Butanal Condensation Chemistry Catalyzed by Br?nsted Acid Sites on Polyoxometalate Clusters

Article abstract of DOI:10.1002/cctc.201601042

The connection of active site structures and their catalytic chemistry during butanal deoxygenation on polyoxometalate clusters with varying H⁺ site densities and identity of central atoms [H_xNa_4?xSiW₁₂O₄₀ (x=0–4) and H_yNa_3?yPW₁₂O₄₀ (y=0–3)] was established with rate assessment, IR spectroscopic, and chemical titration methods. Butanal adsorbs on the H⁺ or Na⁺ ions on polyoxometalate clusters and forms RC=O???H⁺ or RC=O???Na⁺ complexes at 348 K. A portion of the adsorbed butanals on the H⁺ sites converts to surface acetates through their reactions with vicinal framework oxygen atoms, as confirmed from the detection of ν_as(OCO) band at approximately 1580 cm^?1, and remains as the spectator species. Bimolecular reactions of butanals on the remaining H⁺ sites lead to 2-ethyl-2-hexenal as the predominant products, within which a small fraction undergoes sequential cyclization–dehydration to produce aromatics. A trace amount of butanal converts through minor, competitive pathways that form light olefins and dienes. These findings on the connection between active sites and their catalytic chemistry provide mechanistic insights useful for tuning the rates of the various concomitant paths and thus yields towards the different products during deoxygenation reactions.

Full text of DOI:10.1002/cctc.201601042

Accepted Article
Title: Butanal Condensation Chemistry Catalyzed by Brønsted Acid
Sites on Polyoxometalate Clusters
Authors: Yifei Yang, Fan Lin, Honghi Tran, and Ya-Huei (Cathy) Chin
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.201601042
Link to VoR: http://dx.doi.org/10.1002/cctc.201601042
A Journal of
www.chemcatchem.org

10.1002/cctc.201601042
ChemCatChem
FULL PAPER
Butanal Condensation Chemistry Catalyzed by Brønsted Acid Sites
on Polyoxometalate Clusters
Yifei Yang, Fan Lin, Honghi Tran, Ya-Huei (Cathy) Chin* [a]
Abstract: The connection of active site structures and their catalytic
chemistry during butanal deoxygenation on polyoxometalate clusters
with varying H+ site densities and identity of central atoms
[HxNa4-xSiW12O40 (x=0-4) and HyNa3-yPW12O40 (y=0-3)] was
established with rate assessment, infrared spectroscopic, and
chemical titration methods. Butanal adsorbs on the H+ or Na+ ions on
polyoxometalate clusters and forms RC=O∙∙∙H+ or RC=O∙∙∙Na+
complexes at 348 K. A portion of the adsorbed butanals on the H+
sites converts to surface acetates through their reactions with vicinal
framework oxygen atoms, as confirmed from the detection of νas(OCO)
band at ~1580 cm-1, and remains as the spectator species.
Bi-molecular reactions of butanals on the remaining H+ sites lead to
2-ethyl-2-hexenal as the predominant products, within which a small
fraction undergoes sequential cyclization-dehydration to produce
proton transfer weakens the C=O bond and shifts its vibrational
frequencies to lower wavenumbers, as found previously for
acetaldehyde adsorption on H-ZSM-5, during which the C=O
vibrational frequencies shift from 1725 cm-1 to 1686 cm-1.[9]
Separate density functional theory (DFT) calculations confirmed
such perturbations from a bond elongation between the lattice
oxygen atom of ZSM-5 and H+ from 0.97 to 1.04 Å.[11] The
hydrogen-bonded complexes (RC=O∙∙∙H+) are in chemical
equilibrium with their protonated forms (RC+-O-H),[12] because of
small differences in their heats of adsorption. For example, the
energy difference between chemisorbed and physisorbed
acetaldehyde on H-ZSM-5 is ~6 kJ mol−1.[13] Replacing the
Brønsted acid sites with alkali cations leads to aldehyde or ketone
adsorption on these cations as cationic bonded complexes
(RC=O∙∙∙Mn+), as evidenced from the detection of C=O vibrational
bands at ac.1715 cm-1 with infrared spectroscopy during
acetaldehyde adsorption on Li+, Na+, K+, Ca2+, Ni2+, or Al3+
cation-exchanged montmorillonites.[14] Such interactions reflect
the polarization of carbonyl groups by the cationic sites and
weakening of the C=O bond;[10c] these effects may in turn affect
their reactivities in deoxygenation reactions.
aromatics.
A trace amount of butanal converts through minor,
competitive pathways that form light olefins and dienes. These
findings on the connection between active sites and their catalytic
chemistry provide mechanistic insights useful for tuning the rates of
the various concomitant paths and thus yields towards the different
products during deoxygenation reactions.
The mechanism for aldol condensation reactions on solid
Brønsted acid and base site pairs has been well established.[2]
The reaction begins with an initial keto-enol tautomerization of the
aldehyde.[15] First, an aldehyde encounters the acid-base site pair,
during which the proton transfers to the carbonyl oxygen of the
aldehyde, whereas the neighbouring framework oxygen attacks its
alpha-hydrogen, forming an enol.[16] The enol then attacks a vicinal,
protonated aldehyde in a step that creates the inter-molecular C-C
bond and forms a hydroxyl-aldehyde intermediate, which upon
sequential dehydration evolves a larger, unsaturated aldehyde as
the condensation product.[15] The unsaturated aldehyde could also
coordinate to the H+ site as a hydrogen-bonded complex. The
aldol condensation occurs in parallel with transfer hydrogenation
and dehydration that form alkene and diene, respectively.[2] All
three primary reactions occur concomitantly at the catalyst
surfaces and their rates depend on the coverages and identity of
the reaction intermediates. For this reason, we seek to understand
the interactions between aldehydes and solid Brønsted acid sites,
probe the kinetic significance of the various aldehyde-derived
species, and relate these species to the various catalytic events
during aldehyde deoxygenation reactions.
Introduction
Catalytic deoxygenation removes oxygen atoms from
biomass derived oxygenates, thus increasing their heating values
and thermal stabilities as a step towards the sustainable synthesis
of fuels and value-added chemicals.[1] Solid Brønsted acids such
as H-ZSM-5 and H-FAU zeolites are attractive catalysts for the
deoxygenation of light alkanals (R-CHO; e.g., R=C2H5)[1b, or
2]
alkanones (R1C=OR2; e.g., R1=C2H5, R2=CH3)[3] at atmospheric
pressure and without the consumption of external H2. Specifically,
acid-catalyzed deoxygenation of propanal on H-ZSM-5 leads to
alkenes, larger unsaturated aldehydes, and diverse aromatics
(e.g., mono-, di-, and tri-methylbenzenes).[2] Among these
products, unsaturated aldehydes are important intermediates for
the production of fine chemicals used for perfumes,[4] flavors,[5]
and pharmaceutical products.[6] For example, acrolein and
2-ethyl-2-hexenal are attractive intermediates produced via aldol
type C-C bond coupling reactions between acetaldehyde and
formaldehyde[7] and between two butanal molecules,[8]
respectively. These reactions require the initial activation of
aldehydes at Brønsted acid sites (H+).
Small aldehydes or ketones adsorb at Brønsted acid sites via
their carbonyl oxygen atoms, forming hydrogen-bonded
complexes (RC=O∙∙∙H+), as concluded from infrared (IR) and
nuclear magnetic resonance (NMR) studies and quantum
chemical calculations (e.g., acetaldehyde on H-ZSM-5[9] and
acetone on H-ZSM-5[10]). The adsorption causes protonation of
the carbonyl groups and polarization of the adsorbates.[10c] The
Polyoxometalate clusters are a class of solid acid catalysts
with well-defined Keggin structure and chemical compositions of
H+8-n[Xn+Y12O40]n−8·zH2O (with X=P, Si, or B; Y=Mo, W, or V; n=
oxidation state of X, z=0-8). Their precise atomic connectivity
allows the accurate tuning of their acid strength and site
concentrations, by varying the identity of the central atoms (X) or
substituting their protons (H+) with other cations (e.g., Na+, K+),
while maintaining the overall structural integrity. They are effective
catalysts for the deoxygenation of light oxygenates, as reported
for 1,2-propylene glycol dehydration on H4SiW12O40,[17] glycerol
dehydration on H3PMo12O40 and H3PW12O40 clusters,[17-18] and
2-butanol dehydration on H8-nXn+W12O40 (X=P, Si, Al, and Co)
clusters.[19] Their structures also do not impose local, molecular
scale confinements on the active sites, thus eliminating the effects
[a]
Dr. Y.Yang, F. Lin, Prof. H. Tran, Prof.Y-H.Chin,
Department of Chemical Engineering and Applied Chemistry,
University of Toronto, Toronto, Ontario, M5S 3E5, Canada
Fax: +416 978 8605
E-mail: cathy.chin@utoronto.ca
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601042
ChemCatChem
FULL PAPER
arising from van der Waals interactions prevalent in acidic,
microporous crystalline solids. These structural and compositional
flexibilities (of polyoxometalate clusters) and their lack of local
confinements have allowed detailed, rigorous assessments of
kinetics and their connection to the acid strength, as demonstrated
previously for the case of 2-butanol dehydration.[19] Their lack of
local site confinements promotes the aldol condensation much
more effectively while suppressing hydride transfer and
dehydration reactions that form light alkenes and alkadienes,
when comparing with microporous crystalline catalysts (H-ZSM-5
and H-FAU zeolites). For example, the overall carbon selectivity to
butene and butadiene is ~3 % on H4SiW12O40, much less than the
40 % on H-MFI and 38 % on H-FAU during butanal reactions at
573 K (space velocity: 0.045, 0.0033, and 0.0074 s‒1 on
H4SiW12O40, H-MFI, and H-FAU, respectively).[20] For these
reasons, polyoxometalate clusters are an important class of
catalysts for selective transformation of small aldehydes to larger,
more valuable oxygenates.
A series of HxNa4-xSiW12O40 (x=0-4) or HyNa3-yPW12O40
(y=0-3) polyoxometalate clusters supported on SiO2 with varying
densities of Brønsted and Lewis acid sites are synthesized
according to the methods described in Experimental Section.
HxNa4-xSiW12O40
(x=0-4)
or
HyNa3-yPW12O40
(y=0-3)
polyoxometalate clusters are denoted hereinafter as HxNa4-xSiW
(x=0-4) and HyNa3-yPW (y=0-3), respectively. Butanal titrations at
348 K show that most of the H+ and Na+ sites remain accessible to
butanal molecules, with near unity butanal-to-cation ratios (0.78 to
0.96 on HxNa4-xSiW and 0.84-0.98 on HyNa3-yPW, respectively,
Table 1). A very small portion of the H+ and Na+ sites is
inaccessible to the titrant, likely due to cluster aggregation.
Figures 1a and 1b show infrared absorption spectra for pyridine
adsorption on the series of HxNa4-xSiW (x=0-4) and HyNa3-yPW
(y=0-3) samples, respectively, with varying Brønsted acid site
densities, χi (Equation 6, in Experimental Section), at 348 K,
together with a spectrum for pyridine adsorption on bare SiO2
(without polyoxometalate clusters, in Figure 1b) as the reference.
The absorption bands at 1637 cm-1, 1540 cm-1, and 1490 cm-1
correspond to the Py-H+ stretching (Py denotes pyridine) of
pyridinium ions adsorbed at the Brønsted acid sites (H+)[22] and the
bands at 1611 cm-1, 1490 cm-1, and 1450 cm-1 to Py-L stretching
(L denotes Lewis acid site) of pyridine coordinatively bonded to
Lewis acid sites of polyoxometalate clusters.[22b] Pyridine
adsorption on bare SiO2 leads to bands at 1595 cm-1 and 1445
cm-1, which correspond to the binding of pyridine at weak Lewis
acid sites at SiO2 surfaces. These peaks associated with SiO2
disappear after dispersing the polyoxometalate clusters onto SiO2
surfaces, caused apparently by the occupation of these sites by
the clusters.
Previous kinetic studies[21] have shown that polyoxometalate
clusters are selective for aldol condensation, but the details of
aldehyde adsorption and activation remain unestablished. Here,
we probe the adsorption and activation of butanal, a model
compound of aldehydes, on HxNa4-xSiW12O40 (x=0-4) and
HyNa3-yPW12O40 (y=0-3) polyoxometalate clusters, with different
site densities using combined infrared spectroscopic and kinetic
studies. Our findings show that butanals initially adsorb on the H+
sites as H+-bonded complexes (RC=O∙∙∙H+), a portion of which
converts to surface acetate by reactions with a vicinal framework
oxygen atom and remains as spectators during catalysis. The
other portion of H+-bonded butanals undergoes aldol
condensation to produce 2-ethyl-2-hexenal, which may undergo
sequential condensation and cyclization-dehydration, leading to
cyclic oxygenates or aromatics. On Na+ sites, butanals adsorb
onto the Na+ sites as inactive Na+-bonded complexes
(RC=O∙∙∙Na+) and these complexes do not participate in the
reactions. These systematic studies on the adsorbed species and
their catalytic roles shed light into the catalytic details and the
connection between site structures and their rates and
selectivities.
Pyridine uptakes on the series of samples were determined
from separate pyridine chemical titrations (in flow mode) at 473 K,
during which pyridine was introduced to and adsorbed onto the
catalysts. The cumulative pyridine uptakes for each
polyoxometalate cluster (휆ads, Equation 7) reflect the amounts of
pyridine adsorbed per cluster at saturation. These uptakes are
summarized in Table 1. After the titration experiments, the overall
amounts of pyridine desorbed from the catalysts (휆des,overall
,
Equation 9a) were measured from the sequential temperature
programmed desorption experiments and are also summarized in
Table 1. The uptake values 휆ads, which reflect the molar ratio of
pyridine per polyoxometalate cluster, are 3.13 for H4Na0SiW and
2.68 for H3Na0PW clusters. Exchanging a portion of the H+ sites
with Na+ ions in HxNa4-xSiW decreased the nominal H+ site
densities (χP, Equation 6) from 4 to 0 and the related 휆ads values
from 3.13 to 0.93. Similarly, exchanging the H+ with Na+ on
HyNa3-yPW decreased the nominal H+ densities from 3 to 0 and in
Results and Discussion
Quantifications of Brønsted and Lewis acid sites on
supported HxNa4-xSiW and HyNa3-yPW clusters by pyridine
adsorption and temperature programmed desorption
Table 1. Nominal H+ site densities and pyridine/butanal uptakes on HxNa4-xSiW (x=0-4) and HyNa3-yPW (y=0-3) clusters
Pyridine desorption during temperature programmed desorption
Nominal H+
(χi, i=Si,P,
mol/molcluster
Pyridine uptake
(휆ads
mol/molcluster)
experiments (mol/molcluster
)
Butanal uptake
(훼but, mol/molcluster)
Butanal-to-
cation ratio
Sample
,
Total
(휆des,overall
Low Medium
High
[d]
)
[a]
[b]
[c]
)
(휆des,1
0.92
0.83
0.36
0.28
0.51
0.28
0.36
0.54
0.77
)
(휆des,2
)
(휆des,3
0.31
0.27
0.13
B.D.
B.D.
0.27
0.23
B.D.
B.D.
)
H4Na0SiW
H3Na1SiW
H2Na2SiW
H1Na3SiW
H0Na4SiW
H3Na0PW
H2Na1PW
H1Na2PW
H0Na3PW
4
3
2
1
0
3
2
1
0
3.13
2.63
1.86
1.07
0.93
2.68
1.70
1.22
0.98
2.72
2.55
1.39
0.82
0.51
2.12
1.67
0.96
0.77
1.49
1.46
0.90
0.54
B.D.
1.57
1.08
0.42
B.D.
3.33
3.30
3.85
3.43
3.10
2.93
2.62
2.53
2.64
0.83
0.82
0.96
0.86
0.78
0.98
0.87
0.84
0.88
[a] obtained by integrating the pyridine desorption peak over the entire temperature range; [b], [c], and [d] obtained by decoupling the pyridine desorption peak to
three temperature ranges of 500-650 K, 550-750 K, and 673-790 K, respectively; B.D.: below the detectable limit.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601042
ChemCatChem
FULL PAPER
On both sample series, pyridine desorption began at similar
temperatures (~500 K), irrespective of their H+ site densities (χi,
i=Si or P, Equation 6) and the identity of the central atom (Si or P).
Pyridine desorption, however, completed at temperatures that
varied with the H+ site densities (χP, Equation 6), when comparing
within the same series of either HxNa4-xSiW or HyNa3-yPW clusters.
Pyridine desorbed from these samples in three different
temperature ranges of 500-650 K, 550-750 K, and 673-790 K. Our
previous study shows that thermal treatment at 673 K may reduce
the H+ site densities on H3PW12O40 (by 17 %) via dehydration.[21]
We note that the pyridine desorption above 673 K occurs together
with a concomitant reduction in the H+ site density, as the Keggin
clusters dehydrate while maintaining their structural integrity.[23]
turn the 휆ads values from 2.68 to 0.98.
The decomposition of polyoxometalate clusters (into WO3 and
[23]
SiO2
)
takes place above ~800 K, according to the
thermogravimetric-differential scanning calorimetric analyses
(TG-DSC) of H4SiW12O40 catalysts in our previous work.[21] The
onset temperature of decomposition is consistent with those
previously reported based on differential thermal analyses (DTA)
and X-ray diffraction (XRD) analyses (803 K for H4SiW12O40 and
773 K for H3PW12O40).[23] Since most pyridine molecules desorb
below 800 K (Fig. 2), the pyridine-TPD profiles reflect mostly the
binding strengths of pyridine molecules on the intact clusters.
Figure 1. Normalized infrared absorption spectra taken after pyridine adsorption
(0.88-1.32 μmol) on (a) HxNa4-xSiW (x=0-4) and (b) HyNa3-yPW (y=0-3) at 473 K
and then cooling to 348 K for measurements.
Figures 2a and 2b show pyridine desorption rates during
temperature programmed desorption from the series of
HxNa4-xSiW (x=0-4) and HyNa3-yPW (y=0-3) clusters, respectively.
These desorption rates reflect strictly the contributions from the
polyoxometalate clusters, free of those from SiO2 supports,
because He purge at 473 K for 30 min, conducted before the
temperature programmed experiments, removed all weakly
adsorbed pyridines on SiO2 surfaces. In addition, pyridine
desorption peaks in the temperature programed desorption
profiles are ascribed to those adsorbed on the Brønsted acid sites
only, because most of the pyridines adsorbed on the Lewis acid
sites (1450 cm-1) were easily removed, either by purging with
flowing He (Figure S1, Supporting Information) or by evacuating
(Figure S2, Supporting Information) the samples at 473 K. There
are only a small amount of strong Lewis sites which can bind to
pyridines. In fact, in our previous work, the Brønsted-to-Lewis acid
site ratio on H4Na0SiW sample was determined to be ~15:1, based
on the infrared spectra of pyridine adsorption followed by He
purging.[21]
4
3
2
1
0
1.5
1.0
0.5
0.0
HxNa4-xSiW
HyNa3-yPW
HxNa4-xSiW
HyNa3-yPW
b
a
des,overall

des,2-3
0
1
2
3
4
0
1
2
3
+
)-1)
-1
i (mol H (mol
des,overall(mol (mol
) )
cluster
cluster
Figure 3. (a) The total amount of pyridine desorbed (휆des,overall , between
500-790 K, Equation 9a) and the amount of pyridine desorbed from the high
temperature ranges ( 휆des,2−3
, between 550-790 K, Equation 9b) during
temperature programmed desorption, plotted as a function of the nominal H+ site
densities (χi, i=Si or P, Equation 6) and (b) infrared absorption band intensity at
1540 cm−1, which corresponds to pyridinium ion at the Brønsted acid sites,
plotted as a function of the overall amount of H+ sites (휆des,overall, Equation 9a)
determined by temperature programmed desorption of pyridine.
It has been established that H+ sites on the polyoxometalate
clusters are chemically equivalent to each other[24] and their acid
strengths are dictated by the enthalpies required to remove them
to non-interacting distances from the clusters, as defined by the
deprotonation enthalpy (∆HDPE, which equals the heat of reaction
for HnXY12O40 → H+ + Hn-1XY12O40−; X=P, Si, or B, Y=Mo, W, or V,
n=oxidation state of X; ∆HDPE values are 1087 kJ mol-1 for
H3PW12O40 and 1105 kJ mol-1 for H4SiW12O40 clusters[19a, 25]).
However, when a portion of the H+ sites on a cluster is partially
replaced by Na+ ions or is occupied by pyridines (as C5H5NH+),
the deprotonation enthalpies for the remaining H+ sites would
increase and acid strength would commensurately decrease. Na+
and pyridinium ion (C5H5NH+) have lower electron affinities (496[26]
and 590 kJ mol‒1,[27] respectively) than H+ (1312 kJ mol‒1).[26] Thus,
replacing H+ sites by Na+ ions or adsorbing pyridine onto the H+
Figure 2. Pyridine desorption rates during temperature programmed desorption
of pyridine on (a) HxNa4-xSiW (x=0-4) and (b) HyNa3-yPW (y=0-3), after pyridine
adsorption followed by purging in flowing He for 30 min at 473 K (heating rate=
0.033 K s-1).
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601042
ChemCatChem
FULL PAPER
sites, transforming them into C5H5NH+ ions, would result in less
extent of charge delocalization on polyoxometalate anions
Pyridine also desorbs from H-Beta and H-ZSM-5 zeolites at
similar temperature ranges,[28] on which H+ sites are weaker and
deprotonation enthalpies are higher at 1195 kJ mol-1 and 1200 kJ
mol-1,[25] respectively. Pyridine desorbs from H-Beta (Si/Al=13) in
the temperature ranges of 453-623 K, 473-873 K, and 673-1023 K
and from H-ZSM-5 (Si/Al=84) in the ranges of 423-573 K and
673-973 K. The temperature ranges are higher for the zeolites
than polyoxometalate clusters (Figure 2), likely because the heats
of pyridine desorption (∆Hdes, which reflect the heats of reaction for
PyH+→Py + H+) in these microporous crystalline structures (5.6 Å
and 6.6-7.7 Å for H-beta and 5.5 Å for H-ZSM-5) contain
significant enthalpy contributions arising from van der Waals
interactions between pyridines and the local pores and cages of
the zeolites.
The overall pyridine desorption rate profiles for HxNa4-xSiW
(x=0-4) and HyNa3-yPW (y=0-3) in Figure 2 were decoupled into
three temperature ranges of 500-650 K, 550-750 K, and 673-790
K and their corresponded amounts of pyridine desorbed, denoted
as 휆des,1, 휆des,2, and 휆des,3 (Equation 8), are summarized in Table
1. Within each series of samples, both the total amount of pyridine
desorbed (휆des,overall, Equation 9) and the amount of pyridine
desorbed at the higher temperatures (휆des,2−3) increase linearly
with the nominal H+ site densities (χi, Equation 6), as shown in
Figure 3a. The dotted lines in Figure 3a are the linear fits for
휆des,overall and 휆des,2−3; these fits run in parallel, an indication that
the last H+ site on the polyoxometalate clusters is weak (because
of the neighboring Na+ ions), thus it cannot be substituted by a Na+
ion ( 휆des,1 , with pyridine desorption temperatures between
500-650 K). The integrated intensities for the pyridine absorption
bands at 1540 cm-1 and 1450 cm-1 in Figure 1 reflect the site
densities for Brønsted and Lewis sites,[29] respectively. The
integrated intensities at 1540 cm-1 increase proportionally with the
total amount of pyridine desorbed during the temperature
programmed desorption (휆des,overall, Equation 9, Figure 2) for both
HxNa4-xSiW and HyNa3-yPW clusters, as shown in Figure 3b,
irrespective of the chemical identity of their central atoms. This
linear, direct relation suggests that pyridine molecules are
desorbed from the Brønsted acid sites during the temperature
programmed desorption experiments.
4‒
(SiW12O40 and PW12O403‒) to these cations. The lower extent of
delocalization causes the anions to bind to the remaining H+ sites
more strongly. We hypothesize that this extent of charge
delocalization leads to the multiple desorption peaks during the
temperature programmed desorption (Fig. 2): during the
desorption of a pyridine (C5H5N) molecule, the pyridinium ion
(C5H5NH+) converts back to a H+ ion, which interacts with the
anion more strongly than for the case of C5H5NH+ ion-anion
interactions. This increase in the anion-H+ interactions decreases
the extent of electron donation from the anion to the rest of the H+
ions and thus causes the strength of the remaining H+ sites on the
polyoxometalate clusters to increase concomitantly. Such
changes in acid strength lead to the higher desorption
temperature required for the removal of next pyridine molecules
from the polyoxometalate clusters. For the Na+ exchanged
samples, the anion-Na+ interactions are weaker than the anion-H+
interactions. As a result, the anion retains most of its negative
charge and is able to donate the electron density into the
remaining H+ sites, making them weaker. These weaker H+ sites
give the pyridine desorption peak at the lower temperatures.
Butanal chemical titration and infrared absorption studies on
HxNa4-xSiW (x=0-4) and HyNa3-yPW (y=0-3) clusters supported
on SiO2 support
Butanal uptakes determined from chemical titrations carried
out on the series of HxNa4-xSiW (x=0-4) and HyNa3-yPW (y=0-3)
clusters at 348 K give the average number of butanal molecules
adsorbed per polyoxometalate cluster, 훼but (Equation 11), on
these samples. The 훼but values, as summarized in Table 1, are
nearly independent of the H+ densities (χi) and the related
-1
H+-to-Na+ ratios: they range from 3.10 to 3.85 molbutanal molcluster
-1
for HxNa4-xSiW clusters and 2.53 to 2.93 molbutanal molcluster for
HyNa3-yPW clusters. These 훼but values suggest that butanal
adsorbs on both H+ and Na+ sites of the HxNa4-xSiW and
HyNa3-yPW clusters.
Figure 4. Infrared absorption spectra, taken during adsorption of butanal (6
pulses total, 0.32 μmol per pulse, labeled 1st, 3rd, and 6th pulse on the figure) to
saturation coverages and then evacuation under dynamic vacuum at 348 K, on
H4Na0SiW (a, b, c) and H3Na0PW (d, e, f) clusters. The evacuation times (3-50
min) are indicated on the figure. Region (a, d) 3600-4000 cm−1 corresponds to
the region of hydroxyl stretching, Region (b, e) 2650-3050 cm−1 to methyl
stretching, and Region (c, f) 1320-1770 cm−1 to carbonyl stretching, surface
acetate stretching, and methyl bending, respectively.
On H4Na0SiW and H3Na0PW clusters, evacuation under
dynamic vacuum (0.01-0.1 mbar) at 348
K after butanal
adsorption leads to the infrared absorption spectra in Figure 4.
The spectra show three regions of: (i) hydroxyl stretching at
3745-3725 cm-1 (Figures 4a and 4d), (ii) methyl stretching at
3000-2700 cm-1 (Figures 4b and 4e), and (iii) stretching of
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601042
ChemCatChem
FULL PAPER
adsorbed carbonyl (C=O) and surface acetate (OCO) at
1750-1550 cm-1 and methyl bending at 1465-1380 cm-1 (Figures
4c and 4f). Details of the absorption band assignments are
summarized in Table 2, together with those reported in the
H3Na0PW (Figure 4f) suggest the formation of surface acetate
[νas(OCO), Structure C, Scheme 1], as found also during
acetaldehyde adsorption on oxide surfaces (CeO2[30] and TiO2[31]).
Acetaldehyde adsorption at 473 K led to the characteristic surface
acetate band at 1542 cm-1 on CeO2 and at 1562 cm-1 on TiO2.[30-31]
The coverages of surface acetate are determined from the
νas(OCO) band intensities at 1580 and 1577 cm-1 for the entire
series of HxNa4-xSiW (x=0-4) and HyNa3-yPW (y=0-3) clusters,
respectively, as shown in details later in Figures 7c and 7f. The
νas(OCO) band intensities appear to relate directly to the amount
of pyridine desorbed between 673-790 K during the temperature
programmed desorption of pyridine, 휆des,3 (Equation 8), as
depicted in Figure 5. The linear relation between the results from
two independent characterization techniques, i.e. the infrared
absorption band intensities for surface acetate at 1580 cm-1 or
1577 cm-1 and the amounts of pyridine desorbed from 673-790 K
during the temperature programmed desorption experiments,
confirms that both the surface acetates and the strongly
chemisorbed pyridine molecules preferentially occupy an identical
type of sites. Such relation appears to remain so, irrespective of
the chemical identity of the central atom of polyoxometalate
clusters.
[30]
literature for acetaldehyde adsorption on various oxides (CeO2
and TiO2[31]) and zeolites (H-ZSM-5[9] and Na-Y[32]). Butanal
adsorption on polyoxometalate clusters leads to the detection of
C=O stretching vibration [νbut.(C=O)] at ~1707 cm-1 (Figures 4c
and 4f) through binding its carbonyl oxygen on the Brønsted acid
site (H+) as a RC=O∙∙∙H+ complex (Structure A, Scheme 1),[9]
which causes a red-shift of ~43 cm-1 from that of its gas phase
molecule (Table 2). This shift is a result of the partial proton
transfer to the carbonyl group, which polarizes and weakens the
C=O bonds of the adsorbed butanal.[10c, 16]
Scheme 1. Adsorption configurations of butanal on HxNa4-xSiW (x=0-4) and
HyNa3-yPW (y=0-3) clusters. The
wavenumbers corresponded to the carbonyl and acetate group stretching are
indicated in the scheme.
R denotes the -CH2-CH3 group. The
Scheme 2. The proposed pathway for surface acetate formation on HxNa4-xSiW
(x=2-4) and HyNa3-yPW (y=2-3) clusters during butanal adsorption at 348 K.
The surface acetate formation may involve an initial butanal
protonation which forms C4H8O-H+ (Structure 2.1, Scheme 2) as
the most abundant surface intermediate. The preferential
adsorption and protonation of aldehydes on H+ sites have been
previously confirmed from butanal titration, which gives C4H8O:H+
molar ratios of 1.0, 1.01, and 1.1 on MFI (348 K), H-FAU (448 K),
and H4SiW12O40 (348 K),[20] respectively, and consistent with the
first-order rate dependence for the primary, propanal aldol
condensation rates with propanal pressure.[2] An adjacent lattice
oxygen (Oox, Lewis basic site) acting as a nucleophile attacks the
carbonyl carbon of the protonated butanal, resulting in the
adsorbed acetate (C4H8O-H+-Oox, Structure 2.2),[33] as depicted in
Scheme 2. The H+ and vicinal lattice oxygen (Oox) acting as a
Brønsted acid-base site pair; the site pair catalyzes several
classes of reaction, such as acetaldehyde keto-enol
tautomerization on ZSM-5[15] and butanol dehydration on
H8-nXn+W12O4 (X=P, Si, Al, and Co) polyoxometalate clusters.[19a]
The surface acetate species remain stable only at low
temperatures (<348 K) and easily desorb to evolve into heavy
aromatics during the sequential heating stage, as evidenced from
the changes in the νas(OCO) band (at 1580 cm-1) on H4Na0SiW
clusters to the ν(C=C)aromatic band (1600 cm-1) in Section S4 of
Supporting information. In fact, in our previous work, we have
demonstrated that only cokes but no acetate species remain on
the H4Na0SiW clusters after butanal reactions at high
Figure 5. Normalized infrared absorption band intensities of surface acetate
(1580 cm−1 for HxNa4−xSiW and 1577 cm−1 for HyNa3−yPW clusters), plotted as a
function of the H+ site densities determined during temperature programmed
desorption of pyridine between 673 K and 790 K (휆des,3, Equation 8).
The bands at 1654 cm-1 [νdi.(C=O)] and 1632 cm-1 [ν(C=C)]
are indicative of adsorbed 2-ethyl-2-hexenal, the aldol
condensation product,[30] as it interacts with the H+ site via its
carbonyl oxygen atom (Structure B, Scheme 1). The bands
detected at 1580 cm-1 on H4Na0SiW (Figure 4c) and 1577 cm-1 on
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601042
ChemCatChem
FULL PAPER
temperatures (523-573 K) and that the coke formation poisons all
the strong H+ sites and a portion of medium and weak H+ sites.[21]
cm−1, which correspond to the different modes of methyl stretching
from the adsorbed butanal, 2-ethyl-2-hexenal, and surface acetate
(Structures A, B, and C, respectively, in Scheme 1), and the
concomitant evolvement of bands related to their bending
vibrations in 1500-1350 cm−1 in Figure 4. The intensities of these
bands, including hydroxyl stretching [ν(OH)] at 3740 cm-1 and
3725 cm-1, methyl stretching [ν(CH3)] at 2754-2970 cm-1, carbonyl
stretching [ν(C=O)] at 1707 cm-1 and 1654 cm-1, acetate stretching
[νas(OCO)] at 1580 cm-1 and 1577 cm-1, and methyl bending
[δ(CH3)] at 1500-1350 cm-1, initially increased as the exposure to
butanal increased (Figure 4, labeled 1st, 3rd, and 6th pulses, each
with 0.32 μmol of butanal). Above a threshold butanal exposure,
the band intensities became unchanged because the surface
reached saturation. Sequential evacuation under dynamic vacuum
at 348 K removed the butanal molecules (Structure A, Scheme 1)
selectively, as shown from the marked decrease in the band
intensities for the C=O vibration of butanal at ~1707 cm-1 in
Figures 4c and 4f. In contrast, band intensities related to the
adsorbed 2-ethyl-2-hexenal [e.g., the ν(C=O) at 1654 cm-1 and
ν(C=C) at 1632 cm-1] and surface acetate [νas(OCO) at 1580 cm-1
or 1577 cm-1] remained relatively stable at 348 K for the entire
evacuation period (50 min) under dynamic vacuum, as shown in
the same profiles in Figure 4. During the evacuation, a band at
1510 cm-1, which corresponds to the C=C stretching in light
aromatics [ν(C=C)aromatic], began to emerge at ~10 min. This band
reflects the formation of aromatics from the sequential
condensation, ring-closure, and dehydration reactions of
alkenals,[1b, 2] as described in next.
Isothermal butanal adsorption at 348 K on H0Na4SiW and
H0Na3PW clusters, which contain predominantly Na+ ions with a
small fraction of residue H+ sites (휆des,overall, Equation 9a, residue
H+ site densities of 0.51 and 0.77 mol molcluster−1, respectively)
gives absorption spectra unlike those of the fully protonated
H4Na0SiW and H3Na0PW clusters. The spectra of adsorbed
butanal, taken under dynamic vacuum and shown in Figure 6,
show absorption bands at 1720-1715 cm−1 on both H0Na4SiW and
H0Na3PW clusters. These bands, undetected on H4Na0SiW and
H3Na0PW clusters (Figure 4), reflect the formation of butanal-Na+
complex [RCH2(CH=O)∙∙∙Na+ complex, Structure D, Scheme 1]
with a red-shift of ~30 cm−1 from that of the gas phase butanal
(Table 2). The shift may reflect electron transfers from the
carbonyl groups to polyoxometalate frameworks and thus the
weakening of the C=O bonds in butanals.[14] The new feature
emerged at 1690 cm−1 may reflect the C=O vibrational band of
2-ethyl-2-hexenal, formed from the aldol condensation reactions
on H+ sites and migrated to bond to the Na+ sites (Structure E,
Scheme 1). The Na+ sites are inactive in catalyzing the aldol
condensation, as confirmed in our previous work on butanal
reaction on Na+ exchanged MFI zeolites at 573 K.[20] The negative
Figure 6. Infrared absorption spectra, taken during adsorption of butanal (7
pulses total, 0.32 μmol per pulse, labeled 1st, 3rd, and 7th pulse on the figure) to
saturation coverages and then evacuation under dynamic vacuum, on H0Na4SiW
clusters (a, b, c) and H0Na3PW clusters (d, e, f) at 348 K. The evacuation times
are provided in the figure. Region 3600-4000 cm−1 (a, d) corresponds to the
region of hydroxyl stretching, Region 2650-3050 cm−1 (b, e) to methyl stretching,
and Region 1320-1770 cm−1 (c, f) to carbonyl stretching, surface acetate
stretching, and methyl bending.
Butanal adsorption and the formation of adsorbed
2-ethyl-2-hexenal and surface acetate (OCO) at the Brønsted acid
sites perturb the hydroxyl stretching bands at 3740 cm-1 on
H4Na0SiW and 3725 cm-1 on H3Na0PW clusters, as evidenced by
the appearance of negative absorption bands in Figures 4a and
4d. These events also lead to the series of bands at 2754-2970
Table 2. Infrared absorption band (in cm-1) assignments for butanal adsorption on HxNa4-xSiW (x=0-4) and HyNa3-yPW (y=0-3) clusters at 348 K and acetaldehyde
adsorption on oxides and zeolites at 313 K
[30]
[31]
C4H8O (g)[a]
H4Na0SiW
H0Na4SiW
H3Na0PW
H0Na3PW
CeO2
TiO2
Na-Y[32]
H-ZSM-5[9]
Band
2972-2894
2970-2935
2880
2970-2935
2970-2935
2880
2970-2935
2960
--
2969
2914
2759
1718
1686
1636
1562
--
3050
--
νas(CH3)
νs(CH3)
2805
2880
2880
--
--
2710
2750
2754
2745
2750
2727
1729
1686
--
2750
--
ν(CH)
1750
1707
1720
1706
1715
1722
1707
1630
1602
--
νbut.(C=O)
νdi.(C=O)
ν(C=C)
--
1654
1690
1654
1690
-
--
1632
--
1632
--
--
--
1580
--
1577
--
1542
--
--
νas(OCO)
ν(C=C)aromatic
δ (CH3)
--
1510
--
1510
--
--
--
1462-1394
1460-1380
1460-1380
1460-1380
1460-1380
1440-1372
1355
1428-1349
1430-1397
[a] Infrared spectrum of gas phase butanal was measured at room temperature and shown in Figure S3.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601042
ChemCatChem
FULL PAPER
peak at 1625 cm−1, which corresponds to the bending vibration of
water, emerged because skeleton water was being eliminated
from the H0Na4SiW and H0Na3PW clusters as butanal
adsorbed,[34] as these clusters contained crystal water even after
the initial pretreatment.[35] In the hydroxyl region (Figures 6a and
6d), a trace amount of butanal adsorbed on the silanol group of
SiO2 supports and led to the weak negative peak at 3745 cm−1.
The assignment of this peak was confirmed from infrared
absorption experiments of butanal adsorption on bare SiO2
support, also shown in Figure 7d.
vacuum at 348 K, after butanal adsorption on HxNa4−xSiW clusters (x=0-4) (a, b,
c) and HyNa3−yPW clusters (y=0-3) (d, e, f) to surface saturation. Region
3600-4000 cm−1 (a, d) corresponds to the region of hydroxyl stretching, Region
2650-3050 cm−1 (b, e) to methyl stretching, and Region 1320-1770 cm−1 (c, f) to
carbonyl stretching, surface acetate stretching, and methyl bending.
Figure 7 shows the infrared absorption spectra for butanal
adsorption on the series of HxNa4−xSiW (x=0-4) and HyNa3−yPW
(y=0-3) clusters, respectively, with different H+ site densities (χi,
Equation 6) and therefore different Na+ site densities, taken under
dynamic vacuum at 348 K. The H+ densities are related to the Na+
site densities, because the sum of site densities (both H+ and Na+)
equals 4 for HxNa4−xSiW (x=0-4) and 3 for HyNa3−yPW (y=0-3)
clusters, as required by the overall site balance relation. Butanal
adsorption on both series of HxNa4−xSiW and HyNa3−yPW clusters
with high H+ site densities (χi>2) leads to the formation of surface
acetate [νas(OCO) at 1580 cm−1 or 1577 cm−1, Figures 7c and 7f].
As the nominal H+ density (χi, Equation 6) decreases and the Na+
density concomitantly increases, the δ(HOH) peak at 1625 cm−1
gradually increases because more butanals adsorb at the Na+
sites. This decrease in the nominal H+ site density (χi) leads also
to the decrease in the νas(OCO) band intensity and the
concomitant increase in the νbut.(C=O) band intensity. The
disappearance of the νas(OCO) band indicates that surface
acetate forms only at the H+ site and not at the Na+ site. When the
nominal H+ site density (χi) varies, the νdi.(C=O) bands
corresponded to 2-ethyl-2-hexenal remain unchanged, as shown
in Figure S6. In addition, the characteristic C=O stretching bands
for adsorbed butanal shift from ~1706 cm−1 (bound on H+ sites) to
~1720 cm−1 (bound on Na+ sites) for both HxNa4−xSiW (x=0-4) and
HyNa3−yPW (y=0-3) clusters with decreasing H+ site density.
Similarly, the vibrational frequencies of the νdi.(C=O) bands of
adsorbed 2-ethyl-2-hexenal [e.g. 1654 cm−1 for H4Na0SiW
clusters] also shift to higher wavenumbers. These changes in
wavenumbers reflect the decreasing extent of interactions
between the cation (H+ or Na+) and the carbonyl groups.
Rates and selectivities of butanal deoxygenation on
supported HxNa4-xSiW and HyNa3-yPW clusters
The reaction pathways for butanal deoxygenation on
HxNa4−xSiW (x=0-4) and HyNa3-yPW (y=0-3) clusters are proposed
Figure 7. Infrared absorption spectra, taken during evacuation under dynamic
Table 3. The rate constants and carbon selectivities to various products during butanal deoxygenation on HxNa4-xSiW and HyNa3-yPW clusters at 573 K
Carbon selectivities in products [e] (mol.%)
Large alkenals (>C8) Large aromatics (>C8)
Rate [a]
Rate constant [c]
(10-3 mol (molcluster s)-1)
(10-2 mol (molcluster s kPa)-1)
Samples
C4
olefin
f]
Other [
[b]
[d]
[d]
[b]
[b]
푟Dehy
푘
푘
Inter,C8
푟
Inter
푟
Intra
Total
83.1
C8
C12-16
18.5
Total
14.3
C8
C12-16
12.7
Inter,C4
1.95
0.76
0.31
0.25
0.37
2.06
1.08
0.56
0.91
4.61
2.21
1.30
2.20
1.23
4.26
1.81
1.74
0.87
64.6
1.6
2.3
3.5
3.0
2.1
1.5
0.6
1.1
1.0
0.5
0.3
0.3
1.0
1.5
0.5
0.3
0.1
0.5
0.2
H4Na0SiW
H3Na1SiW
H2Na2SiW
H1Na3SiW
H0Na4SiW
H3Na0PW
H2Na1PW
H1Na2PW
H0Na3PW
18.2
7.5
2.8
2.4
3.5
20.2
9.9
5.1
7.7
0.74
0.43
0.12
0.02
0.01
0.15
0.09
0.03
0.01
0.25
0.19
0.11
0.10
0.12
0.16
0.16
0.09
0.07
82.5
80.3
83.4
91.6
84.2
88.0
90.1
94.5
70.0
67.4
74.5
86.2
61.6
77.0
80.0
87.6
12.5
12.9
8.9
13.7
15.7
13.0
6.4
2.5
2.8
1.6
0.6
1.0
1.4
1.2
0.6
11.2
12.9
11.4
5.8
5.4
22.6
11.0
10.1
6.9
14.9
10.8
8.4
13.9
9.4
7.2
4.8
4.2
[a] Rates for butanal reactions on polyoxometalate clusters were measured in 1.1 kPa butanal at 573 K during steady-state reaction (time-on-stream=305 min),
subscript cluster=HxNa4-xSiW or HyNa3-yPW, space velocity=0.02 molbutanal (molcluster s)-1; [b] 푟Inter, 푟Dehy, and, 푟
refer to the rates for primary inter-molecular
Intra
C=C bond formation (Step a1), isomerization-dehydration (Step b1), and intra-molecular C=C bond formation (Step c1), respectively, in Scheme 3; [c] Rate
constants for butanal reactions on polyoxometalate clusters were measured in 1.1-4.4 kPa butanal at 573 K during steady-state reaction (time-on-stream=155-605
min), subscript cluster=HxNa4-xSiW or HyNa3-yPW, space velocity=0.02-0.08 molbutanal (molcluster s)-1; [d] 푘Inter,C and 푘Inter,C refer to the first-order rate constants
4
8
for the primary (Step a1) and the secondary (Step a2) inter-molecular C=C bond formation, respectively, in Scheme 3; [e] Selectivities for butanal reactions were
measured during steady-state reaction (time-on-stream=305 min), defined as the amount of carbon in a product divided by the total carbon in all products; [f] Other
species includes C7H14O, C7H14, and some unidentified species (e.g., C8+ species).
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601042
ChemCatChem
FULL PAPER
in Scheme 3, which are consistent with those proposed for butanal
reactions on H-FAU and H-MFI[20] and also for propanal reactions
on H-MFI.[2] Polyoxometalate clusters catalyze butanal
deoxygenation via step-wise inter-molecular C=C bond formation
(Steps a1-a3) as the predominant reaction path (>80 % carbon
selectivities). The series of reactions occurs at H+ sites via the
initial attack of protonated butanal by another butanal in its
tautomeric form (butenol). These coupling reactions repeat
multiple times and lead to a systematic increase in its molecular
during sequential heating to above 423 K, indicated by the
decrease in [νas(OCO)] bands at 1580 cm−1 and concomitant
increase in [ν(C=C)aromatic] bands at 1600 cm−1 (Figure S5a). The
light aromatics (e.g., at 1510 cm−1) occupying the H+ sites could
undergo sequential dehydrogenation (not shown in Scheme 3) to
produce heavier aromatics (e.g., that leads to bands at 1600 cm−1,
Figure S5a). The formation of heavier aromatics during the initial
contact of polyoxometalate clusters to the feed leads to the
decrease in the available H+ sites after reaction at 573 K, as
size
from
butanal
(C4)
to
2-ethyl-2-hexenal
(C8),
measured by the temperature desorption of pyridine. The site
−1
2,4-diethyl-2,4-octadienal (C12), 2,4,6-triethyl-2,4,6-decatrienal
(C16), and their isomers [e.g., 2-methyl-5-ethyl-cyclopentanone
(C8)], because each inter-molecular C=C bond formation event
incorporates a butanal unit (four C atoms) into the growing carbon
chain and removes a H2O from the molecule. These reactions
produce near exclusively oxygenates and aromatics with their
carbon numbers equal exactly the multiple of butanal units (C4n,
where n=2, 3, or 4). Unlike butanal reactions on microporous
crystalline solids (H-MFI and H-FAU),[20] alkylation/dealkylation
reactions of aromatic products remain kinetically inconsequential
on polyoxometalate clusters (the aromatic products are almost
exclusively C4n species, n=2, 3, or 4), because these reactions are
prevalent only in pores and cages of molecular dimensions.
+
density becomes stable at 1.54±0.32 molH (molcluster
)
for
reaction times above 30 min, as reported elsewhere.[21] Infrared
spectra of pyridine adsorption on spent HxNa4−xSiW catalysts
indicate that only Brønsted acid sites [with ν(Py-H+) band at 1540
cm-1] and almost no Lewis acid sites [with ν(Py-L) band at 1450
cm-1] remain accessible to pyridine titrants after butanal reactions
at 573 K for 30 min (Figure S7).
Polyoxometalate
clusters
also
catalyze
the
isomerization-dehydration (Step b1) and intra-molecular C=C
bond formation by hydride transfer (Step c1) reactions that form
butadiene (C4H6) and butene (C4H8), respectively. These
reactions are, however, at least two orders of magnitude smaller
than the inter-molecular C=C bond coupling (푟Intra /푟Inter =0-0.06,
푟Dehy /푟Inter =0.008-0.04, 573 K, as shown by the selectivity data in
Table 3). These rate ratios significantly smaller than unity are
unlike those found on confined, microporous H-MFI or H-FAU
structures: the rate ratios of intra-molecular to inter-molecular C=C
bond formation, 푟Intra /푟Inter , are 1.5 and 2.4 on H-MFI and H-FAU,
whereas those of dehydration to inter-molecular C=C bond
formation, 푟Dehy /푟Inter , are 0.68 and 0.07 on H-MFI and H-FAU,
respectively, at 573 K.[20] These distinct rate ratios between the
unconfined and confined sites indicate that structural
confinements promote the intra-molecular C=C bond formation
and isomerization-dehydration pathways, possibly via solvation of
their respective transition state through van der Waals interactions
in confined voids.[36]
Butanal reactions on all polyoxometalate clusters approached
steady-state for reaction times above 155 min, at which the rates
and selectivities remained unchanged with the time-on-stream, as
shown and reported elsewhere.[21] At steady-state, the
predominant reaction pathway is the inter-molecular C=C bond
formation: the selectivities towards larger alkenals and aromatics
are 97 mol.% at 573 K, whereas those towards butene and
butadiene are less than 4 mol.% selectivities at 573 K for all
HxNa4−xSiW and HyNa3−yPW clusters, as summarized in Table 3.
Next, we correlate these rate and selectivity data to the findings
from butanal and pyridine titration and infrared spectroscopic
studies in order to construct a mechanistic picture for these
reactions.
As butanals were exposed to HxNa4−xSiW (x=0-4) and
HyNa3−yPW (y=0-4) clusters, a portion of the H+ sites were
occupied by surface acetates and aromatic species, as confirmed
from the evolvement of [νas(OCO)] bands at 1580 or 1577 cm−1
and C=C bond stretching of the aromatic rings [ν(C=C)aromatic] at
1510 cm−1 during butanal adsorption and reactions at 348 K
(Figure 7). The acetates remained stable at low temperatures (348
K) but desorbed from the H+ sites to evolve into heavy aromatics
Scheme 3. Reaction pathways for butanal deoxygenation on HxNa4−xSiW
(x=0-4) and HyNa3−yPW (y=0-3) clusters.
The rate of inter-molecular C=C bond formation (Step a1,
Scheme 3) increases proportionally with alkanal pressures, as
shown previously for butanal, pentanal, and hexanal reactions on
H4SiW12O40 polyoxometalate clusters[21] and H-FAU[20] zeolites.
The inter-molecular C=C bond formation step (Step a1, Scheme
3) couples kinetically with intra-molecular C=C bond formation
(Step c1, Scheme 1) and isomerization-dehydration (Step b1,
Scheme 3) steps. The intra-molecular C=C bond formation occurs
together with the inter-molecular C=C bond formation, because it
requires the aromatics, the secondary products, formed from the
later reactions acting as the hydrogen sources in a hydride
transfer step.[20-21] The isomerization-dehydration step couples
with the inter-molecular C=C bond formation step, because they
share the co-adsorbed alkanal-alkenol pairs as the common
surface intermediate.[21] For these reasons, the rates of these
reactions also increase with butanal pressure.[21]
The rates (per H+ site) for the first inter-molecular C=C bond
formation ( 푟Inter,C ,per H+ , Equation 1), which also equal the
4
site-time-yields of 2-ethyl-2-hexenal (C8H14O), are:
푟Inter,C ,per H+ = 푘Inter,C ,per H+푃C H O
(1)
4
8
4
4
where 푘Inter,C ,per H
+
is the first-order rate constant for
4
inter-molecular C=C bond formation of butanal on each H+ site
with a unit of mol(molH+ s kPa)−1. This rate is related to the rate
for 2-ethyl-2-hexenal (C8H14O) formation for each HxNa4−xSiW
cluster (푟Inter,C ,, Equation 2a) via:
4
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601042
ChemCatChem
FULL PAPER
푟
= 푘Inter,C 푃C H O = 휆des,remain푘Inter,C ,per H+푃C H O
(2a)
(2b)
80.3-94.5 % and 4.8-15.7 %, respectively, on HxNa4-xSiW and
HyNa3-yPW, 573 K, space velocity=0.02 molbutanal (molcluster s)-1, as
shown in Table 3]. We note that carbon selectivities to larger
aromatic products are much lower on H0Na4SiW (6.4 %) and
H0Na3PW (4.8 %) than on other HxNa4-xSiW (x=1-4) (13-15.7 %)
and HyNa3-yPW (y=1-3) (8.4-14.9 %) clusters, likely because
H0Na4SiW and H0Na3PW clusters contain only weak H+ sites (as
shown in Fig. 2). These selectivity trends indicate that stronger H+
sites are more reactive than the weaker H+ sites towards
cyclization reactions. The lack of cyclization-dehydration reactions
are in stark contrast to those on microporous crystalline solids
[e.g., carbon selectivities to alkenals and aromatics are 14-16 %
and 20-29 %, respectively, on H-FAU zeolites,[20] 573 K, space
Inter,C
4
4
4
8
4 8
4
푘Inter,C = 휆des,remain푘Inter,C ,per H
+
4
4
Here, 푘Inter,C is the rate constant for inter-molecular C=C bond
4
formation
of
butanal
(per
HxNa4−xSiW
cluster,
mol (molcluster s kPa)−1) and 휆des,remain is the number of H+ site
remained after exposure to reaction mixtures (Equation 10), which
was determined by pyridine titration followed by temperature
programmed desorption of the adsorbed pyridine after the
reactions. As shown in Figure 8, these rate constant values
(푘Inter,C , per HxNa4−xSiW cluster) are directly proportional to the
4
H+ site densities remained after steady-state reactions (휆des,remain),
irrespective of the initial H+ site densities.
velocity=0.03
molbutanal
(molH
+
s)-1],
which
promote
The rates (per H+ site) for the 2,4-diethyl-2,4-octadienal (C12)
formation from the second inter-molecular C=C bond formation
alkylation-dealkylation and lead in turn to aromatic and
hydrocarbon species with varying carbon numbers, influenced
strongly by their pore and cage sizes.[2, 20]
(푟
8 ), a reaction between 2-ethyl-2-hexenal (C8) and butanal
Inter,C
(C4), are related to the first-order rate constant (푘Inter,C8) and the
partial pressure of 2-ethyl-2-hexenal (C8) through:
푟Inter,C ,per H+ = 푘Inter,C ,per H+푃C H
(3)
O
14
8
8
8
푟
= 푘Inter,C 푃C H
= 휆des,remain푘Inter,C ,per H+푃C H
O
14
(4a)
(4b)
Inter,C
O
8
8
8
14
8
8
푘Inter,C = 휆des,remain푘Inter,C ,per H
+
8
8
The values of 푘Inter,C (per HxNa4−xSiW cluster) also increase
8
with the 휆des,remain, as shown in Figure 8. The rate constants for
the second inter-molecular C=C bond formation are higher than
that for the first ones (푘Inter,C > 푘Inter,C4 ), because these rate
8
constants 푘Inter contain the equilibrium constant term for
keto-enol tautomerization ( 퐾taut ) and the rate constant for
alkanal-alkenol pair formation (푘AAP), as shown elsewhere:[21]
Figure 8. First-order rate constants at 573 K for the primary inter-molecular C=C
bond formation leading to C8H14O (푘
, Equation 2b) and the secondary
Inter,C4
푘Inter = 퐾taut푘AAP
(5)
inter-molecular C=C bond formation leading to C12H20O (푘
, Equation 4b)
Inter,C8
on HxNa4−xSiW (x=0-4) clusters, plotted as a function of the remaining H+ site
densities after butanal reactions (휆des,remain, Equation 10) [1.1-4.4 kPa butanal,
space velocity 0.02-0.08 molbutanal (molcluster s)-1].
The larger the aldehyde molecules, the higher the equilibrium
constants for keto-enol tautomerization (e.g., 퐾taut equals to
1.28×10−4 for isobutyraldehyde versus 5.89×10−7 for acetaldehyde
at 298 K); in addition, larger aldehydes interact more strongly with
the polyoxometalate clusters through stronger van der Waals
interactions. Both of these factors favor the enol formation and the
co-adsorption of the enol tautomer onto the lattice oxygen site
adjacent to the protonated alkenal.[37] The scatter in the rate
constant dependence for the secondary condensation reactions,
푘Inter,C8, in Figure 8 is caused by the much lower concentrations
Conclusion
Chemical titrations, infrared spectroscopy, and rate
measurements are used to probe the identities of reactive species
and their involvements in butanal deoxygenation reactions on a
series of polyoxometalate clusters [HxNa4−xSiW12O40 (x=0-4) and
HyNa3−yPW12O40 (y=0-3)] with varying H+ site densities. Butanals
adsorb on these H+ and Na+ sites of the polyoxometalate clusters
and form H+/Na+-bonded complexes (RC=O∙∙∙H+/Na+). A portion of
butanals at the H+ sites reacts with vicinal framework oxygen atom
and forms surface acetate, which remains as spectator species.
The other portion of adsorbed butanal undergoes inter-molecular
C=C bond formation via aldol condensation to produce
2-ethyl-2-hexenal, which then converts into larger, unsaturated
aldehydes and aromatics. Both unsaturated aldehydes and
aromatics are the major products during butanal deoxygenation
on polyoxometalate clusters with carbon selectivities >80 mol.%
towards the unsaturated aldehydes and 6-16 mol.% towards the
aromatics. A linear relation between the rate constants for
inter-molecular C=C bond formation step (per polyoxometalate
cluster) and the densities of H+ sites remained after steady-state
butanal reactions indicates that the rates are directly related to the
H+ site densities in this Brønsted acid catalyzed reaction.
(~1-2 orders of magnitude lower) for 2-ethyl-2-hexenal (C8) than
butanal and therefore larger experimental errors. The linear, direct
relations between these rate constants (푘Inter,C and 푘Inter,C8) and
4
the remaining H+ site densities (휆des,remain, Equation 10) in Figure
8 confirm the direct involvement of H+ sites in the inter-molecular
C=C bond formation step. These results also show that Na+ ions
introduced onto the polyoxometalate clusters, which replace the
Brønsted acid sites, remain kinetically silent for inter-molecular
C=C bond formation.
The concomitant minor, secondary cyclization-dehydration
reactions (e.g., Steps b2 and b3, Scheme 3) form hydrocarbon
species containing either 8, 12, or 16 carbon atoms (e.g., C8H12
cycloalkadienes, C12 aromatics, and C16 aromatics). The carbon
numbers in these products are identical to those formed from
butanal step-wise condensation steps (Steps a1-a3, Scheme 3).
The cyclization-dehydration occurs at much lower selectivities
than the inter-molecular C=C bond formation on polyoxometalate
clusters [e.g., carbon selectivities to alkenals and aromatics are
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601042
ChemCatChem
FULL PAPER
The pyridine-TPD profiles of H4Na0SiW and H3Na0PW show three
peaks centered around 550 K, 650 K, and 730 K, respectively. The
amounts of pyridine desorbed from these temperature ranges reflect the
adsorption strengths of pyridinium ions at the H+ sites. Peak deconvolution
of the TPD profiles by assuming Gaussian distributions leads to the site
densities 휆des,1, 휆des,2, and 휆des,3, categorized according to the desorption
temperature ranges of 500-650 K, 550-750 K, and 673-790 K, respectively.
These values are determined based on the cumulative amount of pyridine
desorbed in each range, obtained from fitting of the overall pyridine
desorption profile to peaks with Gaussian distributions. These values are
then divided by the number of polyoxometalate clusters in the sample:
Experimental Section
Preparation of supported HxNa4-xSiW and HyNa3-yPW clusters on SiO2
supports
H4Na0SiW or H3Na0PW polyoxometalate (POM) clusters were
supported on SiO2 supports at a nominal loading of 0.075 mmolPOM (gSiO2
−1
)
with incipient wetness impregnation method. SiO2 support (GRACE
chromatographic grade, 330 m2 g−1, particle size <75 μm, pore volume 1.2
cm3 g−1) was treated in ambient air at 0.17 K s−1 to 773 K and held at 773 K
for 5 h prior to impregnation at room temperature with a precursor solution,
prepared from dissolving either H4SiW12O40 or H3PW12O40 (Sigma Aldrich,
99.99%, CAS #12027-43-9 or CAS #12501-23-4, respectively) in deionized
water. HxNa4−xSiW (x=0-4) and HyNa3−yPW (y=0-3) polyoxometalate
cumulative mole of pyridine desorbed in temperature range 푖
휆des,푖
=
(8)
mole of polyoxometalate clusters
clusters were supported on SiO2 [0.075 mmolPOM (gSiO )
−1] using the same
where 푖 equals 1, 2, or 3, which corresponds to the temperature ranges of
500-650 K, 550-750 K, or 673-790 K, respectively. The total amount of
pyridine desorbed from fresh samples (휆des,overall, 500-790 K) and the
2
method, but the precursor solution used for impregnation was prepared
from dissolving H4SiW12O40 or H3PW12O40 into an aqueous NaOH solution
(0.3 mol L−1, CALEDON, CAS #1310-73-2) with its H+-to-Na+ molar ratio
adjusted to attain the targeted value [i.e., H:Na ratio of x:(4−x) for
HxNa4−xSiW and y:(3−y) for HyNa3−yPW]. The impregnated samples were
then held in closed vials for 24 h, heated to 323 K at 0.0167 K s−1, and then
held isothermally at 323 K for 24 h in flowing dry air (0.1 cm3 g−1 s−1, zero
grade, Linde). All the samples were denoted as HxNa4−xSiW (x=0-4) and
HyNa3−yPW (y=0-3) and the nominal H+ site density (χi, subscript i=Si or P
for HxNa4−xSiW or HyNa3−yPW, respectively) reported in this manuscript
was defined as:
amount of pyridine desorbed from the high temperature ranges (휆des,2−3
,
550-790 K) are expressed by:
휆des,overall = 휆des,1 + 휆des,2 + 휆des,3
휆des,2−3 = 휆des,2 + 휆des,3
(9a)
(9b)
The spent samples after deoxygenation reaction at 573 K were
subjected to pyridine titration followed by sequential temperature
programmed desorption. The amount of pyridine desorbed from the spent
sample during the temperature programmed desorption reflects the
+
nominal mole of H
remaining H+ site densities, 휆des,remain
:
휒푖 =
(6)
mole of cluster
cumulative mole of pyridine desorbed
mole of polyoxometalate clusters
휆des,remain
=
(10)
Chemical titration with pyridine and butanal and temperature
programmed desorption of pyridine
The experimental protocol for butanal titration is similar to that of
pyridine, except that butanal (Sigma Aldrich, ≥99 %, CAS #123-72-8)
instead of pyridine was introduced into the vaporization zone. The
vaporization zone and reactor were both maintained at 348 K. The
experiment was completed when the difference in butanal molar flow rates
between the reactor effluent and the feed stream was less than 5 %. The
amount of butanal adsorbed on the samples, 훼but, was determined by
dividing the total amount of butanal adsorbed by the number of
polyoxometalate clusters contained within the sample:
Chemical titration with pyridine and its sequential temperature
programmed desorption (TPD) were carried out on the HxNa4−xSiW (x=0-4)
and HyNa3−yPW (y=0-3) samples. For the pyridine titration experiment, a
sample (100 mg) was supported on
a quartz frit in a cylindrical
microcatalytic quartz reactor (8 mm inner diameter). The sample was
treated in-situ under flowing He (0.83 cm3 s−1) at a constant heating rate
(0.083 K s−1) to 473 K and maintained isothermally at 473 K. Liquid pyridine
(Sigma Aldrich, >99.9 %, CAS #110-86-1) was introduced at 3.42×10−8
mol·s−1 via a gas tight syringe (Model 006230, 0.25 cm3, SGE) mounted on
a syringe infusion pump (Model: LEGATO 100, KD Scientific) into a
vaporization zone that was maintained at 391 K and located upstream from
the reactor. In the vaporization zone, pyridine was evaporated and mixed
with a flowing He stream (0.83 cm3 s−1). The pyridine-He mixture was then
introduced to the sample and the amount of pyridine in the effluent stream
cumulative number of mole of adsorbed butanal
훼but
=
(11)
mole of polyoxometalate clusters
The total amount of adsorbed butanal was calculated by subtracting the
cumulative amount of butanal in the effluent stream from that of the inlet
feed stream, quantified using a flame ionization detector (FID).
was quantified with
a
flame ionization detector (FID) without
chromatographic separation. Pyridine titration was completed when the
molar flow rate of pyridine in the effluent stream became identical to that of
the feed stream (3.42×10−8 mol·s−1), at which point the difference in
pyridine molar flow rate between the feed and the effluent stream was less
than 5 %. After the pyridine titration, the sample was purged in flowing He
(0.83 cm3 s−1) at 473 K for 30 min and then the temperature programmed
desorption of pyridine (pyridine-TPD) was conducted. During the
temperature programmed desorption experiments, the temperature was
increased to 923 K at 0.033 K s−1 in flowing He (0.17 cm3 s−1). During the
entire period of the pyridine-TPD experiment, the amount of pyridine in the
effluent stream was quantified with a flame ionization detector (FID) without
chromatographic separation.
Infra-red spectroscopic studies of pyridine and butanal adsorption
Infrared (IR) spectroscopic studies were carried out with a customized,
stainless steel in-situ transmission IR cell equipped with CaF2 windows and
capable for operating between 298 K and 773 K. The cell was mounted in a
Bruker Vertex 70 spectrometer equipped with a mercury cadmium telluride
(MCT) detector. The cell was connected to a vacuum line and a rotary vane
pump (model 101150/11-11, Leroy Somer), which allows evacuation to
less than 0.01 mbar. Powder samples were pressed into a self-supporting
wafer of ~10 mm in diameter and less than 0.5 mm thick, mounted into a
sample holder inside the cell. Spectra were acquired in the absorbance
mode at a resolution of 2 cm−1 and 64 scans per spectrum.
Pyridine uptakes during the pyridine chemical titration (휆ads ) are
defined by the cumulative amount of pyridine adsorbed onto the catalysts
divided by the number of polyoxometalate clusters in the sample,
according to:
Prior to the adsorption of either pyridine or butanal, the cell was
purged with flowing helium (1.67 cm3 s−1) for 10 min, and then heated at
0.167 K s−1 to 573 K for another 30 min under vacuum (0.005-0.05 mbar).
Afterwards, a background spectrum was obtained at 348 K under vacuum.
Pyridine adsorption was carried out at 473 K, by introducing sequential
pyridine pulses (ca. 0.22 μmol each time, ~4 to 6 pulses) until saturation.
The sample chamber was evacuated under dynamic vacuum, which
cumulative number of mole of pyridine adsorbed
휆ads
=
(7)
mole of polyoxometalate clusters
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601042
ChemCatChem
FULL PAPER
[10]
a) R. J. Gorte, D. White, A. I. Biaglow, W. E. Farneth, J. Šepa, J. Catal.
1997, 167, 300-302; b) L. Kubelková, J. Cejka, J. Novakova, Zeolites
1991, 11, 48-53; c) S. Beran, L. Kubelkova, J. Mol. Catal. 1987, 39, 13-19;
d) D. Farcaşiu, J. Catal. 1996, 160, 309-313.
D. H. Barich, J. B. Nicholas, T. Xu, J. F. Haw, J. Am. Chem. Soc. 1998,
120, 12342-12350.
C. M. Nguyen, M.-F. Reyniers, G. B. Marin, J. Phys. Chem. C 2011, 115,
8658-8669.
C. Liu, T. J. Evans, L. Cheng, M. R. Nimlos, C. Mukarakate, D. J.
Robichaud, R. S. Assary, L. A. Curtiss, J. Phys. Chem. C 2015, 119,
24025-24035.
removed the gas phase and physisorbed pyridine, and cooled to 348 K for
infrared spectroscopic measurements. Similar procedures were carried out
for butanal adsorption at 348 K. Butanal pulses (ca. 0.32 μmol each time,
~5 to 8 pulses) were introduced sequentially into the sample compartment
until the catalyst surfaces became saturated, attained when the spectra did
not change with additional pulses. The chamber was then evacuated under
dynamic vacuum (0.01-0.1 mbar), heated step-wise to a temperature
between 348 K and 573 K (348, 373, 423, 473, 523, and 573 K) and then
cooled to 348 K for infrared spectroscopic measurements. Quantitative
comparison of the peak intensities between the catalyst samples was
carried out by spectrum normalization using the vibrational bands
corresponded to the lattice overtone of silica at 1990 cm−1 and 1875 cm−1
following the method demonstrated previously.[22b]
[11]
[12]
[13]
[14]
J. R. Sohn, J. S. Han, J. S. Lim, Applied Surface Science 2005, 252,
858-865.
[15] X. Solans-Monfort, J. Bertran, V. Branchadell, M. Sodupe, J. Phys. Chem.
B 2002, 106, 10220-10226.
[16]
[17]
[18]
[19]
M. Xu, W. Wang, M. Hunger, Chem. Commun. 2003, 722-723.
L. Cheng, X. P. Ye, Catal. Lett. 2009, 130, 100-107.
H. Atia, U. Armbruster, A. Martin, J. Catal. 2008, 258, 71-82.
a) J. Macht, M. J. Janik, M. Neurock, E. Iglesia, Angew. Chem. Int. Ed.
2007, 46, 7864-7868; b) J. Macht, M. J. Janik, M. Neurock, E. Iglesia, J.
Am. Chem. Soc. 2008, 130, 10369-10379.
Assessment of rates and selectivities for butanal deoxygenation on
supported HxNa4-xSiW and HyNa3-yPW clusters
[20]
[21]
[22]
F. Lin, Y.-H. Chin, J. Catal. 2016, 341, 136-148.
F. Lin, Y.-H. Chin, ACS Catal. 2016, 6, 6634-6650.
a) G. Busca, Phys.Chem.Chem.Phys. 1999, 1, 723-736; b) S. Zheng, H.
R. Heydenrych, A. Jentys, J. A. Lercher, J. Phys. Chem. B 2002, 106,
9552-9558; c) E. P. Parry, J. Catal. 1963, 2, 371-379.
Butanal deoxygenation was carried out in a steady-state fixed bed
microcatalytic quartz reactor (8 mm inner diameter) with plug-flow fluid
hydrodynamics at 573 K. The reactor was contained within a resistively
heated furnace with its temperature controlled by a digital feedback
temperature controller. Catalyst samples (25 mg) were supported on a
[23]
[24]
B. K. Hodnett, J. B. Moffat, J. Catal. 1984, 88, 253-263.
J. J. Borrás-Almenar, E. Coronado, A. Müller, M. Pope, Polyoxometalate
Molecular Science, Kluwer Academic Publishers, Boston, 2003.
M. Brändle, J. Sauer, J. Am. Chem. Soc. 1998, 120, 1556-1570
NIST Chemistry WebBook: NIST Standard Reference Database Number
69, http://webbook.nist.gov/chemistry/. (accessed on Oct 11, 2016).
G. O. Pates, L. Guler, J. J. Nash, H. I. Kenttämaa, J. Am. Chem. Soc.
2011, 133, 9331-9342.
a) M. S. Zanuttini, B. O. Dalla Costa, C. A. Querini, M. A. Peralta, Appl.
Catal. A 2014, 482, 352-361; b) G. Rumplmayr, J. A. Lercher, Zeolites
1990, 10, 283-287.
Z. Wu, A. K. P. Mann, M. Li, S. H. Overbury, J. Phys. Chem. C 2015, 119,
7340-7350.
quartz frit and the bed temperature was recorded using
a K-type
[25]
[26]
thermocouple placed in the center (in both the axial and radial directions) of
catalyst bed. Catalysts were treated in-situ under flowing He (Grade 5.0,
−1
Linde, 16.7 cm3 gcat. s−1) by heating at 0.167 K s−1 to 573 K. Liquid
[27]
[28]
butanal was introduced into a vaporization zone using a gas tight syringe
(Hamilton Gastight 1105, 5 mL) mounted on a syringe infusion pump. In the
vaporization zone, butanal was evaporated and mixed with a He (Grade
−1
5.0, Linde, 16.7 cm3 gcat. s−1) purge stream. The partial pressure of
[29]
butanal feed was maintained at a constant value between 1.1 kPa to 4.4
kPa by controlling the infusion rate of liquid butanal. The mixture was fed to
the reactor via heated transfer lines held at 473 K. Reactants and products
were quantified using an on-line gas chromatograph (Agilent 7890A) with
mass spectrometer (Agilent 5975C) equipped with two capillary columns of
(1) Agilent HP-5MS (190091S-433, 30 m, 0.25 mm ID, 0.25 μm film),
connected to a thermal conductivity detector (TCD) and flame ionization
detector (FID) in series, and (2) HP-5 (19091J-413, 30 m, 0.32 mm ID, 0.25
μm film) connected to a mass selective detector (MSD).
[30]
[31]
[32]
J. Raskó, J. Kiss, Appl. Catal. A 2005, 287, 252-260.
J. E. Rekoske, M. A. Barteau, Langmuir 1999, 15, 2061-2070.
J. Szanyi, J. H. Kwak, R. A. Moline, C. H. F. Peden, J. Phys. Chem. B
2004, 108, 17050-17058.
A. Yee, S. J. Morrison, H. Idriss, J. Catal. 1999, 186, 279-295.
A. Bielanski, J. Datka, B. Gil, A. Malecka-Lubanska, A. Micek-Ilnicka,
Phys.Chem.Chem.Phys. 1999, 1, 2355-2360.
A. Bielański, J. Datka, B. Gil, A. Małecka-Lubańska, A. Micek-Ilnicka,
Catal. Lett. 1999, 57, 61-64.
a) A. J. Jones, S. I. Zones, E. Iglesia, J. Phys. Chem. C 2014, 118,
17787-17800; b) R. Gounder, E. Iglesia, Chem. Commun. 2013, 49,
3491-3509.
[33]
[34]
[35]
[36]
[37]
a) Y. Apeloig, D. Arad, Z. Rappoport, J. Am. Chem. Soc. 1990, 112,
9131-9140; b) Y. Chiang, A. J. Kresge, P. A. Walsh, J. Am. Chem. Soc.
1982, 104, 6122-6123; c) J. R. Keeffe, A. J. Kresge, N. P. Schepp, J. Am.
Chem. Soc. 1990, 112, 4862-4868; d) B. L. Newalkar, N. V. Choudary, P.
Kumar, S. Komarneni, T. S. G. Bhat, Chem. Mater. 2002, 14, 304-309.
Acknowledgements
The work was supported by the Natural Sciences and Engineering
Research Council of Canada (NSERC), Canada Foundation for
Innovation (CFI), Valmet, Abellon CleanEnergy; F. Lin
acknowledges Hatch Graduate Scholarship for Sustainable
Energy Research and Ontario Graduate Scholarship for supports.
Received: ((will be filled in by the editorial staff))
Published online: ((will be filled in by the editorial staff))
Keywords: Butanal deoxygenation, Polyoxometalate cluster,
in-situ FT-IR, Aldol condensation
[1]
a) D. Mohan, C. U. Pittman, P. H. Steele, Energy Fuels 2006, 20, 848-889;
b) T. Q. Hoang, X. Zhu, T. Sooknoi, D. E. Resasco, R. G. Mallinson, J.
Catal. 2010, 271, 201-208; c) S. A. W. Hollak, K. P. de Jong, D. S. van Es,
ChemCatChem 2014, 6, 2648-2655; d) G. Li, F. Zhang, L. Chen, C.
Zhang, H. Huang, X. Li, ChemCatChem 2015, 7, 2646-2653; e) J. Dai, X.
Fu, L. Zhu, J. Tang, X. Guo, C. Hu, ChemCatChem 2016, 8, 1379-1385.
F. Lin, Y.-H. Chin, J. Catal. 2014, 311, 244-256.
[2]
[3]
A. G. Gayubo, A. T. Aguayo, A. Atutxa, R. Aguado, M. Olazar, J. Bilbao,
Ind. Eng. Chem. Res. 2004, 43, 2619-2626.
[4]
[5]
C. B. Warren, A. B. Marin, J. F. Butler, Vol. US5665781 A, International
Flavors & Fragrances Inc., The University Of Florida, 1997.
a) W. Pinkenhagen, Vol. US4381410 A, Firmenich Sa, 1983; b) W.
Pickenhagen, A. Velluz, Vol. US4324809 A, Firmenich Sa, 1982.
R. Veltri, G. B. Fodor, Vol. US5098933 A, Theracel Corporation, 1992.
a) E. Dumitriu, N. Bilba, M. Lupascu, A. Azzouz, V. Hulea, G. Cirje, D.
Nibou, J. Catal. 1994, 147, 133-139; b) E. Dumitriu, V. Hulea, N. Bilba, G.
Carja, A. Azzouz, J. Mol. Catal. 1993, 79, 175-185.
[6]
[7]
[8]
[9]
a) H. Tsuji, F. Yagi, H. Hattori, H. Kita, J. Catal. 1994, 148, 759-770; b) H.
E. Swift, J. E. Bozik, F. E. Massoth, J. Catal. 1969, 15, 407-416.
C. D. C. Diaz, S. Locatelli, E. E. Gonzo, Zeolites 1992, 12, 851-857.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601042
ChemCatChem
FULL PAPER
Graphical materials
FULL PAPER
As butanals are exposed to
polyoxometalate clusters, they
undergo inter-molecular C=C bond
formation via an aldol condensation
type mechanism and form larger
alkenals as the predominant
reaction route. A portion of the
butanals adsorbs as surface
acetate, which remains inactive
during the catalytic sojourns.
Y.Yang, F. Lin, H. Tran, Y. Chin*
Page 1 – Page 11
Butanal Condensation Chemistry
Catalyzed by Brønsted Acid Sites
on Polyoxometalate Clusters
This article is protected by copyright. All rights reserved.