Novel derivatives of MCM-36 as catalysts for the reduction of nitrogen oxides from FCC regenerator flue gas streams

Article abstract of DOI:10.1016/j.jcat.2004.06.021

MCM-36 materials containing mixed oxide pillars (MgO-Al₂O ₃-MCM-36, BaO-Al₂O₃-MCM-36, MgO-Al ₂O₃SiO₂-MCM-36, BaO-Al₂O ₃SiO₂-MCM-36) were investigated as additives for the in situ reduction of NO_x formed during the regeneration of coked cracking catalysts in FCC units. The additives were investigated under reaction conditions similar to the oxygen-depleted and the oxygen-enriched zone of the FCC regenerator. MCM-36-type materials with basic mixed oxide clusters in the interlayer galleries showed high NO conversions (～85%) and high N ₂ yields (～80%) under oxygen-deficient reaction conditions. The catalytic activity of MCM-36 with alkaline earth metal aluminum oxide pillars incorporated between the zeolitic layers was higher (MgO-Al₂O ₃-MCM-36: 77% N₂ yield at 650°C) compared to the nonpillared parent material MCM-22 (22%). The materials have been added in regeneration experiments to industrial (coked) FCC catalysts in a fluidized-bed reactor. The materials led to a reduction of NO emissions also in the presence of Pt-based CO combustion promoters (～30% NO reduction). It is speculated that Bronsted acid sites in the zeolite layers of the composite materials catalyze the reduction of NO with NH₃, which is an intermediate in the regeneration of nitrogen-containing carbonaceous deposits on deactivated FCC catalysts.

Full text of DOI:10.1016/j.jcat.2004.06.021

Journal of Catalysis 227 (2004) 117–129
www.elsevier.com/locate/jcat
Novel derivatives of MCM-36 as catalysts for the reduction of nitrogen
oxides from FCC regenerator flue gas streams
a
a
b
Jan-Olaf Barth , Andreas Jentys , Eleni F. Iliopoulou ,
b
a,∗
Iacovos A. Vasalos , Johannes A. Lercher
a
Technische Universität München, Lehrstuhl II für Technische Chemie, 85747 Garching bei München, Germany
b
Center for Research and Technology Hellas, Chemical Process Engineering Research Institute, PO Box 1517, 54006 University City, Thessaloniki, Greece
Received 18 March 2004; revised 23 June 2004; accepted 24 June 2004
Available online 5 August 2004
Abstract
MCM-36 materials containing mixed oxide pillars (MgO–Al O -MCM-36, BaO–Al O -MCM-36, MgO–Al O –SiO -MCM-36, BaO–
2
3
2
3
2
3
2
Al O –SiO -MCM-36) were investigated as additives for the in situ reduction of NOx formed during the regeneration of coked cracking
2
3
2
catalysts in FCC units. The additives were investigated under reaction conditions similar to the oxygen-depleted and the oxygen-enriched
zone of the FCC regenerator. MCM-36-type materials with basic mixed oxide clusters in the interlayer galleries showed high NO conversions
(
∼ 85%) and high N yields (∼ 80%) under oxygen-deficient reaction conditions. The catalytic activity of MCM-36 with alkaline earth metal
2
◦
aluminum oxide pillars incorporated between the zeolitic layers was higher (MgO–Al O -MCM-36: 77% N yield at 650 C) compared to
2
3
2
the nonpillared parent material MCM-22 (22%). The materials have been added in regeneration experiments to industrial (coked) FCC
catalysts in a fluidized-bed reactor. The materials led to a reduction of NO emissions also in the presence of Pt-based CO combustion
promoters (∼ 30% NO reduction). It is speculated that Brønsted acid sites in the zeolite layers of the composite materials catalyze the
reduction of NO with NH , which is an intermediate in the regeneration of nitrogen-containing carbonaceous deposits on deactivated FCC
3
catalysts.

2004 Elsevier Inc. All rights reserved.
Keywords: Fluid catalytic cracking (FCC); NOx reduction; Oxidative regeneration; Pillared zeolites
◦
1
. Introduction
residues, primarily in the 315–650 C boiling point range,
into lighter fuel products and petrochemical feedstocks [3,4].
The FCC process consists of a riser reactor and a regener-
ator in which the deactivated cracking catalyst is regener-
ated by burning off the coke deposited on the FCC catalyst
during the catalytic cracking reaction. The high-temperature
flue gas stream at the exit of a regenerator contains O2, N2,
CO, CO2, H2O, SOx, and NOx. Depending upon feed ni-
trogen levels and regenerator conditions NOx concentrations
are typically in the range of 50–500 ppm [5,6]. The NOx
emissions from the regenerator of a FCC can contribute up
to 50% of the total NOx emissions in a refinery and con-
sist mainly of NO, which is formed in the regenerator, while
NO2 is formed only after NO is being released to the air.
Fluid catalytic cracking (FCC) is one of the largest ap-
plications of heterogeneous catalysis and is a key process in
a modern refinery. Gasoline from FCC units contributes to
approximately 35% of the US gasoline pool [1]. The produc-
tion of petrochemical-based chemicals from FCC and related
processes such as DCC (deep catalytic cracking) and CPP
(
catalytic pyrolysis process) is gaining increasing impor-
tance due to the worldwide shortage in production capacities
for lower olefins [2]. Worldwide more than 300 FCC units
are operated, converting vacuum gas oil and high boiling
*
2
N O in the exit gas of an FCC regenerator, if any, exists typ-
Corresponding author. Fax: +49/89/28913544.
E-mail address: johannes.lercher@ch.tum.de (J.A. Lercher).
ically in very low concentrations [7]. NO is formed in the
0021-9517/$ – see front matter  2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2004.06.021

118
J.-O. Barth et al. / Journal of Catalysis 227 (2004) 117–129
FCC process by the oxidation of nitrogen-containing com-
pounds in coke (fuel NOx) [5,6,8], while the reaction of air-
derived nitrogen and oxygen produces negligible amounts of
NO (thermal NOx).
uncalcined precursor of MCM-22 and intercalation of vari-
ous (aluminum) oxides leads to derivatives of MCM-36 with
particular physicochemical properties [23]. These nanocom-
posite materials are characterized by zeolitic microporosity
in the zeolite phases and mesoporosity in the pores and gal-
leries generated by oxide clusters in the interlayer space.
In addition, the modular structure of the materials consists
of the microporous zeolite phases containing Lewis- and
Brønsted-type acid sites, which can be modified by ion ex-
change [24]. As the incorporated mixed oxides show basic
properties bifunctional catalyst can be created [23]. Com-
pared with other mesoporous materials such as MCM-41-
type materials, materials based on MCM-36 are character-
ized by a much stronger acidity [17,19].
Regulations that confine the emissions of nitrogen ox-
ides (NOx) from the regenerator of a FCC unit are becoming
more and more stringent due to environmental concerns and
to the use of heavier feedstocks [1,3]. NOx control technolo-
gies for the FCC unit should reduce NO emissions without
adversely affecting the cracking activity and selectivity of
the FCC catalyst or increasing the emissions of other pol-
lutants. Current NOx control strategies include (see Ref. [9]
and references therein) (i) hydrotreating of the feed to min-
imize NOx precursors (polycyclic aromatic hydrocarbons of
carbazole or pyridine type) in the FCC unit, (ii) flue gas
treatment in a tail-end unit with either NH3 or hydrocar-
bon injection [10], and (iii) modification of the design or
the operating conditions of the regenerator, as, e.g., the de-
velopment of a countercurrent regenerator or a two-stage
regenerator. Other modifications of the FCC are the addi-
tion of a vaporizable fuel to the upper portion of the FCC
regenerator, or the utilization of a spent catalyst distribu-
tor in the regenerator, and (iv) the use of catalytic additives,
which is a simple, cost-effective method that is applicable in
existing FCC units without significant modifications in the
regenerator design or the unit operating conditions. These
additives should operate at high temperatures and in the
presence of other gases such as CO, CO2, O2, and H2O. Ma-
terials that reduce NOx emissions under the conditions of
the regenerator of a FCC unit have been previously patented
and/or reported by several research groups [3,6,9,11–13].
Typically, most of the modern refineries add CO combustion
promoters (Pt-based compounds that accelerate the oxida-
tion of CO) into their FCC catalyst inventories. However,
CO combustion promoters increase NO emissions through
the oxidation of nitrogen-containing intermediates such as
HCN and NH3 [5,6,8,14,15]. Aiming at the reduction of
NO emissions several methods have been proposed, such
as steam treating of conventional CO combustion promot-
ers [3,16].
The scope of this work was to study the catalytic perfor-
mance of NO reduction additives based on MCM-36 under
x
reaction conditions that reproduce those in the regenerator
of a typical FCC unit. The catalysts were investigated under
reaction conditions simulating the oxygen-depleted and the
oxygen-rich zone of an FCC regenerator operating under full
burn conditions. Oxygen-deficient reaction conditions are
encountered in the oxygen-depleted zones of the dense and
the diluted phase of the FCC fluidized bed. The incomplete
combustion of coke leads to the formation of a reductive at-
mosphere in parts of the dense phase of the fluidized bed
mainly due to the formation of CO and some hydrocarbons
not desorbed in the stripper [25–27]. Simultaneously, in the
bottom region of the regenerator near the air inlet an oxygen-
rich zone is encountered. The MCM-36-type materials were
tested as catalysts for the reduction of NO with CO as well
as in regeneration experiments mixed with spent FCC cat-
alysts in a fluidized-bed reactor. The compatibility of the
additives with a conventional Pt-based CO combustion pro-
moter was also investigated. The reaction mechanism that
leads to the reduction of NO by CO to N2 was studied by IR
spectroscopy.
2
. Experimental
Consequently, there is an urgent need for the development
of catalytic NOx reduction additives that can either oper-
ate in the presence of a CO combustion promoter or have
the ability to simultaneously reduce NO and CO emissions.
In this contribution we discuss the use of novel derivatives
of the molecular sieve MCM-36 as catalytic additives for
the reduction of NO from FCC regenerator flue gas streams.
These materials are characterized by a nanocomposite struc-
ture consisting of layers of zeolite MCM-22 (MWW type)
and intercalated mixed oxides such as MgO–Al2O3, BaO–
Al2O3, MgO–Al2O3–SiO2, and BaO–Al2O3–SiO2 [17–21].
Mixed oxides and spinel-type materials containing MgO and
BaO function as effective catalysts for the decomposition of
2
.1. Synthesis of catalytic additives
Nanocomposite materials based on MCM-22 were syn-
thesized according to the procedure described in Ref. [21].
All catalytic additives were thoroughly characterized and
identified as derivatives of MCM-36 as described in Ref.
[
23].
2
.2. Coked FCC catalysts
Coked (spent) FCC catalysts were obtained from an FCC
unit using no CO combustion promoter (elemental analysis
C, 1.31 wt%; H, 0.055 wt%; N, 233 ppm). The deactivated
catalyst samples were extracted from a DCR (Davison cir-
culating riser) unit after steam stripping.
◦
NO at high temperatures (> 500 C) and are used as SOx
transfer reagents for the reduction of SOx emissions from
FCC units [3,22]. Exfoliation of the zeolite layers of the

J.-O. Barth et al. / Journal of Catalysis 227 (2004) 117–129
119
2
.3. Catalytic measurements
2.6. Temperature-programmed desorption (TPD) of NO2
Temperature-programmed desorption of NO2 was used
for the investigation of surface species such as nitrates and
nitrites on the samples. The sample (∼ 100 mg ) was de-
The catalytic activity of the samples was studied in a
continuous flow system using a fixed-bed quartz reactor of
mm inner diameter containing 0.1 g of catalyst (particle
size 300 µm). The reactant gas consisted of either 1000 ppm
NO, 5% CO, and 2.5% O2 or 600 ppm NO, 1.4% CO, and
.7% O2 while in both cases He was used as the carrier gas.
4
−3
gassed by evacuation to 10 mbar, followed by heating at a
◦
◦
rate of 10 C/min to 450 C and maintaining at that temper-
◦
ature for 60 min. After cooling to 100 C, the sample was
0
equilibrated with 10 mbar of 1% NO2 in He for 60 min.
A total flow of 100 ml/min resulting in a space velocity
−3
−1
Subsequently, the system was evacuated to 10 mbar for
of 38,000 h was used. The reaction products were ana-
lyzed with a chemiluminescence analyzer (NO–NO2 analy-
sis, Thermo Environmental Instruments, TEI-42-C) and a
gas chromatograph (HP) equipped with a TCD detector.
A MS 5A column was used for the separation of N2, O2, and
CO and a Porapack Q column for N2O and CO2 partition.
Before the reaction, the catalysts were activated in He flow
◦ ◦
80 min and heated to 600 C at a rate of 10 C/min. Dur-
1
ing the temperature ramp, mass spectra of the gas phase were
◦
collected at 5 C/min intervals using a Balzers QMG 420
mass spectrometer to determine the rate of desorption.
◦
3. Results
at 600 C for 1 h. Kinetic measurements were carried out at
◦
◦
temperature intervals of 25 C between 500 and 750 C. The
catalysts were equilibrated at each temperature for 165 min.
3
.1. Materials
Table 1 summarizes the textural properties and the con-
2
.4. FCC regeneration experiments
centrations of the acid sites (determined from temperature-
programmed desorption of ammonia [23]) of the derivatives
of MCM-36 investigated. A detailed physicochemical char-
acterization and investigation of the acid–base character of
the nanocomposite materials can be found in Refs. [21] and
Regeneration experiments were carried out in a bench-
scale reaction unit using a protocol that simulates the regen-
erator of a typical fluid catalytic cracking unit [28]. The unit
consists of the feed gas system, a fluidized-bed reactor, a fur-
nace, and the gas analysis system. A detailed description of
this unit can be found in Ref. [28]. The total flow rate was
[
23].
3
.2. Kinetic measurements
1
000 ml/min. The feed composition was 2% O2 in N2. Me-
chanical mixtures of spent catalyst (from a commercial FCC
unit using no CO combustion promoter) with 1 wt% of can-
didate additives were loaded in the reactor and regeneration
All materials were tested as catalysts for the reduction of
NO with CO under reaction conditions simulating the FCC
regeneration process. CO is the dominating reducing agent
in the regenerator, due to the incomplete combustion of coke.
Fig. 1 shows an example for the reduction of NO with CO
over MgO–Al2O3-MCM-36 under conditions similar to the
oxygen-depleted (Fig. 1a) and oxygen-rich (Fig. 1b) phase
of an FCC regenerator operated under full burn conditions.
N2, NO2, and N2O were formed as reaction products, while
CO2 was generated by the subsequent oxidation of CO. In
◦
was carried out at 700 C. The total reactor loading was 10 g
in all experiments. To investigate the compatibility of the
additives with conventional Pt-based CO combustion pro-
moters the regeneration experiments were repeated loading
the reactor with mechanical mixtures of 1 wt% of the candi-
date additives, 0.1 wt% of CP-3, a commercial CO promoter
supplied by Grace-Davison, and 98.9 wt% spent catalyst.
◦
the temperature region between 600 and 750 C high yields
2
.5. IR spectroscopy
(
50–84%) of N2 were obtained under reaction conditions
similar to the oxygen-depleted dense and diluted phases of
IR measurements were performed with a Bruker IFS-88
the fluidized bed of the FCC regenerator (0.5% O ). In the
2
spectrometer in an in situ flow cell. The spectrometer was
used in the transmission-absorption mode with a resolution
presence of higher oxygen concentrations (2.4% O ), which
are encountered in the lower parts of the dense phase, sig-
2
−1
of 4 cm . The IR cell consisted of a stainless-steel chamber
equipped with CaF2 windows and a resistance-heated fur-
nace in which the samples pressed into thin, self-supporting
nificantly lower yields (10–40%) of N were measured. In
2
the first case considerable amounts of N O were formed
2
◦
with a maximum yield of 50% at 575 C, whereas under
2
wafers of approximately 5 mg/cm were placed. The sam-
the reaction conditions with high oxygen concentration only
◦
ples were activated in situ in a flow of He at 450 C (heating
17% N O was formed with its maximum shifted to higher
2
◦
◦
rate 10 C/min) for 60 min. After activation, the system was
cooled to the adsorption temperature (200 C). The amounts
temperatures (625 C). In contrast to this observation, un-
◦
der oxygen-rich conditions (2.4% O ) higher yields of NO
2
2
of 1000 ppm NO in He, 1000 ppm NO + 3% O2 in He, and
were obtained. For both cases the conversion of CO and
◦
5
% CO in He were adsorbed at 200 C in successive adsorp-
yield of CO increased with rising temperature and reached
2
◦
tion steps.
a constant maximum degree of conversion at 575–600 C.

120
J.-O. Barth et al. / Journal of Catalysis 227 (2004) 117–129
Table 1
Elemental composition, BET surface areas, and acid site concentrations (determined from temperature-programmed desorption of NH ) of derivatives of
3
MCM-36 and the MCM-22 parent material
2
Sample
Si (wt%)
Al (wt%)
Na (wt%)
Mg (wt%)
Ba (wt%)
SBET (m /g)
c
(mmol/g)
acid sites
MCM-22
45.40
27.92
9.33
39.90
40.06
3.85
21.28
35.90
4.32
0.25
< 0.10
< 0.10
< 0.10
< 0.10
–
0.64
–
0.24
–
–
–
2.20
–
0.24
432
348
390
757
689
0.79
0.91
0.31
0.42
0.38
MgO–Al O -MCM-36
2
3
BaO–Al O -MCM-36
2
3
MgO–Al O –SiO -MCM-36
2
3
2
BaO–Al O –SiO -MCM-36
4.48
2
3
2
(a)
(a)
(b)
(b)
Fig. 1. NO reduction with CO over MgO–Al O -MCM-36: (a) 700 ppm
Fig. 2. NO conversion over MCM-22 and derivatives of MCM-36:
(a) 700 ppm NO, 1.4% CO, 0.5% O , (b) 1000 ppm NO, 4.4% CO, 2.4%
2
3
NO, 1.4% CO, 0.5% O , (b) 1000 ppm NO, 4.4% CO, 2.4% O .
2
2
2
O .
2
◦
The comparison of the conversion of NO in the reaction
with CO over various MCM-36 derivatives and the pre-
cursor material zeolite MCM-22 is shown in Fig. 2. High-
est conversions of NO (∼ 90%) were obtained for MgO–
Al2O3-MCM-36 and BaO–Al2O3-MCM-36 in the presence
of lower oxygen concentrations (Fig. 2a). Under conditions
simulating the oxygen-rich phase of the regenerator (Fig. 2b)
MgO–Al2O3-MCM-36, BaO–Al2O3-MCM-36, and MCM-
ite materials in the region of 500–600 C is mainly due to
the formation of NO2.
The yields of N2 formation as a function of tempera-
ture over MCM-36 derivatives and MCM-22 are shown in
Fig. 3. Under oxygen deficient conditions the yields of N2
increased with increasing temperature. The highest yields of
N (75–83%) were obtained for MgO–Al O –SiO -MCM-
2
2
3
2
36 and MgO–Al O -MCM-36, whereas for MCM-22 and
2
3
2
2 reach similar degrees of conversion (32–40%) in the
2 3 2 2
BaO–Al O –SiO -MCM-36 the lowest yields of N were
◦
range of 675–750 C, while for MgO–Al2O3–SiO2-MCM-
3
obtained in the temperature window characteristic of the
◦
6 and BaO–Al2O3–SiO2-MCM-36 the conversion of NO
FCC regenerator (600–750 C). In the presence of an excess
decreased with increasing temperature. The high conver-
sions of NO over the mixed oxide intercalated nanocompos-
of oxygen (Fig. 3b) generally lower yields of N were ob-
served with a maximum of 40% for MgO–Al2O3-MCM-36.
2

J.-O. Barth et al. / Journal of Catalysis 227 (2004) 117–129
121
(a)
(a)
(b)
Fig. 4. Yield of N O over MCM-22 and derivatives of MCM-36:
2
(b)
(a) 700 ppm NO, 1.4% CO, 0.5% O , (b) 1000 ppm NO, 4.4% CO, 2.4%
2
O .
2
Fig. 3. Yield of N over MCM-22 and derivatives of MCM-36: (a) 700 ppm
2
NO, 1.4% CO, 0.5% O , (b) 1000 ppm NO, 4.4% CO, 2.4% O .
2
2
Interestingly, MCM-22 showed higher yields in the range
◦
◦
of 650–750 C (34% at 750 C) compared to the three other
MCM-36 derivatives. Highest selectivities to N2 were ob-
tained for MgO–Al2O3-MCM-36 (94% selectivity at 86%
◦
NO conversion, 700 C) and MgO–Al2O3–SiO2-MCM-36
86% selectivity at 86% NO conversion, 700 C), while the
◦
(
BaO containing pillared zeolites and the nonpillared parent
material MCM-22 showed lower selectivities (71–80%).
As already pointed out N2O is one of the by-products
in the reaction of NO with CO over MCM-36 type cata-
lysts. Fig. 4 shows the yield of N2O as a function of tem-
perature under the two different reaction conditions. Lower
oxygen concentrations (0.5 vs 2.4%) favored higher yields
Fig. 5. Time-on-stream measurement of the reaction of NO with CO over
MgO–Al O -MCM-36 in the absence or presence of 5% H O (initial con-
2
3
2
centrations of reactants: 690 ppm NO, 1.4% CO, 0.5% O ).
2
of N2O. The highest yield of N2O (∼ 50%) was detected at
◦
5
75 C for MgO–Al2O3-MCM-36 followed by BaO–Al2O3-
◦
MCM-36 (∼ 40%, 610 C). Note that for BaO–Al2O3–
SiO2-MCM-36 the maximum of N2O formation was found
at higher temperatures than for MgO-containing MCM-36
der oxygen-deficient conditions the lowest yields of N2O
were obtained for the unmodified MCM-22 zeolite (21%,
◦
650 C). In the presence of higher oxygen concentrations
◦
derivatives with SiO2/Al2O3 pillars (650 vs 625 C). Un-
similar yields of N2O were obtained for BaO–Al2O3-MCM-

122
J.-O. Barth et al. / Journal of Catalysis 227 (2004) 117–129
Fig. 6. Difference of IR spectra measured after adsorption of 1000 ppm NO
Fig. 7. Difference of IR spectra measured after adsorption of 1000 ppm NO
and 5% O2 on MgO–Al2O3-MCM-36 (T = 200 C).
◦
◦
on MgO–Al O -MCM-36 (T = 200 C).
2
3
3
6 and its MgO derivative (∼ 17%), while the lowest yields
were determined for the SiO2-modified materials.
Fig. 5 shows the time-on-stream behavior of MgO–
Al2O3-MCM-36 compared in the absence and presence of
◦
5
% water over 20 h at 500–675 C. To investigate the detri-
mental effects of water (steam) on the catalyst the tempera-
◦
ture was increased by steps of 25 C and the catalysts were
maintained at each temperature for 165 min. Under water-
free conditions the pillared zeolites showed stable catalytic
activity, while in the presence of 5% water the catalytic ac-
◦
tivity decreased after ∼ 5.5 h at 550 C with duration of
the experiment and significantly lower conversions of NO
◦
(
∼ 34% at 675 C) were observed.
3
.3. IR spectroscopy: Adsorption of NO, O2, and CO on
MgO–Al2O3-MCM-36
To investigate the intermediate species formed on the sur-
face during the reduction of NO with CO in the presence of
oxygen NO and O2 were coadsorbed followed by adsorp-
tion of CO. The differences in the IR spectra obtained after
adsorption of 1000 ppm NO on MgO–Al2O3-MCM-36 are
shown in Fig. 6. The formation of nitrite and nitrate species
(
bridged bidentate nitrate, monodentate nitrate on Al, linear
nitrite on Al) was observed (for an assignment of the band
−1
see Table 2) [29,30]. The decreasing band at 1479 cm in-
dicated that carbonates (resulting from the exposure of the
sample to air) on the basic material were removed upon NO
adsorption. Note that nitrates were formed, although the ad-
sorption was carried out in the absence of oxygen.
Fig. 8. Differences of IR spectra measured during reaction of 5% CO on
MgO–Al O -MCM-36 after exposure of the catalyst to 1000 ppm NO and
2
3
◦
5
% O (T = 200 C).
2
After adsorption of 1000 ppm NO in the presence of
% O2 predominantly nitrate species were formed on the
trates and nitrites decreased in intensity (negative bands),
−1
3
while bands at 1378, 1384 1588, 1610, and 1637 cm
−
1
◦
nanocomposite materials (Fig. 7). Bands at 1575 cm char-
first increased in intensity, reached a maximum at ∼ 350 C,
−1
acteristic of monodentate nitrates on Al and 1552 cm
bidentate nitrate on Al close to Mg) were observed, whereas
and finally decreased with increasing temperature. Partially
−
1
(
overlapping bands at ∼ 2280–2240 and 2227 cm were
−1
◦
nitrites (1459, 1463 cm : linear nitrite on Al) were only de-
tected in minor concentrations under oxidative conditions.
After adsorption of NO and O2 the sample was equili-
brated with 5% CO in He. Fig. 8 shows differences in the
IR spectra measured during heating of the catalyst from 200
observed above 250 C, which increased in intensity with
increasing temperature. These IR bands can be attributed
−
1
−1
to bicarbonate (1445, 1646 cm , shoulder at 3610 cm ),
−1
−1
uni-(1410, 1540–1580 cm ) and bidentate (1378 cm )
carbonate species adsorbed on basic sites of the MgO–
Al2O3-MCM-36 material. Bands at 2930 (not shown) 1610
◦
to 500 C in the presence of CO. Bands indicative of ni-

J.-O. Barth et al. / Journal of Catalysis 227 (2004) 117–129
123
Table 2
Assignment of IR bands for species on MgO–Al O -MCM-36
2
3
−1
Species
Structure
Remarks
Position /(cm
1630
)
Bridged bidentate nitrate
Monodentate nitrate
High coverage
Low coverage
1588
1575
1552
Bidentate nitrate on Al close to Mg
Linear nitrite on Al
Bicarbonate
Al–O–N–O
1459
1445, 1646
Unidentate carbonate
1410, 1479
Bidentate carbonate
Isocyanates
1378
Mg–N=C=O
Al–N=C=O
2240
2260
N O
2
2227
Formate
ν (sym)
ν (asym)
δ (CH)
2930
1610
1384
and 1384 cm 1 indicated the presence of formate species
−
from the nonpillared parent material MCM-22 is compared
in Fig. 10. In comparison with MCM-22, significantly more
NO2 was adsorbed on the basic oxide pillars of the compos-
on the MgO–Al2O3 clusters [31]. The broad bands around
280–2240 cm are indicative of the degenerate stretch-
−1
2
◦
ing vibration of surface isocyanate species adsorbed on
the MgO–Al2O3 pillars (Al–NCO, 2260; Mg–NCO, 2240
cm ) [29,32]. The narrow peak centered at 2227 cm is
tentatively ascribed to the N–N stretching vibration of N2O
formed during the reaction [29].
During reaction of CO on the catalyst at 500 C nitrate
and nitrite bands continued to decrease (negative bands in
Figs. 8 and 9) accompanied by increasing intensities of the
ite material. The first maximum of desorption at ∼ 200 C is
tentatively assigned to the desorption of physisorbed NO2,
−1
−1
◦
while the maximum at 310 C and the less intense maxima at
◦
440 and 510 C are tentatively attributed to the decomposi-
tion of monodentate and (bridged) bidentate nitrate species.
Interestingly, some nitrate species decompose in vacuum
(the TPD experiment was performed at a base pressure of
◦
−3
◦
10 mbar) at elevated temperatures above 500 C, suggest-
ing the presence of nitrates with a high thermal stability.
−1
peaks characteristic of isocyanate species (2260 cm ) and
−
1
N2O (2227 cm ).
3
.5. FCC regeneration experiments
3
.4. Temperature-programmed desorption of NO2
To evaluate the behavior of the MCM-36-type additives
NO2 was adsorbed on MgO–Al2O3-MCM-36 to study
during the regeneration process of the FCC catalysts, 1 wt%
of nanocomposite material was added to a coked crack-
ing catalyst from an industrial FCC unit and regenerated at
700 C with 2% O2 in a fluidized-bed reactor. These con-
ditions were selected to represent a full-burn regenerator
operation [28]. During regeneration NO was formed from
the thermal stability of nitrate surface species formed on
the catalysts. The samples were exposed to NO2 to simu-
late the interaction with NOx species under oxidative con-
ditions. The temperature-programmed desorption of NO2
◦
(
fragment NO: m/e = 30) from MgO–Al2O3-MCM-36 and

124
J.-O. Barth et al. / Journal of Catalysis 227 (2004) 117–129
Table 3
Conversion of NO and CO during the oxidative regeneration of a coked FCC catalyst in a fluidized-bed reactor in the presence of 1 wt% of MCM-36-type
catalysts or a conventional Pt-based CO combustion promoter (CP-3). Using CP-3 the amount of NO released increased by −300%
Sample
NO conversion as compared to pure spent catalyst (% )
CO conversion as compared to pure spent catalyst (% )
MgO–Al O -MCM-36
14
21
5
2
2
3
BaO–Al O -MCM-36
2
3
MgO–Al O –SiO -MCM-36
12
3
2
3
2
BaO–Al O –SiO -MCM-36
18
1
2
3
2
CP-3
−300
86
Fig. 10. Temperature-programmed desorption of NO (fragment NO,
2
m/e = 30) from MgO–Al O -MCM-36 and MCM-22.
2
3
min after the start of the experiment. Only when the max-
imum CO concentration diminished, the coke nitrogen was
converted to NO. Dishman et al. measured the concentra-
tion profile of combustion gases during regeneration and
observed a similar behavior [33]. In the presence of BaO–
Al2O3-MCM-36 21% less NO was released and nearly all
CO was consumed in the region of the second CO maxi-
mum of the base case, while only very little effect on CO
formation was observed in the first 10 min of the experiment.
Similar observations were made for all other MCM-36-type
additive investigated in this study.
Fig. 9. Differences of IR spectra measured during reaction of 5% CO on
MgO–Al O -MCM-36 after exposure of the catalyst to 1000 ppm NO and
2
3
◦
5
% O (T = 500 C).
2
oxidation of the nitrogen-containing species of the spent cat-
alyst. Table 3 provides a comparison of the reduction of
NO in the presence of the various additives or a conven-
tional Pt-based CO combustion promoter (CP-3). Compared
to spent catalyst only (base case) all materials, except CP-3,
showed a reduction of NO emissions. Best results (18–21%
NO conversion) were achieved for BaO–Al2O3-MCM-36
and BaO–Al2O3–SiO2-MCM-36, whereas for the MCM-36-
type derivatives containing MgO only 12–14% conversion
was obtained. However, all composite materials showed a
reduction of NO emissions, while CP-3 led to significant
increase in NO emissions by ∼ 300%. In contrast to the
Pt-based CO combustion promoter, the MCM-36-type ma-
terials facilitated only very small conversions of CO to CO2.
The comparison of the evolution of CO and NO emissions
during the regeneration experiment in the fluidized-bed re-
actor for the base case (absence of any additive) and in the
presence of 1 wt% BaO–Al2O3-MCM-36 added to the coked
cracking catalyst is shown in Fig. 11. For the base case, the
majority of CO was released in the first 8 min of the exper-
iment, while a significantly smaller amount was determined
in the following 35 min with a second maximum after 18
min. Interestingly, the formation of NO was observed in par-
allel to this second CO evolution with a maximum 18–20
In most of the industrial FCC regeneration units operating
under full-burn conditions the CO emissions are commonly
controlled by using Pt-based CO combustion promoters in
the FCC catalyst inventories. As pointed out such additives
might cause a significant increase of the NO emissions.
Therefore, we have tested the compatibility of our catalysts
with a commercial CO promoter (CP-3) to achieve a simulta-
neous reduction of CO and NO emissions in a experiment in
which the spent catalyst was mixed with 0.1 wt% CP-3 and
the regeneration was carried out under the same conditions
◦
(700 C, 2% O2 in N2). Table 4 exemplifies the performance
of the additives in the presence of 0.1 wt% CP-3. A reduction
of CO emissions by 51–86% compared to the base case (pure
spent catalyst) was achieved. In the presence of the MCM-
36-type derivatives less NO was released compared to the
experiment with CP-3 only. Best results were obtained for
BaO–Al2O3–SiO2-MCM-36, which lowered the NO emis-
sions by 31% compared to the Pt-based CO combustion pro-
moter. However, the additive also led to a decrease of CO

J.-O. Barth et al. / Journal of Catalysis 227 (2004) 117–129
125
Table 4
Conversion of NO and CO during the regeneration of a coked FCC catalyst in a fluidized-bed reactor in the presence of 1 wt% of MCM-36-type catalysts and
.1 wt% of a conventional Pt-based CO combustion promoter (CP-3)
0
Sample
CO conversion as compared to pure
spent catalyst (%)
NO conversion as compared to
regeneration in the presence of CP-3 (%)
CP-3
86
86
69
57
51
−
MgO–Al O -MCM-36/CP-3
BaO–Al O -MCM-36/CP-3
MgO–Al O –SiO -MCM-36/CP-3
BaO–Al O –SiO -MCM-36/CP-3
6
2
3
16
23
31
2
3
2
3
2
2
3
2
dense and the diluted phase of the FCC regenerator. Higher
oxygen concentrations inhibit the reduction of NO by CO, as
the CO/O2 reaction is favored at the expense of the CO/NO
reaction [34,35]. Under reaction conditions similar to the
oxygen-rich zone, which is encountered in the bottom region
◦
of the regenerator near the air grid (500–625 C; oxygen ex-
cess), NO is mainly converted to NO2 over the MCM-36
derivatives (see Fig. 1b). In comparison to zeolite MCM-22
(
“parent material”) an increase in the catalytic performance
was observed when MgO–Al2O3 and BaO–Al2O3 oxide pil-
lars were incorporated between the zeolitic layers, suggest-
ing that the mixed oxide clusters are the catalytically active
sites. The conversion of NO and yield to N2 increased with
increasing Mg and Ba concentrations in the additives. As a
direct oxidation of NO to NO2 is not favored thermodynam-
ically at high temperatures, high yields to NO2 can only be
attributed to the decomposition of surface nitrite and nitrate
species formed on the basic oxide clusters (MgO–Al2O3,
BaO–Al2O3) in the interlayer galleries. The temperature-
programmed desorption of NO2 from MgO–Al2O3-MCM-
(a)
3
6 (Fig. 10) has shown that such ionic NOx species still
◦
decompose in the high-temperature region above 500 C.
The formation of surface nitrite and nitrate species (mono-,
bidentate, bridged nitrates) was observed by IR spectroscopy
on the basic MCM-36-type materials (cf. Figs. 6 and 7). The
interaction of NO with the pillared zeolites led to the re-
moval of carbonate species (negative band at 1479 cm in
Fig. 6) present on the basic oxide clusters and the formation
of different ionic NOx surface species, which are probably
located on the pillars between the zeolite layers. Interest-
ingly, nitrates were also formed in the absence of O2, sug-
gesting that an oxidation of NO by reactive oxygen species
is present in the mixed oxide clusters in the interlayer gal-
leries. In the presence of 3% O2 mainly nitrate species were
observed on the surface with bidentate nitrates on Al close
(b)
−1
Fig. 11. Formation of (a) CO and (b) NO during the oxidative regenera-
tion (2% O ) of a spent FCC catalyst in the absence or presence of 1 wt%
2
BaO–Al O -MCM-36 in a fluidized bed reactor.
2
3
conversion to ∼ 51% compared to 86% in the sole presence
of CP-3.
4
. Discussion
−1
to Mg (1552 cm ) being the dominant species.
The in situ IR experiments showed that with increasing
temperature the presence of CO led to a removal of nitrite
and nitrate species from the catalyst surface. Simultaneously,
formate and various (bi-) carbonate species were observed,
which can only result from an oxidation of CO during the
reaction with the surface NOx species. The concentration
of carbonate and formate species adsorbed on basic sites
4
.1. Reaction mechanism
The results demonstrate the high catalytic activity of
MCM-36-type materials containing mixed alkaline earth
aluminum oxide pillars in the reduction of NO with CO
to N2 and CO2, in the presence of oxygen. Generally, the
conversion of NO and yields of N2 and N2O were signif-
icantly higher under oxygen-deficient reaction conditions,
which are encountered in the oxygen-depleted zones of the
◦
reached a maximum at ∼ 350 C, which can be explained
by a thermal desorption of the weakly bound species. Note

126
J.-O. Barth et al. / Journal of Catalysis 227 (2004) 117–129
that for MgO–Al2O3-MCM-36 and BaO–Al2O3-MCM-36
two types of basic hydroxyl groups were detected by CO2
adsorption: (i) Mg(Ba)–OH and (ii) AlIV–OH groups on the
spinel-type oxide clusters [23,36]. The latter basic sites were
attributed to hydroxyl groups on aluminum ions in a tetra-
hedral coordination sphere of the defect spinel-type oxide
pillaring clusters.
over the oxide pillars in the interlayer galleries. It is well
◦
known that at temperatures above 400 C N O decomposes
2
over basic oxides such as MgO, CaO, and BaO (cf. Refs.
[39–42] and references therein).
In the presence of excess oxygen (cf. Fig. 4b) signif-
icantly lower yields of N O were found for all materials
2
investigated. In general, higher oxygen concentrations de-
For the understanding of the reaction mechanism of NO
reduction with CO over the MCM-36-based catalysts the for-
crease the availability of the reductant CO and additionally
favor the oxidation of N O to NO and NO , inhibiting the
2
2
mation of isocyanate species and N2O at temperatures above
deNO reaction. Interestingly, under oxygen-deficient reac-
x
◦
3
∼
00 C is important. The partially overlapping IR bands at
2260 and 2240 cm (Al–NCO and Mg–NCO, respec-
tion conditions the maximum of the N O formation occurs
2
−1
over MgO-containing pillared zeolites at lower tempera-
tively) indicated that the –NCO intermediates are most prob-
ably adsorbed on the MgO–Al2O3 pillars [29,32]. Simulta-
neously with the observation of isocyanate species the for-
mation of N2O was revealed by the intense band of the N2O
tures compared to the more basic BaO–Al O -MCM-36 and
2
3
BaO–Al O –SiO -MCM-36 materials, which could result
2
3
2
from a higher coverage of basic sites with nitrate and car-
bonate species, thus inhibiting the N O decomposition.
2
−1
stretching vibration at 2227 cm . Isocyanates such as Al–
NCO and Mg–NCO species may be formed by the reaction
of nitrite and nitrate species with CO. The formation of the
N–N bond during the reduction of NO with CO leading to
N2 and N2O, is tentatively ascribed to the reaction of sur-
face isocyanates with NO:
Hori et al. have shown that N O can also be easily re-
2
duced over MgO by adsorbed formate species. The inter-
action of CO with basic hydroxyl groups of MgO leads to
formates which subsequently reduce N2O to N2 [31],
MgO
+
Mg(OH)2 + CO −−−−→ (MgOH)
,
(4)
Al–NCO + NO → N2 + CO2,
Al–NCO + NO → N2O + CO.
(1)
(2)
This assumption is supported by the observation that
HNCO reduces NO, e.g., in diesel exhaust, at temperatures
above 400 C. Acke et al. have shown by isotope-labeling
MgO
+
(
MgOH)
+ N2O −−−−→ CO2+ N2
◦
transient experiments that isocyanates adsorbed on Pt sup-
ported on Al2O3 reduce NO to N2 and N2O [37]. Schießer
et al. have demonstrated by IR spectroscopy a correlation
between the intensity of the –NCO band on Pt–Al-MCM-41
and the conversion of NO [38]. Similarly to the materials
discussed here, the isocyanates detected were mainly ad-
sorbed on anionic defect sites of the Al2O3-containing sup-
port material. Interestingly, we have observed the formation
of –NCO and N2O in the absence of a noble metal, indicating
that metal clusters are not a prerequisite for the reaction of
NO with isocyanate intermediates at elevated temperatures
+ Mg(OH)2.
(5)
◦
At temperatures below 500 C formates were observed
on MgO–Al2O3-MCM-36 during the reaction of CO with
nitrites and nitrates adsorbed on the catalyst surface, as
indicated by characteristic IR bands at 2930, 1610, and
−
1
1
384 cm . However, it is speculated that at elevated tem-
peratures typical of the FCC process decomposition of ni-
trous oxide over the oxide pillars is the dominating reaction.
Generally, the higher catalytic activity of MgO–Al2O3-
MCM-36 at lower temperatures is attributed to an increased
resistance to poisoning by CO , which is a major reaction
product. MCM-36 materials with the more basic BaO pillars
will be more easily deactivated by the formation of carbon-
ates than the MgO-containing MCM-36 materials. Conse-
quently, the reduction of NO by CO over MCM-36-type
materials might be explained as a two-step process involv-
ing the formation of nitrous oxide as an intermediate
◦
(
above 500 C). N2O is a common by-product in the reaction
2
of NO with CO over noble metal-impregnated oxides. The
formation of N2O was already observed over mixed oxide
materials, such as nonstoichiometric nickel–copper spinel
manganites and Al-MCM-41 impregnated with Pt [35,38].
For MCM-36-type materials with alkaline earth aluminum
oxide pillars significantly higher yields of N2O were ob-
served compared to the nonpillared parent material MCM-
2
NO + CO → N2O + CO2,
N2O + CO → N2 + CO2,
NO + 2CO → N2 + 2CO2.
(6)
(7)
(8)
2
2, indicating that the basic mixed oxide phases enhance the
N2O formation. The generation of N2O is favored by low
oxygen concentrations and higher concentrations of Mg and
Ba (cf. Fig. 5). While the yield of N2 increased continuously
with temperature, the yield of N2O reached a maximum be-
2
Reaction (7) corresponds to the decomposition of nitrous
◦
tween 575 and 675 C, which indicates the decomposition of
oxide and is generally considered as a fast process compared
to reaction (6) [35]. A detailed discussion of the kinetics of
N2O decomposition on oxide catalysts can be found in [43].
N2O,
1
2
N2O → N2 + O2,
(3)

J.-O. Barth et al. / Journal of Catalysis 227 (2004) 117–129
127
In the absence of water, all derivatives of MCM-36 show
stable catalytic activity in time-on-stream measurements, in-
dicating that the composite structure is stable during the
high-temperature reactions. In the presence of water, how-
ever, the catalytic activity for the reduction of NO with CO
decreased after ∼ 5.5 h (see Fig. 5), which is attributed
to a gradual dealumination of the nanocomposite materials
during the high-temperature reaction in the presence of 5%
H2O.
line with observations of Efthimiadis et al. for other FCC
deNOx additives [28]. The conversion of NO in the regen-
eration experiment in the fluidized-bed reactor is slightly
lower than in the fixed-bed catalytic experiments in the ex-
cess of oxygen (Fig. 1b). All particles in the fluidized bed
(CSTR reactor) show the same extent of regeneration, which
implies that in the beginning of regeneration CO is primar-
ily formed, while NO is emitted when the regeneration has
proceeded. Consequently, at this stage of regeneration high
O2 (1.5–2%) and low CO concentrations (< 0.1%) are en-
countered in the fluidized bed. Under realistic conditions in
a continuously operated industrial FCC regenerator, lower
oxygen and nearly constant carbon monoxide concentrations
are present in the oxygen-depleted zone of the fluidized bed,
which should most probably lead to higher NO conversions.
Nevertheless, the MCM-36-type additives reduce NO even
in the presence of very low CO concentrations, indicating
the presence of a second reduction mechanism.
4
.2. Comparison of lab-scale reactor and pilot plant
kinetic measurements
Under reaction conditions similar to those found in the
regenerator unit of a FCC plant, additives based on MCM-
3
6-type catalysts with mixed MgO–Al2O3 and BaO–Al2O3
oxide pillars are only active in the oxygen-deficient region
of the fluidized bed, where most of the oxygen is consumed
by burning off the coke present on the catalyst. After reac-
During the regeneration of the FCC catalysts HCN and
NH3 are present especially in the oxygen-depleted zones of
◦
tion at 500–575 C for 5–10 h, high water concentrations
◦
the fluidized bed. At temperatures above 700 C HCN is hy-
(
5–8%) found in parts of the regenerator and the stripper
drolyzed to NH3 over basic oxides such as the mixed oxide
clusters of the pillared zeolites [45],
have a negative effect on the catalytic activity of the materi-
als, probably due to steam-induced dealumination reactions
as suggested by 27Al MAS-NMR (spectra not shown). The
performance of the MCM-36-type additives under reaction
conditions typical for a fluidized-bed reactor used for the
regeneration of coked FCC catalysts indicated that the ox-
idative regeneration of carbonaceous deposits on the spent
catalysts leads to O2, N2, CO, CO2, H2O, SOx, and NOx
formation in the flue gas of the regenerator (see Table 3 and
Fig. 11) [6,27,44]. Additionally, NH3 and HCN are found
in lower concentrations (10–200 ppm) as reduced nitrogen-
containing intermediates of coke pyrolysis and precursors
of NOx [8,14,27]. Performing the regeneration experiment
in the presence of CP-3, a commercial Pt-based CO com-
bustion promoter, leads to a decrease of CO formation ac-
companied by a significant increase in NOx emissions by
HCN + H2O → NH3 + CO,
(9)
which reduces NO over Brønsted acid sites found in the ze-
olite layers of the composite materials [46]:
3
NO + 2NH3 → 2.5N2 + 3H2O.
(10)
Various studies claim that NH3 bonded to Brønsted acid
sites of H-form zeolites reacts with NO2-type intermediates
formed on the zeolite surface during selective catalytic re-
duction (SCR) reactions ([46] and references therein).
MCM-36 derivatives with mixed alkaline earth–alumi-
num oxide pillars allow stabilization of reactive nitrogen-
containing intermediates such as NH3 (on the Brønsted and
strong Lewis acid sites), HCN (by basic oxide-catalyzed hy-
drolysis to NH3), and NOx (formation of nitrite and nitrate
species on the mixed oxide pillars) prior to the SCR reac-
∼
300%. As shown earlier by Yaluris et al. and our group
this observation can be attributed to the Pt-catalyzed oxida-
tion of HCN and NH3 to NOx comparable to the Ostwald
process [8,14,27]. In contrast, the pillared MCM-36 mate-
rials based on MCM-22 zeolites yielded a reduction of NO
emissions, while only a very small conversion of CO was
observed. The regeneration experiments showed the typical
sequential burning of coke species from the deactivated cat-
alyst. As can be seen from the oxidative regeneration (see
Fig. 11) NO is mainly formed at the end of the regenera-
tion experiment, when most of the carbon has been burnt off.
All additives led to a reduction of NO emissions in this tem-
perature range, BaO–Al2O3-MCM-36 being the most active.
The decrease of NO emissions was accompanied by a dis-
tinct removal of CO in the region of the second maximum
of CO formation (cf. Fig. 11), which strongly suggests a re-
duction of NO with CO. The relatively low conversions of
NO can be best explained with a significantly lower con-
centration of CO at the end of the experiment and are in
tion to N , which probably occurs on Brønsted acid sites of
2
the MCM-22 zeolite layers. Consequently, the potential of
MCM-36-type additives for the reduction of NO emissions
during the regeneration of FCC catalysts may be attributed
to catalytic active sites for (i) the reduction of NO with CO
and (ii) for the reduction of NO with NH3.
In contrast to the catalytic experiments in the fixed-bed
reactor, in the fluidized-bed reactor the best results were ob-
tained for the BaO-containing nanocomposite materials. The
higher catalytic activity of such pillared zeolites is tenta-
tively explained by an enhanced reactivity for the hydrolysis
of the reaction intermediate HCN to NH over the BaO–
3
Al O pillars [Eq. (9)]. This reaction provides a higher con-
2
3
centration of NH which can act as a reducing agent in the
3
,
reaction with NO. In contrast to the fixed-bed experiments
deactivation by poisoning with carbonate species does not
seem to play an important role probably due to the lower
CO2 concentrations in our regeneration experiments.

128
J.-O. Barth et al. / Journal of Catalysis 227 (2004) 117–129
From a practical point of view it is important to note that
involving the formation of nitrous oxide as an intermedi-
the MCM-36-type additives can be used together with a Pt-
based CO combustion promoter (Table 4). Compared to the
pure spent catalyst (base case) the NO emissions have in-
creased due to the oxidation of reduced nitrogen containing
intermediates by the noble metal-containing CO promoter.
However, in comparison with sole CP-3 lower NO concen-
trations were measured, when the additives were used ad-
ditionally in the regeneration experiment. As the promoter
significantly lowers the concentration of CO in the regen-
eration experiment, the reduction of NO emissions is best
explained with a direct interaction of the MCM-36 catalysts
with the precursor molecules (NH3, HCN) for the forma-
tion of NOx. As outlined above, Pt-based CO combustion
promoters lead to the oxidation of HCN and NH3, which
are formed during the pyrolysis of nitrogen-containing coke
molecules. It is speculated that the (intermediate) NH3 is
temporarily trapped on Brønsted acid sites in the zeolite
layers or on strong Lewis acid sites in the mixed oxide pil-
lars, thus preventing an immediate oxidation to NO over the
Pt catalyst. This hypothesis is supported from the fact that
MgO–Al2O3–SiO2-MCM-36 and BaO–Al2O3-SiO2-MCM-
ate. In the presence of H O the catalytic activity decreases
due to a gradual dealumination of the composite material
during the high-temperature reaction. The additives show a
2
reduction of NO emissions even in experiments simulat-
x
ing the regeneration of industrial coked FCC catalysts in a
fluidized-bed reactor. In contrast to Pt-based additives, the
nanocomposite materials do not oxidize reduced nitrogen
containing intermediates (NH , HCN), which are generated
3
during the pyrolysis of nitrogen coke species, but lower
the concentration of NO released in the flue gases. It is
x
speculated that during the regeneration of spent FCC cata-
lysts the MCM-36-type additives catalyze, in addition to the
NO + CO reaction, the SCR reaction of NO with NH to N
3
2
on Brønsted acid sites in the zeolite layers or on the pillars in
the interlayer galleries. HCN, which is a key intermediate in
the nitrogen chemistry of the FCC unit, might be hydrolyzed
to NH over the basic oxide pillars in the interlayer galleries.
3
The additives can be used in combination with Pt-based CO
promoters to simultaneously control the level of NO and CO
in the regeneration unit.
3
6 showed the highest catalytic activity for the reduction of
NO in the presence of CP-3, although the concentration of
Mg and Ba is lower than in the other materials. We have
shown earlier by temperature-programmedoxidation of NH3
that both materials possess additional, strong Brønsted acid
sites on the pillars close to the basic mixed oxide clusters
Acknowledgments
Financial support of the European Union (Project G1RD-
CT99-0065-“DENOXPRO”) is gratefully acknowledged.
The authors thank Dr. H. Rhemann (OMV), Dr. R. Hard-
ing, Dr. J. Nee, and Dr. G. McElhiney (GRACE-Davison)
for helpful discussions concerning FCC deNOx technolo-
gies.
[
23]. Consequently, NH3 adsorbed on the acid sites can re-
act immediately with NOx (from the gas phase or adsorbed
on the basic oxide pillars) to N2. The combination of a con-
ventional Pt-based CO promoter and MCM-36 derivatives
allows the simultaneous control of CO and NO emissions in
the FCC regenerator (see Table 4). A further reduction of NO
emissions might be achieved by a variation of the concentra-
tion of additive in the inventory and a continuous operation
mode of the fluidized-bed reactor, which should better sim-
ulate the actual situation in the FCC regenerator.
References
[1] R.H. Harding, A.W. Peters, J.R.D. Nee, Appl. Catal. A 221 (2001)
389.
[
[
2] J.-O. Barth, Erdöl Erdgas Kohle 119 (2003) 86.
3] W.-C. Cheng, G. Kim, A.W. Peters, X. Zhao, K. Rajagopalan, M.S.
Ziebarth, C.J. Pereira, Catal. Rev.-Sci. Eng. 40 (1998) 39.
4] R. Mann, Catal. Today 18 (1993) 509.
[
[
5
. Conclusions
5] A.W. Peters, G. Yaluris, G.W. Weatherbee, X. Zhao, Fluid Crack.
Catal. (1998) 259.
MCM-36-type additives with mixed alkaline earth alu-
[6] X. Zhao, A.W. Peters, G.W. Weatherbee, Ind. Eng. Chem. Res. 36
minum oxide (MgO–Al2O3-MCM-36, BaO–Al2O3-MCM-
(1997) 4535.
3
3
6, MgO–Al2O3–SiO2-MCM-36, BaO–Al2O3–SiO2-MCM-
6) pillars are highly active additives for the reduction of NO
[7] D.A. Cooper, A. Emanuelsson, Energ. Fuel. 6 (1992) 172.
[
8] J.-O. Barth, A. Jentys, J.A. Lercher, in: Proceedings of the 17th World
Petroleum Congress, Rio de Janeiro, 2002, vol. 3, Institute of Petro-
leum, London, 2003, p. 445.
with CO under reaction conditions similar to the oxygen-
depleted zone of the FCC regenerator. The reaction proceeds
via nitrite, nitrate, and isocyanate intermediates, which are
adsorbed on the basic mixed oxide clusters in the interlayer
galleries of the MCM-36-type materials. N2 and N2O are
formed by the reaction of isocyanates with NO. At temper-
atures characteristic of the FCC regeneration process N2O
decomposes over basic oxide clusters in the composite ma-
terials yielding N2. The reduction of NO by CO over MCM-
[9] E.F. Iliopoulou, E.A. Efthimiadis, I.A. Vasalos, J.-O. Barth, J.A.
Lercher, Appl. Catal. B 47 (2004) 165.
[10] E.A. Efthimiadis, A.A. Lappas, D.K. Iatrides, I.A. Vasalos, Ind. Eng.
Chem. Res. 40 (2001) 515.
11] B. Wen, M. He, Appl. Catal. B 37 (2002) 75.
12] B. Wen, M. He, C. Costello, Energ. Fuel. 16 (2002) 1048.
13] A. Corma, A.E. Palomares, F. Rey, F. Márquez, J. Catal. 170 (1997)
[
[
[
140.
[
14] J.-O. Barth, A. Jentys, J.A. Lercher, Ind. Eng. Chem. Res. 43 (2004)
3
6-type materials can be explained by a two-step process
2368.

J.-O. Barth et al. / Journal of Catalysis 227 (2004) 117–129
129
[
15] J.-O. Barth, A. Jentys, J.A. Lercher, Ind. Eng. Chem. Res. 43 (2004)
[28] E.A. Efthimiadis, E.F. Iliopoulou, A.A. Lappas, D.K. Iatrides, I.A.
Vasalos, Ind. Eng. Chem. Res. 41 (2002) 5401.
3097.
[
[
16] D.O. Chessmore, C.E. Rudy Jr., US patent 4,199,435, 1999.
17] W.J. Roth, C.T. Kresge, J.C. Vartuli, M.E. Leonowicz, A.S. Fung, S.B.
McCullen, Stud. Surf. Sci. Catal. 94 (1995) 301.
[29] K.I. Hadjiivanov, Catal. Rev.-Sci. Eng. 42 (2000) 71.
[30] B. Westerberg, E. Fridell, J. Mol. Catal. A 165 (2001) 249.
[31] B. Hori, N. Takezawa, Catal. Lett. 12 (1992) 383.
[32] V.A. Matyshak, O.V. Krylov, Catal. Today 25 (1995) 1.
[33] K.L. Dishman, P.K. Doolin, L.D. Tullock, Ind. Eng. Chem. Res. 37
(1998) 4631.
[34] M. Shelef, K. Otto, H. Gandhi, J. Catal. 12 (1968) 361.
[35] C. Drouet, P. Alphonse, A. Rousset, Appl. Catal. B 33 (2001)
35.
[36] H. Knözinger, P. Ratnasamy, Catal. Rev.-Sci. Eng. 17 (1978) 31.
[37] F. Acke, B. Westerberg, M. Skoglundh, J. Catal. 179 (1998) 528.
[38] W. Schiesser, H. Vinek, A. Jentys, Appl. Catal. B 33 (2001) 263.
[39] S. Xie, J.H. Lunsford, Appl. Catal. A 188 (1999) 137.
[40] E.M. Serwicka, Polish J. Chem. 75 (2001) 307.
[41] M.A. Wójtowicz, J.R. Pels, J.A. Moulijn, Fuel Process. Technol. 34
(1993) 1.
[
[
[
[
18] W.J. Roth, J.C. Vartuli, C.T. Kresge, Stud. Surf. Sci. Catal. 129 (2000)
501.
19] Y.J. He, G.S. Nivarthy, F. Eder, K. Seshan, J.A. Lercher, Micropor.
Mesopor. Mater. 25 (1998) 207.
20] W. Roth, Jerzy, J. Vartuli, Clarke, Stud. Surf. Sci. Catal. 141 (2002)
73.
21] J.-O. Barth, J. Kornatowski, J.A. Lercher, J. Mater. Chem. 12 (2002)
69.
2
3
[
[
22] S. Xie, M.P. Rosynek, J.H. Lunsford, J. Catal. 188 (1999) 24.
23] J.O. Barth, A. Jentys, J. Kornatowski, J.A. Lercher, Chem.
Mater. 16 (4) (2004) 724.
[
[
[
24] J.-O. Barth, R. Schenkel, J. Kornatowski, J.A. Lercher, Stud. Surf. Sci.
Catal. 135 (2001) 136.
25] L. Rheaume, R.E. Ritter, J.J. Blazek, I.A. Montgomery, Oil Gas J. 74
[42] F. Kapteijn, J. Rodriguez-Mirasol, J.A. Moulijn, Appl. Catal. B 9
(1996) 25.
[43] E.R.S. Winter, J. Chem. Soc. A (1968) 2889.
[44] G. Yaluris, A.W. Peters, in: J.G. Reynolds, M.R. Khan (Eds.), Design-
ing Transportation Fuels for a Cleaner Environment, Taylor & Francis,
Philadelphia, 1999, p. 151.
(1976) 103.
26] K. Limbach, S. Tamhankar, S. Ganguly, V. Balse, R. Menon,
R. Ramachandran, Petroleum Technology Quaterly—eptq.com—
The Refining, Gas and Petrochemical Processing Website, http://
www.eptq.com/, 2000.
[
27] G. Yaluris, A.W. Peters, in: 2nd Int. Conf. on Refining Processes,
[45] S. Schäfer, B. Bonn, Chem. Ing. Techn. 71 (1999) 613.
[46] J. Eng, C.H. Bartholomew, J. Catal. 171 (1997) 27.
AIChE Spring National Meeting, Houston, 1999, p. 27.