A study of the kinetics and mechanism of the adsorption and anaerobic partial oxidation of n-butane over a vanadyl pyrophosphate catalyst

Article abstract of DOI:10.1006/jcat.1999.2513

The interaction of n-butane with a (VO)₂P₂O₇ catalyst was studied by TPD and temperature-programmed reaction. At low temperatures (223 and 423 K), n-butane adsorbed as a butyl-hydroxyl pair. For absorption at 223 K, some of the butyl-hydroxyl species recombined resulting in butane desorption at 260 K upon temperature programming. For adsorption at 423 K, the hydroxyl species of the butyl-hydroxyl pair migrated away from the butyl species, forming water that is detected in the gas phase. Butane did not desorb at 260 K after the temperature was reduced to 223 K under the butane/helium from the adsorption temperature of 423 K before temperature programming from that temperature to 1100 K under a helium stream. Anaerobic temperature-programmed oxidation of n-butane yielded butene and butadiene at a peak maximum temperature of 1000 K, the exact temperature at which, upon programming, oxygen evolves from the lattice and desorbs as O₂. This, and the fact that the amount of oxygen desorbing from (VO)₂P₂O₇ at ~ 1000 K is the same as that needed for n-butane oxidation to butene and butadiene, suggested that lattice oxygen as it emerges at the surface is a selective oxidant and that its appearance at the surface is the rate-determining step in the selective oxidation of n-butane. The surface of (VO)₂P₂O₇ on which this selective oxidation occurred had nearly two monolayers of oxygen removed from it by unselective oxidation of n-butane to CO, CO₂ and H₂O at 550-950 K and had about one monolayer of carbon deposited on it at ~ 1000 K. The anaerobic oxidation of n-butane yielded butene and butadiene instead of maleic anhydride, showing that the mechanism of the n-butane oxidation to maleic anhydride is sequential.

Full text of DOI:10.1006/jcat.1999.2513

Journal of Catalysis 189, 253–262 (2000)
doi:10.1006/jcat.1999.2513, available online at http://www.idealibrary.com on
A Study of the Kinetics and Mechanism of the Adsorption and Anaerobic
Partial Oxidation of n-Butane over a Vanadyl Pyrophosphate Catalyst
B. H. Sakakini, Y. H. Taufiq-Yap,† and K. C. Waugh
Department of Chemistry, UMIST, P.O. Box 88, Manchester M60 1QD, United Kingdom; and
†Department of Chemistry, Universiti Putra Malaysia, 43400 UPM Serdang, Malaysia
Received September 23, 1998; revised February 3, 1999; accepted March 29, 1999
ities has been the subject of considerable investigation and
of consequent debate (1–10, 17–20).
The interaction of n-butane with a ((VO)2P2O7) catalyst has been
investigated by temperature-programmed desorption and anaero-
bic temperature-programmed reaction. n-Butane has been shown to
adsorb on the (VO)2P2O7 to as a butyl–hydroxyl pair. When adsorp-
tion is carried out at 223 K, upon temperature programming some
of the butyl–hydroxyl species recombine resulting in butane desorp-
tion at 260 K. However, when adsorption is carried out at 423 K,
the hydroxyl species of the butyl–hydroxyl pair migrate away from
the butyl species during the adsorption, forming water which is
detected in the gas phase. Butane therefore is not observed to des-
orb at 260 K after we lowered the temperature to 223 K under the
butane/helium from the adsorption temperature of 423 K prior to
temperature programming from that temperature to 1100 K under
a helium stream. Anaerobic temperature-programmed oxidation
of n-butane produces butene and butadiene at a peak maximum
temperature of 1000 K; this is exactly the temperature at which,
upon temperature programming, oxygen evolves from the lattice
and desorbs as O2. This, and the fact that the amount of oxygen
desorbing from the (VO)2P2O7 at 1000 K is the same as that re-
quired for the oxidation of the n-butane to butene and butadiene,
strongly suggests (i) that lattice oxygen as it emerges at the surface
is the selective oxidant and (ii) that its appearance at the surface
is the rate-determining step in the selective oxidation of n-butane.
The surface of the (VO)2P2O7 catalyst on which this selective oxida-
tion takes place has had approximately two monolayers of oxygen
removed from it by unselective oxidation of the n-butane to CO,
CO2, and H2O between 550 and 950 K and has had approximately
one monolayer of carbon deposited on it at 1000 K. It is apparent,
therefore, that the original crystallography of the (VO)2P2O7 cata-
lyst will not exist during this selective oxidation and that theories
that relate selectivity in partial oxidation to the (100) face of the
c
In the partial oxidation of benzene to maleic anhydride
over VO2O5/MoO3 (3 : 1) catalyst, it has been argued that
the selective reactive pathway occurred when an electroni-
cally perturbed oxygen molecule (an O2 species produced
by the molecular oxygen chemisorbing on a V4+ site form-
ing an O2 V5+ dimer) added para across the electronically
perturbed benzene ring. This is produced when the benzene
molecule adsorbes on a V5+ site giving a C6H+6 V4+ dimer
(1, 2). Orbital symmetry conservation arguments showed
this to be an allowed reaction (3). The adduct (adduct 1) so
obtained is
and was thought to rearrange to hydroquinone, which after
an identical para addition of O2 and rearrangement pro-
duced maleic anhydride. Subsequent quantum mechanical
calculations by Haber and co-workers showed this to be a
feasible low-energy pathway (4, 5). The fundamental thesis
of these papers was that reaction of the benzene with the
oxygen ions of the V2O5/MoO3 lattice would probably be
unspecific and lead to the formation of CO and CO2.
However, a considerable body of opinion holds that it is
the oxygen of the catalyst that is the oxidant and that selec-
tivity to maleic anhydride derives from the configuration
of the adsorbate imposed on it by the crystal field of the
oxide. In the case of the oxidation of n-butane to maleic an-
hydride over a vanadium pyrophosphate catalyst, it is the
(100) face of that catalyst that is though to be selective by
its imposition of a maleic anhydride-like configuration on
(VO)2P2O7 catalyst cannot apply in this case.
2000 Academic Press
Key Words: partial oxidation; vanadyl pyrophosphate catalyst;
n-butane; kinetics; mechanism; adsorption.
INTRODUCTION
Selectivities (product formed/reactant consumed) of the the adsorbed butane (6–10) (Fig. 1).
order of 60–80% are routinely achieved in the partial oxi-
The mechanism is essentially Mars and Van Krevelen, in
dation of hydrocarbons (benzene, n-butane, butenes, etc.) which the gas phase oxygen is used to replace the catalyst
over vanadium-based catalysts. The origin of these selectiv- (11). Recently it has been reported that the anaerobic
253
0021-9517/00 $35.00
c
Copyright
2000 by Academic Press
All rights of reproduction in any form reserved.

254
SAKAKINI, TAUFIQ-YAP, AND WAUGH
FIG. 1. (a) Plan of the oxygen atoms/ions on the (100) face of (VO)2P2O7. The oxygens A, B, and D are bonded to vanadium and the oxygen C is
bonded to phosphorus. They are of increasing activity A to D. (b) Postulated configuration of the n-butane adsorbed on the oxygen atoms/ions on the
(100) face of (VO)2P2O7 (6–10).
oxidation of n-butane over a vanadium pyrophosphate fluxed for 3 h at 393 K under continuous stirring. During this
catalyst produced selectivities of 80% . These selectivities period the solution changed in color from brown to black.
are roughly 10–20% higher than those achieved in aerobic The mixture was then cooled to room temperature and left
oxidation (12). The observation has found application stirring at this temperature overnight. Ortho-phosphorus
industrially in a plant in which the n-butane is oxidized acid (99% ) was added in such quantity as to obtained a final
anaerobically in one part of the reactor system, after which P : V atomic ratio of 1.03. The resulting solution was again
the oxygen-depleted catalyst is transferred to another reac- heated to 393 K and maintained under reflux with constant
tor for reoxidation (13). The existence of this process would stirring for 2 h. During this time, a reduced vanadium phos-
appear to eliminate chemisorbed oxygen as the oxidant; phate was formed, as indicated by a change in color from
indeed, the fact that 1–2% of the total oxygen content black to blue. Then the slurrywasfiltered, washed, and dried
of the catalyst is used (14), corresponding to more than at 423 K. Figure 2a is the x-ray powder diffraction pattern
monolayer coverage of chemisorbed oxygen, effectively (obtained on a Scintag XDS200) of the material produced
rules out the involvement of chemisorbed oxygen.
showing the main peaks at 2 values of 15.6 , 19.7 , 24.3 ,
Thispaper ispart ofa studyto investigate the originsofse- 27.1 , and 30.4 , all of which are consistent with the material
lectivityin the anaerobicoxidation ofn-butane to maleican- being crystalline VOHPO4 0.5H2O (14–18).
hydride. The nature of the chemisorption of n-butane on the
This vanadyl hydrogen phosphate hemihydrate precur-
(VO)2P2O7 surface has been determined by temperature- sor was then calcined in air for 6 h at 673 K followed by
programmed desorption and temperature-programmed an additional 3 h at the same temperature in a n-butane/air
reaction spectroscopy, techniques that have allowed: mixture (0.75% n-butane, 25 cm3 min 1). Figure 2b is the
(i) evaluation of the adsorption and desorption activation XRD pattern of this material showing the main peaks at
energies of the adsorption process, (ii) explicit definition of
2 value of 22.7 , 28.4 , and 28.9 , consistent with mate-
the structure of adsorbate on the oxide surface, and (iii) de- rial being crystalline (VO)2P2O7. This material is quite dif-
termination of the surface coverage of that adsorbate. The ferent from the poorly crystalline VPO catalysts obtained
mechanism of the selective oxidation has been elucidated by Kiely and co-workers after treating the vanadyl hydro-
by anaerobic temperature-programmed reaction.
gen phosphate hemihydrate for different periods of time in
a n-butane/oxygen helium mixture (1.6, 18, and 80.4% ) at
40 cm3 min 1 (19). This could be due to their higher butane
partial pressure, lower oxygen partial pressure, and higher
flow rate producing a more defective material.
Figure 3 (curve a) is the diffuse reflectance infrared
Fourier transform spectrum (DRIFTS) obtained on a
Nicolet Magna 550infrared spectrometer for the VOHPO4
0.5H2O material;Fig. 3(curve b) isthe DRIFTSspectrum of
the (VO)2P2O7 material. Table 1 lists their absorption bands
and putative assignments. They are corroborative evidence
EXPERIMENTAL
Catalyst Preparation
The catalyst, vanadyl pyrophosphate, (VO)2P2O7 with
P :V = 1.03, was prepared in organic medium. Vanadium
pentoxide (15.00 g from Sigma) was suspended by rapid
stirring into isobutyl alcohol (90 cm3) and benzyl alcohol
(60 cm3). The vanadium oxide–alcohol mixture was re-

ADSORPTION AND ANAEROBIC PARTIAL OXIDATION OF n-BUTANE
255
FIG. 2. (a) X-ray diffraction pattern of the precursor identifying it as VOHPO4 0.5H2O. (b) X-ray diffraction pattern of the material produced by
treating the hemihydrate precursor in air at 673 K for 6 h followed by an additional 3 h in n-butane/air (0.75% n-butane) also at 673 K, identifying it
as (VO)2P2O7.
for the material in curve a being the hemihydrate and that The Apparatus
in curve b being vanadyl pyrophosphate (14–18, 20–22).
The continuous flow multipurpose microreactor system
The surface area of the vanadyl pyrophosphate was mea-
used to conduct total surface area (TSA), temperature-
sured in situ by N2 adsorption at 77 K to be 24 m2 g 1. It
programmed desorption (TPD), temperature-programmed
should be pointed out that the n-butane/air pretreatment
reaction (TPRn), and temperature-programmed oxidation
for 3 h used here is only half the time used by Trifiro and
(TPO) experiments has been described previously (24, 25).
co-workers to produce their most selective morphology of
It is a single-tube reactor (20 cm long, 0.4 cm i.d.) connected
the vanadyl pyrophosphate-based catalyst (23).
via a heated capillary to a quadrupole mass spectrometer
(Hiden Analytical, Warrington, England) that is capable of
monitoring 16 masses continuously with temperature/time.
The Gases
Helium (ECM) was 99.995% pure and was passed
through an Oxyclear catalyst before use to remove trace
quantities of oxygen. Oxygen was supplied both by BOC
and ECM. The n-butane helium mixture (1.85% in helium)
was supplied by BOC; all was 99.995% pure and was used
direct from the cylinder.
The catalyst, in the form of powder (0.50 g), was placed in
the reactor and could be cooled to 77 K by pumping liquid
nitrogen around it for the TSA measurements or heated
by an electrical furnace to 1100 K. A thermocouple was
embedded in the center of the catalyst to allow accurate
temperature readings to be obtained. The temperature of
TABLE 1
Infrared Spectra of VOHPO4 0.5H2O and (VO)2P2O7 Adsorption
Bands (cm 1) and Vibration Assignments
VOHPO4 0.5H2O
(VO)2P2O7
Adsorption band
Assignment
Adsorption band
Assignment
1199
1134
1105
1056
982
934
686
645
548
as(PO2)
ip(P–OH)
}
1240
1220
1144
1117
1096
972
797
746
635
582
}
}
} (PO3)
} (PO3)
}
}
==
(V O)
==
(V O)
(P–OH)
ω(H2O)
(P–OH)
}
==
[V–(O V)
s(P–O–P)
as(PO3)
535
490
} (OPO)
}
}
}
514
459
(PO3)
FIG. 3. Curve (a) is the infrared spectrum of the hemihydrate precur-
sor. Curve b is the infrared spectrum of (VO)2P2O7.

256
SAKAKINI, TAUFIQ-YAP, AND WAUGH
the reactor was controlled in the temperature-programmed where Ed (J mol 1) is the desorption activation energy, Tm
mode by a Newtronics controller. All of the experiments (K) is the peak maximum temperature, R (J mol 1 K 1) is
(TSA, TPD, TPRn, TPO) were carried out in situ in a com- the gas constant, A (s 1) is the desorption preexponential
pletely sealed reactor system.
term, and (K s 1) is the heating rate, for an assumed value
of 1013 s 1 for the desorption preexponential term (27).
There are no sharp desorption peaks for oxygen in the
temperature range 300 to 850 K corresponding to specific
sites for chemisorbed molecular oxygen. There is however
a broad hump in this range that might derive from the des-
orption of oxygen from surface oxide states. The shoulder
at 998 K and the peak at 1023 K correspond to desorp-
tion activation energies of 256 and 282 kJ mol 1. The to-
tal amount of oxygen desorbed between 900 to 1100 K
is 2.2 1020 atoms g 1, which on a catalyst whose sur-
RESULTS AND DISCUSSION
Oxygen Desorption
The oxygen desorption spectrum shown in Fig. 4 was ob-
tained by pretreating the (VO)2P2O7 catalyst by heating it
to 673 K in an oxygen flow (101 kPa, 25 cm3 min 1) and
holding it under that flow at 673 K for 1 h before cooling
it to 77 K. This high-temperature pretreatment in flowing
oxygen removed any last traces of water from the system;
since thispretreatment preceded each experiment and since
the gases were 99.995% pure, no adventitious dosing of wa-
ter or any other compound occurred. In addition since we
have previously shown that N2O oxidation of a highly re-
duced (VO)2P2O7 catalyst produced by reduction in CO
completely replaces all of the anion defects (26), this oxy-
gen pretreatment is likely to produce a near defect-free
(VO)2P2O7 material.
Having dosed on the oxygen, the flow was then switched
to helium (101 kPa, 25 cm3 min 1) and the temperature was
raised (10 K min 1) to 1100 K following the m/z = 32 signal
on the mass spectrometer. Peaks are observed at 83, 91,
and 1023 K with a shoulder at 998 K. The peaks at 83 and
91 K correspond to desorption activation energies of 23 and
25 kJ mol 1. This is physisorbed oxygen. These values are
obtained by solution of the Redhead equation
1
face area is 24 m2 g would correspond to a coverage of
9.2 1014 atoms cm 2. Assuming there are approximately
2
1015 atoms cm on the surface of the (VO)2P2O7 catalyst
and that the proportion of oxygen ions on the surface is
roughly stoichiometric, then the surface oxygen ion pop-
2
ulation is 6.9 1014 ions cm and so a coverage of 9.2
2
1014 atoms/ions cm would correspond to more than a
monolayer of chemisorbed oxygen. It is therefore lattice
oxygen, the total amount desorbing being 1% of the lat-
tice oxygen. This is approximately the amount of oxygen
involved in the industrial anaerobic oxidation of n-butane
over the (VO)2P2O7 catalyst (9).
The Adsorption of n-Butane at 223 K
The vanadyl phosphate catalyst was pretreated in oxygen
(25 cm3 min 1, 1 h, 101 kPa) at 673 K after which the temper-
ature was lowered to 223 K under the oxygen flow. The flow
was then switched to a n-butane He stream (1.85% butane,
101 kPa, 25 cm3 min 1) that was passed over the catalyst for
30 min. No CO, CO2, H2O, or any other partially oxidized
product, e.g., butene, butadiene, was observed during the
dosing of the n-butane onto the catalyst at 223 K. Having
dosed the n-butane on for 30 min at 223 K, the flow was
switched to helium (25 cm 3, 101 kPa) for 30 min at 223 K,
during which time no detectable species (CO, CO2, H2O,
butene, butadiene, or n-butane) desorbed from the cata-
lyst. Temperature programming was begun at the end of the
30-min He flow, producing the desorption profile shown in
Fig. 5.
Ed /RTm
Ed
RTm2
A
=
,
[1]
Now, in addition to the oxygen seen in Fig. 4, water, car-
bon dioxide, carbon monoxide, and n-butane are also ob-
served in the desorption profile. A complete listing of the
amounts of species desorbed and of their peak maximum
temperatures is given in Table 2. The oxygen evolves at
a peak maximum temperature of 970 K; this is the lattice
oxygen seen in Fig. 4.
Water evolves at two peak maxima (370 and 800 K);
carbon dioxide and carbon monoxide evolve coincidently
at a peak maximum temperature of 970 K, the tempera-
ture at which the oxygen evolves from the lattice. A small
FIG. 4. The temperature-programmed desorption spectrum of oxy-
gen from the (VO)2P2O7 catalyst.

ADSORPTION AND ANAEROBIC PARTIAL OXIDATION OF n-BUTANE
257
It is clear that from these results even at an adsorption
temperature of 223 K the adsorption of n-butane is disso-
ciative, forming n-butyl and hydroxyl species (reaction 1).
Some of the water that evolves at the peak maximum tem-
peratures of 370 and 800 K derives from this; the majority
of the water, however, as will be explained below, results
from the oxidative dehydrogenation of the adsorbed butyl
species (the composite reaction 3). These two water peaks
show that the O(s) species are of different strength of bind-
ing at the surface. While it is possible that this surface mi-
gration of the hydroxy groups from the butyl–hydroxyl pair
occurs during the 30-min dosing period of the n-butane at
223 K, it is more likely that the surface migration is acti-
vated and that it occurs upon heating from 223 to 260 K. At
260 K the hydrogen atoms of those hydroxyl species that
have not migrated from the original adsorption site recom-
bine with the n-butyl species, producing butane in the gas
phase.
FIG. 5. The temperature-programmed desorption spectrum obtained
after having adsorbed n-butane at 223 K on the (VO)2P2O7 catalyst.
Several additional conclusions can be drawn from this
desorption profile. They are:
quantity of n-butane evolves at a peak maximum tempera-
ture of 260 K.
From these observations the adsorption process can be
described by the following elementary reactions:
(i) The activation energy for the migration of the surface
hydroxyl species is low so that under steady state operating
temperatures (between 600 to 700 K) they will be highly
mobile on the surface, affording little prospect of the n-
butyl species and the hydrogen atom of the hydroxy species
recombining under steady state conditions. The adsorption
of n-butane therefore is essentially irreversible.
(ii) The recombinative desorption activation of the n-
butyl–hydroxyl pair, calculated from the Redhead equation
(Eq. [1]), for the assumed value of 1013 s 1 for the desorp-
tion preexponential term, is 72 kJ mol 1. The surface res-
idence time, , of an adsorbate at any temperature, T, can
be calculated by use of the Frenkel equation
*
)
1.
2.
C4H10(g) + O(s)
OH(a) + OH(a)
C H9(a) + OH(a)
H O(g) + O(s) + Vo(s)
2
4
*
)
3. C4 H9(a) + 4 1 O(s) → 4 1 H2 O(g) + 4C(a)
2
2
4.
4C(a) + 6O(l) → 2CO(g) + 2CO2(g) + 6Vo(l)
12 O2(g) + Vo(l) ,
*
)
5.
O(l)
where the subscripts (g), (a), and (l) refer to gas phase, ad-
sorbed species, and lattice species; Vo(l) is a lattice oxygen
vacancy.
e1H/RT
=
,
A
where is the residence time (s), 1H is the heat of ad-
TABLE 2
1
sorption (J mol 1), and R is the gas constant (8.314 J K
mol 1). For the residence time of any adsorbate to exceed
the 30 min He purge time at 223 K, the heat of adsorption
must exceed 69 kJ mol 1. The desorption activation energy
of 72 kJ mol 1 for butane obtained from the Redhead equa-
tion easily exceeds this value.
Amounts of Species Desorbed upon Temperature Program-
ming following n-Butane Adsorption on the (VO)2P2O7 Catal-
yst at 223 K
Peak maximum
temperature (K)
Amount desorbed
(molecules g )
1
Species
(iii) Since no water is evolved during the adsorption of
the n-butane, the ionic configuration of the (VO)2P2O7 sur-
face is essentially unchanged and so this result shows that
the adsorbed butyl species is not forced by the crystal field
of the (VO)2P2O7 lattice into a structure that predisposes
it to be oxidized to maleic anhydride as has been suggested
(6, 9).
(iv) The surface O species on the (VO)2P2O7 catalyst ab-
stract the hydrogen atoms from the adsorbed butyl species
leaving carbon on the surface. The residual carbon is oxi-
dized to CO2 by the lattice oxygen species as they emerge
H2O
H2O
CO2
CO
O2
C4H10
340
790
970
950
950
260
4.1 1019
4.2 1019
4.2 1019
7.7 1018
2.4 1020
4.6 1017
Note. Total carbon (CO + CO2) = 5.0 1019 C atoms g
1
2
= 2.1 1014 C atoms cm
or a C4 coverage of 5 1013 butyl species cm 2. Total H2O is 8.3
1019 molecules g 1. The surface carbon calculated from the water evolved
assuming a H2 : C ratio of 1.25 in n-butane is 6.6 1019 C atoms g
.
1

258
SAKAKINI, TAUFIQ-YAP, AND WAUGH
at 1000 K. The purpose of the 6-h activation process in
butane/air at 673 K used to produce the highly selective
morphology (23) may be to remove these active unselec-
tive oxygen species.
The totalamount ofwater evolved is8.3 1019 molecules/
g (catalyst) in two peaks of roughly equal amounts. Since
each water molecule corresponds to the removal of one
oxygen atom from the surface of the catalyst, on the ba-
sis of a surface area of 24 m2 g 1, this corresponds to the
removal of 3.4 1014 oxygen atoms cm 2. Taking the oxy-
2
gen atom/ion population to be 6.9 1014 species cm as
explained previously, the reaction of the hydrogen atoms
of the adsorbed butyl species has removed approximately
50% of the surface oxygen ions as atoms. This is consistent
with the idea that activation in butane/air mixtures at 673 K
for 6 h removes the more active unselective surface oxygen
species.
FIG. 6. The temperature-programmed desorption spectrum obtained
after having adsorbed n-butane at 423 K on the (VO)2P2O7 catalyst.
The butyl surface coverage calculated from the number
of hydrogen atoms in the evolved water is 6.9 1013 species
per cm2. The butane desorbed at 260 K corresponded to a
coverage of 1.9 1012 molecule cm 2 so that the majority of
the hydroxyl species formed by the dissociative adsorption
of the n-butane appear to have migrated from the original
adsorption site.
223 K, the desorption spectrum obtained after dosing at
423 K did not include any n-butane desorption. The only
molecules observed were H2O, CO2, CO, and O2. No se-
lective oxidation products, e.g., butene or butadiene, were
observed. Therefore during the adsorption of the n-butane
and the formation of the butyl–hydroxyl pair, the hydroxyl
species diffuses away from the butyl species completely
(possibly forming the H2O which was observed in the gas
phase at that point) while the configuration ofthe remaining
adsorbed n-butyl species on the (VO)2P2O7 catalyst is not
constrained by the crystal field of the catalyst to be selec-
tively oxidized to maleic anhydride (A complete listing of
the amounts of species evolved and of their peak maximum
temperatures is given in Table 3).
Carbon dioxide (4.2 1019 molecules g 1) and carbon
monoxide (7.7 1018 molecules g 1) are evolved roughly
coincidentally at 1000 K, i.e., the temperature at which
the oxygen evolves from the lattice. No water is evolved at
this temperature and so the surface oxygen species of the
(VO)2P2O7 catalyst appear to have completely dehydro-
genated the adsorbed butyl species before this temperature.
The total carbon coverage is 5.0 1019 C atom g 1. This
would predict a surface butyl coverage of 5.2 1013 butyl
species cm 2 which is only slightly lower than that predicted
from the amount of water evolved. These results however
confirm that the adsorbate is the C4H9 butyl–hydroxyl pair.
Water is evolved at a roughly constant rate in the temper-
ature range 330 to 775 K, the total amount being 5.9 1019
molecules/g catalyst, which is slightly smaller than that
evolved (8.3 1019 molecules/g catalyst) having adsorbed
the n-butane at 223 K. This, however, does not take account
The Adsorption of n-Butane at 423 K
TABLE 3
The catalyst pretreatment prior to the adsorption of
n-butane at 423 K was identical to that adopted for the
pretreatment prior to the adsorption of n-butane at 223 K.
The n-butane was dosed onto the catalyst at 423 K from
the n-butane/He stream (1.85% n-butane, 101 kPa, 25 cm3
min 1) for 30 min. The temperature was then lowered to
223 K under the n-butane/He stream at which point the
catalyst was flushed with He for 30 min before raising the
temperature to 1123 K under that stream. The desorption
profile so obtained is shown in Fig. 6.
In contrast to what was observed upon dosing the n-
butane onto the (VO)2P2O7 catalyst at 223 K, dosing the
n-butane on at 423 K resulted in the production of an un-
quantified amount of H2O in the gas phase during the dos-
ing. In addition, again in contrast to what was observed at
Amounts of Species Desorbed upon Temperature Program-
ming following n-Butane of Adsorption on the (VO)2P2O7 Cata-
lyst at 423 K
Peak maximum
temperature (K)
Amount desorbed
(molecules g )
1
Species
H2O
CO2
CO
No clear maximum
5.9 1019
3.4 1019
5.8 1018
3.2 1020
1042
1018
1013
O2
Note. Total carbon (CO + CO2) = 4.0 1019 C atoms g
1
1.7 1014 C atoms cm
.
2
The surface carbon calculated from the water evolved assuming a H2 :C
ratio of 1.25 in n-butane is 4.8 1019 C atoms g
.
1

ADSORPTION AND ANAEROBIC PARTIAL OXIDATION OF n-BUTANE
TABLE 4
259
of the unquantified amount of water evolved during ad-
sorption at 423 K. Since the water is produced by the dehy-
drogenation of the adsorbing n-butane species, it appears
therefore that the adsorption of n-butane in the n-butyl–
hydroxyl dimer form is not, or is only mildly, activated.
Again the broad range of temperature in which the wa-
ter is evolved upon temperature programming reflects the
range of activities of the surface oxygen atoms.
Amounts of Species Evolved on Anaerobic Temperature Pro-
grammed Reaction of a n-Butane/He Mixture (1.85% n-butane)
over the (VO)2P2O7 Catalyst
Peak maximum
temperature (K)
Amount evolved
1
Species
(molecules g
)
CO
CO
CO
CO2
CO2
H2O
H2O
730
1000
1040
730
960
730
3.5 1019
1
1
15.5 1019
4.8 1019
2.2 1020
7.2 1020
1020
1020
As before, the oxygen evolving from the lattice at
1000 K oxidizes the adsorbed carbon that was produced
by the dehydrogenation of the adsorbed butyl species. The
sum of the CO and CO2 produced corresponds to 4.5
1019 atoms/gcatalyst or a coverage of1.9 1014 C atoms/cm2
20% of a monolayer or a butyl species coverage of 5
1013 species cm 2. Assuming that the water evolution 5.9
1019 molecules/g catalyst results from the dehydrogena-
tion of the adsorbed butyl species, the amount of water
produced would predict a carbon atom coverage of 4.8
1019 C atoms/g catalyst. This is in accordance with the car-
bon atom coverage calculated from the sum of the CO
and CO2 produced by oxidation of the surface carbon
by the lattice oxygen, confirming that the adsorbate is a
butyl species and that the adsorbed butyl species is com-
pletely dehydrogenated by the surface oxygen atoms, the
residual carbon being completely oxidized by the lattice
oxygen.
960
Note. Total carbon evolved at 730 K (CO + CO2) = 3.5 1019 + 15.5
1019 = 1.9 1020 C atoms g 1. Total predicted surface carbon from the
amount of water evolved at 730 K assuming a H2 : C ratio of 1.13 in butyl
is 1.8 1020 C atoms g 1. Total carbon evolved at 960 K (CO + CO2) =
1
1020 + 0.5 1020 = 1.5 1020 C atoms g 1. Total predicted surface car-
bon from the amount of water evolved at 960 K assuming a H2 : C ratio of
1.13 in butyl is 5.8 1020 C atoms g 1. Therefore the carbon deposited on
the surface is 4.3 1020 C atoms g 1 or 1.8 1015 C atoms cm
.
2
ent under the oxygen stream before being switched to the
n-butane/He stream and raising the temperature at 5 K
min 1. A complete listing of the amounts of species evolved
and oftheir peak maximum temperaturesisgiven in Table 4.
This profile is complex.
Anaerobic Temperature-Programmed Reaction
of the n-Butane/He (1.85% n-Butane) Stream
over the (VO)2P2O7 Catalyst
Water (2.2 1020 molecules g 1), CO (3.5 1019 mole-
cules g 1), and CO2 (1.55 1020 molecules g 1) are evolved
coincidently at a peak maximum temperature of 730 K.
This corresponds to a C : H ratio of 1 : 2.3, which is ex-
actly the C : H ratio of a butyl species. This peak there-
fore derives from the complete oxidation of an adsorbed
butyl species. The amount of butyl species adsorbed cal-
culated from the amount of H2O or the sum of the CO
and CO2 evolved is 4.1 1020 species/g or a coverage of
Figure 7 is the temperature-programmed reaction pro-
file obtained by passing a n-butane/He stream (1.85% bu-
tane, 101 kPa, 25 cm3, min 1) over the (VO)2P2O7 catalyst
that had been preoxidized in O2 at 673 K for 1 h. With the
catalyst oxidized, the temperature was lowered to ambi-
2
1014 butyl species cm 1. Were the butyl species to be
adsorbed end-on, this would correspond to roughly 20%
of a monolayer. This value is roughly the surface popula-
==
tion of the protruding oxygen atoms of the vanadyl (V O)
species, which might implicate these in the initial adsorption
step.
The difference in the oxidative behavior of the butyl
species produced by adsorption in the temperature-
programmed reaction mode compared with that of those
produced by isothermal adsorption at 423 or 223 K (the
latter were completely dehydrogenated before the carbon
was oxidized (Figs. 5 and 6), whereas in the temperature-
programmed reaction mode the C and the H of the butyl
species were oxidized simultaneously (Fig. 7)), could de-
rive from the cooling of the lower coverage of butyl species
( 5 1013 butyl species cm 2) adsorbed at 423 to 223 K,
or adsorbing at 223 K, allowing the butyl species that had
been adsorbed initially end-on to lie flat on the surface
FIG. 7. The anaerobic temperature-programmed reaction profile of
n-butane (1.85% in He) over the (VO)2P2O7 catalyst.

260
SAKAKINI, TAUFIQ-YAP, AND WAUGH
and so to become hydrogen bonded to the oxygen ions on
the (VO)2P2O7 surface. Complete initial dehydrogenation
therefore would then be inevitable. In the temperature-
programmed reaction, the end-on bonded butyl species is
simply oxidized sequentially in CH3 and CH2 units, produc-
ing CO, CO2, and H2O coincidentally.
The total amount of oxygen removed from the
(VO)2P2O7 catalyst during the oxidation of this 730 K butyl
species was 1.7 1015 O ions cm 2. Since the oxygen ion sur-
face population is approximately 6.9 1014 ions cm 2, oxi-
dation of the butyl species removes at least two monolayers
of oxygen from the lattice of the (VO)2P2O7 catalyst. This
enormous loss of oxygen will probably be accommodated
by the (VO)2P2O7 catalyst by the creation of a large num-
ber of shear planes (28), creation of which will still allow
some of the original crystallography to exist. Water (7.2
1020 molecules g 1) and CO2 (4.8 1019 molecules g 1) are
evolved at a peak maximum temperature of 960 K with
CO (1 1020 molecules g 1) evolving at the slightly higher
temperature of 1000 K. The C : H ratio of this peak is 1 : 9.6;
even when the higher temperature (1040 K) CO evolving
state (1 1020 molecules g 1) is taken into account this
reduces the ratio only to 1 : 5.8. The surface is therefore
becoming carbided. The difference between the amount of
carbon predicted to be deposited from the water evolved,
i.e.,
FIG. 8. The temperature dependence of the rate of production of
butene and butadiene in the anaerobic temperature-programmed reac-
tion of n-butane over the (VO)2P2O7 catalyst.
and
n-C4H10 + 2O(l) → C4H6 + 2H2O,
where O(l) is the lattice oxygen.
The amount of oxygen required for their production is
3.0 1020 atoms g 1. This is roughly the same as the amount
of oxygen evolving from the lattice upon temperature pro-
gramming in He (Fig. 4). Furthermore the peak maximum
temperature of the oxygen evolution 1020 K coincides with
that of the butene and butadiene evolution while the tem-
perature dependence of the production of the butene and
butadiene closely resembles that of the oxygen evolution
of Fig. 1 and is quite different from the temperature de-
pendence of CO, CO2, or H2O. This suggests that it is the
oxygen that evolves from the lattice in the desorption state
at 1020 K that is responsible for the selective oxidation of
n-butane to butene and butadiene. Other forms of surface
oxygen, bearing in mind that at least two monolayers of this
oxygen have been shown to be involved, appear to be un-
selective. The surface of the operating catalyst producing
butene and butadiene at 1020 K will be considerably differ-
ent from the crystalline form of the (VO)2P2O7 lattice that
was loaded into the reactor. It has lost at least two monolay-
ers of surface oxygen and has approximately a monolayer
or more of carbon deposited on it.
The fact that selective oxidation occurs on this surface
when the 1000 K desorbing oxygen evolves at the surface
suggests that one crystal face (possibly a defected form of
the (100) face) must be free from carbon and that the oxygen
evolution isunique to it. Were carbon deposited on thisface,
it would have been oxidized to carbon dioxide bythe oxygen
evolving from the lattice at 1000 K and so there would be
no balance between the amount of oxygen desorbing at
that temperature and the amount required for the selective
oxidation of n-butane to butene and butadiene.
H2O
10
2
4 = 7.2 1020 0.8 or
5.76 1020 C atoms g
1
,
and that removed as CO and CO2 (2 1020 + 4.8 1019 =
1
2.5 1019 atoms g 1) is 3.26 1020 atoms/g or a carbon
atom coverage of 1.4 1015 atoms cm 2. This is more than
monolayer coverage.
It can be seen from Figs. 4, 5, and 6 that oxygen evo-
lution/desorption from the (VO)2P2O7 lattice occurs be-
tween 830 and 1070 K. No oxygen evolution is observed
here. In addition to the CO, CO2, and H2O being pro-
duced in this temperature range, the anaerobic oxidation of
n-butane produces butene (identified using the m/z = 41
fragment which is unique to butene) and but-1,3-diene
(identified using the m/z = 39 and 54 fragments both of
which have only minor contributions from n-butane and
butene). The butene and butadiene profiles are shown in
detail in Fig. 8, which shows clearly, by the fact that the buta-
diene peak is not a constant fraction of the butene peak and
indeed in the lower temperature range is higher than the
butene peak, that they are both genuine selective products
of the reaction. The amount of butene produced is 1.2 1020
1
1
molecule g and of butadiene is 0.9 1020 molecule g
.
The reactions involved are
n-C4H10 + O(l) → C4H8 + H2O

ADSORPTION AND ANAEROBIC PARTIAL OXIDATION OF n-BUTANE
261
We have shown that the anaerobic temperature-pro- products (butene, butadiene, furan, or maleic anhydride)
grammed oxidation of but-1,3-diene over the (VO)2P2O7 are formed. Therefore under these conditions of gas phase
catalyst produces dihydrofuran, furan, and maleic anhy- composition, temperature flow rate, and weight of catalyst,
dride (29). In unpublished work we have also shown that, in the coadsorbing oxygen totally oxidizes all of the hydrocar-
the same type of anaerobic, temperature-programmed oxi- bonacious adsorbates, nullifying the effect of the selective
dation experiment, dihydrofuran and furan produce maleic lattice oxygen.
anhydride using the same lattice oxygen (28). All of these
occur on a carbided surface using the oxygen that would
have desorbed as molecular oxygen at 1000 K (30). While
CONCLUSIONS
it is clear that the intermediate leading to the formation of
1. n-Butane adsorbs at low temperatures (223 and 423 K)
dihydrofuran, furan, and maleic anhydride must be cyclic,
as a butyl–hydroxy pair probably on a V5+O2 cation/anion
it is difficult to sustain the notion that this is induced by the
site. The adsorption is negligibly or only mildly activated.
crystal field of the original (VO)2P2O7 lattice (6–9) bearing
At both adsorption temperatures the butyl species lies flat
in mind: (i) the possible disruption of the lattice by the loss
on the surface, becoming completely dehydrogentated by
of oxygen and (ii) that a butyl species adsorbed flat on the
the oxygen ions in the surface layers, the remaining carbon
oxide will be immediately dehydrogenated. The mechanism
being oxidized to CO and CO2 at 1000 K by the lattice
by which the end-on adsorbed butyl-type species are trans-
oxygen as it emerges from the bulk.
formed to dihydrofuran and furan is by coiling the other
2. In the anaerobic temperature-programmed oxidation
end group onto the oxygen. The importance of the oxygen
of n-butane over the (VO)2P2O7 catalyst, selective oxida-
desorbing at this temperature is that it ejects the cyclic oxy-
tion to butene and butadiene occurs at 1000 K—the tem-
genated products, dihydrofuran or furan, from the surface.
perature at which oxygen desorbs from the lattice.
3. The total amount of oxygen required for the transient
selective oxidation of n-butane to butene and butadiene is
roughly the same as that which desorbs at 1000 K, sug-
Aerobic Temperature-Programmed Oxidation of n-Butane
(0.75% in Air) over the (VO)2P2O7 Catalyst
The (VO)2P2O7 catalyst was preoxidized in oxygen
(25 cm3 min 1, 101 kPa, 673 K, 1 h) in a fashion identical
to that employed in the aerobic temperature-programmed
oxidation. Having preoxidized the catalyst, the tempera-
ture was lowered to ambient under the oxygen flow before
switching to the hydrocarbon/air flow (0.75% n-butane in
air, 25 cm3 min 1, 101 kPa) and temperature programming
from ambient to 1100 K at 5 K min 1. The temperature
dependence of the evolution of the products is shown in
Fig. 9.
gesting that this is the selective oxidant.
4. The selective oxidation occurs on a catalyst surface
that has lost approximately two monolayers of surface oxy-
gen and has the equivalent of one monolayer of carbon
deposited on it. The active face is carbon free. Some of the
original crystallography of the (VO)2P2O7 catalyst could
still exist, the oxygen loss from the catalyst being accom-
modated for by the formation of extended shear planes.
5. The butyl species adsorbed on the surface during
anaerobic temperature-programmed oxidation at higher
temperatures (>423 K) appears to be held end-on being ox-
idized at 730 K by the oxygen ions in the surface layers of
the (VO)2P2O7 catalyst to CO, CO2, and H2O, which evolve
coincidentally. The lattice oxygen evolving at 1000 K re-
acts with additional butane adsorbing end-on from the gas
phase to form butene and butadiene quantitatively.
6. The amount of oxygen required for the selective ox-
idation (3 1020 atoms/g catalyst) corresponds to roughly
two monolayers of oxygen. This oxygen evolves at 1000 K
on a catalyst already denuded of surface oxygen.
Comparison with Fig. 7 shows that it is completely dif-
ferent from anaerobic oxidation. Significantly no selective
7. The observation here that the anaerobic oxidation of
n-butane produces butene and butadiene, only, and not
maleic anhydride, suggests that the mechanism of the oxi-
dation of n-butane to maleic anhydride is sequential. Cor-
roborative evidence for this conclusion is to be found in
our observation that the anaerobic oxidation of but-1-
ene produces butadiene and furan while that of but-1,3-
diene produces furan and maleic anhydride (9). There ap-
pears to be no direct oxidative route to maleic anhydride
(10).
FIG. 9. The aerobic temperature-programmed reaction profile of
n-butane (0.75% in air) over the (VO)2P2O7 catalyst.

262
SAKAKINI, TAUFIQ-YAP, AND WAUGH
8. There being no other oxygen available, the oxygen 12. Lerou, J. J., and Mills, Ph., in “Precision Process Technology,”
(M. P. C. Weijnen and A. A. Drinkenburg, Eds.), p. 175. Kluwer Aca-
demic, Dordrecht/Norwell, MA, 1993.
13. Lerou, J. J., and Ng, K. M., Chem. Eng. Sci. 51, 1595 (1996).
14. Lerou, J. J., private communication.
that evolves from the (VO)2P2O7 catalyst at 1000 K must
be the selective oxidant used in the industrial partial oxi-
dation of n-butane to maleic anhydride. It will evolve at a
slower rate at 650–750 K, the temperature range used in-
dustrially (13), compensation for which is made for in the
contact times employed which are longer than those used
here.
15. Batis, N. H., Batis, H., Ghorbel, A., Vedrine, J. C., and Volta, J. C.,
J. Catal. 128, 248 (1991).
16. Granados, M. L., Conesa, J. C., and Fernandez-Garcia, M., J. Catal.
141, 671 (1993).
17. Igarashi, H., Tsuji, K., Okuhara, T., and Misono, M., J. Phys. Chem.
97, 7065 (1993).
REFERENCES
18. Johnson, J. W., Johnston, D. C., Jacobson, A. J., and Brody, J. F.,
J. Am. Chem. Soc. 106, 8123 (1984).
1. Lucas, J., Vandervell, H. D., and Waugh, K. C., J. Chem. Soc. Faraday
Trans. 1 77, 15 (1981).
2. Lucas, J., Vandervell, H. D., and Waugh, K. C., J. Chem. Soc. Faraday
Trans. 1 77, 31 (1981).
19. Kiely, C. J., Burrows, A., Hutchings, G. J., Bere, K. E., Volta, J.-C., and
Abon, M., Faraday Discuss. R. Soc. Chem. 105, 103 (1996).
20. Wenig, R. W., and Schrader, G. C., Ind. Eng. Chem. Fundam. 25, 612
(1986).
3. Petts, R. W., and Waugh, K. C., J. Chem. Soc. Faraday Trans. 1 78, 803
(1982).
21. Busca, G., Cavani, F., Centi, G., and Trifiro´, F., J. Catal. 99, 400
(1986).
4. Bielanski, A., and Haber, J., in “Oxygen in Catalysis,” p. 395. Dekker, 22. Bordes, E., and Courtine, P., J. Catal. 57, 236 (1979).
New York, 1991.
23. Busca, G., Centi, G., Trifiro, F., and Lorenzelli, V., J. Phys Chem. 90,
5. Witko, M., Broclawik, E., and Haber, J., J. Mol. Catal. 26, 249 (1984).
1337 (1986).
6. Ziolotkowski, J., Bordes, E., and Courtine, P., J. Catal. 122, 126 (1990). 24. Chinchen, G. C., Hay, C. M., Vandervell, H. D., and Waugh, K. C.,
7. Ziolotkowski, J., Bordes, E., and Courtine, P., Stud. Sci. Surf. Catal. 55,
J. Catal. 103, 79 (1987).
625 (1991).
25. Waugh, K. C., Appl. Catal. 43, 315 (1988).
8. Ziolotkowski, J., Bordes, E., and Courtine, P., J. Mol. Catal. 84, 307 26. Taufiq-Yap, Y. H., Sakakini, B. H., and Waugh, K. C., Catal. Lett. 48,
(1993).
105 (1997).
9. Bordes, E., Catal. Today 16, 27 (1993).
27. Redhead, P. A., Vacuum 12, 203 (1962).
10. Zhang-Lin, Y., Forrissier, M., Sneeden, R. P., Vedrine, J. C., and Volta, 28. Gai, L. P., and Kourtakis, K., Science 67, 661 (1995).
J. C., J. Catal. 145, 256 (1994).
29. Taufiq-Yap, Y. H., Sakakini, B. H., and Waugh, K. C., Catal. Lett. 46,
11. Mars, J., and van Krevelen, D. W., Chem. Eng. Sci., Spec. Suppl. 3, 41
273–277 (1997).
(1954).
30. To be published.