A reaction pathway for the ammoxidation of ethane and ethylene over Co-ZSM-5 catalyst

Article abstract of DOI:10.1006/jcat.1998.2089

The ammoxidation of ethane and ethylene to acetonitrile was studied over a Co-ZSM-5 catalyst with an emphasis on the reaction pathway. We found that the adsorption of ammonia on Co-ZSM-5 is stronger than that on H-ZSM-5. CzH4 adsorption is weak and readily desorbs below 300°C. While the adsorption of C₂H₃N (a reaction product) is very strong in He, its desorption is accelerated with the presence of NH₃. With a specially designed temperature programmed experiment (reaction between the adsorbed NH₃ and gaseous C₂H₄/O₂/He mixture), we observed C₂H₅NH₂ as a reactive intermediate, and this intermediate was demonstrated to be readily converted to C₂H₃N under the ammoxidation reaction conditions. A detailed pathway is offered, whereby C₂H₄ is thought to add on an adsorbed NH₃, forming an adsorbed ethylamine which is subsequently dehydrogenated to form C₂H₃N. We further speculate that an oxidative environment N₂ comes from N-N pairing between the adsorbed NH₃ and an amine (both on a single Co²⁺ site).

Full text of DOI:10.1006/jcat.1998.2089

JOURNAL OF CATALYSIS 176, 495–502 (1998)
ARTICLE NO. CA982089
A Reaction Pathway for the Ammoxidation of Ethane and Ethylene
over Co-ZSM-5 Catalyst
Yuejin Li1 and John N. Armor
Air Products and Chemicals, Inc., 7201 Hamilton Boulevard, Allentown, Pennsylvania 18195
Received October 24, 1997; revised March 18, 1998; accepted March 24, 1998
intermediate for C2H6 ammoxidation. To simplify the dis-
cussion regarding the reaction pathway, in this manuscript,
we use C2H4, instead of C2H6, as a reactant for the ammox-
idation reaction (Eq. [2]).
The ammoxidation of ethane and ethylene to acetonitrile was
studied over a Co-ZSM-5 catalyst with an emphasis on the reaction
pathway. We found that the adsorption of ammonia on Co-ZSM-5
is stronger than that on H-ZSM-5. C2H4 adsorption is weak and
readily desorbs below 300 C. While the adsorption of C2H3N (a
reaction product) is very strong in He, its desorption is accelerated
with the presence of NH3. With a specially designed temperature
programmed experiment (reaction between the adsorbed NH3 and
gaseous C2H4/O2/He mixture), we observed C2H5NH2 as a reactive
intermediate, and this intermediate was demonstrated to be readily
converted to C2H3N under the ammoxidation reaction conditions.
A detailed pathway is offered, whereby C2H4 is thought to add
on an adsorbed NH3, forming an adsorbed ethylamine which is
subsequently dehydrogenated to form C2H3N. We further speculate
C2H6 + NH3 + 1.5O2 → C2H3N + 3H2O
C2H4 + NH3 + O2 → C2H3N + 2H2O
[1]
[2]
Temperature programmed experiments(desorption and re-
action) were carried out over a Co-ZSM-5 catalyst to study
the adsorption features of NH3, C2H3N, and C2H4. We were
able to capture a reactive intermediate, C2H5NH2, with a
specially designed, in situ reaction and confirmed that this
intermediate was readily converted to C2H3N under the
ammoxidation conditions.
–
that an oxidative environment N2 comes from N N pairing between
the adsorbed NH3 and an amine (both on a single Co2+ site).
1998
c
Air Products and Chemicals, Inc.
EXPERIMENTAL
The Co-ZSM-5 catalyst was prepared by cation exchange
of the ammonium form ZSM-5 with cobalt acetate solu-
tion at 80 C, and the detailed procedures for this prepara-
tion were described earlier (2). The ammoxidation reaction
runs with C2H4 were conducted using the same apparatus
as C2H6 ammoxidation (2). The concentration of NH3 was
varied between 2 and 15% , that of hydrocarbon between
5 and 15% and O2 between 1 and 8% . The catalysts were
first pelletized, crushed, and sieved to 20–40 mesh before
loading in the reactor. (The apparent density of the cata-
lyst sample is 0.5 g/cm3.) For plug flow reaction studies, the
total flow rate was 100 ml/min with 0.05 g catalyst, which
renders a space velocity 60000 h 1. A catalyst was routinely
pretreated with flowing helium at 500 C for 1 h before a
reaction run.
INTRODUCTION
Recently, we reported that ethane can be efficiently con-
verted to acetonitrile (Eq. [1]) in the presence of ammo-
nia and oxygen (ammoxidation reaction) over some metal
exchanged zeolite catalysts (1, 2). Among the most effec-
tive catalysts for this reaction are Co2+ exchanged ZSM-5,
beta and NU-87. These catalysts have much higher reac-
tion rate and yield versus most oxide-based catalysts for
the ammoxidation reactions (1, 3, 4). Based on our earlier
reaction studies, we propose a three-step pathway to ac-
count the C2H6 ammoxidation reaction. The first step is ox-
idative dehydrogenation of C2H6 to C2H4; C2H4, a reactive
hydrocarbon and a major product for the reaction, is fur-
ther added to an adsorbed NH3 molecule (on a Co2+ site)
forming an adsorbed C2H5NH2 intermediate; this amine
intermediate is then converted to C2H3N. In this contribu-
tion,we report data to elaborate and further substantiate
this reaction scheme. C2H4 was found to be a hydrocarbon
For C2H4 ammoxidation the major carbon containing
products are C2H3N and CO2, and the nitrogen contain-
ing products are C2H3N, N2, and small amounts of N2O.
The conversion, selectivity and yield are defined as
X
X
Conversion of ethene, X =
yi ni
y E n E +
y i n i
,
1 Current address: Engelhard Corporation, 101 Wood Avenue, Iselin,
New Jersey 08830-0770.
i
i
[3]
495
0021-9517/98 $25.00
1998 by Air Products and Chemicals, Inc.
c
Copyright
All rights of reproduction in any form reserved.

496
LI AND ARMOR
Selectivity of product P (carbon basis),
for 1 h) of Co-ZSM-5, NH3 was first adsorbed on Co-ZSM-5
i
at 25 C. Then a stream of C2H4 (5% )/O2 (5% )/He mixture
was passed through the catalyst at 150 cc/min. Upon stabi-
lization of the flow, the temperature was ramped to 525 C at
a rate of 10 C/min while monitoring relevant mass species
with RGA.
X
Si = yi ni
yi ni ,
[4]
[5]
i
Yield of product P , Yi = X Si ,
i
where yi and yE are the mole fractions of carbon containing
RESULTS
product P and C2H4, respectively; ni and nE are the number
i
of carbon atoms in each molecule of product P and ethene,
respectively, and all the terms were evaluated for the exit
stream.
i
Reaction Studies for C2H4 Ammoxidation
When a stream of C2H4/NH3/O2/He flows through a
Co-ZSM-5 catalyst at T > 350 C, a mixture of products are
produced. Acetonitrile dominates the carbon-containing
products, with small amounts of CO2 formation and trace
amounts of CO (<1% ), HCN (<1% ), and C3H5N (<1% ).
In addition, N2 is a major product of NH3 conversion with a
trace amount of N2O formation. The conversions and selec-
tivities for the C2H4 ammoxidation reaction are presented
in Figs. 1–3 as a function of inlet [C2H4], [NH3], and [O2] in
feed.
Selectivity of product Pj (nitrogen basis),
X
Sj = yj n j
yAnA +
yj n j
,
[6]
j
where yj and yA are the mole fractions of nitrogen contain-
ing product Pj and NH3, respectively; n j and nA are the
number of nitrogen atoms in each molecule of product P
i
and ammonia, respectively.
As shown in Fig. 1, the selectivity of C2H4 to C2H3N is
generally high (50–80% ) and increases with [C2H4]. Con-
versely, the conversion of C2H4 decreases with increasing
[C2H4]. NH3 has a moderate C2H3N selectivity, which is
higher at high [C2H4]. Interestingly, a higher C2H4 partial
pressure increases the utilization of both NH3 and C2H4.
The O2 is almost completely (not shown in Fig. 1) converted
except at the inlet [C2H4] of 15% (80% O2 conversion).
Figure 2 shows the [NH3] dependence. Both C2H4 conver-
sion and selectivityincrease with increasing[NH3]. The NH3
conversion and selectivity are slightly lower at higher inlet
A residualgasanalyzer (or RGA, UTI 100C) wasused for
all the temperature-programmed experiments. This RGA is
capable of monitoring 12 different mass species simultane-
ously as a function of time. TPD experiments were carried
out over Co-ZSM-5 with NH3, C2H3N, C2H4, and C2H5NH2,
respectively. Co-ZSM-5 (0.1 g) was first pretreated in flow-
ing He at 500 C for 1 h and cooled to an adsorption tem-
perature. The adsorption temperatures are 100 C for NH3
and C2H3N and 25 C for C2H4. After the Co-ZSM-5 cata-
lyst was saturated with an adsorbate (in 3000 ppm NH3/He,
3000 ppm C2H4/He, and 1.2% C2H3N, respectively), the
catalyst was flushed with He at the adsorption temperature
with a flow rate of 150 cc/min. When the mass intensity of
the species monitored by RGA is below a reasonable level
(or at a certain cutoff time, e.g., 20 min for NH3), the tem-
perature was ramped to 550 C at a ramp rate of 10 C/min.
Relevant mass numbers were continuously monitored as a
function of time.
For the C2H5NH2 TPD experiment, the Co-ZSM-5 cata-
lyst was contacted with C2H5NH2 vapor at 25 C. (A stream
of He (50 cc/min) flowed through a C2H5NH2 bubbler at
77 C; then it was further diluted with another stream of
He at 50 cc/min just before entering the reactor.) After
flushing the system with He (150 cc/min) for about 20 min,
a stream of NH3/O2/He mixture (3.3% NH3, 3.9% , O2) was
passed through the catalyst bed at 150 cc/min until the RGA
signals were stabilized. The temperature was then raised
from 25 to 525 C in flowing the mixture gas (150 cc/min) at
a ramp rate of 10 C/min. Mass factions at 2, 17, 18, 27, 28,
41, and 45 were monitored simultaneously.
The temperature-programmed reaction between ad-
sorbed NH3 and a gaseous mixture of C2H4/O2/He was con-
ducted over Co-ZSM-5. After pretreatment (500 C in He
FIG. 1. C2H4 ammoxidation as a function of [C2H4]. The reaction run
was carried out at 425 C with 0.05 g Co-ZSM-5, F = 100 cc/min, [NH3] =
10% , and [O2] = 6.5% .

AMMOXIDATION OF ETHANE AND ETHYLENE
497
FIG. 4. Comparison of TPD profiles between NH3 adsorbed on Co-
ZSM-5 and H-ZSM-5.
FIG. 2. C2H4 ammoxidation as a function of [NH3]. The reaction run
was carried out at 425 C with 0.05 g Co-ZSM-5, F = 100 cc/min, [C2H4] =
10% , and [O2] = 6.5% .
Temperature Programmed Desorption
of Ammonia (TPDA)
NH3 levels. Similar to the C2H6 ammoxidation reaction (2),
O2 is a strong driving force for the C2H4 ammoxidation re-
action, and in the absence of O2 no reaction occurs. Both
NH3 and C2H4 conversion are enhanced at higher [O2]. The
C2H4 selectivity only slightly decreased (from 80 to 75% at
[O2] of 3.5 to 8% ), while the selectivity of NH3 is not af-
fected by the O2 level (Fig. 3). Using these partial pressure
dependence data, we obtained the empirical reaction rate
(for C2H3N) orders of 0.62, 0.74, and 1.0 with respect to
[C2H4], [NH3], and [O2], respectively.
Temperature programmed desorption of NH3 over a
Co-ZSM-5 catalyst was conducted and compared with
H-ZSM-5 (see Fig. 4). The NH3 adsorption from the pro-
ton sites of H-ZSM-5 is represented by the higher tem-
perature peak at 400 C. The TPDA profile over the
Co-ZSM-5 catalyst shows a very different picture. The high
temperature peak is not visible; rather, a continuous des-
orption occurred starting from the low temperature peak
and extended beyond 550 C. This Co-ZSM-5 catalyst was
almost fully exchanged with Co2+ (Co/Al = 0.49), and IR
spectroscopy measured with DRIFTS confirmed that there
was no observable acidic OH group (about 3610 cm 1) on
this catalyst. To quantify the NH3 desorption, TPDA exper-
iments were also carried out with a thermogravimetric anal-
ysis (TGA) method. The amount of NH3 desorbed above
285 C was measured as 1 mmol/g (or 0.76 NH3/bulk Al)
and that over Co-ZSM-5 as 1.2 mmol/g (or 1.8 NH3/Co2+).
The IR spectra we collected show that NH3 adsorbs on
Co-ZSM-5 as a Lewis base and Co2+ is an excellent Lewis
acid center. The above TPD results suggest that two NH3
molecules can adsorb on a single Co2+
.
Temperature Programmed Desorption of Ethylene
Ethylene is a major product of the ammoxidation reac-
tion and is thought to be a primary product of C2H6 acti-
vation (1, 2). The adsorption of C2H4 on Co-ZSM-5 is rela-
tively weak, compared to the adsorption of NH3. As shown
in Fig. 5, most of the adsorbed C2H4 molecules desorbed
below 300 C with an intensity much lower than that for
NH3. It is conceivable that in the presence of NH3, or under
the ammoxidation conditions (e.g., high temperatures), the
FIG. 3. C2H4 ammoxidation as a function of [O2]. The reaction run
was carried out at 425 C with 0.05 g Co-ZSM-5, F = 100 cc/min, [C2H4] =
10% , and [NH3] = 10% .

498
LI AND ARMOR
replaced by a stream of NH3/He mixture (3000 ppm NH3)
at 100 C, most of the adsorbed C2H3N molecules were dis-
placed by NH3. As shown in Fig. 6, the overall desorption
intensity of C2H3N was significantly reduced from 1.88 to
0.51 C2H3N/Co2+—a 73% reduction. In addition, the high
temperature desorption peak at 520 C was no longer visi-
ble with flowing the NH3/He mixture. What is remarkable is
that all this reduction in C2H3N coverage occurred at 100 C
with a stream containing only 3000 ppm NH3.
Temperature Programmed Reaction (TPRx):
NH3(a)/C2H4/O2
Previously, we reported that C2H5NH2 might be a re-
active intermediate for C2H3N formation (2). An experi-
ment was designed to capture the intermediate during a
temperature programmed reaction. After a standard cata-
lyst pretreatment (500 C in He for 1 h), NH3 was first
adsorbed on Co-ZSM-5 at 25 C. Then a stream of C2H4
(5% )/O2 (5% )/He mixture was flowed through the catalyst
at 150 cc/min. Upon stabilization of the flow, the temper-
ature was ramped to 525 C at a ramp rate of 10 C/min.
The very high space velocity minimizes the over oxida-
tion of any intermediates. Figure 7 shows some of the mass
fragments monitored. As the temperature reached 350 C,
both O2 and C2H4 consumption began and O2 was depleted
at 450 C. C2H3N started appearing at 350 C, peaked at
440 C and then sharply dropped with increasing temper-
ature. At T > 440 C, C2H5NH2 was observed, slowly tailing
off with increasing temperature. This observation suggests
that C2H5NH2 is indeed an intermediate reaction for the
ammoxidation reaction. The sharp drop in C2H3N inten-
sity beyond 440 C must be caused by the depletion of O2,
slowing down the dehydrogenation of amine.
FIG. 5. Comparison of NH3 TPD and C2H4 profiles over Co-ZSM-5.
coverage of C2H4 on cobalt sites is insignificant. The adsorp-
tion/desorption of C2H4 appears to be reversible because
no other products were observed.
Temperature Programmed Desorption of Acetonitrile
Acetonitrile is not only a product of the ammoxida-
tion reaction but also an excellent Lewis base due to its
exceptional electron donating ability from the N atom.
The adsorption and desorption of C2H3N was studied on
Co-ZSM-5. Figure 6 compares the C2H3N TPD profile
in He and in a mixture of NH3/He (3000 ppm NH3). In
a He flow, the adsorption of C2H3N is very strong, high
intensity desorption was observed throughout the whole
temperature range (100–525 C). However, when He was
Temperature Programmed Desorption/Reaction
of Ethylamine in NH3/O2
We reported earlier that, under steady-state conditions,
C2H5NH5 was oxidized to C2H3N over a Co-ZSM-5 cata-
lyst. This experiment was intended to show that the ad-
sorbed C2H5NH2 is converted to C2H3N under reaction
conditions. After adsorption of C2H5NH2 at room tempera-
ture and flushed with He, an NH3/O2/He mixture waspassed
through the catalyst bed while temperature was ramped to
525 C at 10 C/min. As shown in Fig. 8, C2H3N was observed
and peaked at 320, 390, and 500 C. An C2H5NH2 desorp-
tion peak can be seen around 300 C with a relatively low
intensity.
Ammoxidation of Higher Hydrocarbons
As a way to verify the reaction scheme proposed ear-
lier, ammoxidation reactions with some higher hydrocar-
bons were conducted over the Co-ZSM-5 catalyst. Table 1
shows propane ammoxidation results over Co-ZSM-5. The
FIG. 6. Comparison of C2H3N TPD in a He flow and in an NH3/He
mixture over Co-ZSM-5.

AMMOXIDATION OF ETHANE AND ETHYLENE
499
FIG. 7. Temperature programmed reaction between adsorbed NH3 and gaseous mixture of C2H4 and O2 over a Co-ZSM-5 catalyst.
major products for C3H8 ammoxidation are C2H3N, C3H6, involves an addition of a gaseous C2H4 to the adsorbed
and CO2. Interestingly, only a very small amount of C3H5N NH3 forming an adsorbed ethylamine molecule (Eq. [8]).
was formed, even though the feed here was C3H8 and the The adsorbed amine is dehydrogenated by reacting with its
dominant nitrile product is C2H3N. The ammoxidation of neighboring OH forming an ethylamine anion and H2O as
other hydrocarbons (n-C4H10, iso-C4H10, and n-C5H12) also a by-product (Eq. [9]). This adsorbed amine anion is sub-
produced C2H3N as a dominant nitrile product with very sequently oxidized by a gaseous O2 forming an adsorbed
low quantities of propionitrile formation (5).
pair of C2H3N and OH group and H2O as a by-product
(Eq. [10]). Finally, C2H3N is desorbed (Eq. [11]). These
steps are further explained below.
DISCUSSION
The TPDA experiments suggest that NH3 adsorption is
a necessary and important step for the ammoxidation reac-
tion. At the reaction temperatures, NH3 should cover most
of the sites, because of the weak adsorption of C2H4 on
the Co2+ sites. It is logical to assume that a gaseous C2H4
molecule would add on the adsorbed NH3 molecule at our
reaction temperatures. Gas phase NH3 + C2H4 addition re-
action to form ethylamine is not thermodynamically favor-
able at our reaction temperatures. However, the strongly
adsorbed NH3 can lower the Gibbs free energy for the for-
mation of an adsorbed amine molecule. The formation of
the adsorbed ethylamine is supported by the TPRx exper-
iment with the C2H4/O2 mixture flowing through the NH3
Acetonitrile Formation Pathway
The elementary steps for C2H4 ammoxidation are pro-
posed below (see Fig. 9). First, NH3 is adsorbed on a
hydroxylated Co2+ site (Eq. [7]). The NH3 adsorption is
a reversible process, and its equilibrium is dependent on
the reaction temperature. However, its adsorption should
be the strongest among all the reactants. The second step
TABLE 1
Propane Ammoxidationa on Co-ZSM-5
C3H8
O2
Sel. to
Sel. to
CO2
(% )
Sel. to
C3H6
(% )
Sel. to
C2H4
(% )
Sel. to
C3H5N
(% )
T
( C)
Conv. Conv. C2H3N
(% )
(% )
(% )
375
400
425
450
19
26
29
31
86
97
100
100
25
28
29
29
26
28
28
30
42
35
34
31
5
6
6
7
1
2
1
1
a
FIG. 8. Temperature programmed desorption/reaction of C2H5NH2
Feed composition: 5% hydrocarbon, 10% NH3, 6.5% O2; 0.2 g cata-
over Co-ZSM-5 in a NH3/O2/He stream.
lyst, F = 100 cc/min.

500
LI AND ARMOR
propane, butane, or butene, we obtained C2H3N as a ma-
jor nitrile product with only trace amounts of higher ni-
triles. This implies that the NH3 + olefin addition step fol-
lows a Markovnikov orientation; i.e., NH2 is added to the
-position of the carbon chains. With the formation of C N
–
and simultaneous breaking of a C C bond, C2H3N is ob-
tained. Thus, C2H3N is the most significant nitrile product
regardless of hydrocarbon used. The higher hydrocarbon
ammoxidation runs support the notion that addition and se-
quential dehydrogenation are necessary mechanistic steps
for the ammoxidation reaction.
Based on the elementarysteps(Eqs. [7]–[11]), one can de-
rive a kinetic expression for C2H4 ammoxidation to C2H3N.
For the sake of simplicity, Eqs. [9] and [10] will be combined
as Eq. [12],
–
–
–
–
HO Co C2H5NH2 + O2 → HO Co C2H3N + 2H2O.
[12]
For C2H4 adsorption, we have the following equilibrium
process,
FIG. 9. Acetonitrile formation pathway.
–
–
C2H4 + Co(OH) = C2H4 Co OH.
[13]
adsorbed Co-ZSM-5 catalyst (see Fig. 7). The TPR studies
in Figs. 7– 8 show the onset of the C2H4, NH3, O2 reaction at
350 C with O2 depletion at 450 C. The drop in O2 coincides
with the sharp drop of acetonitrile formation and mirrors
the formation of ethylamine in Fig. 7. Desorption and reac-
tion studies of ethylamine in Fig. 8 show that ethylamine is
converted to acetonitrile in the same temperature region;
ethylamine was thus considered as a reaction intermediate.
A series of sequential dehydrogenation steps are proposed
to account for nitrile formation. The H atoms associated
with N are thought to be more reactive than the ones on
C because of the proximity between N and Co. We spec-
ulate that the dehydration between OH and the adsorbed
C2H5NH2 results in an amine anion adsorbed on Co2+. The
Co2+ is thus balanced by an negative charge generated from
the zeolite lattice and by a negative charge of this amine an-
ion. Subsequent dehydrogenation is proposed to proceed
with O2 insertion and dehydration. Although, we have not
identified the adsorbed amine anion as an intermediate for
C2H3N formation, we do know that C2H5NH2 reacting with
O2 over Co-ZSM-5 does result in C2H3N and that when O2
is depleted C2H5NH2 desorbed. The desorption of C2H3N
from the Co2+ sites is a reversible process. Although the
strength of C2H3N adsorption on Co2+ sites is strong, the
adsorption of NH3 is even stronger. As shown in Fig. 6,
a majority of the adsorbed C2H3N can be displaced by a
low concentration (3000 ppm) of NH3 at 100 C. Therefore,
under the ammoxidation conditions, where a high concen-
tration of NH3 is available, the desorption of C2H3N should
readily occur.
The following assumptions are made: (1) steady-state re-
action; (2) Eqs. [7], [8], [11], and [13] are in equilibrium;
and (3) adsorption site coverage is conserved as shown in
Eq. [14],
+
+
, and
+
= 1,
[14]
NH3
C2H3N
C2H4
0
where,
,
,
are fractional coverage
NH3 C2H3N C2H4
0
for the adsorption of NH3, C2H3N, C2H4, and fraction of
vacant sites, respectively. Note, the coverage for C2H5NH2
and O2 are omitted here, assuming that they are too small
to be significant. Based on the steady-state assumption, the
C2H3N formation rate can be expressed as
r = r12 = k12
pO
.
[15]
C2H5NH2
2
Based on the assumption [2], we have
= K8 pC H
.
4
[16]
C2H5NH2
NH3
2
Similarly, one can express the coverage of other adsorbates
as Eqs. [17]–[19],
= K7 0 pNH
[17]
[18]
[19]
NH3
C2H3N
C2H4
3
= K
0 pC H N
11
2 3
= K13 0 pC H
;
2
4
K7, K8, and K13 are equilibrium adsorption constant for
NH3, C2H5NH2, and C2H4, respectively; and K is the
desorption equilibrium constant for C2H3N. Based on
Eq. [14], one obtains Eq. [20],
11
This proposed pathway is consistent with the results
of ammoxidation of higher alkanes and alkenes. With
= 1
.
0
[20]
NH3
C2H3N
C2H4

AMMOXIDATION OF ETHANE AND ETHYLENE
501
Combining Eqs. [17]–[20], one obtains Eq. [21],
1
=
.
[21]
0
1 + K7 pNH + K 11 pC H N + K13 pC H
3
2
3
2
4
Combining Eqs. [21], [15]–[17], we have the following rate
expression for C2H3N:
k12 K8 K7 p
p
p
NH3 C2H4 O2
r =
.
[22]
1 + K7 pNH + K13 pC H + K 11 pC H N
3
2
4
2
3
This resulting rate expression is consistent with our em-
pirical reaction rate orders, fractional orders with respect
to C2H4 and NH3, and first order with respect to [O2]. Our
TPD experiments with NH3, C2H4, and C2H3N show that
NH3 adsorption is by far the strongest and NH3 should be
the dominant adsorbed species on the catalyst.
FIG. 10. Nitrogen formation scheme.
bon being totally oxidized compared to N2 formation. How-
ever, there is always some C2 hydrocarbons being oxidized
to CO2. This proposal pathway explains our observations
that more N2 than CO2 is formed during the ammoxidation
reaction. The nitrogen pairing between amine and ammo-
Nitrogen Formation Pathway
N2 is a significant by-product for the ammoxidation reac-
tion, and its formation should be minimized. Therefore, it is
important to understand how N2 is formed, and we specu-
late that one obvious N2 formation step is the NH3/O2 reac-
tion (Eq. [23]). After all, NH3 is the only, original source of
nitrogen. A parallel, NH3 oxidation reaction would seem to
be a first logicalguess. However, a test reaction with NH3/O2
(5% NH3 and 3% O2 over 0.1 g Co-ZSM-5 at 450 C) shows
that the rate of N2 formation (3 mmol/g/h) is only 1/6 of
that from the ammoxidation reaction under identical condi-
tions (+5% C2H4). This suggests that NH3 oxidation alone
only accounts for a small fraction of N2 generated. What
about the total oxidation of the adsorbed amine intermedi-
ate? Simple stoichiometry tells us that C2H5NH2 oxidation
would yield four CO2 molecules for every N2 formed, which
is not consistent with our observation that more N2 was pro-
duced than CO2:
–
nia may be similar to the NH3 NH3 pairing except that the
amine intermediate may be more reactive.
Ethane Oxidative Dehydrogenation
For ethane ammoxidation, C2H6 needs to be first acti-
vated, and its activation is likely to be aided by O2. Ethy-
lene is a significant co-product of C2H6 ammoxidation and
is more reactive compared to C2H6 as an ammoxidation
reactant. It is obvious that C2H4 is a primary product and
a reactive intermediate for C2H6 ammoxidation. Thus, the
first step for C2H6 activation would be itsoxidative dehydro-
genation. Specific steps are proposed in Fig. 11 (Eqs. [28],
[29] and [26], [27]). We believe that C2H6 is activated by
sequential oxidative H abstractions. The first H abstraction
is accomplished by dehydration between an OH group and
C2H6. The adsorption of C2H6 on (with) Co2+ is not favor-
able in the presence of NH3. However, this process becomes
feasible at higher temperatures at which some vacant sites
2NH3 + 1.5O2 = N2 + 3H2O.
[23]
Obviously, another pathway for N2 formation must exist,
and this pathway may involve amine intermediates. This
alternative N2 formation pathway is proposed below (see
the equation in Fig. 10), which involves nitrogen pairing
between NH3 and C2H5NH (Eq. [9]) under an oxidative
environment. The adsorbed C2H5NH2 can be converted to
either C2H3N (Eq. [12]) or N2 in the presence of NH3/O2
–
(Eq. [25]). This N2 pathway (Eq. [25]) is essentially a N N
pairing step between these two N containing molecules.
Since the hydrocarbon part of the amine complex is on top
(far away from the Co2+ center) most of C2H4 is released
into gasphase after N2 isformed. C2H4 desorption (Eq. [27])
is thought to be fast, and its desorption is accelerated by the
adsorption of NH3. The gaseous C2H4 may then further
react with an adsorbed NH3 forming another adsorbed
amine molecule (Eq. [8]). As a result, there is less hydrocar-
FIG. 11. Scheme for C2H6 activation.

502
LI AND ARMOR
are available by desorbing NH3. This assumption is con- This reactive intermediate has been identified as C2H5NH2,
sistent with the fact that more C2H4 and C2H3N are pro- which is readily dehydrogenated to form C2H3N. N2 forma-
–
duced at higher temperatures. Assuming that every C2H3N tion mainly comes from N N pairing between the adsorbed
formed requires one C2H4 molecule, C2H6 activation is ex- NH3 and an amine species (both on a single cobalt site) in an
tremely favored at high temperatures. On the other hand, oxidative environment. The parallel NH3/O2 reaction only
our earlier report (2) show that the oxidative dehydrogena- contributes to a minor portion of the N2 formed.
tion of C2H6 to C2H4 is aided by the presence of NH3. Thus
it is possible that C2H6 is also activated on the cobalt sites
that are partially coordinated with NH3. Further studies are
ACKNOWLEDGMENT
needed to depict a detailed picture of the NH3 promotion.
We thank Peter Hohl for his technical assistance for some of the exper-
iments reported here and Air Products and Chemicals, Inc. for permission
to publish this work.
CONCLUSIONS
The adsorption of NH3 is very strong on the Co2+ sites
of the Co-ZSM-5 catalyst, and its adsorption is essential for
REFERENCES
1. Li, Y., and Armor, J. N., Chem. Comm., 2013 (1997).
2. Li, Y., and Armor, J. N., J. Catal. 173, 511 (1998).
3. Catani, R., and Centi, G., J. Chem. Soc., Chem. Comm., 1081 (1991).
4. Aliev, S. M., Sokolovskii, V. D., and Boreskov, G. B., USSR Patent
SU-738657 (1980).
–
the formation of a C N bond. The adsorption of C2H3N
on the cobalt sites is also strong but can be displaced by
a diluted stream of NH3/He mixture at 100 C. During the
ammoxidation reaction, C2H4, formed from the oxidative
dehydrogenation of C2H6, adds to an adsorbed NH3 species.
5. Li, Y., and Armor, J. N., Eur. Pat. Appl., EP 0761654A2 (1997).