Hood 2016

Chapter 1
Progress in Maleic Anhydride Production
David K. Hood and Osama M. Musa
1.1 Introduction
Maleic anhydride (furan-2,5-dione) is a key component to an incredibly diverse group of

chemistries and commercial applications. The ability of maleic anhydride to achieve such
great diversity is a direct result of the maleic anhydride’s industrial complex capacity to
develop economical production routes in the face of ever-increasing governmental
scrutiny, fierce pricing pressures, and tremendous technical requirements (at significant
volumes) coupled to high demand for quality fundamental product properties.
What is it about maleic anhydride that creates such demand, utility, and value?
The two critical success factors are chemistry and economics. Both are truly
connected, intertwined, and inseparable. First, the basic chemical structure of
maleic anhydride is inherently versatile. Maleic anhydride, based upon a five-
member heterocyclic ring, comprises a double bond at the C3–C4 position and
two carbonyl groups, one at the C2 position and the other at the C5 position (see
Scheme 1.1, Left). The double bond is well known to be very reactive. Maleic
anhydride is a powerful electron-accepting monomer due to the electron-deficient
character of double bond. The electronic deficiency originates from the two C ¼ O
substituent group’s strong electron-withdrawing forces.
Upon exposure to water, maleic anhydride can hydrolyze to yield a dicarboxylic
acid structure known as maleic acid. The chemical structure of maleic acid is
presented in Scheme 1.1 (Right). In addition to the reactive double bond, these
carboxylic acid groups are also well known for high reactivity coupled to two
distinct acid dissociation constants. Dehydration of maleic acid enables the refor-
mation of maleic anhydride. Key physical properties for these two chemical
compounds are presented in Table 1.1.
D.K. Hood (*) • O.M. Musa

Ashland Inc., 1005 Route 202/206, Bridgewater, NJ 08807, USA
e-mail: [email protected]
© Springer International Publishing Switzerland 2016 3

O.M. Musa (ed.), Handbook of Maleic Anhydride Based Materials,
DOI 10.1007/978-3-319-29454-4_1
4 D.K. Hood and O.M. Musa
Scheme 1.1 Maleic anhydride and maleic acid, respectively
Table 1.1 Key physical properties for maleic anhydride and maleic acid
Property Maleic anhydride Maleic acid
CAS number 108–31–6 110–16–7
EINECS number 203–571–6 203–742–5
Molecular formula C4H2O3 C4H4O4
Molecular weight (g/mol) 98.06 116.07
Physical state Solid Solid
Color Colorless to white [1] White [2]
Odor Irritating, choking [1] Odorless [2]
Density 1.48 [3] 1.59 [3]
Melting point ( C) 53.58 [1] 132.5 at 101.3 kPa [2]
Boiling point ( C) 200.1 [1] 157.8 at 99.7 kPa [2]
Vapor pressure 0.033 kPa [1] 9 107 kPa at 20 C [2]
Refractive index [d2020 (solid)] NA 1.590 [4]
Heat of formation (kJ/mol) 470.4 [5] 788.3 [6]
Heat of fusion (kJ/mol) 12.26 [7] 26.9 [7]
Heat of sublimation (kJ/mol) 71.5 5.0 [7] 105.4 1.7 [7]
Heat of combustion (kJ/mol) 1391.2 [6] 1358.9 [6]
Specific heat (liquid) (kJ mol1 K1) 1.67 [6] NA
Heat of evaporation (kJ/mol) 54.8 [6] NA
Solubility in water ~400 g/L at 20 C [1] 478.8 g/L at 20 C [2]
pKa NA pK1 ¼ 1.910; pK2 ¼ 6.332 [8]
Biodegradation Readily biodegradable [1] Readily biodegradable [2]
Within these two chemical compounds are a wide variety of possible chemical
transformations. From free radical and charge transfer polymerization processes to
Diels–Alder, esterification, and amidation reactions, these thematic transformations
form the basis of the tremendously flexible synthetic platform provided to the
myriad of chemical technology implementers.
Second, the economics of maleic anhydride are particularly attractive. Maleic
anhydride is largely derived from butane gas, especially in North America, but can
also be produced from benzene. Both production routes remain the most relevant
feedstocks in global product supply. Butane, C4H10, is a colorless, odorless gas that
is normally shipped under pressure in the form of a liquefied gas. Common
synonyms for butane are n-butane, butyl hydride, diethyl, liquefied petroleum
1 Progress in Maleic Anhydride Production 5
Table 1.2 Examples of compositional variation of natural gas with geographical location [11]
Example 1 Example 2 Example 3
State Colorado New Jersey New Mexico
County Cheyenne Offshore Chaves
Field Cheyenne Wells Mid-Atlantic Buffalo Valley
Methane (mole %) 9.2 94.1 89.1
Ethane (mole %) 3.0 2.0 6.6
Propane (mole %) 25.5 0.8 2.2
n-Butane (mole %) 20.9 0.1 0.7
Isobutane (mole %) 10.2 Trace 0.1
n-Pentane (mole %) 3.7 0.0 0.2
Isopentane (mole %) 9.4 0.1 0.3
Cyclopentane (mole %) 1.1 Trace 0.1
Hexanes plus (mole %) 2.6 Trace 0.1
Nitrogen (mole %) 10.0 1.1 0.4
Oxygen (mole %) 0.1 0.0 0.0
Argon (mole %) 0.1 Trace Trace
Hydrogen (mole %) 0.0 0.2 0.0
Hydrogen sulfide (mole %) 1.8 0.0 0.0
Carbon dioxide (mole %) 2.4 1.7 0.2
Helium (mole %) 0.12 0.02 0.04
Heating value 2605 1018 1135
Specific gravity 1.713 0.596 0.639
gas, and methylethylmethane [9]. This chapter will employ the term “C4” to mean
butane. C4 is produced by refining materials that originate from either crude oil or
natural gas. Natural gas is often found in geological shale deposits and tight sands.
Recent advancements in horizontal drilling and well-completion technologies,
particularly in low-permeability formations, have liberated tremendous reservoirs
of supply in large sections of the North American continent [10]. Geographic
regions spanning from Western Canada to Appalachia and the Gulf Coast are
producing significant quantities of natural gas [10]. Natural gas as such is com-
prised of regionally varied amounts of gases, which is often, but not always,
dominated by methane gas. Examples of compositional variations are presented
in Table 1.2.
With existing pipeline infrastructures, very large volumes of natural gas as well
as discrete, refined gases, such as C4, can be transported to key downstream
facilities. C4 is commonly stored in underground caverns, such as the 1MM barrel
(bbl) cavern recently constructed near Neal, West Virginia [12]. More modest
volumes, ranging from 800 bbl to 1600 bbl, are stored in aboveground pressurized
vessels [13]. Thus, economies of global scale and positive supply chain dynamics
exist in significant ways for continued maleic anhydride growth potentials.
In 1982, Trivedi and Culbertson’s Maleic Anhydride published a broad overview
of maleic anhydride production technology in a chapter entitled Production of
Maleic Anhydride [14]. The processes detailed in this chapter spanned the three
main production routes: benzene, C4, and phthalic anhydride. Benzene routes
designed by Scientific Design, Ruhrol-Lurgi, Ruhrol-Bayer, and Sava were
discussed in detail. C4 routes by Mitsubishi and BASF were previewed followed
by a brief discussion of phthalic anhydride by-product sourcing. In 2001, Maleic
Anhydride, Maleic Acid, and Fumaric Acid was published in the Kirk-Othmer
Encyclopedia of Chemical Technology [15]. C4 routes by Huntsman (formerly
Monsanto), ALMA, and DuPont were discussed in detail. This author is fortunate
to use these chapters as a backdrop to efficiently analyze more recent technical
trends and accomplishments that have empowered the maleic anhydride product, as
it nears its ninth decade in commercial production. It is hoped that the reader will
benefit from the wide-ranging discussion, enabling a broader appreciation for the
scope of technologies, challenges, and current solution strategies that exist in
current maleic anhydride production.
1.2 Historical Overview
Historically, the two main commercial routes to maleic anhydride production are by
the thermal oxidation of benzene or C4 feedstocks via heterogeneous catalysis. In
1982, it was reported that the benzene route was favored due to the lack of efficient
catalysts and purification processes for suitable C4 process implementation. How-
ever, it was also noted that increasing environmental regulations on the benzene
route were expected to lead to significant research efforts in the field in order to
develop new, more environmentally sustainable processes. In a time frame of about
15 years, the C4 route evolved into the preferred production process, certainly
within the USA. Today, C4 is the dominant commercial route for maleic anhydride
production and is recognized for its superior economics [16].
Stoichiometry favors the C4 route, where 100 kg of C4 should yield 168.9 kg of
maleic anhydride. Conversely, 100 kg of benzene is only expected to yield 125.6 kg
of maleic anhydride [17]. During the C4 oxidation process, there are a number of
less desirable but significant side products: butadiene, furan, crotonaldehyde, and
C4 isomers. Principally, the main products are maleic anhydride [Eqs. (1.2) and
(1.3)] and unstable intermediates (Eq. 1.1). The reaction schemes for various
oxidation process steps are presented below [18]:
ð12 mÞ
C4 H8 þ O2 ! mCO þ ð4 mÞCO2 þ 4H2 O ð1:1Þ
2
C4 H8 þ 3O2 ! C4 H2 O3 þ 3H2 O ð1:2Þ
ð6 nÞ
C4 H2 O3 þ O2 ! nCO þ ð4 nÞCO2 þ H2 O ð1:3Þ
2
1
CO þ O2 ! CO2 ð1:4Þ
2
1.2.1 Early Maleic Anhydride (via C4) Production Examples
The literature is imprecise as to when the first maleic anhydride (via C4) production
process was fully implemented. Some point to 1975, where Amoco deployed a
Chem Systems process [19]. However, other literatures present published examples
prior to 1975. For example, Mitsubishi Chemical Industries announced a fluid-bed
reactor capable of 18,000 tons/year operating in Mizushima, Japan, as of 1970
[20]. BASF announced a fixed-bed reactor plant capable of 3000 tons/year operat-
ing in Ludwigshafen, Germany, in May 1968 [21]. The earliest process was likely
Petro-Tex in a Scientific Design-engineered plant in 1962 (near Houston, Texas)
[22]. Reportedly, this effort struggled in the early years due to competition from
cheaper benzene and was eventually stopped in favor of the overall benzene
economics [22]. The earliest plants typically state a capability to interchange
feedstock between benzene and C4 enabling economic forces to balance the basic
plant operating conditions.
Because of the basic interchangeability of feedstock, the key production tech-
nologies comprise very similar design concepts and equipment schemes. The two
main production processes employ either fixed-bed or fluid-bed oxidative designs.
Every major maleic anhydride production process is equipped with a reactor, gas
cooler, separator, scrubber, refiner/condenser, and product column collector. To
better understand the role of each step, we will first examine a Halcon–Scientific
Design (HSD) process.
1.2.1.1 Halcon–Scientific Design (HSD) Fixed-Bed Process [23]
Early maleic anhydride production processes were benzene based. The existing
capital equipment was then converted to C4-based processes. A prime example of a
convertible feedstock plant is the Halcon–Scientific Design (HSD) process for
maleic anhydride manufacture. A schematic of this process is presented in
Scheme 1.2.
Interpreting the HSD process, from left to right, either benzene or C4 is fed into a
compressed air line to be presented into a carbon steel reactor comprised of
numerous tubes nominally 3.7 m long and 25 mm in diameter [25]. In this case,
the process is fixed bed, where the C4 and air are mixed, normally 1.0 to 2.0 mol
percent of C4, prior to introduction to the reactor, which houses the catalyst. The
reactor is often positioned in an upflow configuration. The pressure in the operating
reactor, ranging from 20 to 50 psi, [26, 27] is close to atmospheric requiring only
enough pressure to move the gas through the catalyst bed [1800–2800 gas hourly
8
Scheme 1.2 Halcon–Scientific Design (HSD) process for maleic anhydride [24] (Adapted from Environmental Progress, Malow, M., Vol. 4, No. 3, Maleic
Anhydride via Butane Oxidation Substitution of n-butane for benzene in the production of maleic anhydride has led to the solution of a serious environmental
problem, 151–154, Copyright August 1985 with permission from John Wiley and Sons)
D.K. Hood and O.M. Musa
space velocity (GHSV)] [28]. The catalyst is housed in tubes, often numbering in
thousands, and surrounded by a heat transfer fluid system. Typically, the catalyst
tubes are first packed with an inert material, i.e., metal or ceramic balls, to preheat
the gas flow prior to exposure to the catalyst [28]. The reactor heat and heat
exchanger are supplied by molten salts. The temperature range of these salts is
typically between 390 and 430 C. These eutectic salt mixtures comprise salts
such as potassium nitrate, sodium nitrate, and sodium nitrite (i.e., HITEC®)
[25]. Many processes recycle this heat in the form of steam generation for addi-
tional energy supply. Upon exiting the reactor, the process stream enters a gas
cooler, often fabricated from carbon steel [25]. Partial cooling is applied to the
process stream in order to reduce its temperature to just below maleic anhydride
dew point (53–60 C). The cooled materials then enter a separator, constructed of
stainless steel, enabling a crude separation between the partially condensed maleic
anhydride and other gases and uncondensed vapors. The condensed maleic anhy-
dride is destined for the still pot. It is critical to monitor the temperature at this
stage. If the maleic anhydride solidifies, the process can experience blockages
coupled to pressure and temperature challenges. The gas flow continues onto a
scrubber unit, which is constructed of a grade of stainless steel that includes
molybdenum. The gas flow is contacted with a countercurrent of water to absorb
remaining maleic anhydride in the hydrolyzed form of maleic acid [29]. Remaining,
unabsorbed overhead gases are exhausted either to the atmosphere or an incinerator.
Maleic acid is collected and later pumped to the refiner for further purification
processing. In the stainless steel refiner/condenser batch unit, maleic acid solution
is first dehydrated (azeotropically) with xylenes or toluene. This remaining solution
is then combined with the maleic anhydride from the still pot for refining. The crude
maleic anhydride is refined via vacuum distillation. Normally, a forecut at high
reflux is removed followed by collection of product at lower reflux. The molten
maleic anhydride is collected for sale or further converted into pastilles or bri-
quettes for easier handling. Typical product impurities from such processes include
acrylic acid and acetic acid [30].
During the operation of this process, two challenges must be understood and
controlled: hot spots and gas mixtures. First, as gases are fed into the reactor tubes,
these gases are generally cooler than the mixed salts in the heat exchanger, leading
to a cooling effect of the salt mixture in the heat exchanger. A more significant
problem is if the gas temperature in the reactor tubes is greater than the salt mixture
in the heat exchanger. Under this scenario a hot spot can be generated in the catalyst
tubes. If not controlled, these hot spots can negatively affect the catalyst life and
reactor maintenance and decrease the reactor yield. Operational control of the
temperature is challenging because there are normally thousands of tubes and
only a few points of temperature measurement [25]. Hot spots can also lead to
possible runaway process conditions. It is critical that the operator understand the
hot spot effect, employing modeling and empirical observations in detail, to achieve
the maximum operating efficiency [31]. Second, careful monitoring of the gas
mixtures is required, thereby ensuring operation at the appropriate temperatures
relative to the inherent flammability limit of the process gas mixture. Typically, the
reactor inlet is set for the lower flammability limit of butane in air, and the exit line
is set for the lower flammability of the combustible composition at the reactor exit
[25]. For additional safety, the reactors are equipped with rupture disks at both the
reactor inlet and exit areas in case of a runaway event [25].
1.2.1.2 ALMA Fluid-Bed Process
In early 1984, Alusuisse Italia/Lummus Crest (ALMA) began operating a 3000 ton
per year maleic anhydride reactor system in Scanzorosciate, Italy [32]. A key
element to this plant was the implementation of a fluid-bed reactor. A schematic
of this plant is presented in Scheme 1.3:
In this reactor configuration, C4 and air are fed separately, from 2 to 8 volume
percent C4 in air, into the bottom of the reactor for presentation to the fluidized
catalyst bed [34]. The reactor is typically operating at 360–460 C [35]. Fluidized-
bed reactors tend to operate at higher C4 concentrations and process stream
residence times versus typical fixed-bed processes [15]. Heat generated during
the oxidative process is removed from the reactor via steam coils that are in intimate
contact with the fluidized solid materials. Reactors of this type are known to exhibit
Saturated
Steam Drum HP Steam
Cyclones
Off Gas
Reactor Cooler Reaction
Efftuent
BFW
Butane Off Gas Filters
Evaporator
Butane Catalyst Transport

Handling Gas
System
Air Air Compressor Spent Catalyst
Scheme 1.3 Simplified flow sheet of the reaction section of the Lonza ALMA process for maleic
anhydride production [33] (Springer and Topics in Catalysis, Vol. 38, Nos. 1–3, 2006, page
148, “VPO catalyst for n-butane oxidation to maleic anhydride: A goal achieved, or a still open
challenge?,” Ballarini, N.; Cavani, F.; Cortelli, C.; Ligi, S.; Pierelli, F; Trifiro, F.; Fumagalli, C.;
Mazzoni, G.; Monti, T., Fig. 1.1, Copyright 2006 with kind permission from Springer Science and
Business Media)
excellent heat transfer between fluidized solids and steam coils enabling smaller
heat transfer zones as compared to a typical fixed-bed reactor. As a result, hot spots
are typically less problematic. As with fixed-bed processes, the process heat gener-
ated is often recycled to drive reboiling and turbine drives. One interesting compo-
nent to the process is the ability to change catalyst during operation by intermittently
adding a virgin catalyst to maintain stable activities and particle size distributions
(~80–150 μm) [15, 34]. Gas flow patterns are generally back-mixed, a reverse
circulation flow pattern for a material in a process, and are a common challenge
that leads to enhanced impurity profiles and reduced yields [34]. For equipment, a
large volume of space is required above the catalyst fluid bed to enable separation of
solids and gases. Partitioning of the solids is achieved by cyclone and filter separa-
tion. Once separated, the product gas stream is cooled prior to the collection and
refining stages. As in the fixed-bed processes, any tail gases are incinerated prior to
venting. The process employs solvent adsorption to selectively remove maleic
anhydride from the cooled reactor effluent. Typically, the solvent is a cycloaliphatic
acid dialkyl ester such as dibutylhexahydrophthalate [36]. From there, the maleic
anhydride solvent mixture is pumped to the stripper where crude maleic anhydride is
separated as distillate. This material is fed to the light ends column where small
amounts of impure light ends are removed and incinerated. The remaining materials
are fed to the product column where the maleic anhydride product is recovered as
distillate and residuals are pumped back to the stripper. There is also a solvent
purification loop to prevent the buildup of impurities in the solvent. Process impu-
rities typically consist of carbon monoxide and acetic and acrylic acids [34].
1.2.2 Fixed-Bed and Fluid-Bed Comparisons
In analyzing these two established reactor technologies, a comparison scheme

proposed in 1985 by Wellauer is presented in Table 1.3.
Table 1.3 presents a logical map to determine where the key areas of development
are required to improve such processes. For fixed-bed reactors, topics ranging from
Table 1.3 Comparison of reactor technologies [37, 38]

Fixed-bed reactor Fluid-bed reactor
Capital requirements Large Less significant
Design knowledge Clear Emerging
Increasing scale Straightforward Challenging
Ease of operation Straightforward Challenging
Pressure drop Major Minor
Back-mixing Insignificant Challenging
Heat transfer Not good Good
Catalyst attrition Insignificant Challenging
Catalyst pore diffusion Critical Minor
Regeneration of catalyst during operation Challenging Minor
heat transfer, pressure control, catalyst porosity/geometry, as well as catalyst regen-

eration would be expected to be of keen interest to manufacturers that employ such
processes. Likewise, process modeling for increasing production scale, back-mixing,
and catalyst attrition are topics that would be of particular interest to fluid-bed
reactor designers and operators. As we will see later, this map was effective in
identifying the relevant, thematic topics of research that have since emerged.
1.2.3 Industry Response
In 1978, the majority of maleic anhydride manufacturing in the USA was by the
benzene route. Within 8 years, this manufacturing had been largely converted to C4
feedstocks [37]. This shift becomes even more apparent by reviewing the data
presented in Table 1.4.
Despite a trend toward fewer major North American (NA) suppliers (a few major
plant acquisitions are reflected in Table 1.4 by arrows) and reductions in plant
numbers, not only did the feedstock shift but overall industrial production volumes
increased. Such trends suggest that the current facilities have been able to operate
more efficiently and more competitively in an increasingly global marketplace.
The industry has transitioned from benzene to C4. What are the main technical
insights that have emerged to enable a successful transition to C4 feedstocks? How
have the general plant operating efficiencies been enhanced? Beyond the obvious
importance of plant engineering is the requirement for optimized materials, espe-
cially catalysts. It is the catalyst that is the real keystone to the entire maleic
anhydride industry. We must have a deeper, more fundamental understanding of
these specialized materials and their behavior in order to appreciate the current
industrial improvement efforts that are under way. Consequently, we now transition
the discussion into the world of catalytic oxidation science to understand the basic
structure of these materials and their resulting function, remembering to keep in
mind that these materials are critical to all of the manufacturing environments we
have previously discussed.
Table 1.4 Annual production capacity of NA maleic anhydride in the USA (tons/year) [37, 39, 40]
1.3 Catalysis: Progress in Structure and Function
Thermal, oxidative catalysis is an essential component to the success of commercial

maleic anhydride production. It is a complex topic to which significant research
resources, both in industry and academia, are dedicated. To better understand the
fundamental elements of this catalytic process, we will break this topic down into
two distinct discussions: inorganic catalyst structure and proposed C4 catalytic
oxidation mechanisms. With this background, the reader is expected to gain a
more representative insight to the current trends in this field.
1.3.1 Inorganic Catalyst Structure
Vanadium oxides represent the most important catalyst class in C4 oxidation. Vana-
dium found in vanadium oxides can exist in a variety of oxidation states (+5, +4, +3,
and +2). This broad range of oxidation states enables vanadium to form numerous
coordination polyhedral structures that can include tetrahedron, trigonal bipyramid,
square pyramid, distorted octahedra, and rectilinear octahedra [41]. Specifically
toward maleic anhydride production, the key oxidation states are V5+ and V4+ [42]. Per-
haps the most important form of vanadium oxide is vanadyl pyrophosphate which is
considered the common template. Remembering that the typical maleic anhydride
process occurs at high temperature (i.e., ~400 C), where the catalytic precursor is in a
dehydrated state and effectively in a catalyst heat treatment environment, the active
component responsible for the oxidation process is attributed to (VO)2P2O7 [42]. Lit-
erature nomenclature references to (VO)2P2O7 include both vanadyl diphosphate and
vanadyl pyrophosphate [43, 44]. We will employ VPO to refer to this inorganic
composition. VPO is comprised of two polyhedra that are connected through oxy-
gen-bridging: VO6 and PO4. The basic polyhedra structures, which are depicted in the
most common illustration styles, are presented in Scheme 1.4.
VO6 is often referred to as a distorted octahedron, where the centrally located
vanadium atom is coordinated to six oxygen atoms. Other references cite a VO5
square pyramid that is weakly coordinated to another O from a neighboring VO5
square pyramid. The distinct feature of this structure is the vanadyl group
(Scheme 1.4a) that forms a double bond to oxygen, denoted as V ¼ O. This
particular bond is shorter than the others resulting in a distortion of the octahedron.
Bond lengths determined on single green crystals are present in three ranges: ~1.6 A
(V ¼ O, “short”), ~2 A (~coplanar V–O), and ~2.3 A (V–O, “long”) [45]. The
structure for PO4 (Scheme 1.4b) is a tetrahedron where the centrally located
phosphorus atom is coordinated to four oxygen atoms. Bond lengths are found to
be ~1.5 A [45]. Bond valence calculations yield ~4.1 for V and ~4.9 for P,
corresponding to V4+ and P5+, respectively [45].
From these two basic structures, the VPO catalyst architecture begins to
emerge, as illustrated in Fig. 1.1. Vanadium polyhedra are equatorially
Scheme 1.4 VO6 and PO4 crystal structures [41, 43] ( filled blue circle is vanadium; filled yellow
circle is phosphorus; open red circle is oxygen)
Fig. 1.1 Various VO6 and PO4 connective structures in an idealized VPO strand ( filled blue circle
is vanadium; filled yellow circle is phosphorus; open red circle is oxygen)
connected, in a trans-oriented manner, exhibiting an oscillating structure due to

the spatial arrangements of the short V ¼ O and long V–O bonds. The four
nearly coplanar oxygen atoms bonded to the central vanadium atom are linked
to P via either single PO4 groups or double tetrahedral pyrophosphates (P2O7),
where the oxygen atoms are shared [42]. Because V3+ is not known to be
common to mild oxidation catalysis, it is generally not represented. Conse-
quently, two general structures prevail: one single distorted octahedron and a
pair of distorted octahedra [42]. Examples of these various connective structures
are presented in Fig. 1.1 for an idealized VPO strand. Note that Fig. 1.1 presents
these idealized strands in three common illustration styles found in the literature.
Also note that the bond angles are not actualized and these structures are only to
be considered a general representation.
Figure 1.1a presents formal oxygen atoms to highlight oxygen bridging between
the polyhedral, while Fig. 1.1b presents such bridges through implication. Perhaps
least intuitively, Fig. 1.1c eliminates the long V–O bond contribution from the
distorted octahedron, presenting VO6 as a square pyramidal polyhedron to simplify
the vanadium environment for visual interpretation.
From visualizing simple individual strands, a more complete VPO can be confi-
dently presented and investigated. Negating the long V–O bond from VO6, as
previously discussed, and presenting the square pyramidal polyhedron (dark gray)
and including tetrahedral (light gray), PO4 crystals enable the presentation of the
ideal orthorhombic VPO crystal structure in Fig. 1.2.
Upon further examination of Fig. 1.2, one finds crystalline sheets comprised
of VO6 polyhedra with alignments that are further distorted by variations in the
group arrangements of tetrahedral phosphate. Unlike previous results, which
demonstrated a low degree of symmetry and eight distinct V atoms, more
recent results yield improved accuracy and four distinct V atoms. This ideal
structure is expected to differ from commercial VPO catalyst structures due to
additional defects caused by impurities as well as other inherent structural
complications [45].
Fig. 1.2 Polyhedral representation of the crystal structure of (VO)2P2O7 viewed (a) along the
[100] direction and (b) along the [010] direction. The VO5 square pyramids are drawn in dark grey,
the PO4 tetrahedra in light grey. Large circles represent V atoms and medium circles represent P
atoms. Small circles in (a) denote O atoms, which are not shown in (b). [46] (Geupel, S., Pilz, K.,
Smaalen, S. van, Büllesfeld, F., Prokofiev, A., Assmus, W. (2002) Acta Cryst. C58, 9-13.
Reproduced with permission of the International Union of Crystallography)
1.3.2 Proposed C4 Catalytic Oxidation Mechanisms
The catalytic activity of VPO is considered to be located at the interface of the (100)
plane, to where we will focus our attention on its finer surface details. In 1987,
Bordes proposed a model of the VPO surface which is presented in Fig. 1.3.
A key feature to this VPO surface model is its apparent amorphous structure
which is a result of localized γ-VOPO4 micro-domains formed via surface topotactic
reoxidative processes. These features, when coupled to inherent variations in topog-
raphy from the various atomic contortions, account for the irregular structure
[42]. During maleic anhydride production, the catalyst is continuously exposed to
V+5
+4 +4 +5
(VO)2 P2 O7 -VO PO4
Fig. 1.3 Model (100) plane crystal surface of VPO and γ-VOPO4 [42] (Reprinted from Catalysis
Today, 1, Bordes, E., “Crystallochemistry of V–P–O Phases and Application to Catalysis,”
499–526, Copyright 1987 with permission from Elsevier)
oxygen/air, C4/hydrocarbons, heat, and water. Such an environment, when linked to

the inherent trans orientations of the paired distorted VO6 octahedra, suggests a
redox-type coupling for the paired crystals. A closer inspection of Fig. 1.3 reveals five
different oxygen sites. These sites are V ¼ O, V–O, V(1)–O–P, V(2)–O–P, and P–O–P.
Vanadium oxygen vacant sites (denoted by an □ in Fig. 1.3) are considered to be
Lewis acid (electron pair acceptor) sites. Such sites can be reoccupied by oxygen,
water (denoted by an ● which is OH2 free or coordinated to V), or –OH resulting in the
formation of Bronsted acid (proton donor) sites. Additionally, it is also possible for
oxygen from PO4 moieties to be abstracted from the surface (also denoted by an □ in
Fig. 1.3), which can further alter the crystalline framework. As such, there is a slight
excess in PO4 (~P/V ¼ 1.2). Interestingly, note that when the P/V ratio decreases to ~1,
the reactivity rate was found to diminish significantly. This effect can be considered an
indication of PO4’s contribution to the overall catalyst activity enhancement [42].
With this model surface in hand, a visualized interface between VPO crystalline
phases and C4 intermediates has been suggested. Bordes envisioned that C4 acti-
vation was achieved by available oxygen sites at the exposed surfaces of vanadium
and phosphorus. Envisioned arrangements of hydrocarbon orientations are included
in Scheme 1.5.
Scheme 1.5 (1) suggests C4 activation at the V4+ site resulting in hydrogen
abstraction from methylene and methyl groups which protonates oxygen on V ¼ O
and P–O, respectively. The protonations yield V–OH and P–OH moieties at the
catalyst surface and absorbed “ene” hydrocarbons. Ultimately, desorption of water
molecules, resulting from this process, results in new vacancies. These vacancies can
be replenished by two pathways: lower catalyst layers, via gliding processes, and
gaseous oxygen, forming γ-VOPO4 micro-domains. Scheme 1.5 (2) suggests another
possibility that V4+ can abstract both hydrogens resulting in adsorbed butadiene
molecule, depicted in either cis or trans conformations [42]. Scheme 1.5 (3) suggests
that adsorbed butadienes on γ-VOPO4 can cyclize via incorporation of oxygen from a
P–O moiety yielding 2,5-dihydrofuran (DHF) or furan. Finally, Scheme 1.5 (4) pre-
sents the concept of dehydrogenation of furan by acid phosphate groups which can then
receive oxygen which is on top of V5+ leading to the formation of maleic anhydride
Scheme 1.5 Possible molecular orientations during maleic anhydride production [42] (Reprinted
from Catalysis Today, 1, Bordes, E., “Crystallochemistry of V–P–O Phases and Application to
Catalysis,” 499–526, Copyright 1987 with permission from Elsevier)
[42]. These mechanistic concepts were more recently explored by other techniques to
develop a more comprehensive picture to the nature of VPO catalytic behavior.
In 1989, Busca and Centi employed steady-state and transient reactivity measure-
ments, Fourier transform infrared (FTIR) studies, and stopped-flow desorption (SFD)
measurements to study the surface dynamics of absorbed species on the VPO catalyst
surface during C4 oxidation. Some of their key experimental observations include:
(1) Steady-state oxidation studies reveal decreasing selectivity as maleic anhydride
conversion rates increase to greater than ~80 % in combination with enhanced forma-
tion of carbon oxides. (2) C4 oxidation behavior follows a Langmuir–Hinshelwood
dependence on oxygen concentration. (3) Comparing non-steady state pulsed reactor
(anaerobic conditions) flow to steady-state flow reactor (aerobic conditions) presented
similar results suggesting that under these conditions, the C4 oxidation is not affected
by the rate of catalyst surface reoxidation by gaseous oxygen. (4) C4 oxidation is
increased significantly if the catalyst is initially treated at higher temperatures under
inert helium gas flow conditions. (5) FTIR hydrocarbon absorption experiments
produced results enabling the formation of a hydrocarbon reactivity scale where
butadiene (most reactive) > butene > C4 (least reactive). (6) Behavior of C4 hydro-
carbons under aerobic or in vacuo comparative conditions suggests that maleic
anhydride can form on the oxidized VPO surface in the absence of gas-phase oxygen.
(7) Furan was found to strongly absorb to the VPO surface at room temperature while
at higher temperatures exhibits absorbed species similar to DHF and C4 reactive
absorption conditions. (8) Stopped-flow desorption methods revealed much stronger
interactions for butadiene compared to other C4 hydrocarbons. (9) Stopped-flow
desorption measurements suggest furan disproportion hydrogen transfer reactions
from furan to form crotonaldehyde and butadiene [47]. Using these numerous exper-
imental observations, Busca and Centi proposed a reaction mechanism for C4 conver-
sion to maleic anhydride that is presented in Scheme 1.6.
Scheme 1.6 Proposed reaction mechanism of C4 conversion to maleic anhydride with VPO [47]
(Reprinted with permission from Busca, G.; Centi, G., J. Am. Chem. Soc. 1989, 111, 46–54.
Copyright 1989 American Chemical Society)
As presented in Scheme 1.6, there are two main pathways for maleic anhydride
formation, both of which occur predominately via transformation of absorbed species
on the catalyst surface. Route A is believed to be the faster reaction due to higher
reactivity found in the presence of gaseous oxygen. Thus, C4 proceeds to butadiene
which is then converted to dihydrofuran (DHF) via oxygen insertion from Vv ¼ O. This
oxidation is believed to occur via 1,4-electrophilic addition. Once the DHF interme-
diate is formed, the presence of labile oxygen species [O*] can promote its transfor-
mation to γ-crotonolactone (III). Butadiene can also transform directly to
γ-crotonolactone by coupling the insertion of oxygen from both Vv ¼ O and labile
oxygen species [O*]. Upon γ-crotonolactone formation, additional oxidation yields
maleic anhydride. From γ-crotonolactone it is also possible to form phthalic anhydride,
which can result from an enhanced lifetime of this intermediate on the catalyst surface
in the presence of butadiene. Route B similarly proceeds through the formation of the
DHF intermediate. DHF can then be dehydrogenated to yield a furan (I) intermediate.
Furan can yield maleic anhydride when further oxidized by labile oxygen species [O*];
however, due to strong Lewis acid site interactions at the VPO surface, furan can also
serve as a hydrogen transfer agent promoting hydrocarbon surface species (II) that are
potentially capable of forming both crotonaldehyde and carbon oxides [47].
A 1991 study of VPO’s electronic structure and properties by Schiott, Jorgensen,
and Hoffmann developed a mechanism for the formation of maleic anhydride that
focused on the more selective DHF to lactone route to maleic anhydride. Tech-
niques such as density of states (DOS) and crystal orbital overlap populations
(COOP) were employed to suggest models of the VPO surface and interaction
schemes between various VPO sites and C4 molecules. Also incorporated in the
study were +4 and +5 oxidation states for VPO, for structures comprising V2O88,
analogs that incorporate P(OH)3 moieties, and molecular orientations between
organic/inorganic interfaces. Importantly, they point out that in this reaction, bulk
oxygen is not significant; but it is more essential for molecular oxygen to become
Fig. 1.4 Proposed coordinated structures of molecular oxygen to VPO superoxo (Left) and peroxo
(Right) [49] (open red circle is oxygen)
activated from the surface of VPO, an effect that is still not clear. They found only
minor variations for comparative electronic surface properties, with and without
phosphorus, leading them to conclude that the V2O88 structure was a suitable
structural model for the catalytic transformation of DHF [48].
The important intermediate is the formation of DHF, which they suggest is derived
from molecular oxygen inserting from the vanadyl group, in a concerted [2 + 4]
cycloaddition-type process, to the 1,4-positions of butadiene. How molecular oxygen
becomes available for C4 coordination is a feature of the mechanism to explore in
more detail. The molecular oxygen molecule is capable of coordinating with VPO in
two forms, known as superoxo and peroxo. Remembering the trans configuration of
neighboring VO6 distorted octahedra, molecular oxygen coordinates to vanadium
trans to the V ¼ O. The two proposed coordinated structures are presented in Fig. 1.4.
Using extended Huckle approaches, it was not possible to distinguish energy-
level differences between these structures. However, upon further consideration,
where orbitals were rotated, the peroxo form was perceived to be more favorable
[48]. As we will learn later, both structures, when abstracting hydrogen from DHF,
lead to O–O–H surface species.
DHF is thought to be formed by insertion of molecular oxygen from the vanadyl
group (V ¼ O) to butadiene. With this in mind, once in the appropriate configura-
tion of surface species is achieved, electron donation from DHF to the peroxo σ*
orbitals becomes possible. A sequence of various proposed intermediate structures
are described in more detail in Fig. 1.5 [50].
From the Fig. 1.5a intermediate, formation of a new O–O–H bond via hydrogen
transfer is possible (see Fig. 1.5b). A similar process is also envisioned to be
suitable for the superoxo intermediate, yielding an identical structure. From the
Fig. 1.5b intermediate, interaction between the newly formed O–O–H and C1 is
favorable, enabling the formation of the 2-hydroxy derivative of DHF (see
Fig. 1.5c). Rotation leads to Fig. 1.5d, enabling enhanced interaction with the
oxygen atom on the adjoining vanadium atom. Another hydrogen abstraction
from this 2-hydroxy derivative results in the formation of a new vanadium-hydroxyl
functionality pictured in Fig. 1.5e. Concurrently, the oxygen atom on the C1 is
beginning to transition to C ¼ O bond formation. Finally, hydrogen is transferred to
the V–OH resulting in the formation of a V–H2O species and a coordinated
γ-crotonolactone species on the VPO surface, depicted in Fig. 1.5f [48]. Once
γ-crotonolactone has been formed, there are two required steps envisioned to
Fig. 1.5 Sequence of proposed intermediate structures in transformation of DHF in the presence of
molecular oxygen and the vanadyl pyrophosphate surface [48] (Figs. 5B, 5C, 5D, and 5E adapted with
permission from Schiott, B.; Jorgensen, K.A.; Hoffman, R., J. Phys. Chem. 1991, 95, 2297–2307.
Copyright 1989 American Chemical Society) ( filled blue circle is vanadium; open red circle is oxygen)
complete the formation of maleic anhydride. First, the water molecule (depicted in
Fig. 1.5f) is exchanged for molecular oxygen. Second, the coordinated
γ-crotonolactone rotates 180 enabling another CH2 moiety to be in close proximity
to the activated di-oxygen species derived from reabsorption of molecular oxygen.
Thus, with these two steps completed, the maleic anhydride formation process then
proceeds in a similar manner to the steps previously highlighted in Fig. 1.5,
ultimately resulting in the formation of a second carbonyl moiety, at the
4-position (see Fig. 1.5f), and maleic anhydride [48].
In 2013, M. Cheng and W. A. Goddard employed density functional theory
(DFT) techniques to study the mechanism of C4 oxidation by VPO. They propose a
more essential role for phosphorus, pointedly remarking that unlike previous
studies where phosphorus moieties are presumed to be “innocent linked ligand
(s)” in the catalytic process, the phosphorus–oxo bonds [P ¼ O] are actually highly
reactive and activate C4 by decoupling hydrogen from the methylene groups. The
Scheme 1.7 Schematic of VPO surface oxidation [50] (Reprinted with permission from Cheng,
M.; Goddard, W.A., J. Am. Chem. Soc. 2013, 135, 4600–4603. Copyright 2013 American
Chemical Society)
first step in their proposed process is the molecular oxygen’s oxidation of the VPO
surface. A schematic of this reaction is presented in Scheme 1.7 [50].
In Scheme 1.7, the VPO surface is depicted in a manner that eliminates the oxygen
atoms coordinated to the vanadium atom in order to make the process more clear. In
this reaction scheme, oxygen adds to phosphorus, converting the pyrophosphate (P–
O–P) moiety into two orthophosphate. In concert, V4+ is oxidized to a new V5+ state.
Once VPO has been oxidized, the newly formed P ¼ O moiety is thought to be
capable of abstracting a hydrogen atom from the C4 methylene as depicted in
Scheme 1.8a.
This hydrogen abstraction, which results in the formation of a P–OH moiety in
the catalyst surface, overcomes a potential energy barrier of 0.0 kcal/mol and a
reaction barrier of 13.6 kcal/mol. Such an energy barrier is reported to be consistent
with experimental determinations suggesting this transformation is the origin of the
VPO catalytic cycle. The next step in the proposed mechanism is trapping the
resulting methylene radical by a V–O surface moiety. A subsequent hydrogen atom
abstraction by a neighboring P ¼ O surface group yields a second P–OH and a
reduction of V–O to a V ¼ O moiety (see Scheme 1.8a). This step concludes with
the formation of the intermediate butene. Similar processes are repeated to enable
the formation of the intermediates butadiene (Scheme 1.8b), DHF (Scheme 1.8c),
and γ-crotonolactone (Scheme 1.8d). The conversion of γ-crotonolactone to maleic
anhydride is the final step in the transformation of C4, as presented in
Scheme 1.8e. For each step in this proposed mechanism, M. Cheng and W. A.
Goddard envisioned that the intermediate species desorbs from the catalyst surface
to then re-adsorb to a fresh catalyst surface for the next step in their specific
transformation [50].
From the myriad of fundamental research, we learn that the structural, perhaps
even architectural, catalyst features and surface properties are essential to the
inherent oxidation capabilities. We will find later that, even now, researchers are
continuing to adjust formulations and compositions in addition to catalyst prepara-
tion processes to develop even deeper understanding of these structure/function
relationships to further improve maleic anhydride manufacturing processes.
Beyond purely compositional considerations, maintaining the compositional
Scheme 1.8 Proposed initial mechanistic steps of C4 activation by VOPO4 [50] (Reprinted with
permission from Cheng, M.; Goddard, W.A., J. Am. Chem. Soc. 2013, 135, 4600–4603. Copyright
2013 American Chemical Society)
balance (~P/V ¼ 1.2) of the catalyst has been suggested as critical to the long-term
performance behavior. As a result, strategies to maintain this compositional balance
during production have emerged. Thus, while we continue to learn more about the
basic nature of these materials, our ability to improve manufacturing processes
based on current understandings has been fruitful, and this background should
improve our appreciation for the technical approaches yet to be discussed.
1.3.3 Catalyst Improvement and Implementation Strategies

in Manufacturing
Common improvement strategies in the industrial application of maleic anhydride

catalyst technology include insertion of enhancement agents, co-metal inclusions,
multiple catalytic zones, halide removal, pretreatment and purity, dimensional
stability, morphology and porosity, and shape and regeneration. Note that in the
case of insertion agents, where the discourse may not be obviously linked to the
catalyst material per se, this topic is incorporated into this section due to the
inherent exposure of such chemistries to the catalysts featured in the process.
1.3.3.1 Insertion of Enhancement Agents
Autoignition Suppression Agents
Gas compositions for C4 oxidation processes must be carefully monitored to

minimize the possibility of autoignition. Manufacturing conditions may require
operation within, or at least near, potentially hazardous conditions, wherein the feed
compositions may result in exposure to ignition sources, combustion, and possibly
explosive combustion. The consequences of autoignition include loss of product
yield, productivity, and personnel safety. Without an ignition source, the flammable
process mixtures are still subject to autoignition if exposed to conditions suitable
for oxidation chain reactions, such as prolonged exposure to elevated temperatures
and pressures [51]. In addition to developing a method for estimating the
autoignition temperature of the gaseous flammable mixture, incorporating the
influencing factors such as pressure, concentration, volume, and other conditions,
a method to suppress the autoignition temperature was developed by utilizing a
pyrophoric suppression agent. In deploying the suppression agent, the process
stream contains an autoignition suppression agent in contact with the gases and
catalyst. The autoignition suppression agent is comprised of acidic sites or trivalent
phosphorus which results in associated deposits along the plant walls of the process
stream. These agents maintain contact with oxygen and iron at the wall interface at
temperatures suitable for decomposition enabling the formation of polyphosphate
acidic polymers or trivalent phosphorus as an inhibitor to autoignition [51].
Additions of Process Gases
In addition to air and C4, it is also possible to inject other gases to further modify the
process behavior. For example, additional mixing of a substantially pure carbon
monoxide into the reactor gaseous feed streams, which comprise C4 and an oxygen-
containing gas, in a fixed-bed reactor has been found to improve the process yield
and selectivity. The continuous addition of at least one trialkyl phosphite or trialkyl
phosphate component to this new gas combination, in amounts ranging from 0.5 to
4 ppm based on the weight of elemental phosphorus to the total amount of the
gaseous feed, is also considered. This reactant combination results in a conversion
rate of C4 to maleic anhydride of about 75 % or more [52].
To improve the C4 conversion rates and maleic anhydride yields, another
approach has been to operate the oxidation process in the presence of an inert
gas, i.e., nitrogen, helium or neon, and molecular oxygen (in the form of air).
Critically, the thermal conductivity of the inert gas mixtures is designed to be not
less than 800 (104Wm1K1) at ~427 C (700 K). The inert gas is recovered by
separating for reuse as reactor feed. The total concentration of gases in the reactor
feed ranges from 75 to 90 % by volume (inert gas mixtures), 1 to 5 % (C4), and 5 to
25 % (molecular oxygen as air). Using helium in this approach, a process conver-
sion of 83.2 (mol. %), selectivity of 73.9 (mol. %), and yield of 61.5 (mol. %) were
achieved. Helium was recovered at 95.0 (mol. %). In contrast, in the absence of
helium, the process conversion of 81.5 (mol. %), selectivity of 62.0 (mol. %), and
yield of 50.5 (mol. %) were achieved, indicating significant process improvement
for the inert gas strategy [53].
Addition of Phosphates
The reaction temperature profile within the oxidative reaction zone in a fixed-bed
reactor can be controlled by aqueous phosphorus compound additions, even if the
process was operating in a known flammability zone. Consequently, the ability to
control hot spots translated to an enhancement of the catalyst life and production
rates. To achieve these benefits, a phosphorus compound, such as triethyl phosphate
(TEP), in water, is continuously added to the feed stream. The weight ratio of water
to phosphorus ranges from 6500:1 to 35,000:1. The temperature throughout the
entire effective reaction zone is maintained to be no greater than 30–50 F above the
salt bath temperature. Under these conditions, a uniform isothermal method of
operation enabled high product yields using catalysts for extended periods of
time. Using this approach, process conversions of 85 (%), selectivity of 65
(mol. %), and yield of 95 (wt. %) are achievable [54].
A similar improved process for continuously distributing phosphorus agents
[i.e., trimethyl phosphate (TMP)] in fixed-bed reactors, at levels of 7–14 ppmw,
that incorporates a strategy for enhancing the interfacial contact between the liquid
TMP and the gaseous C4 (ranging from 85 to 95 C) by reducing the average size of
the liquid droplets, enhancing the phosphorus agent’s vaporization behavior, was
developed. To reduce the droplet size, the liquid TMP (optionally with water) is
first warmed, typically by wrapping the TMP feed line around the warm C4 pipe.
Then, during continuous insertion, the TMP is passed through an in-line filter
medium which promotes its vaporization and distribution into the gaseous feed
stream. The residence time for this mixture is typically 8–15 s prior to contact with
the air feed. The increasing mixing time further improves the uniformity of the feed
stream prior to contact with the reactor catalyst. In the case of multiple reactors,
similar strategies can be practiced prior to each reactor to permit individual

phosphate level adjustments and good mixing. Reduction in reactor surface wear,
maintenance, and improvements to reactor longevity are benefits of this
approach [55].
Addition of Peroxides
Other oxidative compounds, i.e., hydrogen peroxide, have been found to be bene-
ficial. In one example, where H2O2 was added directly to the C4/air feedstock, an
equivalent yield was observed coupled to a lower salt bath temperature (~20–30 F
lower). The preferred amount of H2O2 ranged from 5 to 1000 ppmw. A lower salt
bath temperature can prolong the catalyst lifetime, resulting in longer production
runs between maintenance of the catalyst bed. While not limited to H2O2, H2O2 has
provided additional benefits such as additional water during decomposition and
oxygen supply for the oxidation reaction [56].
1.3.3.2 Catalytic Co-metal Inclusions
Another very popular strategy has been to develop new inorganic catalyst compo-
sitions. Catalyst compositions comprising composite metal oxides of molybdenum,
vanadium, tellurium, and certain other types of metals have been found to be
effective in the maleic anhydride oxidation process. Other elements include Nb,
Ta, W, Ti, Al, Zr, Cr, Mn, Fe, Ru, Co, Rh, Ni, Pd, Pt, Sb, Bi, B, In, P, and Ce. In a
specific example, 90 wt% of Mo1V0.3Te0.23Nb0.12On and 10 wt% of SiO2 were
found to exhibit conversion of 96.6 %, selectivity of 37.0 %, and a yield of 35.7 %
when operating the oxidation reactor, where the molar ratio of C4–air ¼ 1:24, at
1000 SV (hr1) at 422 C [57].
In a recent VPO catalyst modification from BASF, compositions comprising
vanadium, phosphorus, iron, and oxygen were developed. The iron, typically iron
phosphate, is present at an atomic ratio of 0.005:<0.05 (iron–vanadium). Proper-
ties of the typical catalysts include BET surface area from >15 to 40 m2/g, pore
volume from 0.15 to 0.4 mL/g, and bulk density from 0.5 to 1.0 kg/L [58]. In
another catalyst approach employing iron, a polynary vanadyl pyrophosphate
composition, such as (VO)Fe2(P2O7)2, was designed. The powder product exhibits
a BET surface area, ranging from 1 to 4 m2/g, depending on the catalyst prepar-
ative procedure [59].
1.3.3.3 Multiple Catalytic Zones
The performance of the oxidation reactor can be modified by deploying at least two
different catalysts, where the catalyst activities are designed to be packed into the
tubes in a gradient manner to enable improved reactor performance and control
Table 1.5 Mixed catalyst loading schedule in a fixed-bed reactor [60]

Bed function % of tube length in reactor Catalyst activity
Preheat 1923 0.91.0
Critical 2329 0.70.9
Downstream 4856 1.0
[60]. By deploying a catalytic gradient, the operator is capable of running the

reactor at an initial (feed gas) hydrocarbon concentration of > 1.5 % by volume,
to maintain a temperature difference between the gas and cooling fluid > 15 C in
the reactor tube area where the gas temperature exceeds the fluid temperature, and a
maleic anhydride productivity of at least about 5.0 lbs. per hour. Critically, during
the reaction, the temperature difference between the gas and cooling fluids does not
exceed 80 C. The catalyst activity and the gas permeability of the bed vary in the
general direction of the gas stream. In this manner, the catalyst activity and the
pressure drop are lower in the critical region, or hot spot. An example of a catalyst
loading schedule is presented in Table 1.5.
Assembly of the catalyst bed in this fashion results in significant improvement in
reactor performance and reduced catalyst degradation [60].
Other researchers found success in designing two successive catalytic reaction
zones. In the first reaction zone, catalysts suitable for producing 1,3-butadiene from
C4 via oxydehydrogenation were employed. Examples of suitable catalysts include
Mo12BiNi8Pb0.5Cr3K0.2Ox and Mo12BibNi7Al3Cr0.5K0.5Ox [61]. In the second reac-
tion zone, catalysts suitable for the oxidation of 1,3-butadiene to maleic anhydride
were deployed. Suitable catalysts include SbMo3V0.1Fe0.2W0.06Ox [62]. In demon-
strating this process, the process was configured so that the first reaction zone was
330 C and the second zone was 400 C. The maleic anhydride yield was deter-
mined to be 62 %. Conversely, operating the process under isothermal conditions of
either 330 C or 400 C resulted in maleic anhydride yields of 32 % or 54 %,
respectively [63].
1.3.3.4 Halide Removal
The concept of halide removal from the catalyst composition has received attention.
To achieve this goal involves utilizing glycol ether solvents and phosphoric acid in
a slurry of vanadium and other metals and metal oxides with no chlorides present.
Water can be introduced during the catalyst preparation and/or presented during the
oxidation process. In a specific example where water was part of the preparative
process, comparing a chloride-containing catalyst to a chloride-free catalyst, both
based upon VPO–Mo compositions, results for a pilot reactor run are presented in
Table 1.6 [64].
Results presented in Table 1.6 demonstrate that the chloride-free catalyst
improves the overall performance of the oxidative reactor, as indicated by greater
conversion, selectivity, and yields, coupled to operating the process at lower reactor
Table 1.6 Performance of chloride-free catalyst in a pilot reactor [64]

Hours on Reactor Chloride-containing Reactor Chloride-free
stream temperature ( C) catalyst (VPO–Mo) temperature ( C) catalyst (VPO–Mo)
1000 406 Conversion 399 Conversion
(mol%) ¼ 84.6 (mol%) ¼ 87.7
Selectivity Selectivity
(mol%) ¼ 62.7 (mol%) ¼ 64.6
Yield Yield
(wt.%) ¼ 89.5 (wt.%) ¼ 95.6
temperatures. All of these features combined to improve the overall efficiency and
economics of the process.
1.3.3.5 Pretreatment and Purity
Not surprisingly, pretreatment of VPO catalysts was shown to boost the activity of
various catalyst compositions. Chemical pretreatment can be accomplished by
contacting the catalyst precursor to dry inert gas to remove any residual oxygen,
then introducing acetic anhydride vapors, at a temperature of 150 C, at 2–5 % by
volume, for about 8 h of exposure. Residual acetic anhydride vapors were removed
by sweeping the system with nitrogen gas. The powder from this product was
formulated with 4 % graphite and pressed into tablets in preparation for calcination.
The calcination process is critical to the overall performance of this strategy. The
atmosphere during this process is typically a mixture of steam, inert gas, and
oxygen. In general, slower heating rates result in better catalyst performance. In
the example provided, the optimal heating rates are <1 C per minute to a
temperature 420 C coupled to isothermal temperature dwell time, suitable for
achieving a vanadium oxidation state no greater than +4.5, after 8 h [65].
In modifying the previous catalyst preparation strategy, a step was added for the
removal of water formed during the catalyst preparation to achieve < 0.5 % residual
water content by volume. The water is removed via an entrainer, such as cyclohex-
ane, toluene, or xylenes. Crude amounts of water were removed by a water
separator followed by additional water removal during reflux. Heating schedules
for calcination are in four distinct steps: step 1, purge the furnace with nitrogen at
20–100 C to reduce oxygen to <5 % by volume; step 2, hold temperature ranging
from 150 to 250 C, heating rate from 5 C to 20 C/min, for an isothermal hold up
to 3 h; step 3, hold temperature ranging from 200 to 300 C, heating rate from 1 C
to 10 C/min, for an isothermal hold up to 3 h; step 4, hold temperature ranging
from 380 to 460 C, heating rate from 0.1 C to 3 C/min, for an isothermal hold
from 2 to 8 h. Mixtures of air, nitrogen, and water vapor (0.1–0.5:0.1–0.5:0.0–0.8
volume ratios) are fed in the furnace during these temperature steps. The treated
catalysts are cooled to < 100 C under nitrogen prior to removal from the furnace
for storage in a tightly closed container. The performance of the catalyst was judged
to be superior to the previous catalyst treatment, particularly in terms of yield

improvement [66].
Phosphorus additions to fluidized-bed catalysts have developed for suitable TEP
insertion techniques. Previous methods demonstrated that vapor-phase addition of
TEP led to increased operating temperatures, which limit reactor throughput and
accelerate catalyst degradation [67]. An alternative approach is to add phosphorus
by impregnating the catalyst with TEP prior to adding the catalyst to the reactor.
Impregnated VPO catalyst enables lower TEP use levels and increased yield of
maleic anhydride at lower operating temperatures. The amount of TEP added to the
catalyst is from about 0.9 mg to about 0.454 kg of TEP per 45.4 kg of total catalyst
bed per day [68].
Lonza researchers report negative performance effects to catalyst activity related
to remaining carbonaceous materials trapped inside the catalyst. To control the
carbon content in the VPO catalyst precursor, and catalyst purity, the catalyst is
prepared in an organic medium comprising (a) isobutyl alcohol or a mixture of
isobutyl alcohol and benzyl alcohol and (b) a polyol, such as 1,3-propandiol and
1,4-butanediol. The preparation of the catalyst precursor in such organic mediums
results in carbon content ranges from 0.7 to 4 wt%. The catalyst precursor is
transformed into the active catalyst by an activation process that includes a heat
treatment that includes applying a temperature of up to 600 C. The performance of
the final catalyst improved for conversion, selectivity and yield [68].
Another process for preparing pure VPO catalyst precursors which exhibit low
residual organic material is based upon the following steps: (1) reacting VPO with
102–110 % strength H3PO4 in the presence of isobutanol in a temperature range
from 80 to 160 C. It includes the option of additional alcohols, such as a primary or
secondary, noncyclic or cyclic, unbranched or branched, saturated alcohol, and
having 3–6 carbon atoms; (2) isolating precipitated product; (3) (a) drying the
precipitate to <5 wt% residual isobutanol or other alcohols; (b) passing gases, at
temperatures ranging from 130 to 200 C, such as nitrogen comprised of 0.1–9 % by
volume of oxygen through the dried precipitate, directly or after isolation. The
powder product exhibits BET surface areas, ranging from 20 to 40 m2/g, pore
volumes ranging from 0.2 to 0.4 mL/g, and bulk density ranging from 0.5 to 1.0 kg/
L, depending on the catalyst preparative procedure [69].
1.3.3.6 Dimensional Stability
Dimensional stability of the catalyst is critical to the ability to implement a

commercially viable catalytic material. The material must be strong enough to
withstand handling, abrasion, and other physical forces. One critical physical
attribute is catalyst expansion. Amoco studied the manufacture of VPO catalysts
and developed techniques to alter their fabrication in order to minimize the ten-
dencies of the catalyst toward expansion. In this approach, the VPO was prepared
with a co-metal, such as molybdenum trioxide (MoO3). Critically, in the prepara-
tive process, water is added, ranging from 2.5 to 3.5 moles H2O per mole of P. This
CATALYST TABLET EXPANSION AS A FUNCTION

OF REACTION H2O / P MOLE RATIO
40
30
TABLET EXPANSION, %
20
10
-10
1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00
H O / P MOLE RATIO
2
Fig. 1.6 Catalyst tablet expansion as a function of reaction water/P mole ratio [70]
water reacts to form phosphoric acid and esters. If the water is outside this range,
the catalyst tablet will exhibit undesirable expansion properties. Additionally, to
achieve the final catalyst form, additives such as inert supports (i.e., alumina,
titania, silicon carbide, kieselguhr, pumice, or silica) and binders (i.e., starch and
polyvinyl alcohol) can be employed to improve physical properties and shape
forming processes. In developing this approach, a tablet expansion test was
designed, where tablets were dimensionally analyzed prior to being placed in a
900 F oven equipped with a humid air stream for 2 h. The tablets are allowed to
cool in a desiccator and are dimensionally reevaluated. Results of these tests are
presented in Fig. 1.6.
The results from these tests indicate that optimal range for the water/P mole ratio
is 2.5–3.5, where the tablet either does not expand or slightly contracts. Thus, by
controlling the water in the catalyst preparation, catalyst expansion can be mini-
mized enabling longer production runs and catalyst cycle life [70].
Other strategies toward improving catalyst expansion properties include the
production of coated catalysts. Coated catalysts are achieved by applying a catalytic
composition in a coatable form to a shaped structural support body. The structural
support is comprised of materials such as aluminum oxide, aluminates, silicon
carbide, silicon oxide, silicates, steatite, duranite, porcelain, or stoneware
[71]. An important feature to this approach is to first calcine the catalyst prior to
applying the coating composition, enabling the binder to be present upon introduc-
tion of the coated catalyst to the oxidation reactor. This approach will enhance the
adhesion of the oxidation catalyst to the support body. For the coating composition,
additives include binder, inert fillers, and promoters. Binders, such as polyvinyl
acetate, are effective at 10–20 % (wt% based on solids) use levels. Inert fillers, such
as SiO2, TiO2, SiC, and graphite, can be added at 0.5–25 % (by wt% based on the
calcined precursor). These inert fillers can aid in activity control of the catalyst.
Promoters, such as (NH4)2MoO4, Li2CO3, and Co(NO3)2·6H2O which are water-
soluble compounds, can also be added to the coating slurry. The coating slurry is
typically applied using a spray coater or coating drum, where support bodies in the
shape of spheres, rings, or saddles are coated in the temperature range of
50–100 C. Using this strategy, the coating mechanical stability was significantly
increased even after removal of the binder during maleic anhydride production.
Thus, sufficient adhesion of the active catalyst composition to the inert support
material existed during the entire maleic anhydride production. Another advantage
of this approach is reduced catalyst dusting enabling easier change-over processes
in the plant [71].
1.3.3.7 Morphology and Porosity
As discussed earlier, VPO is considered to be the active phase. The catalyst

microstructure, or morphology, has been found to exhibit significant influence on
its selectivity and reactivity. Empirically, selectivity is found on platelike particles
versus particles comprising more roselike shapes. By implication, the crystallo-
graphic (100) plane is likely the active surface of the catalyst on these platelike
particles [72]. To accentuate the formation of these platelike catalysts, a synthetic
process was designed. The key features of this product by process are [73]:
1. An aqueous solution of NH2OH·HCl, including H3PO4, is employed to reduce
V2O5 solid, which is indicated by forming a blue solution.
2. The solution is allowed to evaporate to form a pasty mass.
3. The pasty mass is aged to yield a blue solid.
4. The blue solid is washed with boiling water, removing any water-soluble
materials.
5. The remaining solid is dried to form VOHPO4·0.5H2O product.
6. The dry product, VOHPO4·0.5H2O, is ground to a fine powder.
7. The VOHPO4·0.5H2O powder is slurried with a mixture of dimethylformamide
(DMF) and water (H2O).
8. The dispersed VOHPO4·0.5H2O powder is recovered from the slurry and
washed with hot water.
9. The remaining VOHPO4·0.5H2O is dried to a solid, which exhibits an enhanced
(001) plane.
10. The VOHPO4·0.5H2O solid is calcined to yield (VO)2P2O7 with the selectively
exposed (100) plane [73].
Other researchers at Huntsman have sought to affect the catalyst morphology by
controllably altering the catalyst porosity. Their concept is that VPO catalysts
containing a high concentration of pores can enable a more rapid diffusion of
product and reactant gases within its structure, improving the overall activity of the
catalyst. To form pores, a solvent-removable pore-building agent, such as 1,1,1-tris
(hydroxymethyl)ethane, is incorporated in the activated catalyst, at 10 wt%, and
then removed by submersing the pellets in acetone for 6–24 h. By enhancing the
porosity of the catalyst, the maleic anhydride yield also is improved [74].
Other methods for adjusting the surface area of the catalyst, such as incorporat-
ing inorganic promoters, have been found to be useful in enhancing the catalytic
activity. For catalyst produced in organic mediums via precipitation, a typical BET
surface area ranges from 10 to 20 m2/g after activation and rarely beyond 22–23 m2/
g. The incorporation of a copromoter, such as zirconium, enables the catalyst to
develop surface areas greater than 28 m2/g. For example, when Zr
(C4H9O)4·C4H9OH (zirconium butoxide) is incorporated into the catalyst, where
the Zr/V ratio ranges from 0.005 to 0.0375 and is evaluated comparatively, the
conversions improve ~15 %, and yields increase by ~5 %. Such enhancements are
significant in commercial environments [75].
Molybdenum-modified VPO catalyst shaped bodied structures can also yield
improved surface area, activity, selectivity, and productivity. The catalyst shaped
body has a volume > 0.02 mL and a BET surface area > 15 m2/g. The molar ratio of
molybdenum to vanadium ranges from 0.0020 to 0.0060 (Mo/V). Catalyst compo-
sitions tend to slightly lower the maleic anhydride yield %, ranging from 0 to ~2 %,
but also significantly lower acrylic acid production, ranging from 35 to 65 %
reductions in gas stream [76].
Adjusted micropore structures, where the pore diameter ranges from 0.01 to
0.6 μm, can enable improved performance of the catalyst structure and morphology
by reducing the catalyst’s internal resistance to diffusion. The total pore volume of
the catalyst structure is comprised of >55 % micropores which exhibit a volume
of > 0.2 mL/g. In one example, propylene glycol was added to enhance the forma-
tion of micropores. As a result the BET surface area (m2/g) increased from 18.09 to
>30.0 with the pore volume increase > 20 %. The increase in the micropore volume
was found to be > 30 %. It was found that these improved catalyst structures
enabled a 2.1 % yield improvement over the original catalyst structures [77].
1.3.3.8 Shape Considerations in Production
The kinetic rate of catalysis is limited by the ability of the reactant gases to
internally diffuse into the pore structure of the catalyst body. By increasing the
porosity of the catalyst body, the catalytic activity can be enhanced. As a result,
pore formers, i.e., organic compounds, which burn out during later heating and
sintering, are added. However, this strategy is subject to limitation because exces-
sively porous shaped catalyst bodies will lack sufficient mechanical strength to
exhibit commercial utility.
The control of the macroscopic catalyst body shape in a way that optimizes its
performance in the reactor is another approach. For fixed-bed reactors, there are at
least four factors that control the process behavior of the catalyst bed: particle
Fig. 1.7 Recent examples of shaped bodies suitable for VPO catalysts [79, 80]
properties (i.e., shape, size, mass, etc.), container properties (i.e., shape, size,
surface properties), charging techniques (i.e., intensity, speed, methods), and
post-charging (i.e., vibration, etc.) processes [78]. In general, it is known that
catalyst activity increases as the catalyst particle size decreases and exposed surface
area increases. The pressure drop across the fixed catalyst bed tends to increase as
this particle size decreases, as the bed becomes more and more closely packed. As
the pressure drops, the amount of reactant gases that can pass through the bed
becomes more limited. The process gas flow can be improved by increasing the
catalyst size, but the activity of the catalyst can be somewhat compromised as a
result. Consequently, the focus shifts to catalyst shape. Often, the catalyst shape is
designed based upon cylindrical bodies comprising at least two parallel internal
holes which are parallel to the cylinder body axis and go right through the catalyst
structure. An example of this shaped body concept is presented in Fig. 1.7a [81].
Additional examples of common shapes include spheres, cubes, hollow cubes,
solid cylinders, hollow cylinders, single rings, cross-webs, grooved cylinders, pall
rings, Intalox saddles, and Berl saddles [78]. Recently, a shaped body structure
comprised of a solid cylinder with three void spaces running along the cylinder
height to yield three lobes was designed by Huntsman. An example of this shaped
body concept is presented in Fig. 1.7b [82]. Shape and size are optimized to achieve
enhanced surface area properties. Ultimately, a proper optimization of surface area
can contribute to minimization of required reactor volume, leading to reduced
capital expenditures.
Beyond purely shape and size, general mechanical properties are known to be
important to proper shape body selection. In shape body manufacturing, the side
crush strength is a critical factor to quality, since it is directly related to the product
durability during handling, shipment, and catalyst loading [82]. Greater crush
strengths are generally preferred. Another challenge for shape catalyst bodies is
their tendency to exhibit high attrition behavior. The term attrition describes the
amount of catalyst lost to disintegration after some amount of process wear and tear
and, as an aside, is an especially crucial property in fluid-bed reactors. Attrition
measurements are important because shaped bodies that exhibit high attrition
percentages generally yield increased, undesirable pressure drops in the commer-
cial reactor tubes.
Methods employed to load catalysts into reactor tubes are also critical to process
performance. Especially in maleic anhydride production, where thousands of reac-
tors tubes are often utilized, charging the shaped catalyst body must be consistent,
tube to tube, in order to maintain uniform pressure drops. One common technique
for charging a reactor is to fill a properly shaped sock with the catalyst, then
completely slip the sock into the tube. Once fully inserted, the sock is released
from the bottom enabling the catalyst to be placed in the tube in a controlled
manner, minimizing catalyst body motions in the tube. Another common technique
is to “pour” the catalyst into the tube. In one study, fast filling, slow filling, and
“snow storm” filling were comparatively evaluated for effectiveness in bed packing
weight and bed voidage via a “pour” filling process. “Snow storm” filling (SSF) is a
process which involves passing the packing material over wires that have been
either staggered or meshed to enable the falling particle to be interrupted prior to
reaching the bed face. This process causes the particles to radially disperse enabling
the bed face to be filled uniformly and align particles such that more are able
achieve a minimal contact area. Consequently, the SSF process can consistently
achieve greater packing weights and smaller bed voidages [78]. Once the deposition
of catalyst is achieved, further densification of the bed can be achieved by vibration
techniques in a post-deposition procedure.
Once the shaped catalyst bodies are charged into the reactor housing, the
possibility of expansion must be considered. Minimal dimensional change is
preferred. Catalyst compositions and molding processes must be evaluated for
dimensional change. For example, Amoco researchers designed a process for
preparing and shaping solid catalysts that exhibited minimal expansion character-
istics. Upon synthesis, the catalyst formed is a powdery material by grinding and
passing through a 30-mesh screen. This powder was calcined suitably in air or a
nitrogen–air combination at temperatures ranging from 300 to 370 C. The resulting
product is then formed into geometric shapes, such as cylinders, by incorporating
additional ingredients such as lubricants (i.e., graphite, Sterotex, stearic acid, zinc
stearate) and binders (i.e., starch and polyvinyl alcohol). The shaped catalyst is then
heated in a nitrogen atmosphere at temperatures ranging from 343 to 704 C. The
catalyst, in its final geometric shape, must be treated in the inert atmosphere prior to
being exposed to an oxygen-containing gas at an elevated temperature in order to
minimize its expansion characteristics. The specific size and shape of the tablets,
such as right cylinders, are critical since the available void fraction in the reactor is
determined by these parameters. The void fraction must be large enough to avoid
any large pressure drops inside the reactor. To determine if the catalyst exhibits
desirable properties, the catalyst is subjected to a standard expansion test. In a
standard expansion test, the dimensions, length and diameter, of 10 tablets are
determined. Also, the length and diameter of 10 tablets are measured with a caliper.
An average volume is estimated based upon the volume determination of a cylinder.
The tablets are placed in an oven at 482 C in the presence of a humid air stream
Table 1.7 Expansion test results for pretreated tablets exposed to different atmospheres [83]
Pretreatment Volume Hours on Reactor Conversion Selectivity
Tablet ( C) change stream temperature ( C) % (mol) % (mol)
A 427 (air) +2.18 192 422 84 64
B 427 (nitrogen) 2.23 192 422 83 64
C 427 (helium) 5.41 192 421 82 64
for 2 h. The tablets are then removed and allowed to cool inside a desiccator. The
tablet dimensions are remeasured and average volumes are reestimated. By com-
paring the average volume of the tablets, before and after thermal treatment, the
effect of the pretreatment process can be assessed. To determine effects of the
pretreatment process on the catalytic activity, the pretreated tablets were tested in a
mini-reactor test, exposing a charge of pretreated tablets to a feed of C4 in synthetic
air (1.1 mol%) at 1200 h1 VHSV with a continuous addition of 10,000 ppm water.
The data are shown in Table 1.4. The results of the expansion test are presented in
Table 1.7 [83].
Results presented in Table 1.7 indicate that the performance of pretreated
catalyst is equivalent but the expansion characteristics are superior for the catalysts
that were treated in inert atmospheres.
Solid shaped oxidation catalyst was designed by Monsanto to include at least
one void space, which results in a shape exhibiting a volume that is 30–67 % of a
comparable shape comprising no voids. In addition, the shaped catalyst, with at
least one void, exhibits an external geometric surface to volume ratio of > 20 cm1
coupled to a bulk density ranging from 0.4 to 1.4 g/cm3 [84]. Importantly, the
voided shape still must exhibit sufficient mechanical integrity to be applicable in
commercial process. Thus, the side crush strength should range from 13.3 N to
89 N. The effect of shape on the maleic anhydride oxidation process is illustrated in
Table 1.8.
In general, as more void space becomes available, the % conversion improves.
Interestingly, when the external geometric surface to volume ratio is maximized at
27.0, the % conversion and % yield are optimal. Similar trends are followed within
other examples from this research [84].
1.3.3.9 Regeneration
Catalyst materials contribute a significant and necessary expense to the maleic

anhydride process. Enhancing the longevity of catalyst when already loaded in a
reactor is very desirable to plant operators. One approach to improving reduced
catalyst activity is to regenerate, in situ, and stabilize the catalyst. In this approach,
first treatment of the aged catalyst is with a carbon halide compound, such as CCl4.
The treatment enables the old surface of the catalyst to be removed. Typically,
Table 1.8 Comparison of cylinder shape effects on the performance of VP1.15Ox catalyst [84]
Side crush Geometric External Bulk
Void Height Width strength volume surface/vol. density
Tablet space no. (mm) (mm) (N) ratio (cm1) (g/cm3)
A None 3.73 3.18 66.75 100 18.0 0.84
B Center Core 4.76 4.76 35.60 89.0 17.0 0.87
hole
C Grooves, 6 5.36 5.56 48.95 67.0 20.0 0.66
D Grooves, 3 3.97 3.96 35.60 61.0 27.0 0.64
E Grooves, 3 5.56 5.56 35.60 45.0 20.0 0.66
Reactor Wt/Wt
On Reactor charge production
stream catalyst density Bath Hot spot % g maleic
time, charge (kg/m3 temperature temperature conversion % yield anhydride
Tablet hr (kg) 102) ( C) ( C) (mol. %) (mol %) (kg cat hr)
A 3679 0.336 7.96 410 465 74.0 51.0 77.7
B 1657 0.303 7.18 439 490 79.0 50.0 84.5
C 536 0.241 5.71 424 471 77.3 50.0 106.2
D 1500 0.253 5.99 421 474 80.0 53.0 107.3
E 1042 0.261 6.18 424 488 78.0 51.0 100.0
0.01 g to 0.1 g CCl4/g catalyst is added to a nitrogen gas carrier. This composition is
passed through the catalyst bed at 350–500 C, at 100–1000 VHSV for 10–30 min.
Once the catalyst is reactivated, phosphorus compounds, such as TEP, are added to
the catalyst feed stream to optimize the P/V atomic ratio at the active catalyst sites,
thus regenerating the catalyst. It is also possible to add TEP in the presence of water
ranging from 1000 ppm to 40,000 ppm (by weight % of the reactor feed gas stream).
It was found that this approach was superior to those approaches where the catalyst
was regenerated by either process alone. This efficient technique for catalyst
regeneration leads to significant economic savings, since downtime, labor costs,
and requirements for new catalysts are minimized [85].
As we have now seen, there are many techniques and strategies researchers
employ to optimize the catalyst, the process, and the processing window. Interest-
ingly, the material selections, beyond the inorganic catalyst composition itself, are
relatively narrow in scope. The practical awareness of the key variables is high,
such as phosphate addition, shape, composition, mechanical properties, etc. The
critical challenge remains to develop an even more comprehensive understanding
of the interrelationship of these variables to the actual manufacturing processes, and
the corresponding product yield, to enable further improvements to production
processes.
1.4 Strategic Design Improvements in Existing Plants
Returning to the manufacturing facility, a number of strategies have been found to

exist in recently improved processing designs for maleic anhydride plants. More
common examples include additional reactor modifications, gas recycling, and
column preparation and improvement to solvent absorption and distillation pro-
cesses. We’ll examine each of these examples in more detail.
1.4.1 Reactor Design Modifications

1.4.1.1 High-Purity Maleic Anhydride Insertion
Insertion of process improvement agents is a popular strategy. In an effort to

improve the recovery of maleic anhydride from the oxidation reactor, researchers
at Amoco found that the addition of maleic anhydride, with critically a minimum
melting point of 52.5 C and 99.9 wt% + purity, into the reactor effluent stream
and before the stream enters the condenser can increase the recovery of maleic
anhydride from 40–45 wt% to 50–60 wt%. In addition, the fouling rate of the
condenser, due to maleic anhydride deposits on condenser surfaces, is signifi-
cantly reduced, enhancing production rates and reducing plant downtime. The
maleic anhydride is typically injected, by means of a fogging-jet spray nozzle,
into the gas stream at temperatures ranging from 93 to 177 C resulting in a
maleic anhydride gas-phase addition. The amount of maleic anhydride added to
the reactor effluent stream is at least 1 wt% of the maleic anhydride effluent
stream content [86].
1.4.1.2 Fixed-Bed Modifications
Scientific Design (SD) modified the original HSD process by inserting a second
fixed-bed reactor, in a series configuration, to improve plant yield and process
efficiency. This type of reactor configuration is commonly known as a “Series
Reactor.” The first type of series modification involved the critical insertion of a
cooling step to the C4 process after the process stream exits the first reactor. Thus,
the gas effluent from the first reactor is moderated under pressure- and temperature-
controlled conditions to a range of 50–300 C and more closely to 200 C. Then a
new, controlled amount of C4 is added to this process stream prior to the introduc-
tion into the second reactor for further oxidation. Importantly, no additional oxygen
is required at this point, indicating that the oxygen content of the second reactor is
actually lower than the first reactor. The net result is a reduction in the flammability
of the process stream. The new cooling step is essential to preventing “cold flame”
effects. The term “cold flame” is used to describe processes where the C4 reacts
with oxygen, in the absence of a catalyst, resulting in the production of polyglot

mixtures of oxidation products. Such undesirable reactions reduce the process yield
and cause problems in downstream purification processes. Operating two reactors
in series in this manner is reported to increase the effective conversion rate above
what would be possible for a single reactor. The yield of maleic anhydride can
increase 8–10 wt% and reduce waste gases, power consumption, steam usage, and
capital costs. Based on laboratory work, a parallel reactor approach was also
evaluated by operating two reactors at a 20 psig inlet pressure. This approach was
found to demand significantly more power resulting in an unrealistic economic
operation model. The series process typically yields 85–90 % butane conversion
and 94–99 % of maleic anhydride weight yield percent [27].
In a second modification to the Series Reactor concept, SD inserted a maleic
anhydride removal step between the reactors [87]. Thus, once the process stream
from the first reactor is cooled, the process stream can be treated for partial
condensation maleic anhydride products and their removal from the process
stream. Another possibility is the first reactor process stream can be scrubbed
with liquid, i.e., water, removing maleic anhydride formed in the first reactor.
The gases from this intermediate scrubber, which comprise C4 and oxygen, are
subsequently passed, along with additional, new C4 to the second reactor for
further oxidative processing [87]. From there, the process continues to treat the
process stream as previously described in the HSD process. This approach is
reported to demonstrate significant improvements to reducing raw material costs
(particularly reduced C4 consumption), power demands, and general operating
expenses [87]. It was also found that oxygen-enriched air results in improved
conversions and yields. In the case of oxygen-enriched air, the overall conversion
was reported to be 90.54 %, selectivity 70.88 %, and reaction yield 108.30 %
(wt%) [87].
In another modified fixed-bed process, which can include Series Reactor
designs, two successive reaction zones are cooled such that the temperature range
of the first reaction zone is 350–450 C and the temperature range of the second
reaction zone is 350–480 C. Importantly, the temperature differential between the
hottest reaction zone and the coolest upstream reaction zone is > 2 C. In addition,
the hot spot for the second, upstream reaction zone is hotter than any previous hot
spots. Numerous results present a process that is capable of achieving high con-
version, high selectivity, and high yield of maleic anhydride. High space-time yield
(150 g/L*h) coupled to high space velocity of hydrocarbon over the catalyst bed
reportedly enabled several thousand operating hours under stable reaction
conditions [62].
1.4.1.3 Reduction in Fluid-Bed Back-Mixing
Back-mixing is a significant challenge for fluid-bed reactors. Back-mixing

describes an effect where, within a given reactor, some of the reaction products
are recirculated throughout the bed, forward and backward. The back-mixing,
also known as backflow or reverse circulation, results in increased contact time

between the product, catalyst, and process environment. The increased contact
time can enhance degradation effects in the product [34]. In efforts to prevent this
effect in fluidized-bed reactors, a process in which both the gaseous fluid phase
and the fluidized solids move simultaneously through at least two fluid-bed
compartments in a plug flow manner was envisioned [34]. Back-mixing, which
remains a possibility within each compartment, is reduced in the total process
because the fluidized gases and catalyst advance in parallel, plug flow. At the
final stage of the reactor, the solid particles are removed from the fluid phase and
returned to initial reactor compartment [34]. To achieve the proper process flow
conditions, the linear velocity of the fluid phase ranges from 2 to 50 times higher
than the minimum fluidization velocity in each compartment. In addition, the
fluid phase within the free area of the partition(s) exhibits a linear velocity
ranging 0.8–2.0 times greater than inherent entrainment velocity of the solid
particles [34]. The various linear velocities of the system must be carefully
balanced to avoid systematic pressure drops and mechanical attrition of particles
while maintaining plug flow [34].
1.4.1.4 Process Gas Recycling
Gas recycling has been an important development to maleic anhydride production

and offers several ecological advantages leading to more sustainable production
processes. In one approach for fixed-bed reactors, the feed stream is water scrubbed,
under pressures ranging from 121 to 304 kPa, to remove maleic anhydride. The
resulting process stream is then recycled back to the reactor feed stage of the
process, again under at least 121 kPa [88]. Typical processing conditions for this
type of fixed-bed plant design are presented in Table 1.9.
Under the conditions described in Table 1.9, typical conversion rates range from
75 to 90 % and conversion selectivity ranging from 55 to 75 %. Any remaining
exhaust gases are incinerated, producing water and carbon dioxide. The benefit of
recycling becomes apparent when evaluating the specific parameters presented in
Column B. Under those conditions, operating the plant without recycling was found
to reduce process yields from 1.03 to 0.94 Kg. The decrease in performance was
Table 1.9 Processing conditions for a fixed-bed gas recycling plant design [88]
Process parameter A B
Reactor temperature ( C) 370–440 400
Reactor inlet pressure (kPa) 203–608 344
Space velocity (h1) 1000–3000 1500
Total C4 concentration fed to reactor (% volume) 1.6–3.0 2.1
% oxygen in reactor feed (% volume) 11–16 12.2
Total yield (kg of maleic anhydride/kg of C4) 0.9–1.05 1.03
Yield (% mol) NA 60.7
attributed to recovering C4 which would otherwise be lost to incineration and,

curiously, reducing the available oxygen in the reactor to ranges of 11–14 % versus
air, which is 21 % by volume. The net result of recycling is that only a fraction of
the reaction gases are conveyed to the incinerator. Advantages to gas recycling can
include reduced incinerator dimensions, compression energy savings (~30 %), and
a significant reduction of carbon dioxide release into the atmosphere (~270 kg per
ton of maleic anhydride produced) in addition to the previously mentioned yield
improvement [88].
In considering the above approach, it was noted that a disadvantage of the
process was its proximity to the explosive-range boundary conditions [89, 90]. It
was also noted that only by using an oxygen concentration closer to 10 % (volume)
and/or C4 concentrations at the lower end of the range (~1.6 % by volume) was a
safe condition possible [90]. However, since the potential value of exhaust gas
recycling was clearly demonstrated, more efforts to demonstrate safer parameters
were undertaken.
To further develop a recycling approach, where exhaust gases are recirculated
after removing maleic anhydride product, new processing parameters were consid-
ered. These parameters are presented in Table 1.10.
Under the conditions described in Column D of Table 1.10, including addition of
1.5 ppm TEP (by volume) and converting between 50 and 75 % of C4 within each
reactor pass, similar process yields were achieved at 79.4 % recycling rates. Impor-
tantly, this process operates within safer operational parameters [90]. A compara-
tive proximity of these processes to the flammability region is more easily seen in
Fig. 1.8.
Figure 1.8 graphically presents the flammability region of gaseous mixtures
comprising molecular oxygen, C4, and balanced with nitrogen at 100 kPa at
20 C. As previously indicated, Point B is very near the boundary, whereas Point
D is comfortably located some distance from the flammable region.
Table 1.10 Processing conditions for a fixed-bed gas recycling plant design [90]
Process parameter C D
Reactor temperature ( C) 380–440 419
Reactor inlet pressure (kPa) 600–1000 850
Space velocity (h1) 3000–6000 4400
Total C4 concentration fed to reactor (% volume) 1–1.3 1.2
% oxygen in reactor feed (% volume) 8–13 8.7
Total yield (kg of maleic anhydride/kg of C4) NA NA
Yield (%, mol) NA 59.2
Fig. 1.8 Flammability diagram for mixtures of C4 and oxygen balanced with nitrogen [89]
1.4.1.5 Solvent Absorption and Distillation
Efficient separation of the maleic anhydride product from the process stream is an
essential feature to any economical plant process. As we saw earlier, strategies for
separating some maleic anhydride product include partial cooling of the reactor
effluent, cooling to the maleic anhydride dew point, and removing some maleic
anhydride product. Water absorption is also employed to remove maleic anhydride
product from the reactor effluent. However, some have noted drawbacks to water
absorption [91]. These include expensive distillation processes, where water and
dehydration of maleic acid requires significant energy, and loss of product, where
maleic acid conversion to fumaric acid reduces process yields. Equipment fouling
can also increase which leads to more plant maintenance [91]. Conversely, organic
solvent processing can eliminate many of these drawbacks. Previous processes have
demonstrated the effectiveness of maleic anhydride solvent-based recovery opera-
tions [92]. However, solvent methods are also known to accumulate undesirable
by-products in the absorbing and stripping loops. Contaminants such as acrylic
acid, fumaric acids, maleic acids, etc. as well as polymeric tars can build up and
cause significant problems over time.
In one approach to improving solvent processing, the reactor effluent is brought in
contact with an organic solvent, such as dibutyl phthalate. The water found in the
maleic anhydride-enriched solvent is removed by stripping it off with a low-relative
humidity gas and/or adsorbing agent. In the case of a low-humidity gas, the gas is
brought into contact with the enriched solvent at elevated temperature (>80 C) and
between 1 kPA and 200 kPa. These conditions enable the water to evaporate from the
maleic anhydride-enriched solvent. Suitable gases, with water less than 15 %, are air,
carbon dioxide, and/or nitrogen. Adsorbing agents, such as zeolites, are also effective
in the dehydration process. This process is reported to reduce decomposition of the
solvent, improve yields of maleic anhydride, and enable commercial production of
maleic anhydride-enriched solvent streams with water contents less than 0.1 % [91].
Observers found that the process above typically resulted in maleic anhydride
purity in the range of 97–99 %. The main residual was found to be trace levels of
solvent, organic by-products (i.e., acrylic acid and carbonyl compounds), and
water. To reach even higher levels of purity, another process was designed which
incorporates a three-stage distillation process. In a simplified version of this solvent
process, the maleic anhydride effluent from the reactor is brought into contact with
organic solvent in the absorber. From the base of the first absorber, the maleic
anhydride-enriched solvent stream is sent to a first vacuum separation column. In
this first column, the top section is separated under stripping conditions, enabling
the overhead vapors that form condensates to be fed to a second absorber, which is
also countercurrently fed with the same organic solvent. Vapors that do not
condense here are sent to disposal. The condensed solvent phase is sent on to the
first fractionation column. The overhead containing impurities such as acrylic acid
from the first fractionation column is sent directly to the second absorber. The
bottoms of this absorber are returned to the original oxidation reactor absorber for
recycling. The bottoms of the first fractionation column feed to the second frac-
tionation column. It is from this second fractionation column that the high-purity
maleic anhydride is obtained. Purities of 99.0 + % are achieved [93].
In another approach, purification of the oxidation gas stream, which comprises
maleic anhydride, is again absorbed by a solvent, i.e., dibutyl phthalate. The maleic
anhydride is then separated from the solvent by stripping in the presence of
hydrogen gas in a column. The advantage of this approach is that the hydrogen
gas cools the liquid phase which reduces the formation of by-products like fumaric
acids. Another reported benefit is that the overall column diameter can be smaller,
which reveals significant cost advantages [94].
Other research efforts enabled the development of a polymeric tar removal strategy.
The approach was to first cool the effluent from the oxidation reactor and bring it into
contact with the maleic anhydride solvent absorbent, i.e., dibutyl phthalate, in the
absorber. The enriched solvent comprises maleic anhydride and contaminants. Maleic
anhydride is then stripped from enriched solvent resulting in two process streams: one
mainly maleic anhydride and the other comprised of solvent and contaminants. The
solvent/contaminant mixture is brought into contact with agitated water, which is largely
immiscible with the solvent. Polymeric tars present in the solvent are hydrolyzed and
extracted into the aqueous phase. The solvent raffinate phase is substantially reduced of
the polymeric tars and recycled back to the maleic anhydride absorption zone [95].
The formation of tars in the absorber and the degradation of solvent have also
been found to be affected by the presence of catalyst fines. These catalyst fines,
which result from attrition of the shaped catalyst body, exhibit an ability to
promote hydrolytic degradation of solvents such as alkyl phthalates. This degra-

dation leads to tar formation in the solvent circuit. The process of catalyst
attrition, once under way, can become accentuated and self-accelerating as the
fines are continuously removed from a reactor tube. As the numbers of reactor
tube fines become reduced, resulting in greater void spaces in the upper tube
zone, there can be a decrease in the resistance inside the reactor tube. The
increase in the tube gas flow, or gas velocity, coupled to a catalyst loss in the
upper portion can ultimately lead to a fluidization of the fixed-bed or shaped
catalyst expansion leading to greater abrasion effects, against both the tube walls
and neighboring catalyst shaped bodies, further promoting catalyst attrition. To
overcome these issues, two strategies are employed. First, a restraining bed
composed of discrete material substantially denser than the catalyst is placed
on the catalyst column top in each reactor tube. For example, 0.25-inch spherical
Norton Denstone bodies exhibiting a density of 1.44 g/mL are placed on top of
the VPO catalyst with a bulk density of 0.65 g/mL [96]. This will help to
immobilize the catalyst bed. Tension springs, or other mechanical means, can
also be employed to apply as a supplemental immobilizing force to the fixed
catalyst bed. Second, fines typically < 1000 μm are removed from the catalyst
bed by deploying a purging gas that is capable of dislodging and removing
catalyst fines from the bed. A temporary, porous filter in direct communication
with the exiting purge gas is employed to remove the fines from the process
stream. The most effective removal of fines is utilizing the purging gas in an
intermittent manner, effectively pulsing the purge gas in the catalyst bed. By
passing pulsing gas through the tubes, one removes not only the fines capable of
finding a clear exit path but also particles which are entrapped in various crevices
of the catalyst bed. Essentially, this process is a de-dusting of the catalyst
bed [96].
In a recent development, a process was designed by BASF with the intent to
simplify maleic anhydride purification while maintaining high-purity levels. An
important feature to this process is the dividing wall purification column presented
in Scheme 1.9.
The purification column is comprised of three chambers, labeled 1, 2, and 3. The
column is fed with crude maleic anhydride, via inlet #10, into chamber 2A. The
purified maleic anhydride is collected in the chamber labeled 2B. Chamber 2A and
2B are separated from each other. Both chambers are open at the top to chamber
3 and bottom to the lower chamber 1. Within chambers 2A and 2B are a dividing
wall and trays for amassing and dispersing materials, in rectified regions, ensuring
distilling-type flow through the column. Vapors at the top of chamber 3 are fed into
an overhead condenser labeled #4, where the condensate can be recycled.
Remaining gases and liquids are discharged via #13. Material bottoms in chamber
#1 are either recycled or discharged via #12. In Example #1, five theoretical plates
are found in chambers #1 and #3, and chambers #2A and #2B comprised 12 theo-
retical plates, due to packing elements from Montz (A3-750 Hilden). The chamber
pressure was operated at 30 mbar. The temperature of the chamber 1 was 200 C
(4)
(13)
(3)
(11)
(2)
(2A) (2B)
(10)
(1)
(12)
Scheme 1.9 A dividing wall column for maleic anhydride purification [97]
and 3 was 93 C. Dibutyl phthalate solvent absorbent was employed where the feed
rate of the enriched solvent was 1400 g/h and draw rate was 86 g/h. Under these
general conditions, the collected material exhibited the properties found in
Table 1.11.
As illustrated above, the distillation process yields 99.9+ % maleic anhydride,
with very low levels acrylic acid and acetic acid. Additionally, the overall simpli-
fication of the equipment scheme is of great benefit to plant personnel [97].
Table 1.11 Maleic anhydride distillation purity from a divided wall column [97]
Compound Collected result Compound Collected result
Acetic acid NA Monobutyl maleate 0.0185
Acrylic acid NA Monobutyl fumarate NA
Maleic anhydride 99.9162 Phthalic anhydride 0.0027
Citraconic anhydride NA Dibutyl maleate NA
Benzoic acid 0.0001 Dibutyl fumarate NA
Dimeric acrylic acid NA Phthalic acid 0.0014
Maleic acid 0.0070 Monobutyl phthalate 0.0002
Fumaric acid 0.0006 Dibutyl phthalate 0.0528
Citraconic acid NA Unknown 0.0528
Color (APHA) 15 Water NA
1.5 Improving Product Quality and Safety
Efforts to improve the quality of maleic anhydride by reducing color and improving
thermal stability in addition to increasing purity are of critical industrial importance.
The treatment of maleic anhydride with chemical additives can be technically com-
plicated and leads to “impurities” that are intentionally introduced into downstream
maleic anhydride production processes. As a result, strategies that focus on additives
and/or processes continue to be developed in efforts to improve product quality.
1.5.1 Distillation in Designed Atmospheres
To improve the quality of crude maleic anhydride obtained from production of

phthalic anhydride via o-xylene, the material was distilled in the presence of air,
which contains molecular oxygen. The oxygen ratio to crude maleic anhydride
ranges from 1 mmol/mol to 15 mmol/mol and is performed at temperatures ranging
from 60 to 180 C, in vacuo. After separation from the bottoms, comprised of
maleic anhydride, citraconic anhydride, phthalic anhydride, benzoic acid, and other
miscellaneous impurities (at a ratio of 27:18:23:27.5:4.5), the molten maleic anhy-
dride was found to exhibit improved color stability, even over several days [98].
Highly pure, color-stable maleic anhydride with long shelf life stability was also
prepared by fractional distillation of crude maleic anhydride, separating the desired
product from low-boiling and high-boiling materials. The high-boiling materials
are removed by partially condensing the maleic anhydride vapor, so 0.5–15 % by
weight of maleic anhydride vapor is separated off as a liquid pre-condensate. The
remaining vapor, which is highly pure maleic anhydride vapor, is then condensed
into the product. This process is carried out downstream of conventional distillation
of crude maleic anhydride. The process yields a maleic anhydride product which is
color stable, a purity 98 + % by weight, and a color number (Hazen) < 40 on
heating. Stored in liquid form, this product is shelf stable for 6–8 weeks [99].
In another distillation process, the crude maleic anhydride is introduced into the
stripper column while removing the overhead stream from the stripper column.
Meanwhile, the bottom stream is removed from the stripper column and introduced
into the refining column where portions of the overhead low- and middle-boiling
vapors are removed. Low-boiling materials range from 45 to 155 C at 760 mm Hg,
while middle-boiling materials exhibit boiling temperatures ranging from 160 to
285 C at 760 mm Hg. A portion of this overhead vapor originating from the middle-
boiling materials is condensed and directed to a refining column as a reflux. Finally, in
the refining column, a side draw stream is collected and is at least 99.90 wt% maleic
anhydride, while the overhead vapor is removed. The process is operated in an inert
gas, such as nitrogen, and can also employ oxygen (<5 mol%). The inert gas treatment
aids in purging the non-condensable materials from the crude maleic anhydride [100].
1.5.2 Additives for Distillates
An additive approach to product quality improvement, where maleic anhydride is

distilled in the presence of tridecyl phosphate (TDP) ranging from 100 to 200 ppm as
an antioxidant, was developed. After TDP addition, the crude maleic anhydride is
thermally treated at 120–190 C for 1–10 h. After thermal treatment, the distillation
process begins at pressures of 50–100 mm Hg. The bottom temperature of the
distillation tower ranges from 120 to 160 C. To the distillate are the additions of
n-propyl gallate (1–100 ppm), cuprous chloride (0.1–5 ppm), and zinc chloride
stabilizers (0.1–5 ppm). Key to their effectiveness is good mixing. Zinc chloride is
added to improve the maleic anhydride thermal stability. The effect of these stabilizer
packages on maleic anhydride color and thermal stability are presented in Table 1.12.
In Table 1.12 Example B, the specific addition levels of n-propyl gallate
(10 ppm), cuprous chloride (1 ppm), and zinc chloride stabilizers (0.5 ppm) enabled
molten maleic anhydride color (APHA) results as follows when directly compared
to no additive package [101].
The prior additive method was observed to be insufficient in meeting the stringent
quality requirements when solid maleic anhydride is stored for long time periods.
Phenols and catechols additives are known to have discoloration tendencies, espe-
cially if used in large quantities, and may become ineffective over time by exposure to
light and oxygen [102]. As a result, monochloro-substituted aliphatic alcohols, such as
2-chloro-1,3-propanediol and 3-chloro-1,2-propanediol, were found to be effective
stabilizing additives, when used in the presence of cuprous chloride and zinc chloride
stabilizers. Additive use levels for chlorine-substituted alcohols range from 5 to
60 ppm, copper compounds from 0.25 to 1 ppm, and zinc compounds from 0.1 to
0.5 ppm. In a specific example, presented in Table 1.12, additive levels for Example C
were 20 ppm for 5-chloro-1-pentanol, 1 ppm for copper acetate, and 0.2 ppm for zinc
acetate [102]. Again, it was found that the stabilizer package improved the thermal and
color stability of the purified maleic anhydride product.
Another method to improve maleic anhydride production quality is to minimize
the amount of impurities generated in the oxidative reactor stream. For example,
Table 1.12 Results of additives for color and thermal stability [101, 102]
Example Stabilizer Molten color (APHA) Heat-melted color (APHA)
Stored in Stored in
Immediately After Immediately liquid state solid state
after addition 30 days after addition (30 days) (30 days)
A – 5 1 300 >500 >500
B Added 5 10 10 10 15
C Added 5 5 10 NA 35
trialkyl phosphite or trialkyl phosphate can be continuously added to reactor feed

gases, in amounts ranging from 0.5 to 4 ppm (by wt. of elemental phosphorus). The
reactor temperature is ranging from 400 to 440 C, with a maximum temperature of
480 C. Operating the reactor under these conditions results in conversion to maleic
anhydride of ~84 + % and acrylic acid formation of < ~1.5 % [103].
1.5.3 Removal of Impurities via Precipitation
BASF employed a precipitation concept, where the reduction of fumaric acid

deposits on plant equipment was envisioned to significantly improve product
efficiency, reduce downtime and cleanup efforts, as well as improve product
quality. Fumaric acid is formed from the isomerization of maleic acid, which is
formed from the reaction of water and maleic anhydride. A key element to this
process is a controlled precipitation of fumaric acid, where maleic anhydride-
enriched phthalate-derived absorbent in temperatures ranging from 100 to 150 C
is cooled to a range of 30 to 70 C. Fumaric acid exhibits a temperature-dependant
solubility limit in dibutyl phthalate, the maleic anhydride-absorbing solvent in this
case. As a result, cooling the solvent liquor enables significant amounts of fumaric
acid precipitate to be removed from the purification and production process [104].
1.6 Safety
Mitsubishi Chemical has worked to establish safer operating processes for

fluidized-bed reactors by studying a variety of process parameters and their inter-
relationships. This effort has enabled the determination of a safety index, F. The
safety index, F, is defined in Eq. (1.5): [105]
F ¼ FL FR > 0 ð1:5Þ
where FL is determined by Eq. (1.6):
FL ¼ C=CT =Co ð1:6Þ

where C is concentration (volume %) of flammable gas in the reaction product gas.

Due to the numerous gases in this feed stream, C is actually
X n the summation of each
concentration flammable gas, i.e., C1, C2, etc. or C ¼ C . Co represents the
i¼1 i
concentration of oxygen gas (volume %) in the reaction product gas as represented
in Eq. (1.7):
Concentration ðVol: %Þof Oxygen in Reaction Product Gas

Co ¼ 100 ð1:7Þ
ð100 CÞ
CT is defined as the concentration of the flammable gas components in the

reaction product gas which can be completely combusted with the amount of
oxygen present in the gas feed. It is determined from Eq. (1.8):
100
CT ¼ h nnX n Xn o o i; ð1:8Þ
1þ i¼1
ð RixC i Þ= i¼1
Ci =C o 100
and FR is determined from Eq. (1.9):
FR ¼ 2:319 105 T 2 1:688 102 T þ 3:288 þ ðP 0:15Þ 0:3 ð1:9Þ
where T is the reaction product gas temperature ( C) and P is the gauged reaction
pressure in MPa. In numerous experimental examples, using these relationships, it
was possible to discern between explosive and nonexplosive processes, providing
another technique for improving the safe operation of fluidized-bed reactors.
Substantial data was also reported for a variety of experimental process conditions
coupled to safety index data in support of this research effort [105].
To further develop the fluid-bed process, process parameters were developed
that were constrained to empirical mathematical relationships. First, a safety coef-
ficient (α) was developed, where α ranges between 0 and 10. The safety coefficient
is determined by equations presented in (1.10):
α ¼ 10:51 þ 51:22D 35:35D2 E where

C 100A
D¼ and E ¼ ð1:10Þ
BþC 100 B C
where the reactor exit gas stream is comprised of, in volume %, a maleic anhydride
concentration > 2 %, an oxygen concentration (A), a hydrocarbon concentration
(B), and a CO concentration (C) that satisfies the constraints of 0 < α < 10. Using
these relationships enhances the overall process safety and efficiency of the fluid-
bed maleic anhydride plant [106].
1.7 Other Production Routes
Within the known reactor technologies, numerous challenges remain unmet. While
there is continuing focus on hot spot control, minimization of catalyst thermal
degradation, catalyst attrition, and avoidance of thermal runaway episodes, the
basic conditions and challenges of current reactor technology are relatively stable
relying on operation in safe operation reaction zones to achieve economically
viable results. Within these confines, there has been renewed effort to design
even more capable processes for maleic anhydride production. Among these pro-
cesses are circulating fluidized-bed reactors, microchannel technology, and bio-
based production.
1.7.1 Circulating Fluidized-Bed (CFB) Reactor
Other processes are known for maleic anhydride production. Perhaps the most well-
known alternative process is the circulating fluidized-bed (CFB) reactor developed
and commercially demonstrated by DuPont [107]. Scheme 1.10 presents a sche-
matic of the CFB reactor.
Interpreting Scheme 1.10 beginning at the reactor feed, C4 and nitrogen gases in
tandem with the catalyst rise through the riser pipe in plug flow. Upon their exit, the
gases are separated from the catalyst in the separator/stripper, where any carbona-
ceous species are removed. The product stream continues on as in the other
processes once completely separated. The reduced catalyst, on the other hand, is
fed to the fluid-bed regenerator to be reoxidized to then return to the base of the
riser. In this manner, the process is continuously deploying “fresh” catalyst. The
CFB approach sought to eliminate several challenges for typical fixed-bed and
fluid-bed systems. As with fluid-bed processes, the CFB heat transfer is very good,
but, in this case, the system operates in plug flow within the riser resulting in
minimal back-mixing. One key improvement is the reduction in catalyst attrition,
which results in selectivity losses when attrition becomes severe, a common issue in
fluid-bed reactors. The selectivity of the catalyst to C4 to maleic anhydride is
preferred to be in an inert atmosphere, such as nitrogen. Note that the process in
the riser is substantially free of oxygen, but in the fluid-bed regenerator, where the
catalyst is oxidized (regenerated), there is 20 mol% oxygen and 80 mol% helium.
But before the regenerated catalyst returns to the riser, the catalyst must be stripped
to minimize the amount of gas-phase oxygen when the catalyst communicated with
C4. In one specific example, where the CFB process is operated at 438 C, the
residence time for C4/nitrogen [3:97 (mol %)] gases in catalyst contact contained in
the riser was 4 s, and the fluid-bed regenerator feed gas was O2/helium [20:80
(mol %)] with a residence time of 3 s, enabling 83.3 % conversion and 60.8 %
selectivity [108]. In 1996, an 82,000 tons/year CFB plant was started in Asturias,
Spain, for producing THF using this process for C4 to maleic acid [109].
Separator/Stripper
Product
Reduced
Catalyst
Fluid Bed
Regenerator
H.P.Steam Riser
Air
Oxidized
Catalyst
Reactor Feed
Riser Reactor
Scheme 1.10 Circulating fluidized-bed (CFB) reactor schematic [107] (Reprinted from Chemical
Engineering Science, 54, Contractor, R.M., “Dupont’s CFB technology for maleic anhydride,”
5627–5632, Copyright 1999 with permission from Elsevier)
1.7.2 Microchannel Reactor
In 2007, BASF employed a microchannel reactor for C4 oxidation demonstrating

that a reactor could be safely operated in the explosive range and produce an above-
average yield of maleic anhydride [110]. Because the microchannel reactor is
dimensionally so small, where in at least one direction the reactor is <3 mm, the
ability of fire to spread is not possible, implying a much safer process. With
flammability issues minimized or significantly reduced, there is a significant increase
in the design freedom for the processing parameters: gaseous ratios (organic/
oxygen), reactor pressure design beyond maximum explosion pressures, and improved
mass/heat transfer (enabling suppression of hot spots) [110]. A general comparison
of fixed-bed and microchannel process parameters is presented in Table 1.13.
Table 1.13 Comparison of fixed-bed and microchannel process parameters [88, 110]
Fixed-bed B
Process parameter (see Table 1.9) Microchannel
Reactor temperature ( C) 400 380–440
Reactor inlet pressure (MPa abs) 0.344 ~0.101
Space velocity (h1) 1500 3000–8000
Total C4 concentration fed to reactor (% volume) 2.1 2–20
% oxygen in reactor feed (% Volume) 12.2 20–98
Total yield (kg of maleic anhydride/kg of C4) 1.03 NA
Yield (%, mol) 60.7 ~70 %
Scheme 1.11 Transformation of maleic anhydride to dimethyl maleate [111]
1.7.3 Bio-based Reactants
To generate even more value from existing maleic anhydride production facilities,
plants capable of producing other commercially relevant compounds than can be
derived from maleic anhydride, such as butane-1,4-diol (B1D), γ-butyrolactone
(BLO or GBL), and tetrahydrofuran (THF), are envisioned. To accomplish this
additional production, an important intermediate step is to react maleic anhydride
with an alkyl alcohol, such as methanol, to ultimately produce dimethyl maleate.
This chemical reaction scheme is presented in Scheme 1.11.
Once the dimethyl maleate is formed, the maleated material can undergo a
vapor-phase hydrogenation using a copper chromite catalyst or a promoted copper
catalyst [111, 112]. The result of this key hydrogenation step is the production of
B1D, GBL, and THF [111].
In 2013, Arkema revealed a process for maleic anhydride production based on
butanol via a fermentation process versus the now more traditional C4 approach. In
developing the bio-based strategy, organic matters, such as sugars, starches, and
cellulosics, as well as vegetable matter, which are all renewable raw materials, are
fermented in the presence of microorganisms, such as Closlridium and Closlridium
acelobuzylicum. Among the mixture of products produced are butanol and acetone,
which are separated, usually by distillation, to isolate butanol. Butanol is then
oxidized to maleic anhydride at temperatures ranging from 300 to 600 C using
traditional VPO catalysts. To determine the actual source of the maleic anhydride
products, maleic anhydride prepared from renewable sources contains 14C, which is
a component of the Earth’s atmosphere. Thus, isotopes are convenient composi-
tional feature to determine the ultimate source of the maleic anhydride
product [113].
Acknowledgments The authors would like to express their sincere gratitude to Frank Fusiak,
Larry Lake, Jude Ruszkay and Laurence Senak, all of Ashland, Inc., and Professor Robert D. Pike,
of The College of William and Mary, for their many helpful conversations/suggestions and
thorough review of the manuscript prior to publication.
References
1. Ashland Product Stewardship Summary, Maleic Anhydride, 16 Sept 2013

2. DSM Product safety summary, Maleic Acid, 8 Aug 2011
3. The Merck index, 9th edn. Merck & Co, Rahway, NJ (1976)
4. Lohbeck K, Haferkorn H, Fuhrmann W, Fedtke N (2012) Maleic and fumaric acids.
Ullmann’s Encycl Ind Chem 22:145–155
5. Wilhoit RC, Shiao D (1964) J Chem Eng Data 9(4):595–599
6. Lohbeck K, Haferkorn H, Fuhrmann W, Fedtke N (2012) Maleic and fumaric acids.
Ullmann’s Encycl Ind Chem 22:145–155
7. Acree W, Chickos JS (2010) J Phys Chem Ref Data 39(4)
8. Dahlgren G, Long FA (1960) J Am Chem Soc 82(6):1303–1308
9. OSHA (1992) n-Butane
10. Paper #1-13 Natural Gas Liquids (2011) Working document of the National Petroleum
Council North American Resource Development Study made available 15 Sept 2011
11. Moore BJ, Sigler S (1987) Analyses of natural gases, 1917–85. Bureau of Mines, Information
Circular 9129
12. Marathon Petroleum Corp (2012) Annual Report
13. Ashland Performance Materials Neal Plant (2011) Title V Fact Sheet R30-09900009-2012 to
the State of West Virginia, 2011
14. Trivedi BC, Culbertson BM (1982) Maleic anhydride. Plenum Press, New York
15. Felthouse TR, Burnett JC, Horrell B, Mummey MJ, Kuo Y-J (2001) Maleic anhydride, maleic
acid, and fumaric acid. In: Kirk-Othmer encyclopedia of chemical technology. Wiley,
New York
16. Mirasol F (2012) US chemical profile: maleic anhydride. ICIS.com. Accessed 23 Apr 2012
17. Trivedi BC, Culbertson BM (1982) Maleic anhydride. Plenum Press, New York, p 17
18. Sunderland P (1976) Ind Eng Chem Prod Res Dev 15(2):90–99
19. Chemical Weekly (2010) 11 May 2010, pp 185–190
20. Hydrocarbon Process 176 (1971)
22. Malow M (1980) Hydrocarbon Process 149–153
23. Malow M (1985) Environ Prog 4(3):151–154
24. Malow M Environ Prog 4(3):151–154
25. Burnett JC, Keppel RA, Robinson WD (1987) Catal Today 1:537–586
26. Doshi BM (1999) Production of maleic anhydride. US 6,194,587 B1
27. Padia AS, Click, GT (1994) Two stage butane maleic anhydride process. US 5,360,916
28. Padia AS, Bortinger A, Mazzoni G, Monti T (2009) Reduction of acrylic acid in the
production of maleic anhydride. US 7,619,099 B2
29. Schmidt J (1978) Two stage maleic anhydride process. US 4,116,983
30. Daniel C (1984) Manufacture of maleic anhydride. US 4,429,136
31. Wellauer TP, Cresswell DL, Newson EJ (1986) Chemical Eng Sci 41(4):765–772
33. Ballarini N, Cavani F, Cortelli C, Ligi S, Pierelli F, Trifiro F, Fumagalli C, Mazzoni G, Monti
T (2006) Topics Catal 38(1–3)
34. Suciu GD, Paustian JE (1987) Process for preventing backmixing in a fluidized bed vessel.
US 4,691,031
35. Suciu GD, Stefani G, Fumagalli C (1986) Production of maleic anhydride. US 4,594,433
36. Neri A, Sanchioni S (1982) Process for the continuous separation of maleic anhydride from
process gases. US 4,314,946
37. Emig G, Martin F (1987) Catal Today 1:477–498
38. Wellauer TP (1985) Optimal policies in maleic anhydride production through detailed reactor
modelling. Dissertation, ETH Zurich
39. Emig G, Martin F (2007) ICIS Chem Bus:58, 17–23 Sept
40. ICIS (2012) US chemical profile: maleic anhydride
41. Zavalij PY, Whittingham MS (1999) Acta Cryst B55:627–663
42. Bordes E (1987) Catal Today 1:499–526
43. Bordes E (1987) Catal Today 1:499–526
44. Boudin S, Guesdon A, Leclaire A, Borel M-M (2000) Int J Inorg Mat 2:561–579
45. Nguyen PT, Hoffman RD, Sleight AW (1995) Mater Res Bull 30(9):1055–1063
46. Geupel S, Pilz K, van Smaalen S, Büllesfeld F, Prokofiev A, Assmus W (2002) Acta Cryst
C58:9–13
47. Busca G, Centi G (1989) J Am Chem Soc 111:46–54
48. Schiott B, Jorgensen KA, Hoffman R (1991) J Phys Chem 95:2297–2307
49. Schiott B, Jorgensen KA, Hoffman R (1991) J Phys Chem 95:2297–2307
50. Cheng M, Goddard WA (2013) J Am Chem Soc 135:4600–4603
51. Keppel RA, Mitchell SF, Mummey MJ (1998) Supression of autoignition in maleic anhydride
production. US 5,734,066
52. Padia AS, Sachs H, Millman M (2009) Yield improvement in the production of maleic
anhydride. US 7,589,217 B2
53. Takayasu O, Nakajima A, Hashiba H (2001) Method for production of maleic anhydride. US
6,194,588 B1
54. Taheri H (1992) Continuous process for the production of maleic anhydride from a C4-
hydrocarbon feedstock. US 5,117,007
55. Burnett JC, Alumbaugh WH, Reeves LA (2001) Method and apparatus for improving the
uniformity of distribution of a phosphorous-containing agent throughout a maleic anhydride
catalytic reactor. US 6,300,505 B1
56. McCandless HA, Cearley JL, Taheri H (1990) Process for the manufacture of maleic
anhydride utilizing peroxides to improve catalytic activity. US 4,950,769
57. Ushikubo T, Oshima K (1995) Method for the production of maleic anhydride. US 5,449,792
58. Dobner C, Duda M, Raichle A, Wilmer H, Rosowski F, Holzle M (2008) Catalyst and
methods for producing maleic anhydride. US 2008/0227992 A1
59. Hibst H, Glaum R, Benser E (2010) Polynary vanadyl pyrophosphate. US 2010/0105927 A1
60. Ebner JR, Keppel RA, Mummey MJ (1999) High productivity process for the production of
maleic anhydride. US 6,005,121
61. Yamamoto H, Yoneyama N (1983) Process for producing conjugated diolefins. US 4,423,281
62. Weiguny J, Storck S, Tenten A, Ruppel (2004) Method for producing maleic acid anhydride.
US 6,803,473 B2
63. Hibst H, Noe R, Exner KM, Duda M (2006) Method for producing maleic anhydride. US
6,989,454 B2
64. Haddad MS, Meyers BL, Taheri H, Wolfe PA (1998) Catalysts for the production of maleic
anhydride by the oxidation of butane. US 5,792,722
65. Bertola A, Cassarino S, Nsunda VK (2001) Process for the preparation of improved
vanadium-phosphorus catalysts and use thereof for the production of maleic anhydride. US
6,174,833 B1
66. Groke D, Bosch R, Lotz J, Eberle HJ (2002) Method for producing catalysts for synthesizing
maleic anhydride by means of gas phase oxidation. US 6,407,030
67. Haddad MS, Goeden GV (2009) Phosphorus addition process for improvement of catalysts
suitable for maleic anhydride production. US 7,629,286 B2
68. Albonetti S, Cavani F, Ligi S, Mazzoni G (2005) Vanadium/phosphorous mixed oxide
catalyst. US 6,956,004 B2
69. Altwasser S, Dobner CK, Wilmer H, Rosowski F (2012) Maleic anhydride synthesis catalyst
precursor and process for its preparation. US 2012/0149919 A1
70. Haddad MS, Eryman WS (1992) Maleic anhydride catalysts. US 5,134,106
71. Groke D, Ruedinger C, Eberle HJ, Jering R, Bosch R (2000) Process for producing coated
catalysts for the synthesis of maleic anhydride by gas-phase oxidation. US 6,048,987
72. Okuhara T, Misono M (1993) Catal Today 16:61–67
73. Datta A, Dasgupta S, Agarwal M (2004) Process for preparing vanadyl pyrophosphate
catalyst. US 6,774,081
74. Forkner MW (2012) Maleic anhydride catalyst and method for its preparation. US 8,143,461
B2
75. Andrews WJ, Ebner JR, Felthouse TR (1994) Catalysis for the production of maleic anhy-
dride containing vanadium-phosphorous oxide with selected promoter elements. US
5,364,824
76. Felthouse TR, Keppel RA, Schaefer CJ (1999) Molydenum-modified vanadium-phosphorous
oxide catalysts for the production of maleic anhydride. US 5,945,368
77. Shan Z (2011) High pore volume vpo catalyst for maleic anhydride production. US 2011/
0201830 A1
78. Afandizadeh S, Foumeny EA (2001) Appl Thermal Eng 21:669–682
79. Dobner CK, Altwasser S, Wilmer H, Rosowski F (2011) Catalyst and method for producing
maleic anhydride. US 2011/0257413 A1
80. Melde LE, Smith WA (2011) Modified trilobe shape for maleic anhydride catalyst. US 2011/
0201829 A1
81. Dobner CK, Altwasser S, Wilmer H, Rosowski F (2011) Catalyst and method for producing
maleic anhydride. US 2011/0257413 A1
82. Melde LE, Smith WA (2011) Modified trilobe shape for maleic anhydride catalyst. US 2011/
0201829 A1
83. Haddad MS, Meyers BL, Eryman WS (1992) Maleic anhydride production. US 5,095,125
84. Ebner JR, Keppel RA (1992) Shaped oxidation catalyst structures for the production of
85. Edwards RC (1990) Process for preparation of maleic anhydride. US 4,918,201
86. Ceisel SC, Conrad JF, Lestan EM, Nelson AP (1990) Process for the continuous separation of
maleic anhydride from process gases. US 4,941,895
87. Doshi BM (2001) Production of maleic anhydride. US 6,194,587 B1
88. Ruggieri R, Cassarino S (1997) Process to recycle exhaust gases from n-butane conversion
into maleic anhydride. US 5,688,970
89. Palmer DA, Holzhauer JK (1987) Process for production of maleic anhydride. EP0099431 B1
90. Duda M, Machhammer O, Weck A (2008) Method for the production of maleic anhydride.
US 7,345,167 B2
91. Bertola A, Ruggieri R (1991) Process of recovery of maleic anhydride from reaction gaseous
mixtures. US 5,069,687
92. White JE (1978) Recovery of maleic anhydride. US 4,118,403
93. Ruggieri R, Conni S (1996) Process and plant for the purification of raw maleic anhydride
recovered from gaseous reaction mixtures. US 5,585,502
94. Windecker G, Weck A, Steiniger M, Bredehoft JP, Bender B, Knappik P, Hautte DV, Pierre
B, Haffner B (2008) Method for separating maleic anhydride from mixtures containing the
same by means of stripping. US 2008/0182996 A1
95. Brown HC, Alumbaugh WH (1997) Process for the preparation and recovery of maleic
anhydride. US 5,631,387
96. Mummey MJ, Keppel RA, Reeves AL (2000) Prevention of catalyst attrition and tar forma-
tion in the manufacture of maleic anhydride. US 6,046,343
97. Windecker G, Kaibel G, Steiniger M, Buntrock T, Weiguny J, Hautte DV, Lurquin T (2013)
Process for obtaining maleic anhydride by distillation. US 8,461,356 B2
98. Michl G, Seubert R, Schmidt JE (1990) Obtaining maleic anhydride not prone to discolor. US
4,921,977
99. Zimmerling D, Schmidt JE, Seubert R, Fischer K, Sauer F (1990) Obtaining very pure maleic
anhydride from crude maleic anhydride. US 4,961,827
100. Brown HC, Alumbaugh WH, Burnett JC (2000) Process for the purification of maleic
anhydride. US 6,120,654
101. Sugawara H, Okawa T (1991) Process for producing maleic anhydride. US 5.026,876
102. Sugawara H, Ohkawa T (1992) Method of preventing discoloring of maleic anhydride. US
5,157,130
103. Padia AS, Bortinger A, Mazzoni G, Monti T (2009) Reduction of acrylic acid in the
104. Windecker G, Weiguny J, Weck A, Dahlhoff E, Weibker WS, Heilek J, Krug T, Freyberger R
(2013) Process for separating off fumaric acid and other minor components during the
105. Iizuka Y, Ihara T, Yamauchi Y, Sawano M, Sawaki I (2002) Process for the production of
106. Sawaki I, Suwa H, Ishimura Y, Iizuka Y, Izumi M (2003) Process for the preparation of
107. Contractor RM (1999) Chem Eng Sci 54:5627–5632
108. Contractor RM (1991) Vapor phase catalytic oxidation of butane to maleic anhydride. US
5,021,588
109. Contractor RM (1999) Chem Eng Sci 54:5627–5632
110. Wilmer H, Maurer T, Rosowski F, Schuch R (2007) Process for preparing maleic anhydride
in a microchannel reactor. US 2007/0270597 A1
111. Sutton DM, Hiles AG, Backes AF (2003) Process for the simultaneous production of maleic
anhydride and its hydrogenated derivatives. US 6,620,949
112. Scarlett J, Wood MA (1995) Process for the production of hydroxy compounds by hydroge-
nation of oxygenates. EP0656336A1
113. Dubois JL (2013) Manufacture of maleic anhydride from renewable materials, maleic
anhydride obtained, and uses thereof. US 8,394,973

Hood 2016

Uploaded by

Copyright:

Available Formats

Hood 2016

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hood 2016

Uploaded by

Copyright:

Available Formats

Chapter 1

Progress in Maleic Anhydride Production

David K. Hood and Osama M. Musa

Maleic anhydride (furan-2,5-dione) is a key component to an incredibly diverse group of

D.K. Hood (*) • O.M. Musa

© Springer International Publishing Switzerland 2016 3

Scheme 1.1 Maleic anhydride and maleic acid, respectively

1.2 Historical Overview

1.2.1 Early Maleic Anhydride (via C4) Production Examples

1.2.1.1 Halcon–Scientific Design (HSD) Fixed-Bed Process [23]

1.2.1.2 ALMA Fluid-Bed Process

Butane Catalyst Transport

Air Air Compressor Spent Catalyst

1.2.2 Fixed-Bed and Fluid-Bed Comparisons

In analyzing these two established reactor technologies, a comparison scheme

Table 1.3 Comparison of reactor technologies [37, 38]

heat transfer, pressure control, catalyst porosity/geometry, as well as catalyst regen-

1.2.3 Industry Response

1.3 Catalysis: Progress in Structure and Function

Thermal, oxidative catalysis is an essential component to the success of commercial

1.3.1 Inorganic Catalyst Structure

connected, in a trans-oriented manner, exhibiting an oscillating structure due to

1.3.2 Proposed C4 Catalytic Oxidation Mechanisms

(VO)2 P2 O7 -VO PO4

oxygen/air, C4/hydrocarbons, heat, and water. Such an environment, when linked to

1.3.3 Catalyst Improvement and Implementation Strategies

Common improvement strategies in the industrial application of maleic anhydride

1.3.3.1 Insertion of Enhancement Agents

Autoignition Suppression Agents

Gas compositions for C4 oxidation processes must be carefully monitored to

Additions of Process Gases

similar strategies can be practiced prior to each reactor to permit individual

1.3.3.2 Catalytic Co-metal Inclusions

1.3.3.3 Multiple Catalytic Zones

Table 1.5 Mixed catalyst loading schedule in a fixed-bed reactor [60]

[60]. By deploying a catalytic gradient, the operator is capable of running the

1.3.3.4 Halide Removal

Table 1.6 Performance of chloride-free catalyst in a pilot reactor [64]

1.3.3.5 Pretreatment and Purity

to be superior to the previous catalyst treatment, particularly in terms of yield

1.3.3.6 Dimensional Stability

Dimensional stability of the catalyst is critical to the ability to implement a

CATALYST TABLET EXPANSION AS A FUNCTION

1.3.3.7 Morphology and Porosity

As discussed earlier, VPO is considered to be the active phase. The catalyst

1.3.3.8 Shape Considerations in Production

Catalyst materials contribute a significant and necessary expense to the maleic

1.4 Strategic Design Improvements in Existing Plants

Returning to the manufacturing facility, a number of strategies have been found to

1.4.1 Reactor Design Modifications

Insertion of process improvement agents is a popular strategy. In an effort to

1.4.1.2 Fixed-Bed Modifications

with oxygen, in the absence of a catalyst, resulting in the production of polyglot

1.4.1.3 Reduction in Fluid-Bed Back-Mixing

Back-mixing is a significant challenge for fluid-bed reactors. Back-mixing

also known as backflow or reverse circulation, results in increased contact time

1.4.1.4 Process Gas Recycling

Gas recycling has been an important development to maleic anhydride production

attributed to recovering C4 which would otherwise be lost to incineration and,