Electronic activity relationship for methacrolein formation over 4th period transition metal phosphomolybdates

Article abstract of DOI:10.1007/s10562-012-0917-y

Phosphomolybdate compounds have been investigated for their structural characteristics and oxidation activity toward isobutane. The phosphomolybdates were synthesized from phosphomolybdic acid and the fourth period transition metal cations Cr³⁺, Mn²⁺, Fe³⁺, Fe ²⁺, Co²⁺, Ni²⁺ Cu²⁺, Cu⁺ and Zn²⁺. Two compounds were synthesized per transition metal: where (i) all the protons had been replaced by the particular transition metal, and (ii) only partial proton replacement leaving a single proton per phosphomolybdate. X-ray diffraction analysis has shown that a primitive cubic phase is apparent with some of the transition metal phosphomolybdates. Each solid was exposed to isobutane using the anaerobic low-pressure steady-state technique. Category 1 exponential-like distributions of methacrolein were observed with all the transition metal phosphomolybdates, except the lower oxidation state iron and copper salts, Fe_1.5[PMo₁₂O ₄₀] and Cu₃[PMo₁₂O₄₀] respectively. Activation energies ranged from 51.31 ± 0.27 kJ mol^-1 (Cr[PMo₁₂O₄₀]) to over 200 kJ mol^-1 (Zn _1.5[PMo₁₂O₄₀]). Phosphomolybdates with counter cations which are one or two electrons deficient from either a 3d ⁵ or 3d ¹⁰ configuration (in this case 3d ³, 3d ⁸ or 3d ⁹) had the lowest activation barriers for methacrolein formation. A computational investigation into HOMO-LUMO band gap energies agrees with the association. The presence of protons also enhances Category 1 product formation and may be attributed to migration of H⁺ through the bulk of the solid. Graphical Abstract: [Figure not available: see fulltext.]

Full text of DOI:10.1007/s10562-012-0917-y

Catal Lett (2013) 143:61–70
DOI 10.1007/s10562-012-0917-y
Electronic Activity Relationship for Methacrolein Formation
Over 4th Period Transition Metal Phosphomolybdates
Shane M. Kendell
Trevor C. Brown
•
Ngoc Ha Nguyen
•
Received: 14 August 2012 / Accepted: 23 September 2012 / Published online: 9 October 2012
Ó Springer Science+Business Media New York 2012
Abstract Phosphomolybdate compounds have been
investigated for their structural characteristics and oxida-
tion activity toward isobutane. The phosphomolybdates
were synthesized from phosphomolybdic acid and the
methacrolein formation. A computational investigation into
HOMO–LUMO band gap energies agrees with the asso-
ciation. The presence of protons also enhances Category 1
product formation and may be attributed to migration of
3
?
2?
3?
?
fourth period transition metal cations Cr , Mn , Fe
,
H through the bulk of the solid.
2
?
2?
2?
2?
?
2?
Fe , Co , Ni Cu , Cu and Zn . Two compounds
were synthesized per transition metal: where (i) all the
protons had been replaced by the particular transition
metal, and (ii) only partial proton replacement leaving a
single proton per phosphomolybdate. X-ray diffraction
analysis has shown that a primitive cubic phase is apparent
with some of the transition metal phosphomolybdates.
Each solid was exposed to isobutane using the anaerobic
low-pressure steady-state technique. Category 1 exponen-
tial-like distributions of methacrolein were observed
with all the transition metal phosphomolybdates, except
the lower oxidation state iron and copper salts, Fe1.5
Keywords Transition metal ꢀ Oxidation ꢀ Processes and
reactions ꢀ Phosphomolybdate ꢀ Isobutane ꢀ Methacrolein
1 Introduction
Heteropoly compounds such as phosphomolybdates have
been used extensively in catalysis [1]. These substrates
exhibit unique catalytic properties which are easily tuned
[2]. One important application is in the alternative routes to
methyl methacrylate. The overwhelming majority of
methyl methacrylate is manufactured via the acetone cya-
nohydrin process (ACH). This production route has several
deleterious aspects including hazardous reactants and
by-products. Intuitively, an alternative route would be the
esterification of methacrylic acid with methanol; however
methacrylic acid is also produced via ACH. A fesible
solution to avoiding the ACH process relies on the one-step
catalytic oxidation of isobutane to give methacrolein,
whence methacrolein can be readily oxidized to metha-
crylic acid for the esterification.
[
PMo O ] and Cu [PMo O ] respectively. Activation
12 40 3 12 40
-1
energies ranged from 51.31 ± 0.27 kJ mol (Cr[PMo O ])
12 40
-
to over 200 kJ mol (Zn1.5[PMo O ]). Phosphomolyb-
1
12
40
dates with counter cations which are one or two electrons
5
10
deficient from either a 3d or 3d configuration (in this
3
case 3d , 3d or 3d ) had the lowest activation barriers for
8
9
S. M. Kendell (&)
Department of Physical Sciences, School of Science
and Mathematics, Howard Payne University,
Brownwood, TX 76801, USA
Phosphomolybdic acid derivatives have shown high
activity for the production of methacrolein from isobutane,
and varying the metal cation both in type and quantity has a
marked effect on the activation energy for isobutane
selective oxidation to methacrolein; however, there has
been difficulty in establishing a trend between activity and
metal cation type and quantity [3–6]. It has been well
established that catalytic reactions involving heteropoly
e-mail: shane.kendell@gmail.com
N. H. Nguyen
Department of Physical Chemistry, Hanoi National University
of Education, Hanoi, Vietnam
T. C. Brown
School of Science and Technology, University of New England,
Armidale, NSW, Australia
123

6
2
S. M. Kendell et al.
compounds occur via bulk mechanisms, whereby the
reactants and products migrate through the bulk of the
material rather than reacting on the surface (Bulk-type (I)),
or the reactions occur on the surface but the rate is con-
trolled by electron and proton migration through the bulk
vacuum and heated to 200 °C. Then the compounds are
further heated via a temperature-programming (5 °C min
to 500 °C whilst under vacuum. This serves to remove or
decompose residual nitrate, chloride or carbonate.
-1
)
Two phosphomolybdates were prepared for each metal
cation: (i) fully substituted salts that contain no protons,
and (ii) partially substituted compounds that contain a
single proton per Keggin unit. Exceptions to this are the
lower oxidation state iron and copper compounds which
were found to be inactive (hence the proton containing
derivatives were not examined), and additional Cu(II) salts
were synthesized to more thoroughly examine the effect of
proton content for cupric phosphomolybdates.
(
Bulk-type (II)) [2, 7–9]. In either case the reaction rate is
surface area independent.
This paper examines the activity of selected fourth
period transition metal substituted phosphomolybdates for
isobutane oxidation into methacrolein. The fourth period
transition metals were chosen due to their similarity in
atomic radius and mass, but differing electronic configu-
rations, such that a relationship between electronic prop-
erties and oxidation activity might be established.
2.2 X-Ray Diffraction (XRD) Study
2
Experimental
Past investigations have demonstrated that the X-ray dif-
fraction data of phosphomolybdic acid is modified
depending upon extent of hydration [10, 11]. Prior to XRD
analysis, the phosphomolybdates were prepared by gradual
heating till 300 °C and then calcining in air at this tem-
perature for four hours. Powdered X-ray diffraction mea-
surements were performed on a D5000 Diffractometer,
equipped with a copper anode X-ray tube. The radiation
used for the analyses consisted of CuKa X-rays with a
wavelength of 154.06 pm. Samples were systematically
2
.1 Phosphomolybdate Substrates and Their Synthesis
The phosphomolybdates investigated in this work were
fully and partially substituted compounds of the form
M H [PMo O ] and are listed in Table 1. For each metal
x
y
12 40
species a fully substituted (no hydrogen) phosphomolyb-
date was initially prepared and tested. All showed activity
except the ferrous and cuprous species, which have been
labeled as such in Table 1. Aside from the inactive metal
species, compounds containing one hydrogen per phos-
phomolybdate anion were also prepared and tested. The
copper(II) compounds (labeled in Table 1) have been
studied previously by Kendell et al. [5] and data has been
sourced from this work. Phosphomolybdates were prepared
by the stoichiometric addition of metal nitrate, chloride, or
carbonate to a solution of approximately 10 g (accurately
weighed) of phosphomolybdic acid dissolved in about
o
scanned with 2h angles ranging from 1.3–65.3 , using
o
.02 increments. The XRD investigation was performed
0
using the facilities at James Cook University, Australia.
2.3 Temperature-Programmed Reaction Spectroscopy
The selective oxidation of isobutane was analyzed using a
low-pressure steady-state technique, which has been
detailed previously [4–6]. Briefly, heterogeneous reactions
occurred in a Pyrex Knudsen cell reactor. A reactant feed-
stream consisting of an equimolar mixture of isobutane and
argon (calibrating standard) was flowed into the Knudsen
cell. The flow rate of this feed-stream was carefully con-
trolled by a variable leak valve and maintained at less than
4
0 mL of water. All solutions were heated and evaporated
with stirring until volume was reduced by about half. The
slurries were then dried in an oven at 120 °C. The resulting
solids were ground to a fine powder in preparation for
loading into the Knudsen cell reactor. When the samples
are first loaded into the Knudsen cell they are held under
Table 1 The counter cations used for synthesis of the transition metal phosphomolybdates
Counter cations (for M [PMo12
H
x y
O40])
a
b
b
Cr0.67
H
MnH
Fe(III)0.67
H
Fe(II)1.5
CoH
NiH
Cu(II)0.5
H
2
Cu(I)
3
ZnH
Zn
b
Cr
Mn1.5
Fe(III)
Co1.5
Ni1.5
Cu(II)H
b
0.5
Cu(II)1.25
H
b
Cu(II)1.5
3
-
x y 40
Each x and y for M H denotes the specific metal and protons molar quantities per mole of [PMo12O ]
a
Inactive compounds
b
Compounds from previous work by Kendell et al. [5]
1
23

Electronic Activity Relationship for Methacrolein Formation
63
1
.2 9 10 molecules per second with a Knudsen cell
5
4
solutions were used to investigate longer wavelengths
pressure of 7–10 Pa; this yielded molecular flow of the
gases which minimized gas-phase collisions, hence pre-
vented secondary reactions from occurring and isolated
intermediates. Oxygen was not included in the feed stream,
instead oxidations were facilitated by lattice oxygen in the
substrate, consistent with a Mars-van Krevelen mechanism
([400 nm) where absorption was weak, whilst the
-
5
1.00 9 10 solutions allowed accurate intensity determi-
nation for the strong absorption that occurs with shorter
wavelengths (\400 nm).
[
12].
Products formed, together with argon and unreacted
3 Results and Discussion
isobutane, exit the Knudsen cell via an aperture. These
gases then pass directly into a quadrupole mass spec-
trometer, where they are identified and a mass spectral
signal recorded. The Knudsen cell reactor is enclosed in a
programmable furnace. The mass spectral signals and
reactor temperatures are simultaneously recorded every
3.1 XRD Analysis
The structure of phosphomolybdic acid consists of a tri-
clinic phase, as previously discussed by Langpape et al.
[10]. The addition of transition metal cause disruption to
the triclinic phase and a primitive cubic phase develops
with some of the salts. Inactive copper(I) and iron(II)
compounds happen to exhibit the most clearly resolved
primitive cubic spectra, whilst Cr[PMo O ] and
2
5
9 s. Furnace temperature linearly increases at a rate of
-
1
°C min . A linear heating rate is difficult to maintain at
temperatures below 100 °C, therefore all experiments are
conducted from 100 °C to a maximum of 400–500 °C.
A total of five TPRS experiments were conducted for
each phosphomolybdate. In preparation for each experi-
ment, the reactant mixture of isobutane and argon is flowed
for about two hours. During this time the flow rate is
carefully monitored to ensure that it is molecular and a
steady-state is established. Phosphomolybdic acid decom-
position has been reported to occur at temperatures above
1
2 40
Zn1.5[PMo O ] also display resolvable primitive cubic
12 40
spectra. The unit cell dimensions for these compounds are
given in Table 2. The remaining phosphomolybdates,
however, show peaks that nearly correspond to primitive
cubic but they are shifted and numerous other peaks often
occupy the spectra; unit cell dimensions for these com-
pounds could not be determined. The addition of a proton
to the phosphomolybdate structure causes the spectrum to
become more disordered regardless of the transition metal.
This may be due to the coexistence of the primitive phase
and the triclinic phase of the pure acid. The XRD spectra of
the fully substituted phosphomolybdates are shown in
Fig. 1. Previous work by Kendell et al. has shown that the
primitive cubic phase is observed with aluminum, but body
centered cubics are observed with phosphomolybdates
synthesized from group 1 cations of the periodic table, as
well as cerium and lanthanum salts [3, 4].
400 °C [13–18]. Therefore, in order to eliminate most of
the water, with minimal degradation of the phosphomo-
lybdate structure, the first experiments were terminated at
4
00 °C. The following four experiments were allowed to
achieve 500 °C before cessation. The first four experiments
were conducted with reactant flowing at a constant rate
throughout. The fifth TPRS experiments were prepared in
the same manner as the previous four, with reactant feed
stream flowed over the sample for two hours, however, the
reactant flow is stopped and the cell evacuated for one hour
prior to commencing. Thence the fifth experiments were
conducted from 100–500 °C without a gas-phase reactant
stream. This allows an analysis of desorption processes and
product formation from isobutane pre-adsorbed to the
sample, rather than in the gas-phase. Some color change
was observed with the substrates during the temperature-
programmed experiments, which denotes mild reduction
and some structural transformation; however, this is ben-
eficial and is actually necessary to achieve maximum
activity [6, 19–23].
3.2 Product Formation
Products observed over the phosphomolybdates were
methacrolein, lactone, acetic acid and carbon dioxide and
water. Propene and methacrylic acid were not detected
with any of these compounds despite thorough analysis.
Products are formed via two different reaction processes,
labeled Categories 1 and 2 [6]:
Table 2 Unit cell dimensions of transition metal phosphomolybdates
˚
a (A)
2
.4 UV Analysis
Compound
Fe(II)1.5[PMo12
O
40
]
13.302(12)
13.224(17)
13.360(4)
13.337(7)
-
4
-5
-1
Solutions (1.00 9 10 and 1.00 9 10 mol L ) of the
fourth period transition metals were prepared and analysed
using via a Perkin Elmer Lamda 25 UV–Vis spectropho-
Cu(I)
Cr[PMo12
Zn1.5[PMo12
3
[PMo12
O
40
]
O
40
]
O
40
]
-
4
tometer, with a 1 nm bandwidth. The 1.00 9 10
M
123

6
4
S. M. Kendell et al.
Fig. 1 XRD spectra for the fully substituted transition metal cations.
Cu(I) [PMo12 40] and Fe(II)1.5[PMo12 40] display the sharpest
patterns
Fig. 2 Methacrolein formation over Cr[PMo12O40]. This product
profile shows both Category 2 desorptions (lower temperature) and
Category 1 exponential rise
3
O
O
Category 1: Distributions where formation increases
with increasing temperature, in a manner consistent with
an exponential product evolution and are only observed
whilst gas-phase isobutane is present.
Methacrolein formed via Category 1 and 2 mechanisms.
Category 1 methacrolein was observed with all com-
pounds. There were large variations in product profiles,
with some salts evolving methacrolein at less than 200 °C,
while others required high temperatures ([400 °C) before
significant amounts could be detected. Category 2 meth-
acrolein formation occurs with about a third of the phos-
phomolybdates. Figure 2 shows the methacrolein profile
Category 2: Profiles that consist of several rises, falls and
inflections (although sometimes a single peak is
observed), usually overlapping and only resolvable into
individual peaks by extensive modeling using Gaussian
curves; this Category of distribution is observed with or
without gas-phase isobutane flowing over the substrate
and are consistent with desorption.
over Cr[PMo12O40], with the desorption from Category 2
processes being dominate till *250 °C, thenceforth the
exponential contribution from Category 1 becomes signif-
icant and continues till the cessation temperature of
In order to resolve the Category 2 distributions from the
Category 1 contributions, Category 2 product data is collected
in the absence of gaseous isobutane (i.e. the fifth TPRS
experiment). Overall Category 2 product profiles are a com-
bination of peaks. The product intensities of each peak as a
function of temperature are resolved byGaussian distributions:
5
00 °C. Lactone is observed with the majority of salts and
only occurs via Category 2 product distributions. Acetic
acid also forms exclusively by a Category 2 process and is
detected with four phosphomolybdates. Table 3 shows the
Category 2 products associated with the various phospho-
molybdates, and the number of Category 2 peaks as
resolved by Gaussian curves.
!
2
ꢁ
ðT ꢁ pÞ
Ip ¼ h exp
ð1Þ
2
2
w
3
.3 Kinetics
Parameters for the height (h), position (p) and width
w) dictate the dimensions of each Gaussian simulation. A
(
Category 1 methacrolein production profiles have been
modeled extensively in previous work [5, 6, 26].
These methods were successfully applied to Category 1
methacrolein formations observed over the transition
phosphomolybdates.
non-linear least squares analysis is performed on the
summation of the individual simulations, in order to
maximize the fit to the raw data.
The Fe(II) and Cu(I) phosphomolybdates, Fe [PMo O ]
12 40
1.5
and Cu [PMo O ] did not exhibit significant product for-
3 12 40
Reliable techniques have also been developed to accu-
rately analyze Category 2 products, with kinetic parameters
for isobutane oxidation over lanthanum and cerium salts
[26] and over the pure acid [6]. However the Category 2
profiles observed over the transition metal compounds are
not as clearly resolved as those in the previous studies,
resulting in large error ranges when determining kinetic
parameters. Moreover, in comparison to the Category 1
mation, either for Category 1 or 2 distributions. The lack of
activity is likely due to reduction of the molybdenum atoms in
the phosphomolybdate anion, from Mo(VI) to Mo(V), with
corresponding oxidation of Fe(II) and Cu(I) to Fe(III) and
Cu(II) respectively. This was evidenced by the formation
of a blue solution during synthesis [24, 25]. Henceforth,
Fe [PMo O ] and Cu [PMo O ] are not discussed.
1.5
12 40
3
12 40
1
23

Electronic Activity Relationship for Methacrolein Formation
65
Table 3 This table presents the number of bell-shaped peaks asso-
ciated with Category 2 methacrolein (MA), lactone (LAC) and acetic
acid (AA) production
Catalyst
MA
LAC
AA
Cr0.67H[PMo12
O
40
]
–
4
6
–
6
–
5
1
–
–
–
–
–
–
7
–
–
2
4
–
4
–
4
1
4
3
5
4
5
5
4
–
–
–
4
–
7
–
–
–
5
–
–
–
–
–
7
–
Cr[PMo12
O
40
]
MnH[PMo12
O
40
]
Mn1.5[PMo12
O
40
]
Fe0.67H[PMo12
O
40
]
40
Fe[PMo12O ]
CoH[PMo12
O
40
40
40
40
]
Co1.5[PMo12
O
]
NiH[PMo12
O
]
Ni1.5[PMo12
O
]
Fig. 3 Methacrolein formation and associated isobutane coverage
over Ni1.5[PMo12O ]
40
Cu0.5
CuH[PMo12
Cu1.25 0.5[PMo12
Cu1.5[PMo12
40
2 40
H [PMo12O ]
40
O ]
H
40
O ]
O
]
ZnH[PMo12O
40
]
Eapp
InIp ¼ ꢁ
RT
ꢀ
ꢁ
n
act
Zn1.5[PMo12
O
40
]
þ ðm þ nÞInhR þ In bapAappN
ð3Þ
The parameters b, A
and N are the respective
act
app
-
1
heating rate (5 °C min ), apparent pre-exponential factor
and number of active sites. The order of reaction is denoted
by n and the value m is introduced to account for any
deviations from linearity. A full derivation of the reactant
coverage and kinetic considerations has been described
elsewhere [5]. A multiple linear regression is performed,
mechanism, methacrolein yield via the Category 2 process
is far less (*1 %). Therefore the Category 2 distributions
have been excluded from analysis. Henceforth, all discus-
sion on kinetics of methacrolein formation will refer to
Category 1 methacrolein exclusively.
using the module in MS Excel, with the Y data as lnI , and
P
3
.3.1 Coverage
-
1/RT and lnhR as the X inputs, yields slopes E
and
app
ꢀ
ꢁ
n
act
These distributions are consistent with a Langmuir
adsorption isotherm. A simplified first approximation to the
(m ? n), and In bapAappN
as the intercept.
Equation 3 was applied to the methacrolein distribu-
tions for the 2nd, 3rd and 4th TPRS experiments. Table 1
displays the kinetic data associated with methacrolein for-
mation, including the temperature range in which the mod-
eling was conducted. The A parameter in the log A
fractional coverage for the reactant isobutane (h ) of these
R
distributions is given by:
Ir=aR
hR ¼
ð2Þ
1
0
1
þ Ir=aR þ Ip=ap
n
act
expression is the product of b, a , A and N .
P
app
Here, IR and IP are the reactant and product mass
spectral intensities, while the a and a denote sensitivity
factors that relate these intensities to flow rate, in
Methacrolein formation kinetics varies significantly for
the transition metal compounds. This occurs both with
?
cation selection and the presence of H . The Arrhenius
R
P
-
1
molecules s . The complete derivation for this coverage
model with the low-pressure steady-state apparatus has
been published previously [5]. Figure 3 shows the reactant
fractional coverage for methacrolein formation, as well as
the corresponding product profile. As can be seen, coverage
diminishes less than 7 % throughout the TPRS experiment.
parameters associated with methacrolein are expressed in
Table 4, where Eapp is the apparent activation energy and
T range refers to the temperature span used in the model-
ling procedure. Errors associated with the kinetic values are
given as parenthetic terms in Table 4. The high coverage of
isobutane throughout a temperature-programmed experi-
ment does not impact on the kinetics at all. Rather, this is a
reflection that the rate of adsoprtion of reactant is large
relative to the formation rate of product. Moreover, the
Arrhenius plots were linear, indicating that there were no
coverage-related deviations and the kinetic model was
appropriate.
3
.3.2 Methacrolein Arrhenius Parameters
The product intensity is related to the apparent activation
energy (Eapp) and reactant coverage through an Arrhenius-
type equation:
123

6
6
S. M. Kendell et al.
3
.4 Investigation into Activity Relationships
Methacrolein Formation
A bar graph of the activation energies for each of fully
substituted 4th period transition metal phosphomolybdates
is shown in Fig. 4, where the metal and its oxidation state,
rather than the complete phosphomolybdate formula, is
used to distinguish each substrate. Above each bar in the
figure is the 3d occupancy of the counter cation. The
As no correlation exists between the structural character-
istics of the phosphomolybdates (as determined by XRD)
and the activity parameters for methacrolein formation,
several other properties were investigated in order to
establish characteristics of the compounds that influence
activity.
5
7
activation barriers are highest with the d and d metals and
become progressively lower as the electron configuration
8
increases to d and d , or decreases to d . However an
9
3
1
0
increase is associated with the d configuration. Previous
work has suggested that partially filled d-bands are nec-
essary for rapid chemisorption and catalytic activity
[40, 41], hence this may explain why a large activation
barrier occurs with Zn1.5[PMo O ]. The most active
3
.4.1 Electronic Transitions
Previous research has shown that heteropoly compounds,
such as phosphomolybdates, undergo reaction mechanisms
in the bulk [2, 7, 27–29]. Work on the oxidative dehy-
drogenation of isobutyraldehyde and isobutyric acid to
yield methacrolein and methacrylic acid respectively show
that products are formed via a Bulk-type (II) mechanism
12
40
compounds are thus those that are either one or two elec-
5
trons deficient from d or d electron occupancy.
10
The electron configurations of chromium and copper in
1
5
1
10
the ground state are [Ar]4s 3d and [Ar]4s 3d respec-
2
4
2
9
[
15, 30, 31]; the Bulk-type (II) mechanism involves rapid
tively, rather than [Ar]4s 3d and [Ar]4s 3d as predicted
by the Aufbau principle, and is due to the enhanced sta-
bility achieved by having half or completely filled d orbi-
tals [42]. Therefore it follows that metal ions deficient of
electron and proton migration through the catalyst bulk,
with reaction rate determined by these charge transfers [2].
The bulk redox properties of the substrate dictate the
effectiveness of the Bulk-type (II) process [7]. Catalytic
activity has also been described in terms of electronic
transitions, d-band energies and semiconductor principles
5
10
electrons from the d or d configurations would achieve a
stabilising effect by the addition of the required electrons.
Considering oxidation of isobutane necessitates reduction
of the phosphomolybdate structure, this stabilising effect
may impact markedly upon activity. On a per electron
basis, this stabilisation should be more pronounced with
atoms that require fewer electrons to achieve half or
complete d orbital occupancy. Furthermore, it has also
been shown that d-bands become lower in energy from left
to right across a period in the Periodic Table due to
increasing nuclear charge [37, 43]. This may account for
[
32–35], and are associated with chemisorption bond
strengths [36, 37]. Many reactions depend entirely on the
presence of d-band holes and are independent of structural
effects [38]. It has been demonstrated that transition metals
influence activity by charge-transfer within the Keggin
structure [39].
Table 4 Arrhenius parameters for methacrolein formation over
transition metal phosphomolybdates
8
9
low activation barriers corresponding with the d or d
5
metals rather than the d or d .
7
-1
)
Compound
E
app (kJ mol
log10A
T range (°C)
Cr0.67H[PMo12
O
40
]
58.22 (0.30)
51.31 (0.27)
96.3 (1.2)
196 (9)
6.32 (0.02) 330–493
5.85 (0.02) 312–504
8.66 (0.08) 417–503
Cr[PMo12
O
40
]
MnH[PMo12
O
40
]
Mn1.5[PMo12
O
40
]
14.7 (0.6)
7.13 (0.06) 382–489
19.8 (0.4) 370–417
5.90 (0.08) 366–457
15.1 (0.2) 409–462
466–500
Fe0.67H[PMo12
O
40
]
78.2 (0.8)
253 (6)
40
Fe[PMo12O ]
CoH[PMo12
O
40
40
40
40
]
58.0 (1.1)
196 (3)
Co1.5[PMo12
O
]
NiH[PMo12
O
]
78.2 (0.5)
100.6 (0.6)
68.1 (1.0)
60.8 (0.5)
52.9 (1.7)
77 (2)
7.29 (0.04) 363–474
9.12 (0.05) 306–421
6.95 (0.08) 362–478
6.43 (0.04) 332–473
5.87 (0.13) 401–469
Ni1.5[PMo12
O
]
Cu0.5
CuH[PMo12
Cu1.25 0.5[PMo12
Cu1.5[PMo12
40
2 40
H [PMo12O ]
40
O ]
H
40
O ]
O
]
7.6 (0.8)
6.87 (0.05) 382–485
16.7 (0.5) 374–418
354–468
ZnH[PMo12O
40
]
72.5 (0.7)
209 (6)
Fig. 4 Bar graph of apparent activation energy for each transition
metal. Shown above each bar is the 3d electron occupancy
Zn1.5[PMo12
O
40
]
1
23

Electronic Activity Relationship for Methacrolein Formation
67
3
The Cr ion is associated with a lower activation bar-
?
below 190 nm however the spectrophotometer is incapable
of those measurements. The inactive compounds Fe1.5[P-
Mo O ] and Cu [PMo O ] display the lowest absorp-
2
rier than Cu despite requiring an additional electron to
?
gain the stabilising configuration and being five atoms left
1
2
40
3
12 40
2
in the period. Also, the Mn exhibits a lower activation
?
tion intensities at the low wavelengths around 190 nm.
Despite the suggestion of a relationship between pairs of
phosphomolybdates and absorbance readings at low
wavelengths, there is no clear connection between the
activation barriers and absorbance intensities at any
wavelength.
3
?
5
barrier than Fe , despite both existing with d occupancy.
This indicates that another property(s) contributes to
activity, such as cation radius, charge, electronegativity etc.
4
3?
2?
It may be that a d metal ion, such as Mn or Cr , would
3
?
yield an even lower activation barrier than Cr , however
this was not investigated.
It has been reported that the low wavelength peaks are
due to charge transfer between oxygen and the addenda
metal atoms of the phosphomolybdate structure—in this
case the Mo=O and Mo–O–Mo bonds [46]. Therefore, the
pair-wise association would be related to this phenomenon.
Transitions involving the d electrons have been shown to
absorb at around 500 nm [44], however no evidence of this
could be found at wavelengths in this vicinity. Overall, the
results from the UV absorption analysis did not confirm the
existence of d orbital transitions that would support the
hypothesis outlined in the preceding section. However, this
may not be unexpected as Fournier et al. have shown that
UV–Vis spectra are unreliable predictors of catalyst
activity [47, 53].
3
.4.2 UV Absorption Analysis
Heteropoly compounds have previously been studied using
ultraviolet (UV) spectroscopy [19, 44–51]. Correlations
between UV reflectance spectroscopy and reaction rates
and activity have been reported [22, 52]. The phospho-
molybdates in this chapter were analysed by UV absorption
spectroscopy. This was performed in order to investigate a
correlation with absorption energy and kinetic parameters
for methacrolein formation and to verify the d electron
relationship established in the preceding subsection. No
meaningful absorbance information could be found for
wavelengths greater than 400 nm however wavelengths
below 400 nm showed strong absorption.
3.4.3 Computational Investigation
Overlaid plots of the absorbance spectra for the
-
.00 9 10 M solutions are presented in Fig. 5. The
5
1
The phosphomolybdate salts of the fourth period transition
metals, excluding the inactive Fe(II) and Cu(I) compounds,
were analysed using computational methods, in order to
establish evidence for a d-band relationship with activity.
Density functional studies have previously been performed
on vanadium substituted phosphomolybdates to determine
HOMO–LUMO energies [54]. Calculations were conducted
by Ngoc Ha Nguyen from Hanoi National University of
Education, Hanoi, Vietnam. Non integral stoichiometries
of metal cation to phosphomolybdate anion, e.g. Ni1.5[P-
Mo O ], were not used and only salts of the form
spectra show that there appears to be a relationship
between some of the solutions at *192 nm: Fe[PMo O ]
1
2 40
and Co1.5[PMo O ] have similar absorbance intensities
12 40
as do the pairs Ni1.5[PMo O ] and H [PMo O ], and
12 40
12
40
3
Mn1.5[PMo O ] and Zn [PMo O ]. From the graph it
12
40
1.5
12 40
also appears that Cr[PMo O ] and Cu [PMo O ] may
12 40
1
2
40
1.5
have a similar absorbance intensity at some wavelength
1
2 40
M[PMo O ], when M is a trivalent metal cation and
12 40
MH[PMo O ], when M is divalent, were analysed. These
2 40
1
phosphomolybdates were structurally optimised using
CP2 K software (Fig. 6) and band gaps were determined
using MOPAC 2007 software.
Computational results for the band gap energies for the
compounds are shown in Table 5, along with the activation
energies for methacrolein formation. Previous work has
shown that protonation of the phosphomolybdate anion
does not significantly alter the HOMO–LUMO gap [31],
therefore the band gap energies have been simultaneously
listed for the partially and fully substituted salts.
A correlation is observed with band gap and activation
energies for the fully substituted phosphomolybdates when
separately comparing ?2 and ?3 metal cations, exemplified in
Fig. 7. The proton containing phosphomolybdates do not
Fig. 5 Ultraviolet absorption spectroscopic data plots for phospho-
molybdic acid (PMA) and the fourth period transition metal salts
-5
(
1.00 9 10 M)
123

6
8
S. M. Kendell et al.
Fig. 7 Plot of theoretical band gap energies versus experimentally
determined apparent activation barriers for methacrolein formation.
The ?2 transition metals are shown in red and the ?3 in turquoise
Fig. 6 Optimized structure of Cr[PMo12O40]. The chromium is
shown in purple, oxygen atoms in red and molybdenum atoms in
green
proton to the phosphomolybdate anion, with the exception
of the chromium compounds. For example, the activa-
Table 5 HOMO-LUMO band gap energies associated with the
fourth period transition metals investigated in this work
tion energy is lowered from 100.6 ± 0.6 to 78.2 ±
-1
-
1
Compound
E
app (kJ mol
)
Band gap (eV)
5.86
0.5 kJ mol
,
when comparing Ni1.5[PMo O ] to
12 40
NiH[PMo O ]. The cupric salts show that the highest
12 40
Cr0.67H[PMo12
O
40
]
58.22 (0.30)
51.31 (0.27)
96.3 (1.2)
196 (9)
activation barrier occurs with Cu1.5[PMo O ] and the
12 40
Cr[PMo12
O
40
]
lowest with Cu1.25H0.5[PMo O ]. The copper(II) series
12 40
MnH[PMo12
O
40
]
6.20
7.47
6.46
4.51
4.91
also demonstrates that increased proton substitution with
Mn1.5[PMo12
O
40
]
the CuH[PMo O ] and Cu H [PMo O ] salts, coin-
0.5
Fe0.67H[PMo12
O
40
]
78.2 (0.8)
253 (6)
12 40
2
12 40
cides with respective increases in energies of activation.
Therefore an optimum proton concentration is achieved
with Cu1.25[PMo O ]. Similar activity trends with proton
40
Fe[PMo12O ]
CoH[PMo12
O
40
40
40
40
]
58.0 (1.1)
196 (3)
12
40
Co1.5[PMo12
O
]
content have been reported [30, 31, 55–57]. This demon-
strates that Brønsted acidity is a major factor in the
determination of activity. In the case of the chromium
phosphomolybdates, the protonated salt is less active than
NiH[PMo12
O
]
78.2 (0.5)
100.6 (0.6)
68.1 (1.0)
60.8 (0.5)
52.9 (1.7)
77 (2)
Ni1.5[PMo12
O
]
Cu0.5
CuH[PMo12
Cu1.25 0.5[PMo12
Cu1.5[PMo12
40
2 40
H [PMo12O ]
40
O ]
3
?
the very active Cr salt. However there may be a more
active composition with a proton concentration of less than
one per Keggin unit. Therefore the optimum transition
metal to proton ratio depends on the electronic properties
of the transition metal and the effect of proton migration;
phosphomolybdates consisting of metal cations with
favorable electronic properties will reach maximum
activity with high extents of substitution, whilst com-
pounds containing less active metals will benefit by a
greater proton concentration.
H
40
O ]
O
]
ZnH[PMo12
O
40
]
72.5 (0.7)
209 (6)
6.94
Zn1.5[PMo12
O
40
]
show such a connection with activity. The HOMO–LUMO
band gap calculations are preliminary and do not take into
account bulk effects nor do they specifically apply to isobutane
oxidation. Nevertheless they indicate a d-band correlation, as
proposed above or a related electronic transition mechanism.
More work needs to be conducted using computational
methods before a conclusive argument can be presented.
?
Previous work has postulated that H is beneficial in
protonating the organic component [31], or creating reac-
tive Brønsted acid centers [58]. However, phosphomolyb-
dic acid exhibits very high conductivity, which is due to
rapid proton migration [59–62]. This is consistent with the
proton mobility dependence intrinsic to Bulk-type (II)
catalysis, as proposed by Misono [2]. The migration of
protons in catalytic reactions has also been supported by
computational studies [63]. If electronic properties of the
3
.4.4 Partially Substituted 4th Period Transition Metal
Compounds
Table 5 also shows that the activation barrier for methac-
rolein formation is lowered by the addition of a single
1
23

Electronic Activity Relationship for Methacrolein Formation
69
transition metals dictate their effectiveness than it would
?
seem likely that H would predominately enhance activity
transitions supporting the d-band hypothesis, however a
computational investigation into HOMO–LUMO band gap
energies agree with the association. The presence of protons
also enhances Category 1 product formation and may be
through charge distribution throughout the bulk of the
solid, rather than facilitating carbocation or acid center
formation.
?
attributed to migration of H through the bulk of the solid.
Isobutane oxidation also produced Category 2 bell-
shaped profiles of methacrolein, lactone (3-methyl-2-oxe-
tanone) and acetic acid. Unfortunately these distributions
were not able to be resolved in a straight-forward manner
for kinetic analysis. The quantities of each product was
unable to explained, and as such no correlation between
susbstrate properties and Category 2 products has been
proposed.
4
Conclusion
Phosphomolybdate compounds have been investigated for
their structural characteristics and oxidation activity toward
isobutane. The phosphomolybdates were synthesized from
phosphomolybdic acid and the fourth period transition
3
?
2?
3?
2?
2?
2?
2?
metal cations Cr , Mn , Fe , Fe , Co , Ni Cu
,
?
2?
Cu and Zn . Two compounds were synthesized per
transition metal: where (i) all the protons had been replaced
by the particular transition metal, and (ii) only partial
proton replacement leaving a single proton per phospho-
References
2?
?
1. Okuhara T, Mizuno N, Misono M (1996) Adv Catal 41:113–252
molybdate, with the exception of Fe and Cu derivatives
which were fully substituted only.
2
3
. Misono M (2001) Chem Commun 13:1141–1152
. Kendell S, Brown TC (2011) Catal Lett 141:1767–1785
X-ray diffraction analysis has shown that a primitive
cubic phase is apparent with some of the transition metal
phosphomolybdates: Cu [PMo O ], Fe1.5[PMo O ],
4. Kendell SM, Alston A-S, Ballam NJ, Brown TC, Burns RC
2011) Catal Lett 141:374–390
. Kendell SM, Brown TC, Burns RC (2008) Catal Today
31:526–532
6. Kendell SM, Brown TC (2010) React Kinet Mech Catal
99:251–268
(
5
3
12 40
12 40
1
Cr[PMo O ], Zn [PMo O ]. The other phosphomo-
1
2
40
1.5
12 40
lybdates show peaks that nearly correspond to primitive
cubic but they are shifted due to another phase which
introduces other, more random peaks. The addition of a
proton to the phosphomolybdate structure causes the spec-
trum to become more disordered regardless of the transition
metal. This may be due to the coexistence of the primitive
phase and the triclinic phase of the pure acid.
7
. Misono M, Mizuno N, Mori H, Lee KY, Jiao J (1991) In:
Grasselli RK, Sleight AW (eds) Structure-activity and selectivity
relationships in heterogeneous catalysis. Elsevier, Amsterdam,
pp 87–97
8. Mizuno N, Watanabe T, Mori H, Misono M (1990) J Catal
23:157–163
1
9
. Okuhara T, Hashimoto T, Misono M, Yoneda Y (1983) Chem
Lett 12:573–576
Each solid was exposed to isobutane using the anaerobic
low-pressure steady-state technique. Category 1 exponen-
tial-like distributions of methacrolein were observed
with all the transition metal phosphomolybdates, except
1
1
1
0. Langpape M, Millet JMM, Ozkan US, Boudeulle M (1999)
J Catal 181:80–90
1. Black JB, Clayden NJ, Gai PL, Scott JD, Serwicka EM, Goo-
denough JB (1987) J Catal 106:1–15
2. Mars P, van Krevelen DW (1954) Chem Eng Sci 3:41–59
the lower oxidation state iron and copper salts, Fe1.5
13. Chang TH (1995) Journal of the Chemical Society. Faraday Trans
91:375–379
[
PMo O ] and Cu [PMo O ] respectively. The synthe-
12 40 3 12 40
1
4. Fournier M, Feumi-Jantou C, Rabia C, Herv e´ G, Launay S (1992)
J Mater Chem 2:971–978
5. Misono M (1987) Catal Rev Sci Eng 29:269–321
6. Rocchiccioli-Deltcheff C, Aouissi A, Bettahar MM, Launay S,
Fournier M (1996) J Catal 164:16–27
7. Moffat JB (2001) Metal-oxtygen clusters. Kluwer Academic,
New York
8. West SF, Audrieth LF (1955) J Phys Chem 59:1069–1072
9. Cavani F, Etienne E, Mezzogori R, Pigamo A, Trifiro F (2001)
Catal Lett 75:99–105
sis of the ferrous and cuprous derivatives resulted in
reduced phosphomolybdates which accounted for their
inactivity. Accurate kinetic parameters were determined for
the Category 1 methacrolein distributions, with activa-
1
1
-
1
1
tion energies ranging from 51.31 ± 0.27 kJ mol
(Cr
-
PMo O ]) to over 200 kJ mol (Zn1.5[PMo O ]). A
1
[
1
2
40
12 40
1
1
correlation between the d electron configuration of the
fourth period transition metals and the activity of the
phosphomolybdates has been established. Phosphomolyb-
dates with counter cations which are one or two electrons
2
2
0. Cavani F, Mezzogori R, Pigamo A, Trifiro F, Etienne E (2001)
Catal Today 71:97–110
1. Wienold J, Timpe O, Ressler T (2003) Chem A Eur J
5
10
deficient from either a 3d or 3d configuration (in this case
3
9
:6007–6017
3
8
d , 3d or 3d ) had the lowest activation barriers for
9
22. Melsheimer J, Mahmoud SS, Mestl G, Schl o¨ gl R (1999) Catal
Lett 60:103–111
23. Nowinska K, Waclaw A, Sopa M, Klak M (2002) Catal Lett
methacrolein formation. Conversely, compounds in which
5
7
10
the counter cation had a 3d , 3d or 3d configuration,
required substantially higher activation energies to form
methacrolein. UV analysis failed to identify electronic
7
8:347–352
2
4. Greenwood NN, Earnshaw A (1984) Chemistry of the elements,
1st edn. Pergamon Press, Oxford
123

7
0
S. M. Kendell et al.
2
2
5. Barrows JN, Jameson GB, Pope MT (1985) J Am Chem Soc
47. Fournier M, Louis C, Che M, Chaquin P, Masure D (1989)
J Catal 119:400–414
48. Youn MH, Kim H, Jung JC, Song IK, Barteau KP, Barteau MA
(2005) J Mol Catal A Chem 241:227–232
1
07:1771–1773
6. Kendell S, Alston A, Brown T (2009) Chem Prod Process Model
:5
4
2
2
7. Komaya T, Misono M (1983) Chem Lett 12:1177–1180
8. Okuhara T, Kasai A, Hayakawa N, Misono M, Yoneda Y (1981)
Chemistry Letters 391–394
9. Okuhara T, Kasai A, Hayakawa N, Yoneda Y, Misono M (1983)
J Catal 83:121–130
49. Song IK, Kim HS, Chun MS (2003) Korean J Chem Eng
20:844–849
50. Bentrup U, Bruckner A, Kant M, Kolf S, Dingerdissen U, Jansen
S, Maschmeyer D, Siegert H, Zanthoff W (2006) Erd o¨ l Erdgas
Kohle 122:OG145–OG148
2
3
3
3
3
3
3
3
3
3
0. Mizuno N, Tateishi M, Iwamoto M (1996) J Catal 163:87–94
1. Hu J, Burns RC (2000) J Catal 195:360–375
2. Garner WE (1957) Adv Catal 9:169–186
51. Ilkenhans T, Herzog B, Braun T, Schlogl R (1995) J Catal
153:275–292
52. Nondek L, Mihajlova D, Andreev A, Palazov A, Kraus M,
Shopov D (1975) J Catal 40:46–51
53. Weber RS (1994) J Phys Chem 98:2999–3005
54. Wang J, Hu C, Jian M, Zhang J, Li G (2006) J Catal 240:23–30
55. Deusser LM, Petzoldt JC, Gaube JW, Hibst H (1998) Ind Eng
Chem Res 37:3230–3236
56. Langpape M, Millet JMM, Ozkan US, Delich e` re P (1999) J Catal
182:148–155
3. Hauffe K (1955) Adv Catal 7:213–257
4. Beeck O (1950) Discuss Faraday Soc 8:118–128
5. Haber J, Stone FS (1963) Trans Faraday Soc 59:192–206
6. Hammer B (2006) Top Catal 37:3–15
7. Hammer B, Norskov JK (1995) Nature 376:238–240
8. Dowden DA, Reynolds PW (1950) Discuss Faraday Soc
8
:184–190
3
9. Marchal-Roch C, Millet JMM (2001) Comptes rendus de
l’Academie des sciences. series II. fascicule C. 4:321–329
0. Wang KP (1982) Chin J Phys 20:63–66
57. Liu-Cai FX, Pham C, Bey F, Herve G (2002) React Kinet Catal
Lett 75:305–314
58. Dimitratos N, Vedrine J (2006) Catal Commun 7:811–818
59. Nakamura O, Kodama T, Ogino I, Miyake Y (1979) Chem Lett
1:17
60. Pandey K, Lakshmi N (1999) J Mater Sci 34:1749–1752
61. Hardwick A, Dickens PG, Slade RCT (1984) Solid State Ion
13:345–350
4
4
4
4
1. Dowden DA (1952) Ind Eng Chem 44:977–985
2. Melrose MP, Scerri ER (1996) J Chem Educ 73:498–503
3. van Bokhoven JA, Miller JT (2007) X-ray Absorption Fine
Structure - XAFS13: 13th International Conference. 88:582–584
4. Hiskia A, Mylonas A, Papaconstantinou E (2001) Chem Soc Rev
4
3
0:62–69
62. Nakamura O, Ogino I, Kodama T (1981) Solid State Ion
3–4:347–351
63. Mioc UB, Petkovic M, Davidovic M, Peric M, Abdul-Redah T
(2008) J Mol Struct 885:131–138
4
4
5. Bardin BB, Davis RJ (1999) Appl Catal A 185:283–292
6. Guo Y, Yang Y, Hu C, Guo C, Wang E, Zou Y, Feng S (2002)
J Mater Chem 12:3046–3052
1
23