only SiC in the reactor showed neglible ethane and ethene conversions in both
studies.

3.3.4. Elementary steps of ODH

Transient experiments, as presented in the previous chapter, allow a decoupling of the

oxygen activation, transport and hydrocarbon activation steps. As the oxygen loading
of the catalyst was performed at the same conditions but different temperatures, the
amount of stored oxygen for one temperature can be determined. The concentration
of the active intermediate can be determined by integrating the amount of the whole
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 77

ethene formed after a step response, this value corresponds to the amount taken up
by the catalyst at a certain temperature. Additionally, the initial ethene formation
rate after a step response can be determined, which provides exact information for the
hydrocarbon activation step. By varying ethene concentration the reaction order of the
hydrocarbon activation step can be determined.
Figure 30 shows that the initial ethene formation rate after a step increases linearly with
the ethane partial pressure, the reaction order as approximately 1 for all investigated
temperatures.

-0.5
ln rinitial(C2H4) / ln ( mol s gCat. )
-1

-1.0
n=0.99
-1

-1.5

-2.0
n = 1.02
-2.5

-3.0
n = 0.96
-3.5

3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8

ln pC2H6 / ln (mbar)

Figure 30.: Reaction orders of ethane activation after step experiment ( V: T = 575◦C;
◦ ◦
• : 600 C; □: 625 C)

Transient step experiments allow a decoupling of the different elementary steps, thus
kinetics of a single reaction step can be determined. Here, the kinetics of the hydrocarbon
activation step could be probed. In the step experiment, the oxygen activation step
was kept unchanged and the melt was saturated with oxygen, thereby forming the
equilibrium concentration of active intermediates. As the active intermediate is formed
during oxygen loading, oxygen activation is not rate determining after switching the
effluent stream from oxygen to ethane in a step experiment. An evaluation of the
initial ethene formation rates with different ethane partial pressures thus allows the
determination of the true reaction order in ethane. Here, the reaction order is one
compared to 0.5 in steady state operation mode, showing that ethane activation is an
elementary step. Furthermore, it shows that reactivity linearly correlates with collision
probability of ethane with an active intermediate species on the surface of the melt.
Reaction orders for complex rate equations (i.e. Mars-van-Krevelen type rate equation)
can depend on several kinetic parameters and partial pressures, thus they can change
for different regimes.[189]
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 7
8
3.3.5. Kinetic model

3.3.5.1. Derivation of rate equation for ethane ODH

In Chapter 2, a comprehensive mechanistic model with 9 single steps was derived. For
kinetic modeling, the mechanistic model is simplified, as kinetic information of each
of the 9 steps cannot be measured. Thus, ethane ODH is formally divided into two
elementary steps:

0.5 O2 + [Cl−] → [OCl−] (20)

[OCl − ] + C2H6 → C2H4 + H2O + [Cl − ] (21)

Eq. 20 represents the oxygen activation step, eq. 21 represents the hydrocarbon activa-
tion step.
Rate equations can be written as following (eq. 22 for the oxygen activation step, eq.
24 for the hydrocarbon activation step):

√
r1 = k1· p O2 · [Cl− ] (22)

r2 = k2· p C2H6 · [OCl − ] (23)

A chloride balance completes the equation system:

[Cl −] + [OCl − ] = 1 (24)

In steady state, both steps must have the same rate:

√
k1· p O2 · [Cl−] = k2· p C H
2 6
· [OCl −] (25)

Now, [Cl−] can be determined:

[Cl−] = k2· pC2H6 · [OCl −] (26)

√
k1 · pO2

Now, [OCl-] can be calculated (inserting expression for [Cl−] in chloride balance (eq.
24):
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 79

[OCl−] + k2· pC2H6 · [OCl −] = 1 (27)

√
k1· pO2

√ √
k1 · p O2 − k1· p O2 · [OCl −] = k2· pC2H6 · [OCl − ] (28)

√ √
k1 · p O2 = [OCl − ] · (k2· pC H + k1 · p O
2 6
) (29)
2

√
− k1 · pO
[OCl ] = √ 2 (30)
k2· pC2H6 + k1 · pO2

[OCl−] can now be inserted into eq. 24:

√
k1 · pO · k2 · pC2 H6
2
r2 = rtotal = √ (31)
k1 · pO2 + k2 · pC2H6

3.3.5.2. Reaction network including side reactions

For kinetic modeling, the following reactions are taken into consideration:

1. C2H6 + 0.5 O2 → C2H4 + H2O (32)

2. C2H6 + 3.5 O2 → 2 CO2 + 3 H2O (33)

3. C2H6 + 2.5 O2 → 2 CO + 3 H2O (34)

4. C2H4 + 3 O2 → 2 CO2 + 2 H2O (35)

5. C2H4 + 2 O2 → 2 CO + 2 H2O (36)

Both ethane and ethene can undergo total oxidation, thus both parallel (reactions 2 and
3) and consecutive (reactions 4 and 5) side reactions are possible. The reaction network
of ODH and its side reactions is summarized in Figure 31.
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 80

CO2

k2 k4

k1
C2H6 C2H4

k3 k5

CO

Figure 31.: Scheme of ethane ODH and its side reactions

Ethane is directly transformed to ethene via ODH, and both products can undergo
total oxidation towards carbon monoxide and dioxide. Carbon monoxide has only been
observed at high temperatures, mainly in the experiments with ethene in the feed. Thus,
reaction 3 has been neglected in the modeling, as CO is only assumed to form directly
from ethene. The modeling strategy involves two steps: At first, experiments with
ethene and oxygen are taken into consideration. This allows to determine the kinetics
of reactions 4 and 5 decoupled from the rest of the reaction network. As CO and CO 2
are assumed to form via a gas phase mechanism which is not known in detail, reactions
4 and 5 are modeled with power law equations.

b
r4 = k4 · paO2 · pC2H4 (37)

r5 = k5 · pcO2 · pC2H4
d
(38)

The exponents are the average of the temperature dependent reaction orders in ethene
and oxygen. Both rate constants are parametrized with activation energies and pre-
exponential factors.
In a second step, experiments with ethane and oxygen in the feed are taken into con-
sideration. The fitted parameters obtained in the first step (activation energies and
pre-exponential factors of reactions 4 and 5) are used for the parameter fitting of reac-
tions 1 and 2as fixed values. Now, parameters for reactions 1 and 2 can be fitted. The
kinetic model for ODH is based on the chemistry and elementary steps suggested in
Chapter 2. It contains two components: Oxygen activation (taking place at the inter-
face between support and melt) and hydrocarbon activation (occurring at the surface of
the melt). The equation derived above is similar to a Mars-van-Krevelen mechanism as
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 81

reported in literature.[24, 190] This model contains the separation of oxygen and hydro-
carbon actvation. Thus, two rate constants are needed in this equation. Rate constants
are parametrized each with an activation energy and a pre-exponential factor, which are
optimized.
√
k1,1 · pO2 · k1,2· pC2H6 (39)
r1 =
k1,1 · √p 2
+ k1,2· pC2H6
O

CO2 formation, not assumed to proceed via the OCl − species and assumed to form via a
gas-phase mechanism, has been modeled with a power law equation. For the exponents of
oxygen and hydrocarbon partial pressures, the average of the experimentally determined
activation orders at different temperatures has been used.

r2 = k2 · peO2 · pC2H6
f
(40)

CO formation from ethane (reaction 3) is neglected, as CO mainly forma from ethene

(see Figure 28).

3.3.6. Modeling results

3.3.6.1. Total oxidation of ethene

For reaction 4 (CO2-formation from ethene), the following reaction orders have been
used for the power-law-model: 0.7 for ethane, 0.41 in oxygen (average of experimentally
determined temperature dependent reaction orders). For reaction 5 (CO-formation from
ethene), the following reaction orders have been used for the power-law-model: 0.34 for
ethane, 0.1 in oxygen (average of experimentally determined temperature dependent
reaction orders). Figure 32 shows the parity plots for CO2 and CO formation rates.
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 8
2

25 25

20 20
-1

-1
rsim(CO2) / mol min gCat.

rsim(CO) / mol min gCat.

-1

-1
15 15

10 10

5 5

0 0
0 5 10 15 20 25 0 5 10 15 20 25
r (CO2 ) / mol min-1 gCat.-1
exp r exp(CO) / mol min-1 Cat.
g -1

(a) (b)

Figure 32.: Parity plots of CO2 and CO formation rates (experiment with oxygen and
ethene in feed)

Despite a change in the reaction order of both reactants, CO 2 formation rates are de-
scribed well by the power-law-model. For CO however, no good fit could be obtained.
Compared to CO2 , CO formation rates are smaller and are not that fundamental for
the modeling of the whole reaction system. Table 9 presents the optimized parameters
for the formation of both carbon oxides.

Table 9.: Optimized parameters for total oxidation reactions with ethene as pre-cursor

C2H4 formation CO2 formation

EA,1 A1 EA,2 A2 EA A-
kJ mol−1 - kJ mol−1 - kJ mol−1 -
107.1 1.81E9 163.9 8.8E11 283.5 4.4E18

3.3.6.2. ODH of ethane and total oxidation of ethane towards CO2

For reaction 2 (CO2-formation from ethane), the following reaction orders have been
used for the power-law-model: 0.79 for ethane, 0.41 in oxygen (average of experimentally
determined temperature dependent reaction orders). Figure 33 shows the parity plots
of ethene formation via ODH and CO 2 formation.
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 8
3

50
70

60 40
-1

-1
rsim(C2H4) / mol min gCat.

rsim(CO2) / mol min gCat.

50
-1

-1
30
40

30 20

20
10
10

0 0
0 10 20 30 40 50 60 70 0 10 20 30 40 50
-1 -1
r (C2H4 ) / mol min g
exp Cat. r exp(CO 2) / mol min-1 Cat.
g -1

(a) (b)

Figure 33.: Parity plots of ethene and CO 2 formation rates (experiment with oxygen
and ethane in feed)

For both products, experimental and simulated data fit reasonably well. The Mars-
van-Krevelen model seems to describe a complex cascade of reaction steps well, despite
the change in reaction orders of both reactants in temperature. Table 10 presents the
optimized values for ODH and total oxidation of ethane towards CO2 .

Table 10.: Optimized parameters for ethene and CO2 formation with ethane as pre-
cursor (EA,1 and A1 relate the to the rate constant k1,1, EA,2 and A2 relate
the to the rate constant k1,2)

C2H4 formation CO2 formation

−1 −1
EA,1 / kJ mol A1 / - EA,2 / kJ mol A2 / - EA / kJ mol −1 A / -
163.9 8.8E11 107.1 1.81E9 283.5 4.4E18

The activation energy for the oxygen activation is significantly higher than the one
for the hydrocarbon activation, consistent with results presented in Chapter 2. Thus,
the oxygen activation has the higher activation barrier compared to the hydrocarbon
activation. Furthermore, oxygen activation shows a higher temperature dependence
compared to the hydrocarbon activation. It lies in the range of the activation energy
of the formation of the mixed oxygen isotope (190 kJ /mol) as showm in the TPIE
experiment (see Chapter 2). Thereby, the activation energy of the oxidation step is
validated by kinetic modeling. Under differential conditions with elevated flows through
the reactor, the activation energy for CO2 production is higher than for the case of lower
flows. Total oxidation of ethane towards CO2 is assumed to be a gas phase reaction, gas
dynamics and back-mixing phenomena can play a role, which can explain the difference.
At higher space velocities as used for the kinetic measurements, the formation of CO2
CHAPTER 3. KINETIC INVESTIGATIONS OF THE ODH OF ETHANE OVER
SUPPORTED ALKALI CHLORIDE CATALYSTS 8
4
can behave differently.

3.4. Summary and Conclusions

Ethane ODH reactivity over supported alkali chloride catalysts is neither influenced by
water nor by CO 2, thus the reaction order of those molecules is zero. Reaction orders
were lower for oxygen compared to ethane in the ODH reaction. Based on a complex
mechanistic model presented in Chapter 2, a rate equation for ODH could be derived,
both addressing an oxidation and hydrocarbon activation step. Furthermore, a complex
kinetic model including ODH and its side reactions was used to fit kinetic parameters
to experimental data for ODH, total oxidation of ethane and total oxidation of ethene.
This model describes that data well. The activation energy of the oxidation step of
the ODH reaction is higher compared to the one of the hydrocarbon activation step,
which is consistent to previous results. Activation energies for total oxidation reactions,
however, are higher compared to the ones of ODH, thereby reflecting the high olefin
selectivities for catalysis over supported alkali chloride catalysts.
4. Tailoring of novel, highly
selective supported melt
catalysts for the oxidative
dehydrogenation of ethane

Novel supported alkali chloride catalysts were investigated for the oxidative dehydro-
genation of ethane as target reaction. Various factors have been varied to study the
influence of certain properties on the catalytic performance: The influence of different
oxides as supports, the influence of a different anion (bromide instead of chloride) in
the melt and the influence of different cations in a chloride eutectic system. Also, dcat-
alysts prepared with different preparation methodes (chloride and nitrate precursors)
have been investigated tested in long-term behavior. Metal oxides which could undergo
a solid state reaction with chloride do not seem beneficial for the ODH performance, as
substantial deactivation occurs. Especially supports forming volatile chlorides are not
suitable for efficient ODH catalysts, as they would lead to irreversible chloride deple-
tion. Bromides also tend to catalyze ODH, even with increased activities, but lower
olefin selectivities. Thus, the formation of an oxo-halide species as catalytically active
intermediate is also possible with other halogenides. Regarding the overlayer, ternary
and quaternary chloride eutectic systems were explored, some of them with melting
points below 300◦C. Two of those catalysts show olefin selectivities up to 98%, being
among the highest selectivities reported for ethane ODH. This work consists of charac-
terization of the catalysts coupled with reaction studies, describing and discussing the
requirements for good and efficient supported melt catalysts for ethane ODH.
CHAPTER 4. TAILORING OF NOVEL, HIGHLY SELECTIVE SUPPORTED MELT
CATALYSTS FOR THE OXIDATIVE DEHYDROGENATION OF ETHANE 86

4.1. Introduction

The growing market of light olefins spurs the development of novel technologies to selec-
tively produce them from alternative feedstocks. Oxidative dehydrogenation of ethane
towards ethene is a promising route, allowing very high olefin selectivities and lower pro-
cess temperatures compared to stream cracking, the industrially well-established process
for ethene production. Supported alkali chloride catalysts have been reported to be an
excellent system for the oxidation of ethane, especially allowing outstandingly high selec-
tivities (up to 95%).[51, 67] Specific feature of this class of catalysts is an alkali chloride
layer on top of a solid support. This overlayer is molten under reaction conditions,
thus rearranging dynamically. As alkali chlorides per se have a high melting point, eu-
tectic mixtures of different alkali chlorides have been employed so far to decrease the
melting point, i.e. LiCl-KCl, LiCl-NaCl, LiCl − SrCl 2 . It has been reported that a
low melting point of the overlayer leads to high olefin selectivities, thus being the aim
for future catalyst development. In a previous chapter , the mechanism of ODH over
LiKCl/MgDyO as a model system has been reported: Oxygen is absorbed by the molten
overlayer of the catalyst, diffuses through it and is activated at the interface between
the solid support and the melt. There, highly active intermediate species are formed
(i.e. hypochlorite), which diffuse back to the surface of the overlayer. Here, the interme-
diate species catalyzes the C-H bond activation, leading to the cleavage of C-H bonds
and subsequent ethene formation. It could thus be concluded that the ODH reaction on
supported alkali chloride catalysts proceeds via a Pseudo Mars-van-Krevelen mechanism
with spatial separation of the activation of oxygen and the activation of ethane. The
mobility of the active species within the melt and site isolation is the key for the out-
standingly selectivities that can be achieved. This chapter focuses on the design of novel
catalysts with even higher activities and selectivities. This aim has been approached
by varying both the overlayer and the support, while the support is known to mainly
govern activity, while the overlayer tunes the selectivity. Thus, different novel supports
and eutectic overlayers will be employed. The required features of support and overlayer
will be examined and discussed. Thus, several factors for molten alkali chloride ODH
catalysts will be introduced and discussed, and the requirements for tailoring new and
efficient ODH catalysts will be defined.
CHAPTER 4. TAILORING OF NOVEL, HIGHLY SELECTIVE SUPPORTED MELT
CATALYSTS FOR THE OXIDATIVE DEHYDROGENATION OF ETHANE 87

4.2. Experimental

4.2.1. Catalyst preparation

Catalysts were prepared by wet impregnation in two different ways for different studies.
One part of the catalysts was prepared via the addition of alkali chlorides to a slurry
of the support material in 100 ml deionized water as described in Chapter 2. The ref-
erence catalyst (LiKCl supported on MgO) was prepared by adding the alkali chlorides
(1.04 g LiCl (Aldrich, Reagent Plus) and 1.19 g KCl (Merck, 99.995%) to a slurry of
100 ml deionized water and 6.69 g MgO. For the study with different supports, 0.166
mol support were used for each synthesis (13.25 g TiO 2 Rutile (Aldrich, 99.99%), 9.97
g SiO2 (Aldrich,99.5%), 38.47 g WO 3 (Aldrich, >99%)). 1.02 g LiCl (Aldrich, Reagent
Plus) and 1.19 g KCl (Merck, >99.995%) were added to the slurry. For the study with
different halides, 6.45 g MgO (Aldrich, 99.9%) and 0.6 g Dy2O3 (Aldrich,99,9%) were
used as support materials. For the overlayer, 2.08 g LiBr (Aldrich, >99%) and 1.9 g
KBr (Merck, 99.5%) were added to the slurry.
Another part of catalysts was prepared via a method using nitrates as pre-cursors.[51, 67]
The reference catalyst (LiKCl supported on MgO) was prepared by adding the alkali
nitrate precursors (1.66 g LiNO3 (Aldrich, 99.99%) and 1.62 g KNO 3 (Merck, 99.995%)
to a slurry of 100 ml deionized water and 6.69 g MgO. Additionally, 1.02 g of NH 4 NO 3
(Aldrich, 99.99%) and 1.57 ml of HCl (Aldrich, ACS reagent 37%) were added to the
slurry as chloride source. For the study with different chloride eutectica, 6,64 g MgO
(Aldrich, >99%) and 0.6 g Dy2O3 (Aldrich,99,9%) were used as support materials, the
chloride source remained unchanged. The different eutectic systems were established
by mixing different alkali metal nitrates (LiNO3 (Aldrich, 99.99%), KNO 3 (Aldrich,
99.99%), NaNO 3 (Aldrich, 99.995%), RbNO3 (Aldrich, 99.7%), CsNO 3 (Aldrich,
99.99%), Sr(NO3)2 (Aldrich, >99%), Ba(NO3)2 (Aldrich, >99%)) . The compositions
can be found in Table 11.

4.2.2. Elementary analysis

Solid catalysts were dried and solved in different acidic media. Cl − was quantified by
titration with AgNO3 , using a Metrohm Titrando 904. Mg, Li and K were quantified
by AAS (Agilent AAS280FS). Ti, Si and W were quantified by photometry (Shimadzu
UV-160 UV Spectrometer).