Advances and future trends in selective

oxidation and ammoxidation catalysis

Robert K. Grasselli
1
Department of Chemical Engineering and Materials Science, University of Delaware, Newark, DE 19716-2543, USA
Institute of Physical Chemistry, University of Munich, Sophienstrasse 11, D-80333 Munich, Germany
Abstract
A review of the current status and future trends in oxidation catalysis is presented. The topics entail the (amm)oxidation of
propylene to acrylonitrile using complex multicomponent mixed metal molybdates and antimonates as catalysts, the
ammoxidation of propane to acrylonitrile using V-Sb-oxides or V-Te-Nb-Mo-oxides, the oxydehydrogenation of light parafns
using Ni-Co-molybdates or Mg-V-Sb-oxides, dehydrogenation of propane combined with selective hydrogen oxidation using
Group VIII elements, oxidation of propane to acrylic acid using complex molybdates, and the (amm)oxidation of xylenes to
terepthalic acid or terepthalonitrile using Fe-Mo-modied ZSM-5 catalysts.
The respective reaction mechanism of the prevailing selective catalytic oxidations are discussed, as are the specic
functionalities of the various elements contained in the mixed metal catalysts. Some new approaches to the rational design and
synthesis of improved catalysts are offered. # 1999 Elsevier Science B.V. All rights reserved.
Keywords: Selective oxidation; Ammoxidation catalysis; Oxydehydrogenation; Molybdates; Antimonates; Phosphates;
Olens; Parafns; Aromatics
1. Introduction
Selective catalytic oxidation and ammoxidation
processes of hydrocarbons comprise approximately
one quarter of the value produced by all catalytic
processes world wide. They contribute hence signi-
cantly to the gross national product of industrial
countries and afford future opportunities for develop-
ing countries.
These processes are not only important from a
commercial standpoint and their contribution to the
well-being of modern mankind, but they also present
the opportunity for signicant fundamental research.
Such research aims at the molecular level understand-
ing of catalytic behavior, forming a basis for rational
design of catalyst improvement of existing processes
and the discovery of totally new catalysts and pro-
cesses.
This paper limits itself to the chemistry of selective
oxidation and ammoxidation processes of large
volume chemicals and intermediates, addressing their
current status and presenting the author's views of
future opportunities. Although only the catalysis of
large volume chemicals is addressed, the principles
advanced are also applicable to the catalysis of ne
chemicals production.
Catalysis Today 49 (1999) 141153
1
Tel.: +49-89-5902-261; fax: +49-89-5902-602; e-mail:
[email protected]
0920-5861/99/$ see front matter # 1999 Elsevier Science B.V. All rights reserved.
PI I : S0 9 2 0 - 5 8 6 1 ( 9 8 ) 0 0 4 1 8 - 0
2. Discussion
2.1. Ammoxidation of olefins
The ammoxidation of propylene to acrylonitrile
commands the foremost commercial position among
all ammoxidation processes [1,2]:
CH
2
=CHCH
3
NH
3
3a2O
2
(air)

catalyst
CH
2
=CHCN 3H
2
O
More than 10 billion pounds of acrylonitrile are
produced worldwide by this SOHIO (now BP Amer-
ica) invented process. Acrylonitrile is a versatile pet-
rochemical intermediate used extensively in the
production of acrylic bers, resins, rubbers and
specialty products.
As is well documented in [1,2], ammoxidation is a
six-electron redox process (Fig. 1) requiring catalysts
able to activate propylene through the abstraction of
an -hydrogen, followed by NH-insertion, subsequent
rearrangements of the chemisorbed activated surface
species with additional hydrogen abstraction, and
culminating in the desorption of the so formed acry-
lonitrile from the catalyst surface. The catalyst which
is reduced in this propylene ammoxidation process is
continuously regenerated in situ by gaseous oxygen
present in the air of the feed gas mixture. This occurs
by facile lattice oxygen transport from a surface
reoxidation site, where dioxygen is built into the
catalyst, to the reduced active center as shown in
Fig. 1. In this manner the catalytic cycle of ammox-
idation is completed: one molecule of propylene and
ammonia, and one and one half dioxygens are con-
sumed with the production of one molecule of acry-
lonitrile and three molecules of water, and the
catalysts' active center is reconstituted to its original
fully oxidized form.
The most commonly used catalysts for this reaction
are either those based on mixed metal molybdates or
mixed metal antimonates. These systems continue to
dominate the eld.
2.1.1. Multicomponent complex mixed metal
molybdates
Over the past 20 years several generations of multi-
component molybdates have been developed, and the
compositions are becoming ever more sophisticated
and complex, giving ever better yields of the desired
acrylonitrile product. The last three catalyst genera-
tions, summarized according to their empirical for-
mulae, are presented in Table 1.
These compositions have been ne tuned for max-
imum acrylonitrile yield production in uid bed reac-
tors, the reactors of choice in commerce, and each
element contained within these compositions has its
designated function and performs accordingly.
For example, it is by now well known that the
primarily function of bismuth [1,2] in these composi-
tions is its ability to activate the surface adsorbed
propylene through the abstraction of an -hydrogen.
The function of molybdenum is to chemisorb propy-
lene and activate the ammonia molecule by generating
chemisorbed NH-species which are capable of insert-
ing into the chemisorbed allylic intermediate, creating
the acrylonitrile precursors. The function of iron is to
serve as an efcient redox couple (Fe
3/2
), capable of
efcient lattice oxygen transfer to the BiOMo active
site in its 3 oxidation state. In its 2 state it is
capable of efcient dioxygen chemisorption, its reduc-
tive transformation to lattice oxygen (O
2
) and hence
its incorporation into the lattice.
Since it is very difcult to maintain a sufcient
number of Fe
2
surface sites in an overall oxidizing
gaseous atmosphere, it is necessary to stabilize the
Fe
2
state structurally. To this end divalent elements
such as Ni, Co, Mg and Mn form stable isostructural
molybdates with Fe
2
molybdates, and by virtue of
Fig. 1. Generalized mechanistic cycle for alkane ammoxidation
[2].
142 R.K. Grasselli / Catalysis Today 49 (1999) 141153
solid solutions stabilize the Fe
2
state. Thus Ni, Co,
Mg and Mn have the function of providing the host
structure for Fe
2
in these multifunctional, multiphase
catalysts.
In a similar manner, Fe
3
is stabilized through its
limited solubility in the -bismuth molybdate phase,
the latter phase providing the host structure for Fe
3
.
The Cr
3/2
is a redox couple of its own, generally,
effective at a temperature higher than that of the iron
couple. But it also acts as a structural diluent to iron
and booster of the Fe
3
state in the Bi-molybdate
matrix.
The role of the alkali is to annihilate the most acidic
cracking sites of the compositions, to serve as a spacer
and contact enhancer of the two functionally distinct
but epitaxially matched catalyst phases. The two
phases are composed of the Fe
3
containing cataly-
tically active -bismuth molybdate phase and the Fe
2
containing (Ni, Co, Mg, Mn) molybdate phase. These
two phases must cooperate with each other, and there-
fore, they must be in utmost proximity of each other;
they would be ineffective if remotely separated from
each other. For this reason it is of interest that the most
effective multiphase oxidation catalysts have at least
one face which is for all practical purposes epitaxially
essentially matched (principle of phase cooperation
[7]).
A schematic representation of such a phase coop-
eration between two functionally distinct phases of
typical multicomponent multiphase catalysts enumer-
ated above is shown in Fig. 2.
The necessity of having two functionally distinct
phases to attain high acrylonitrile yield efciency is
illustrated in Fig. 3 [8]. It is obvious that the Fe
3
containing Bi-molybdate active phase, as well as the
Fe
2
containing M
2
-molybdate regeneration phase,
must be present at the same time to attain high
acrylonitrile product yields.
Future trends. While a great deal has already been
achieved in the propylene ammoxidation catalysis,
there is still ample room for further improvement.
Since the process is not thermodynamically limited,
there is another 20% acrylonitrile to be conquered.
Approaches that might be fruitful towards this end
would include attempts to discover catalysts capable
of activating ammonia at lower temperatures; thus
allowing for lower temperature process operation
and hence anticipated higher acrylonitrile yields and
less byproducts caused by cracking reactions and
waste forming CO
x
processes. Another approach
might be aimed at discovering single phases which
might be capable of accommodating the two above
described distinct catalytic functions within a single
phase, by propriciously placing these functionalities in
an ordered and optimally designed spatial arrange-
ment. There is also ample latitude left in optimizing
the texture of these complex ammoxidation catalysts.
Surely there are other approaches to be pursued to
achieve higher desired product yields and environ-
mentally more optimal product mixes.
Table 1
Acrylonitrile in-tank yield
(mole % based on feed propylene)
Reference
K
a
(Ni,Co)
9
Fe
3
BiPMo
12
O
x
SiO
2
~75 [3,4]
(K,Cs)
a
(Ni,Co,Mn)
9.5
(Fe,Cr)
2.5
BiMo
12
O
x
SiO
2
~7880 [5]
(K,Cs)
a
(Ni,Mg,Mn)
7.5
(Fe,Cr)
2.3
Bi
0.5
Mo
12
O
x
SiO
2
>80 [6]
Fig. 2. Epitaxial match at the (0 1 0) face of Bi
2
Mo
3
O
12
and -
FeMoO
4
[7].
R.K. Grasselli / Catalysis Today 49 (1999) 141153 143
2.1.2. Multicomponent complex mixed metal
antimonates
The pinnacle of commercial antimonate catalysis
was reached some years ago with the introduction of
the USb
3
O
10
0.8Sb
2
O
4
SiO
2
[911] catalyst which
used
235
U depleted uranium. However, its commercial
dominance was rather short, not because of its product
yield structure or catalyst life problems, but rather
because of presumed environmental concerns owing
to minimal residual radiation from the depleted ura-
nium contained in the catalyst.
The successor catalyst became an Feantimonate
system. This latter FeSb oxide based catalyst has
undergone extensive study and compositional changes
over the years, with concomitant improvements in
acrylonitrile yield. Without these changes the com-
plexity of the composition has also increased greatly.
Catalysts currently used on commercial scale can be
described by the following empirical formula:
Na
03
(CuY MgY ZnY Ni)
04
(VY W)
0X051
Mo
0X12X5
Te
0X25
Fe
10
Sb
1320
O
x
SiO
2
(cf. [12,13]). They typically yield ~75% in-tank
acrylonitrile.
As is the case with the complex molybdates
described above, the individual functions of the var-
ious elements contained in the complex antimonates
are also well understood, but will not be described
here, because of space constraints.
Although the in-tank acrylonitrile yields of the
above antimonate catalyst compositions are respect-
able, they do not measure up to the productivities of
the best molybdate catalysts discussed. In addition, the
antimonate based catalysts are much more fragile than
the molybdate based systems under commercial
operation. Therefore, it is imperative to operate the
antimonate systems under rigidly controlled operating
conditions, in order to prevent possible plant upsets
and irreversible reduction which leads to the destruc-
tion of the antimonate catalysts. In contrast, the later
generation molybdate based catalysts are almost
indestructible and can easily withstand inadvertent
plant upsets, including severe reductions. It is for
these reasons that the complex mixed metal molyb-
dates are currently the industrially preferred catalysts
for the ammoxidation of propylene to acrylonitrile.
Future trends. The greatest advance in antimonate
catalysts would be to improve their redox stability and
irreversible structural changes caused by severe reduc-
tion. This problem has been partially addressed by
incorporating efcient redox components such as Fe
molybdates and Fetungstates to the compositions.
All advanced Feantimonates contain some molybde-
num, and mostly also some tungsten. It is difcult to
match these phases, and special preparational and
Fig. 3. Relationship between catalytic activity and phase composition of K
1
a
M
2
b
M
3
10b
Bi
x
Mo
y
O
z
aSiO
2
[8].
144 R.K. Grasselli / Catalysis Today 49 (1999) 141153
calcination techniques are required to achieve these
goals. Another desirable aim would be to replace the
tellurium by another element having similar selectiv-
ity imparting properties in antimonates since tellurium
is easily reducible and thus rather volatile and poiso-
nous. Attempts to replace the Te with Bi have not fully
achieved the desired goal. The search is still on. In
uid bed reactors, a raised sparger (i.e., propylene and
ammonia feed inlets) and lowered air inlet, constitut-
ing an auto-regenerative zone [14], are useful engi-
neering expedients towards improved catalyst stability
and life.
2.2. Ammoxidation of paraffins
Among parafn ammoxidations, the ammoxidation
of propane is furthest advanced, although not as yet
fully commercialized. BPAmerica (formerly SOHIO)
has announced on several occasions that it is com-
mercializing the process [15,16]; however thus far this
has not been realized, although the process is allegedly
in pilot plant stage at BP.
The process is an eight-electron oxidation and
requires a catalyst or combination of catalysts which
are even more complex than those for the ammoxida-
tion of olens (e.g., propylene):
CH
3
CH
2
CH
3
NH
3
2O
2
(air)

catalyst
CH
2
=CHCN 4H
2
O
There is ample incentive to develop an ammoxidation
process based on propane rather than propylene, since
the price and availability of propane are substantially
more favorable than those of propylene. The task is
difcult from a commercial standpoint, since the well-
established and well-functioning propylene based pro-
cess is constantly being further improved, and the
catalysts therefore constantly made better, more ef-
cient and environmentally friendlier. It is simply a
moving target, difcult to hit, yet not out of reach and
will be achieved in due time.
Many catalysts have been tested for the ammoxida-
tion of propane to acrylonitrile, and the most effective
of them fall, as do propylene ammoxidation catalysts,
into two main classes: they are either antimonates or
molybdates.
Among the antimonates, the most studied belong to
the family VSb
x
M
y
, where M can be many different
elements, with the most frequently used dopants being
W, Te, Nb, Sn, Bi, Al, and Ti [1719]. Virtually all of
these antimonates possess the rutile or trirutile struc-
ture. Some of the highest acrylonitrile yielding cata-
lysts are summarized in Table 2.
The molybdate family is represented by VMo
x-
M
y
O
z
, where M is most often Bi or Te [22,23]. Some
catalysts are of the scheelite structure [24], others are
multiphase and of more complex structure. Among
these molybdates, the highest acrylonitrile yields are
claimed for a catalyst composition having the empiri-
cal formula V
0.3
Te
0.23
Nb
0.12
MoO
x
supported on SiO
2
[25]. A calcination step in nitrogen at 6208C appears
to be crucial for these compositions to attain the
proper catalytic solid state structure and herewith
the desired acrylonitrile yield. A maximum yield of
55.1% acrylonitrile at 86.7% conversion and 63.5%
selectivity has been reported [25]. It is unclear at this
time what the desired structure of this composition
might be, it has variously been claimed to be a ``K-
phase'' [26] and then again an ``M-phase'' [27]. The d-
spacing of the main X-ray diffraction lines are stated
to be located at: 4.02 (1 0 0), 3.16 (20150), 2.48 (5
60), 2.00 (240), 1.82 (240). Quite a latitude of
intensities are recorded and herewith possible differ-
ences in catalytic performance. The underlying struc-
ture is not all that different from that of Thorsteinson's
V
0.51
Nb
0.13
MoO
x
, doped by Te [28]. Unfortunately,
thus for it has been rather difcult to independently
reproduce the Mitsubishi claimed performance, pri-
marly because of the difculties connected with the
preparation of the catalyst.
The Vantimonate based catalysts have been thor-
oughly studied structurally, kinetically and mechan-
istically by several research groups. They are
multiphase, containing VSbO
4
with rutile structure
and -Sb
2
O
4
as a spectator phase, a ready source of
excess antimony during the redox process. Working
catalysts contain also supra-surface antimony sites
Table 2
Conversion Selectivity Yield
VSb
5
W
0.5
Te
0.5
Sn
0.5
O
x
SiO
2
/
Al
2
O
3
[13,17,18]
68.8 56.7 39
VSb
5
Bi
2
Fe
2
O
x
Al
2
O
3
[20] 39 84 32
VSb
5
Sn
5
[21] 14 60 8
R.K. Grasselli / Catalysis Today 49 (1999) 141153 145
[29] on top of the VSbO
4
phase as a result of antimony
migration from the spectator phase during catalytic
reaction.
Extensive kinetic studies (Fig. 4) [30] show unequi-
vocally that propylene is a desorbed primary product
and the main precursor of acrylonitrile from propane.
These and other supporting data suggest that propane
is activated by a VO
v
moiety [31,32] of the catalyst
by abstracting the rst hydrogen from the methylene
group of the molecule in the rate limiting reaction of
the overall process. The desorbed propylene reacts
further on adjacent Sb
3
OSb
5
surface sites,
whereby the Sb
3
O site is responsible for the rate
limiting abstraction of an -hydrogen from the pro-
pylene molecule and an Sb
5
=NH site for the nitrogen
insertion into the chemisorbed allylic surface com-
plex, which after rearrangement and additional hydro-
gen loss desorbs as product acrylonitrile [33]. The
reoxidation of the catalyst occurs by lattice oxygen
(O
2
) migration from the reoxidation site to the active
site with anion vacancy migration in the opposite
direction. The reoxidation sites contain V
4
moieties
capable of dioxygen dissociation and its incorporation
as lattice oxygen into the catalyst structure. This
process keeps the lattice oxygen concentration at a
steady state during the redox process of propane
ammoxidation.
The VSb oxide based parafn ammoxidation cat-
alysts are multiphase in nature, containing a parafn
activating phase (VSbO
4
) and an olen ammoxidation
phase (e.g., Sb
3
OSb
5
overlayers on top of
VSbO
4
; or an excess Sb containing solid solution of
a VSbO
x
; or FeSb
a
O
x
; SnSb
a
O
x
; TeSb
a
O
x
; etc.). The
matching of these phases still needs to be improved to
improve ammoxidation yields.
The Vmolybdate based catalysts are also multi-
phase in nature. Here too a VO
v
moiety of a VMo
x
O
y
phase is responsible for the propane activation and
propylene is the primary rst formed product. The
propylene reacts further on an olen ammoxidation
co-phase such as TeMo
x
O
y
or BiMo
x
O
y
to form acry-
lonitrile, as is customary of such phases [33,34]. Here
too, it is of great importance to bring these two
functionally diverse phases into concert. Further
matching is still needed to improve yields.
Future trends. The approach thus far has been rather
empirical in the search for propane ammoxidation
catalysts. The catalysts reported in the patent literature
contain elements with parafn activating properties
(e.g., V) as well as elements of olen activating
properties (e.g., Bi, Te), and O and/or NH-inserting
properties (e.g., Sb, Mo). What is missing is the
mechanistic cognoscence, that these functions should
not simply be tossed at random into a pot, but that a
rational approach needs to be worked out to sequence
these functionalities either in a single phase if possi-
ble, in a solid solution as the next best possibility, or
lastly by utilizing the two distinct functionalities in
two separate catalytic phases and bringing them into
spatial proximity, best through the choice of phases
having at least one epitaxially similar plane. The two
phases must be also temperature matched, that means
that their respective conversion efciencies must be
optimized to a common compromise operating tem-
perature. A third requirement is that the two phases be
chemically compatible with each other, so that neither
poisons the other. (For example, both are either anti-
monates or molybdates; it is generally unwise to mix
antimonates with molybdates and vice versa, because
they cross poison each other. There are a few excep-
tions to this general rule.) Rather than using two
physically distinct particles, it might be possible to
prepare the two phases separately in two pots, and
combining their ingredients in a paste state or at some
convenient intermediate step during the drying and/or
calcination.
A rational catalyst design along the lines outlined
should result in improved parafn ammoxidation
catalysts. It will be also necessary to keep the
concentration of the propylene intermediate low dur-
ing the process to prevent its destruction by the
parafn activating catalyst. This means, the parafn
activating catalyst must be surrounded by ample olen
Fig. 4. Kinetic reaction network for propane ammoxidation on V
SbAl oxide based catalysts [30]. Legend: ACN acrylonitrile;
AcCN acetonitrile.
146 R.K. Grasselli / Catalysis Today 49 (1999) 141153
ammoxidation catalyst (or its function) and/or the
efciency of the latter one must be increased at the
temperature of operation. Ultimately, it will be desir-
able to invent catalysts capable of activating the
parafn as well as the ammonia at lower temperatures,
resulting in higher useful product selectivities.
Another possible approach might be to generate
propyl radicals in the gas phase in a pre-reactor and
quenching them by an olen ammoxidation catalyst in
a subsequent reactor. Several studies are underway in
this direction to produce propylene and/or oxygenates
from propane. This approach could be adopted to the
propane to acrylonitrile reaction.
2.3. Oxydehydrogenation of paraffins
Conventionally practiced dehydrogenations of par-
afns, particularly those for the dehydrogenation of
propane and isobutane to the respective olens, are
commercially well established. They are endothermic
processes and are thermodynamically limited. Best
known are the Star (Phillips Petroleum), Caton
(Lummus), Oleex (UOP), Linde-BASF and Snam-
progetti-Yarsintez processes. The Star process uses a
supported Pt catalyst and a xed bed, the Caton a
chromiaalumina catalyst and a xed bed, the Oleex
a supported Pt catalyst and a moving bed, the Linde-
BASF a chromiaalumina catalyst and a xed bed, and
the Snamprogetti-Yarsintez a doped chromiaalumina
catalyst and a uid bed.
Among these, the Oleex, Caton and Snampro-
getti-Yarsintez processes are commercially practiced.
The greatest advantage of these processes is that they
are net producers of byproduct hydrogen. Their great-
est disadvantages are their thermodynamically limited
olen yields, high frequency of regeneration to burn
off coke from the catalyst surface, and the high energy
input requirements because of the inherent endother-
micity of the process.
For these reasons it would be desired to invent
an oxydehydrogenation process, which is inherently
not thermodynamically limited for the production of
olens from parafns. An illustrative example would
be the two electron conversion of propane to pro-
pylene:
CH
3
CH
2
CH
3

1
2
O
2
(air)
catalyst
CH
2
=CHCH
3
H
2
O
A large number of catalysts have been investigated
for this reaction, with the results being rather halting
thus far and no large breakthrough looming immedi-
ately on the horizon.
A large majority of the catalysts studied are based
on the chemistry of vanadium oxides. V
2
O
5
-based
catalysts give propylene yields between 8% and 20%
and are greatly dependent on the loading and the
support. The upper yield range is attained with
MgO supported catalysts [35], while SiO
2
, Al
2
O
3
,
TiO
2
, La
2
O
3
, Sm
2
O
3
or Bi
2
O
3
supported systems
are inferior [36]. Magnesium vanadates have been
also investigated [37,38] giving similar yields as
V
2
O
5
on MgO, Mg
2
V
2
O
7
is implicated as the active
phase in the former systems and might also play a role
in the latter systems. Recently, a study was reported on
the system MgV
x
Sb
y
O
z
[3941].
V
2
O
5
/Nb
2
O
5
catalysts have been also investigated,
with V
2
Nb
23
O
62
presumed to be the active phase
[42,43]. Propylene yields of 12% at a relatively high
selectivity of 70%, and at high productivity were
reported; no partially oxidized products were
observed. V-ZSM-5 has also been reported to yield
allegedly 25% propylene from propane, however, at
very low productivity.
The most noteworthy propane oxydehydrogenation
catalysts not based on vanadium chemistry are those
based on Comolybdate and Nimolybdate. CoMoO
4
/
SiO
2
variously doped with P, Ni, Mg, and/or Fe are
reported to yield 15% propylene [44] with high pro-
ductivity and no production of undesirable partial
oxidation byproducts. Unsupported Nimolybdates
were studied as monophases, with -NiMoO
4
yielding
13.2% propylene (20.9% conv., 63.1% sel.) and -
NiMoO
4
yielding 9.3% propylene (24% conv., 37.5%
sel.) [45,46]. In the latter systems, the -phase appears
to be the more selective and overall preferred catalytic
phase.
Recently, a comprehensive study of propane oxy-
dehydrogenation over silica supported molybdate
based catalysts of the formula AMoO
4
, where A=Ni,
Co, Mg, Mn, and Zn, has been reported [47]. The
reaction network and kinetics of a Ni
0.5
Co
0.5
MoO
4
/
SiO
2
system [48] reveal that propylene is the sole
primary product of the reaction, that based on isotope
studies the abstraction of a methylene hydrogen from
propane is the rate limiting step, and that the max-
imum yield obtained is 16% propylene (34% conv.,
R.K. Grasselli / Catalysis Today 49 (1999) 141153 147
47% sel.). These catalysts convert also effectively the
rst formed propylene product to acrolein (k
propylene
/
k
propane
=3.5), but even more efciently the acrolein on
to CO
x
(k
acrolein
/k
propylene
=13). For these reasons it is
proposed [49] that NiComolybdates are more likely
to nd utility as parafn activation catalysts in com-
bination with known selective olen conversion cat-
alysts for the direct production of oxygen (e.g., acrylic
acid) or nitrogen (e.g., acrylonitrile) containing pro-
ducts, rather than as oxydehydrogenation catalysts for
the production of olens (e.g., propylene). A sche-
matic representation of this concept is illustrated in
Fig. 5.
Future trends: VO
x
/MgO and MgVSbO
x
based
systems, as well as NiComolybdate systems hold
some promise as parafn oxydehydrogenation cata-
lysts. However, because in addition to their parafn
activating properties they also posses oxygen inserting
properties, their utility will be better harnessed by
combining them with established olen conversion
catalysts to produce unsaturated acids or nitriles.
Matching of operating temperature and chemical com-
patibility of these two catalyst systems needs still to be
optimized, as well as the desire to contain the two
distinct functionalities in a single solid state structure
or solid solution. Ideas along these lines have been
developed in Section 2.
Toachieveselectiveoxydehydrogenationof parafns
without the production of undesirable partial oxidation
products, it might be useful to turn to noble metal
catalysts, e.g., Pt based catalysts, and to carry out the
reactionat veryshort contact times. Someattemptsalong
these lines have already surfaced in [50].
2.4. Dehydrogenation with selective hydrogen
oxidation
An alternative to parafn oxydehydrogenation is a
compromise between conventional dehydrogenation
and oxydehydrogenation; it is dehydrogenation with
concomitant selective hydrogen oxidation, illustrated
for propane as follows:
CH
3
CH
2
CH
3

catalyst-A
CH
2
=CHCH
3
H
2
H
2

1
2
O
2

catalyst-B
H
2
O
A similar concept has been worked out for the
ethylbenzene conversion to styrene by UOP in their
Smart process.
Fig. 5. Proposed mechanism of propane ammoxidation using a paraffin activating catalyst, e.g., Ni
a
Co
b
M
c
MoM
x
; and a multicomponent
mixed metal molybdate olefin ammoxidation catalyst, e.g., Cs
a
K
b
Ni
c
Mg
d
M
e
Bi
f
Sb
g
Mo
h
O
x
; where M=Ce, Cr, and/or Fe [4749].
148 R.K. Grasselli / Catalysis Today 49 (1999) 141153
Suitable catalysts for the rst step (catalyst-A) are
those selected from Group VIII noble metals, e.g., Pt;
while those effective for the second step (catalyst-B)
are those selected from selectively poisoned Group
VIII metals, e.g., Cs
x
Pt or Bi
x
Pt, where x is a small
fraction of one. Such catalysts are generally supported
on Al
2
O
3
, sometimes additionally doped with Sn, to
make the catalyst steam stable and claimed to have a
95% hydrogen combustion selectivity to H
2
O
[51,52].
More recently, base metal catalysts containing ele-
ments selected from the metals Bi, In, W, Mo, etc.
have been reported to be effective for the selective
combustion of hydrogen (9099% selectivity), with-
out attacking and combusting the produced olen [53
56].
High yields of olens, well in excess of equilibrium,
are attainable by staging dehydrogenation/selective
hydrogen combustion/dehydrogenation reactors; or
zones. The oxygen driven combustion of hydrogen
overcomes the conventional dehydrogenation limita-
tion. This process has also appeal by the possibility of
a perfect heat balance of the overall process (essen-
tially isothermal).
Future trends. The dehydrogenation/selective hy-
drogen combustion process has potential until an
effective oxydehydrogenation process is discovered.
Engineering innovations are desired to optimize heat
balancing (sheet and tube reactors are possibilities).
Base metal oxides commingled with dehydrogenation
catalysts or functionalities therefrom, wherein the
oxygen for the hydrogen combustion is carried out
by the metal oxide as lattice oxygen, would be desir-
able innovations. No oxygen plant would be required
with such a catalyst system, resulting in substantial
savings. The catalyst system would be simply regen-
erated periodically with air.
2.5. Oxydehydrogenation of n-butane to maleic
anhydride
The commercially most successful parafn oxida-
tion process is the 14-electron oxidation of n-butane to
maleic anhydride:
CH
3
CH
2
CH
2
CH
3
7a2O
2
(air)

catalyst
C
4
H
2
O
3
4H
2
O
Several companies, e.g., BP-UCB, ALMA,
DuPontMonsanto, Misubishi, Denka-Scientic
Design, have commercialized the process. Interest-
ingly, all of the companies use essentially the same
catalyst, vanadyl pyrophosphate (VO)
2
P
2
O
7
, which
was originally invented by Chevron [57]. Various
improvements are claimed through selective doping,
however, the improvements appear minor compared to
the effectiveness of the base catalyst.
The catalyst is most frequently used unsupported
even in uid bed reactors, because virtually all sup-
ports studied lower the selectivity of the catalyst. An
elaborate calcination process is required to attain
sufcient hardness and attrition resistance of the
unsupported catalyst. The only exception appears to
be the silica (di-silicic acid) coated system developed
by DuPont.
Reactors used for the process are xed bed, uid
bed and riser reactors. While the latter, used by
DuPont, is claimed to give the highest maleic anhy-
dride selectivities, the complexity of the reactor and
the large amount of solids handling are less than
desirable in practice. For these reasons, the ease of
operation of a simple uid bed makes it the preferred
choice of industry.
The solid state structure of (VO)
2
P
2
O
7
and its redox
mechanism in n-butane oxidation have been studied
extensively by many researchers in the past 10 years
[5862]. Recently, a molecular level mechanism [63]
has been worked out which stresses the site isolation
(an important concept developed some 35 years ago
for achieving selectivity in oxidation catalysts [64]).
Site isolation is also innate to the VPO structure,
where four VOV dimers form domains, isolated
structurally from other domains by phosphate groups,
so that each domain can operate independently from
neighboring domains, is itself self-sufcient to pro-
duce one molecule of maleic anhydride, before it
needs to be regenerated, and the process can start
all over again (Fig. 6). These dimeric VOVsites can
assume interchangeably four different electronic con-
gurations, wherein the formal valance of the vana-
dium can assume a 5

, 4

and/or 3

state. The relative

ratios of the four possible electronic states are dictated
by the feed and product ratios of reductants and
oxidants, and the reaction conditions. The proposed
mechanism constitutes a good basis for contemplating
future catalyst improvements.
R.K. Grasselli / Catalysis Today 49 (1999) 141153 149
Future trends. It appears that VPOs unique structure
and redox properties make it a difcult system to be
displaced by a newer and more efcient system.
Selective doping of the parent compound is in vogue
and might lead to marginal improvements. A radically
different approach is to search for alternative struc-
tures which might yield to rational catalyst design.
Such an approach has been taken recently and some
encouraging preliminary results have materialized
therefrom [6567]. This eld has just begun to attract
interest and is deemed to have excellent potential for
innovative catalyst design, leading ultimately to com-
mercially viable catalytic processes.
2.6. Oxidation of propane to acrylic acid
Many attempts have been made to convert propane
directly to acrylic acid. The catalysts developed until
nowfall far short of the requirements needed to build a
commercial plant. There is no competition to the well-
established two-stage xed bed process to convert
propylene to acrolein, and further, without separation
in a second catalytic stage to acrylic acid.
Catalysts studied for the direct conversion of pro-
pane to acrylic acid include VP oxides [68], MnP
oxides [69], BiVMo oxides [70], BiVNbSbMo
oxides [71]. None of these catalysts give acrylic acid
yields in excess of about 10%. An exception appears to
be the Mitsubishi reported catalyst of the empirical
composition MoV
0.3
Te
0.23
Nb
0.12
O
x
[72], which is
claimed by the patentees to yield 48% acrylic acid
(80.1% conv., 60.5% sel.). Unfortunately, several
independent laboratories in the US and in Europe
have had difculties duplicating the results claimed
by Mitsubishi, primarily because of the difculties
encountered with the exact selection of catalyst start-
ing materials and its involved preparation.
Future trends. If the claims of Mitsubishi can be
veried, the yields are impressive and would serve as a
basis for an eventual commercial process to convert
propane to acrylic acid. In the absence of such a
verication the eld is wide open. Some claims are
being made that useful leads towards this goal are
being made using heteropoly acids and/or salts thereof
as catalysts, as are specically doped VPO catalysts.
No doubt the incentive exists to discover a catalyst or a
combination of catalysts for the conversion of propane
to acrylic acid. The driving force is the much lower
price of propane and its abundance, the same scenario
as that for the ammoxidation of propane to acryloni-
trile. As an aside, the ambient temperature functiona-
lization of propane to isopropanol using Cu
chlorophthalocyanines encapsulated in X and Y zeo-
lites [46] shows promise to attack parafns in a
selective manner at low temperatures and might be
a lead for further pursuit.
2.7. Oxidation of i-butane to methacrylic acid
The direct oxidation of i-butane to methacrylic
acid is a desirable goal. Thus far, no suitable catalysts
have been discovered, as is the case with propane
oxidation.
The majority of the studied catalysts are based on
heteropoly acids or salts thereof [7375], e.g.,
H
m
X
0X51X5
Y
0X21X5
Z
03
P
11X2
Mo
12
O
n
where X = V, As, Cu, Y =alkali, Z = Sb, Sn, Group
VIII elements.
It is presumed that heteropoly compounds having
intact Keggin structures are the active components of
these catalysts. The yields of acrylic acid obtained
with these catalysts lie below 10%, hence well below
commercially attractive levels.
Future trends. The eld is wide open.
Fig. 6. Schematic representation of (VO)
2
P
2
O
7
surface structure.
The arrows represent the possible pathways for facile exchange of
surface bound oxygen, either monoatomic or diatomic, between the
active sites. The ``site isolation'' [64] due to the diffusion barrier
created by the pyrophosphate group is clearly shown by these
arrows [63].
150 R.K. Grasselli / Catalysis Today 49 (1999) 141153
2.8. (Amm)oxidation of xylenes
In the well-established area of selective o-xylene
oxidation to phthalic anhydride, catalyst research
continues along both compositional and preparational
lines. The prevailing compositions are still based on V
oxides supported on TiO
2
(anatase), but are strongly
doped, particularly with elements such as Sb, Nb, and
Cs. As an illustrative example, Nippon Shokubai [76]
discloses among others compositions of the following
empirical formula: VSb
0.11
Nb
0.050.5
P
00.2
Cs
00.2
O
x
/
TiO
2
, with active phase loadings in the vicinity of a
monolayer. Improved selectivities, longer catalyst life
and catalyst ruggedness against plant upsets are
claimed. From the composition one can surmise that
VSbO
4
is formed as one of the phases and hence
modies the base V
2
O
5
/TiO
2
composition. The mod-
ication has both electronic, as well as, structural
consequences. Electronically, the addition of Sb low-
ers the oxidation state of vanadium; structurally, it
interposes itself between VOV chains and thus
helps to isolate VO moieties from each other (site
isolation at work). When excess antimony is present,
the presence of supra-surface antimony sites [29] can
be postulated on top of the V
2
O
5
/TiO
2
or VSbO
4
/TiO
2
active phase. The result is a more selective
catalyst.
These solid state dopings and their catalytic con-
sequences are another manifestation of the importance
of the concept of site isolation in selective (amm)ox-
idation catalysis. This concept, rst conceived in the
early 1960s [64], continues to pervade modern
approaches to catalyst design and their improvement
in efciency. It is amply practiced in industrial as well
as in fundamental studies of selective catalytic
(amm)oxidation.
Recent studies indicate that p-xylene can be pre-
ferentially oxidized from a mixture of xylenes to
terephthalic acid (or precursors thereof), or ammox-
idized to terephthalic nitrile (or its precursors) using
ZSM-5-based catalysts.
Amoco [77,78] reports preferential selective oxida-
tion of p-xylene over o- and m-isomers, using an Fe
Mo oxide modied, partially deborinated B-ZSM-5.
Yields are reasonably good.
German researchers [79] report that p-xylene can be
ammoxidized with large preference over the other two
isomers to terephthalonitrile, using a Cu-ZSM-5.
Both results constitute good leads for future
research.
Future trends. Catalyst compositions capable of
(amm)oxidizing preferentially p-xylene over the other
two isomers will be hotly pursued. Current results are
not sufcient for commercialization. Also, materials
other than ZSM-5, materials which are more steam
stable, will have to be utilized as the basis to achieve
preferential isomer selectivity. The utilization of uid
beds has never gained favor in xylene oxidation
although there is actually no fundamental reason to
stay away from them. Perhaps, some clever new
engineering innovation in uid beds might nd applic-
ability.
3. Summary
An overview of the current status of several com-
mercially important heterogeneous catalytic (amm)-
oxidation processes reveals that catalysts for these
reactions are rather complex in chemical composition
and physical makeup. It reveals further that the under-
standing of their functions is being ever better under-
stood, and because of this better understanding, the
catalytic systems are becoming even more complex in
nature as they also become more efcient. Future
trends aim greatly towards the discovery and devel-
opment of parafn activation catalysts and their use
towards producing useful functionalized chemical
intermediates. Catalysts towards this end are com-
posed of parafn activating elements contained in a
common structure with selective olen oxidation com-
ponents, or they are composed of two distinct phases,
best situated in intimate proximity of each other
(preferably epitaxially matched), or as a worst case
scenario in two physically separate phases; again best
when in proximity of each other. Site isolation pays an
important role within both of the respective phases,
their mutual cooperation is imperative, and their rela-
tive placement to each other and/or respective spatial
sequencing is paramount to high catalytic overall
efciency; and a subject matter which needs further
study. The application of combinatorial chemistry also
shows promise.
(Amm)oxidation catalysis of light hydrocarbons
remains a lucrative eld of study and continues to
hold exceptionally good future commercial rewards in
R.K. Grasselli / Catalysis Today 49 (1999) 141153 151
the conversion of olens, aromatics, and in particular
parafns to useful functionalized end products.
Acknowledgements
This paper was written during the author's v. Hum-
boldt Prize stay at the University of Munich, Germany.
References
[1] R.K. Grasselli, J. Chem. Ed. 63 (1986) 216 and references
therein.
[2] R.K. Grasselli, in: G. Ertl, H. Knoezinger, J. Weitkamp
(Eds.), Handbook of Heterogeneous Catalysis, 4.6.6. Am-
moxidation, WileyVCH, 1997, p. 2302, and references
therein.
[3] R.K. Grasselli, H.F. Hardman, US Patent 3 642 930 (1972).
[4] R.K. Grasselli, H.F. Hardman, US Patent 4 503 001 (1985).
[5] R.K. Grasselli, D.D. Suresh, H.F. Hardman, US Patent
4 139 552 (1979).
[6] D.D. Suresh, M.S. Friedrich, M.J. Seely, US Patent 5 212 137
(1993).
[7] J.F. Brazdil, M. Mehicic, L.C. Glaeser, M.A.S. Hazle, R.K.
Grasselli, in: M.L. Deviney, J.L. Gland (Eds.), ACS Symp.
Ser. 288 (1985) 26.
[8] R.K. Grasselli, Appl. Catal. 15 (1985) 127.
[9] R.K. Grasselli, J.L. Callahan, J. Catal. 14 (1969) 93.
[10] R.K. Grasselli, D.D. Suresh, J. Catal. 25 (1972) 273.
[11] F.G. Freire, J.M. Herrmann, M.F. Portela, in: V.C. Corberan,
S.V. Bellon (Eds.), Proceedings of the New Dev. Sel. Oxid. II,
Benalmadena, Spain, 1993, p. 31.
[12] Y. Sasaki, T. Nakamura, Y. Nakamura, K. Moriya, H. Utsumi,
S. Saito, US Patent 4 370 279 (1983).
[13] R.K. Grasselli, G. Centi, F. Trifiro, Appl. Catal. 57 (1990)
149.
[14] J.L. Callahan, E.C. Milberger, R.K. Grasselli, US Patent
3 427 343 (1969).
[15] European Chemical News, vol. 28, 6 November 1989.
[16] Chem. Eng. News, vol. 13, 29 May 1995.
[17] A.T. Guttmann, R.K. Grasselli, J.F. Brazdil, D.D. Suresh, US
Patent 4 797 381 (1989).
[18] A.T. Guttmann, R.K. Grasselli, J.F. Brazdil, D.D. Suresh, US
Patent 4 788 317 (1988).
[19] J.P. Bartek, A.T. Guttmann, US Patent 4 797 381 (1998).
[20] G. Blanchard, G. Ferre, US Patent 5 336 804 (1994).
[21] S. Albonetti, G. Blanchard, P. Burattin, S. Masetti, F. Trifiro,
in: R.K. Grasselli et al. (Eds.), Proceedings of the Third
World World Congress on Oxidation and Catalysis, Stud.
Surf. Sci. Catal. San Diego, 110 (1997) 403.
[22] D.D. Suresh, J.F. Brazdil, L.C. Glaeser, US Patent 4 760 159
(1988).
[23] Y.C. Kim, W. Ueda, Y. Moro-oka, Catal. Today 13 (1992)
673.
[24] J.S. Kim, S.I. Woo, Appl. Catal. A 110 (1994) 207.
[25] K. Oshama, A. Kayo, T.K. Umezawa, K. Kiyono, I. Sawaki,
EP 529 853 (1992).
[26] K. Wada (Mitsubishi Chem. Co.), Private Communication,
GRC-Chemistry of Hydrocarbon Resources, Hawaii, 1994.
[27] M. Vaarkamp (Mitsubishi Chem. Co.), Private Communica-
tion, Europacat-II, Maastricht, 1995.
[28] E.M. Thorsteinson, T.P. Wilson, F.G. Young, P.H. Kasai, J.
Catal. 52 (1978) 116.
[29] R. Nilsson, T. Linblad, A. Andersson, J. Catal. 148 (1994)
501.
[30] R. Catani, G. Centi, F. Trifiro, R.K. Grasselli, Ind. Eng.
Chem. Res. 31 (1992) 107.
[31] R.K. Grasselli, in: J. Bonelle et al. (Eds.), Surf. Prop. Catal.
Non-Metals, Reidel, Dordrecht, 1983, p. 273.
[32] A Andersson, S.L.T. Andersson, G. Centi, R.K. Grasselli, M.
Sanati, F. Trifiro, in: L. Guczi et al. (Eds.), Proceedings of the
10th International Congress on Catalysis A, Budapest, 1992,
p. 691.
[33] J.D. Burrington, C.T. Kartisak, R.K. Grasselli, J. Catal. 87
(1984) 363.
[34] R.K. Grasselli, J.D. Burrington, Adv. Catal. 30 (1981) 133.
[35] M.A. Chaar, C. Patel, H.H. Kung, J. Catal. 109 (1988)
463.
[36] A. Corma, J.M. Lopez-Nieto, N. Paredes, M.J. Rajadell, M.
Perez, Proceedings of the XIII Symposium on Iberoamerican
Catalysis, Segovia, Spain, vol. 75, 1992.
[37] D.S.H. Sam, V. Soenen, J.C. Volta, J. Catal. 123 (1990) 417.
[38] M.C. Kung, H.H. Kung, J. Catal. 134 (1992) 668.
[39] J.N. Michaels, D.L. Stern, R.K. Grasselli, Catal. Lett. 42
(1996) 135.
[40] J.N. Michaels, D.L. Stern, R.K. Grasselli, Catal. Lett. 42
(1996) 139.
[41] D.L. Stern, J.M. Michaels, L. DeCaul, R.K. Grasselli, Appl.
Catal. A 153 (1997) 21.
[42] R.H.H. Smits, K. Seshan, J.R.H. Ross, in: P. Ruiz, B. Delmon
(Eds.), New Dev. Sel. Oxid. Het. Catal. 72 (1992) 221.
[43] R.H.H. Smits, K. Seshan, J.R.H. Ross, in: P. Ruiz, B. Delmon
(Eds.), Proceedings of the Symp. Catal. Sel. Oxid., ACS
Preprints, Washington, DC, 1992, p. 1121.
[44] H.F. Hardman, US Patent 4 255 284 (1981).
[45] C. Mazzochia, C. Aboumard, C. Daigne, E. Tempesti, J.M.
Herrmann, G. Thomas, Catal. Lett. 10 (1991) 181.
[46] C. Mazzochia, E. Tempesti, C. Aboumard, US Patent
5 086 032 (1992).
[47] D.L. Stern, R.K. Grasselli, J. Catal. 167 (1997) 550.
[48] D.L. Stern, R.K. Grasselli, J. Catal. 167 (1997) 560.
[49] D.L. Stern, R.K. Grasselli, in: R.K. Grasselli, S.T. Oyama,
A.M. Gaffney, J.E. Lyons (Eds.), Proceedings of the Third
World Congress on Oxidation and Catalysis, San Diego, Stud.
Surf. Sci. Catal. 110 (1997) 357.
[50] L.N. Schmidt, C.T. Goralski, in: R.K. Grasselli, S.T. Oyama,
A.M. Gaffney, J.E. Lyons (Eds.), Proceedings of the Third
World Congress on Oxidation and Catalysis, San Diego, Stud.
Surf. Sci. Catal. 110 (1997) 491.
[51] T. Imai, D.-Y. Jan, US Patent 4 788 371 (1988).
[52] T. Imai, R. J. Schmidt, US Patent 4 886 928 (1989).
152 R.K. Grasselli / Catalysis Today 49 (1999) 141153
[53] P.A. Agaskar, R.K. Grasselli, J.M. Michaels, P.T. Reischman,
D.L. Stern, J.G. Tsikoyiannis, US Patent 5 430 209 (1995).
[54] P.A. Agaskar, R.K. Grasselli, J.M. Michaels, P.T. Reischman,
D.L. Stern, J.G. Tsikoyiannis, US Patent 5 527 979 (1996).
[55] P.A. Agaskar, R.K. Grasselli, J.M. Michaels, P.T. Reischman,
D.L. Stern, J.G. Tsikoyiannis, US Patent 5 530 171 (1996).
[56] P.A. Agaskar, R.K. Grasselli, J.M. Michaels, P.T. Reischman,
D.L. Stern, J.G. Tsikoyiannis, US Patent 5 563 314 (1996).
[57] K. Katsumoto, D.M. Marquis, US Patent 4 132 670 (1979).
[58] G. Centi, F. Trifiro, J. Ebner, V. Franchetti, Chem. Rev. 88
(1988) 5.
[59] J. Ziolkowski, E. Bordes, P. Courtine, J. Catal. 122 (1990)
126.
[60] G. Centi, Catal. Today 16 (1993) 5.
[61] D. Dowell, J.T. Gleaves, Y. Schuurman, in: R.K. Grasselli et
al. (Eds.), Proceedings of the Third World Congress on
Oxidation and Catalysis, San Diego, Stud. Surf. Sci. Catal.
110 (1997) 199.
[62] G.J. Graham, A. Burrows, S. Sajip, C.J. Kiely, K.E. Bere, J.-
C. Volta, A. Tuel, M. Abon, in: R.K. Grasselli, S.T. Oyama,
A.M. Gaffney, J.E. Lyons (Eds.), Proceedings of the Third
World Congress on Oxidation and Catalysis, San Diego, Stud.
Surf. Sci. Catal. 110 (1997) 209.
[63] P.A. Agaskar, L. DeCaul, R.K. Grasselli, Catal. Lett. (1994)
339.
[64] J.L. Callahan, R.K. Grasselli, AIChE J. 9 (1963) 755.
[65] P.A. Agaskar, R.K. Grasselli, US Patent 5 354 722 (1994).
[66] P.A. Agaskar, R.K. Grasselli, US Patent 5 550 308 (1996).
[67] P.A. Agaskar, R.K. Grasselli, D.J. Buttrey, B. White, in: R.K.
Grasselli et al. (Eds.), Proceedings of the Third World
Congress on Oxidation and Catalysis, San Diego, Stud. Surf.
Sci. Catal. 110 (1997) 219.
[68] M. Ai, Catal. Today 12 (1992) 679.
[69] Y. Takita, H. Yamashita, K. Moritaka, Chem. Lett. (1989)
1733.
[70] Y.-C. Kim, M. Udea, Y. Moro-oka, Catal. Today 13 (1992)
673.
[71] J.P. Bartek, A.M. Ebner, J.F. Brazdil, US Patent 5 198 580
(1993).
[72] Mitsubishi Chem. Co, EP 0608838A2 (1994).
[73] S. Yamamatsu, T. Yamaguchi, EP 425 666 (1998).
[74] K. Nagai, Y. Nagaoka, H. Sato, M. Ohsu, EP 418 657
(1990).
[75] T. Kuroda, M. Okita, JP 04 128 247 (1991).
[76] Y. Nakanishi, Y. Akazana, N. Ikeda, H. Himeji, T. Suzuki, S.
Ohtsu, DE 3 045 624 (1985).
[77] J.S. Yoo, D.A. Donohue, M.S. Kleefisch, P.S. Lin, S.D.
Elfline, Appl. Catal. A 105 (1993) 83.
[78] J.S. Yoo, P.S. Lin, S.D. Elfline, Appl. Catal. A 106 (1993)
259.
[79] K. Beschmann, L. Riekart, Chem. Ing. Technol. 65 (1993)
1231.
R.K. Grasselli / Catalysis Today 49 (1999) 141153 153