Liquid phase aldol condensation reactions with MgO-ZrO2 and shape-selective nitrogen-substituted NaY

Article abstract of DOI:10.1016/j.apcata.2010.10.023

The aldol condensation reactions of furfural/hydroxymethylfurfural (furfurals) with acetone/propanal in water-methanol solvents were studied over the solid base catalysts MgO-ZrO₂, NaY and nitrogen-substituted NaY (Nit-NaY). The reactions were conducted at 120°C and 750 psig of He in batch reactors. Nit-NaY exhibited catalytic activity for aldol condensation comparable to MgO-ZrO₂ and much higher than that of NaY, indicative of the increased base strength after replacing the bridging oxygen with the lower electronegativity nitrogen over Nit-NaY. The aldol condensation of furfurals with acetone produces two different products, the monomer and the dimer. The monomer is formed from reaction of furfurals with acetone. The dimer is formed from reaction of the monomer with furfurals. MgO-ZrO₂ had a higher selectivity towards dimer formation. In contrast, Nit-NaY was more selective towards the monomer product due to the cage size in the FAU structure, indicating that Nit-NaY is a shape selective base catalyst. Increasing the water concentration in the feed solution or increasing the feed concentration led to both increased catalytic activity and dimer selectivity. The Nit-NaY catalyst was not stable and lost catalytic activity when recycled due to leaching of the framework nitrogen. Different characterization techniques, including XRD, high resolution Ar adsorption isotherm, basic sites titration, CO₂ TPD-MS, TGA and ²⁹Si SP MAS NMR, were used here to characterize the fresh and spent catalysts. The results show that Nit-NaY maintains only part of the FAU-type crystal structure. Furthermore, the base strength over Nit-NaY was found to be between that of Mg²⁺-O ^2- pair and Mg(OH)₂. The reaction mechanism over Nit-NaY was discussed.

Full text of DOI:10.1016/j.apcata.2010.10.023

Applied Catalysis A: General 392 (2011) 57–68
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Liquid phase aldol condensation reactions with MgO–ZrO2 and shape-selective
nitrogen-substituted NaY
Wenqin Shena, Geoffrey A. Tompsetta, Karl D. Hammonda, Rong Xinga, Fulya Doganb, Clare P. Greyb,
W. Curtis Conner Jr.a, Scott M. Auerbacha,c, George W. Hubera,∗
a Department of Chemical Engineering, University of Massachusetts, 686 N. Pleasant St, Amherst, MA 01003-3110, United States
b Department of Chemistry, State University of New York at Stony Brook, Stony Brook, NY 11794-3400, United States
c Department of Chemistry, University of Massachusetts, Amherst, MA 01003-9336, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 20 July 2010
Received in revised form 13 October 2010
Accepted 20 October 2010
Available online 28 October 2010
The aldol condensation reactions of furfural/hydroxymethylfurfural (furfurals) with acetone/propanal
in water–methanol solvents were studied over the solid base catalysts MgO–ZrO2, NaY and nitrogen-
substituted NaY (Nit-NaY). The reactions were conducted at 120 ◦C and 750 psig of He in batch reactors.
Nit-NaY exhibited catalytic activity for aldol condensation comparable to MgO–ZrO2 and much higher
than that of NaY, indicative of the increased base strength after replacing the bridging oxygen with the
lower electronegativity nitrogen over Nit-NaY. The aldol condensation of furfurals with acetone produces
two different products, the monomer and the dimer. The monomer is formed from reaction of furfurals
with acetone. The dimer is formed from reaction of the monomer with furfurals. MgO–ZrO2 had a higher
selectivity towards dimer formation. In contrast, Nit-NaY was more selective towards the monomer
product due to the cage size in the FAU structure, indicating that Nit-NaY is a shape selective base catalyst.
Increasing the water concentration in the feed solution or increasing the feed concentration led to both
increased catalytic activity and dimer selectivity. The Nit-NaY catalyst was not stable and lost catalytic
activity when recycled due to leaching of the framework nitrogen. Different characterization techniques,
including XRD, high resolution Ar adsorption isotherm, basic sites titration, CO2 TPD-MS, TGA and 29Si SP
MAS NMR, were used here to characterize the fresh and spent catalysts. The results show that Nit-NaY
maintains only part of the FAU-type crystal structure. Furthermore, the base strength over Nit-NaY was
found to be between that of Mg2+–O2− pair and Mg(OH)2. The reaction mechanism over Nit-NaY was
discussed.
Keywords:
Nitrogen-substituted zeolite
MgO–ZrO2
Aldol condensation
High-throughput reactor
Shape-selective base catalyst
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
occlusion with basic salts or metals [8–11], organic base grafting
[12–15] and nitridation [16–23]. These methods all have their lim-
Solid base catalysts have attracted much attention because of
their potential to replace corrosive homogeneous base catalysts
in industrially important reactions such as isomerizations, alky-
lations, condensations, additions, cyclizations, transesterifications
and oxidations [1–5]. The use of solid catalysts provides a more
environmentally benign process and reduces the cost of process-
ing due to the ease of separation and recycling. Base strength,
stability, and shape or size selectivity are the most desired prop-
erties for solid base catalysts. Shape selective acidic zeolites are
the backbone of the petroleum industry [6]. However, shape selec-
tive basic catalysts have not been used commercially yet. Basic
nanoporous materials can be prepared by various methods includ-
ing ion-exchange with alkali metal cations [5,7], impregnation or
itations for preparing shape-selective basic zeolites. For example,
ion-exchange methods produce only weak bases [24]. Impregna-
tion or occlusion may cause the blockage of zeolite pores [25].
Organic base grafting may only modify the basicity of the sur-
face or cause the blockage of the micropores of zeolites due to the
large organic groups, and thereby does not produce the desired
shape selectivity [26–28]. Nitridation of zeolites has been discussed
as a promising option for the generation of truly shape selec-
tive base catalysts [29,30]. Nitrided zeolites have been prepared
by treating zeolites with ammonia at elevated temperatures. This
results in an increased base strength of the framework due to the
substitution of bridging oxygen with the lower electronegativity
nitrogen [29,30]. Nitridation has been performed on a range of
porous materials including mesoporous MCM-41 [31–33] and SBA-
15 [34,35], aluminophosphates (AlPOs) [17,36], silicon substituted
aluminophosphates (SAPOs) [37,38] and zeolites [17,18,39,40]. The
nitrogen-substituted materials have been evaluated by Knoeve-
∗
Corresponding author. Tel.: +1 413 545 0267; fax: +1 413 545 1647.
E-mail address: huber@ecs.umass.edu (G.W. Huber).
0926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.10.023

58
W. Shen et al. / Applied Catalysis A: General 392 (2011) 57–68
Scheme 1. Base catalyzed aldol condensation of furfurals with acetone or propanal.
nagel condensations of benzaldehyde with molecules containing
active methylenic groups with different pKa values and showed
basic activity and stability [17,18,32,36,38,40]. However, there are
no reports on the shape selectivity over these nitrogen-substituted
materials.
as MgO, CaO, Mg–Al-oxide, and MgO/ZrO2 prepared by impregna-
tion. However, there is no preference to the monomer or the dimer
product by using either homogeneous NaOH or the heterogeneous
MgO–ZrO2 catalyst. This paper will present the first report to apply
nitrogen-substituted NaY for the reactions from bio-renewable
sources to generate value-added chemicals and fuels. The objec-
tives of this research are (1) to demonstrate the increased catalytic
activity by nitrogen substitution over NaY using aldol condensation
of furfurals with acetone or propanal as probe reactions and (2)
to demonstrate that the nitrogen substituted zeolites can be shape
selective base catalysts. The Nit-NaY catalyst will be compared with
the known non-shape selective MgO–ZrO2 catalyst using different
reaction conditions.
A sustainable biomass economy requires the development of
different integrated steps for the conversion of biomass-derived
feedstocks into a range of fuels and chemicals [41]. Many of these
steps involve basic catalysts. One key step in a biorefinery con-
cept to produce diesel and jet fuel range alkanes involves C–C bond
forming reactions by aldol condensations. Biomass can be depoly-
merized into C5 and C6 sugars by hydrolysis-based approaches.
However, to make diesel and jet fuel range alkanes that are between
C8 to C15, C–C bond forming reactions must occur. One approach
to form these larger alkanes involves first a dehydration step to
produce furfurals. The furfurals are then reacted with an aldehyde
or ketone by aldol condensation to form C8 to C15 alkane precur-
sors. These larger compounds then undergo hydrodeoxygenation
reactions to produce the final alkanes [42–45]. Base catalysis is the
key step in C–C bond forming reactions to produce the chemicals
or fuels with targeted molecular weight.
The products from the condensation of furfural with ace-
tone or propanal are shown in Scheme 1. 5-Hydroxymethyl
furfural (HMF) will undergo the same chemistry as furfural. Ace-
tone reacts with one molecule of furfural (or HMF) to form
furfural (or HMF) acetone monomer 4-(2-furyl)-3-buten-2-one
(or 4-[5-(hydroxymethyl)-2-furanyl]-3-buten-2-one), followed by
condensation with another furfural (or HMF) to form furfural (or
HMF) acetone dimer 1,5-di-2-furanyl-1,4-pentadien-3-one (or 1,5-
bis[(5-hydroxlmethyl)-2-furanyl]-1,4-pentadien-3-one). Propanal
reacts with one furfural (or HMF) to form furfural (or HMF) propanal
monomer 3-(2-furyl)-2-methylacrolein (or 3-[(5-hydroxymethyl)-
2-furanyl]-2-methylacrolein) and the monomer can further react
with another propanal molecule to produce furfural (or HMF)
propanal dimer 5-(2-furyl)-2,4-dimethyl-2,4-pentadienal (or 5-
[(5-hydroxlmethyl)-2-furyl]-2,4-dimethyl-2,4-pentadienal).
Aldol condensation of furfurals with acetone has been stud-
ied in homogeneous ethanol/water solutions [46] and water/THF
biphasic system [42] catalyzed by NaOH. It has also been stud-
ied in aqueous solutions catalyzed by solid base catalysts followed
by hydrogenation [43] and the combined aldol condensation with
hydrogenation over bi-functional catalyst 5 wt.% Pd/MgO–ZrO2
[44]. MgO–ZrO2 showed the highest catalytic activity and excel-
lent recycling ability compared with other solid base catalysts such
2. Experimental
2.1. Catalyst preparation
MgO–ZrO2 was prepared by precipitation of Mg(NO3)2·6H2O
(Aldrich) and ZrO(NO3)2·6H2O (Aldrich) aqueous solution with
NaOH as described elsewhere [44,47,48]. Typically, 50.9 g of mag-
nesium nitrate and 4.04 g zirconyl nitrate were dissolved in 1 L of
deionized water. A 25 wt.% NaOH aqueous solution in a separation
funnel was dropped into the mixed solution with vigorous stir-
ring until the pH was equal to 10 as monitored by a pH meter. The
resulting gel was aged at room temperature for 72 h and separated
by vacuum filtration. The precipitate was thoroughly washed with
deionized water at least four times and dried in an oven at 120 ◦C
overnight. The MgO–ZrO2 catalyst was obtained by calcination of
the resulting dry precipitate in 100 mL/min air flow at 600 ◦C for
3 h with a 3 h heating ramp. NaY (CBV100 with a Si/Al ratio of 2.55,
Zeolyst International) was used as a starting material to prepare
nitrogen-substituted NaY. In a typical synthesis, about 5 g NaY was
loaded into a quartz tube with a 1 in. outer diameter supported by
a quartz frit. Before nitridation, NaY was first dried under a flow
of 100 mL/min nitrogen at 110 ◦C for 2 h with a 0.5 ◦C/min ramp
rate, and then dehydrated at 400 ◦C for 6 h with a 0.5 ◦C/min ramp
rate. It was found by us previously that high ammonia flow rates
are crucial for substantial nitrogen substitution while maintain-
ing acceptable microporosity and crystallinity [21]. The NaY was
nitrided in a flow of 1 L/min ammonia (Airgas, anhydrous, 99.99%)
at 850 ◦C with a 1.5 ◦C/min ramp for 24 h. A slow heating rate was
used here to remove water gradually and avoid dealumination of
the zeolite [22]. The resulting material was cooled to room tem-

W. Shen et al. / Applied Catalysis A: General 392 (2011) 57–68
59
perature under nitrogen flow and stored in a desiccator, labeled as
Nit-NaY.
ated only based on the disappearance of furfural (or HMF) and the
selectivity to aldol products. The definitions are as follows:
moles of furfural consumed
moles of furfural input
2.2. Catalytic reaction and analysis methods
furfural disappearance (%) =
monomer selectivity (%) =
× 100
The catalysts were evaluated in a high throughput parallel auto-
matic reactor system (CAT 24, HEL Group). The system temperature,
pressure, and gas flow rate were controlled and monitored through
WinISO E670 software. In a typical run, 2 mL reactant solution, a
suitable amount of catalyst, and a stir bar were put into a glass tube
reactor with a total capacity of 3 mL. All of the prepared glass tube
reactors (maximum 24) were loaded into a high pressure chamber.
Inside the chamber, stainless steel cooling tips were inserted into
the top of each glass tube, effectively preventing the mixing of vapor
between glass reactors. Before starting the reaction, the chamber
was purged with helium for 10 min and pressurized to 500 psig.
Next, the chamber was heated to 120 ◦C. More helium was intro-
duced to maintain the final pressure at 750 psig. The reaction was
kept at 120 ◦C for 16 h or 24 h and stirred at 850 rpm. After reac-
tion, the whole chamber was quenched in an ice bath immediately
to stop the reaction. The resulting mixture inside each glass tube
reactor was transferred to a 25 mL sample bottle. The glass reactor
was washed three times with methanol (HPLC grade) to dissolve
the products deposited on the glass wall. Samples for HPLC anal-
ysis were filtered by a syringe filter (Waters 0.45 ␮m PES syringe
membrane filter).
The stability test was conducted in a Parr batch reactor (model
#4566) with an external temperature and stirring controller
(Model #4835). A 75 mL reactant solution (5 wt.% total organ-
ics, furfural/acetone = 1 by moles, methanol/water = 2 by volume)
and 250 mg catalyst were added to the reactor. The reactor was
first pressurized to 200 psig with helium and then depressurized,
repeating three times to replace all air in the reactor while checking
the reactor for leaks. Next, the reactor was pressurized to 500 psig,
heated to 120 ◦C, and purged with additional helium to maintain the
reaction pressure at 750 psig. The reaction was kept at 120 ◦C for
48 h and was sampled every 6–8 h (about 1 mL per sample) from
the sampling port during the run. The samples were diluted ten
times by methanol and filtered by using a syringe filter before anal-
ysis. After the first run, the reactor was cooled in an ice bath. The
spent catalysts were collected by vacuum filtration and washed
thoroughly with methanol. After drying in air, the spent catalyst
was placed into the reactor for a second run under the exact same
operating conditions.
moles of monomer
moles of furfural consumed
× 100
2 × moles of dimer
dimer selectivity (%) =
× 100
moles of furfural consumed
selectivity of aldol product (%)
moles of monomer + 2 × moles of dimer
=
× 100
moles of furfural consumed
= monomer selectivity + dimer selectivity
moles of monomer
moles of furfural input
monomer yield (%) =
dimer yield (%) =
× 100
moles of dimer
moles of furfural input
× 100
For the product distribution:
moles of monomer
moles of monomer + moles of dimer
monomer (%) =
dimer (%) =
× 100
moles of dimer
moles of monomer + moles of dimer
× 100
2.3. Characterization
X-ray diffraction (XRD) patterns were measured with a Philips
X’Pert professional diffractometer using a Cu K˛ source with a
˚
wavelength of 1.54 A. An accelerating voltage of 45 kV and a current
of 40 A were used. A slit width of 0.5◦ was used on the source. Scans
were collected using an X’Celerator detector.
The high resolution Ar adsorption isotherms of NaY and Nit-NaY
were conducted at Quantachrome Laboratories using commercially
available Autosorb instruments. The raw data was reduced by using
the Autosorb software (version 1.53, Quantachrome instrument).
The BET surface areas and micropore volumes of NaY and Nit-NaY
were calculated by multipoint BET method and Vt-method in the
relative pressure range of 0.02–0.3, respectively. The BET surface
areas of the fresh and spent MgO–ZrO2 catalyst were calculated by
using multipoint BET method of the N2 adsorption isotherms in the
relative pressure range of 0.01–0.5. The pore size distributions of
NaY and Nit-NaY were obtained by DFT method using a built-in DFT
kernel (Ar at 87K-Zeolite/Silica: Cylinder, pore, NLDFT ads. branch
model).
An additional experiment was designed to test if any homoge-
neous reaction occurred due to leached chemicals from the catalyst.
The catalyst was first refluxed at the reaction temperature in water
for 2 h and was quickly filtered at that temperature. Next, the exact
amount of reactants and methanol to maintain the methanol/water
volumetric ratio of 2 as we did in the stability test were added. The
mixture underwent the reaction at 120 ◦C for 48 h and sampled
every 4–8 h.
The reactants and their condensation products were identi-
fied by gas chromatography/mass spectrometry (GC/MS; Shimadzu
GC/MS-2100 with a DB-5 column from Alltech) if practical or by
1H NMR and analyzed by HPLC (Shimadzu, with an SPD-20AV
UV detector and a reverse phase C-18 column from Agilent). The
monomer and dimer standards were prepared by NaOH catalyzed
aldol condensation at room temperature and purified by chro-
matography. Furfural and HMF were analyzed at UV wavelength
of 254 nm and the monomer/dimer products were analyzed at a
UV wavelength of 330 nm; acetone and propanal are transpar-
ent to ultraviolet light. The side products are difficult to identify
since both the reactant (furfural or HMF) and aldol products are
bio-degradable. Some of the degradation products are insoluble in
methanol. Therefore, the performance of the catalysts was evalu-
Carbon dioxide chemisorption and temperature programmed
desorption (TPD) were conducted by
a thermo-gravimetric
analyzer (TA Instruments, SDTQ600) combined with a mass spec-
trometer (eXtorr Inc.) operating at 10−5 Torr. In a typical run,
approximately 30 mg of the fresh or spent catalyst was placed into
an alumina sample container, weighed, and equilibrated at 50 ◦C.
The sample was then heated to 300 ◦C under 100 mL/min helium
for 1 h with a 10 ◦C/min ramp to degas and then cooled to 50 ◦C until
steady weight was established. The second gas, CO2, was intro-
duced into the chamber with a flow rate of 60 mL/min for about
3 h. After chemisorption equilibrium was established, the CO2 flow
was stopped and the physically adsorbed CO2 was removed under
helium flow until no further weight change was observed. Next,
the sample was heated from 50 ◦C to 800 ◦C with a heating rate of

60
W. Shen et al. / Applied Catalysis A: General 392 (2011) 57–68
20 ◦C/min and maintained at 800 ◦C for 40–80 min. The CO2 mass
spectrum for the effluent was simultaneously monitored. The back-
ground control experiments were conducted in helium (without
the CO2 adsorption step) for calculation of the weight loss due to
the decomposition of the catalyst itself. For the spent MgO–ZrO2
catalyst, an additional run was conducted in which the catalyst
was first calcined in a flow of air at 600 ◦C for 1 h, followed by CO2
chemisorption and TPD with the same procedure as above to eval-
uate the change of basic sites after two runs of the aldol reaction at
120 ◦C for a total 91 h. The CO2 uptake was calculated based on both
the weight changes during chemisorption and TPD. The weight gain
due to CO2 chemisorption was equal to the weight difference before
chemisorption and after the release of the physically adsorbed CO2.
The weight loss due to the CO2 TPD was equal to the difference of
total weight loss during TPD between samples with and without
the CO2 chemisorption step. The CO2 uptakes were also calculated
based on the CO2 desorption peak from the mass spectra. The peak
areas were calibrated by using 1 mL STP CO2 of a standard volume
loop.
Fig. 2. The high resolution Ar adsorption isotherms and the pore size distribution
of NaY and Nit-NaY. The results indicate that the Nit-NaY catalyst maintains pore
size distribution and part of the crystal structure with the loss of pore volume and
3
˚
surface area to a degree. The unit of dV(w) is cm /A/g.
Basic site titration was conducted under inert (helium) atmo-
sphere according to the method used by Vidruk et al. [49]. Typically,
catalyst powder (50 mg) was placed into a flask containing 50 mL
o-xylene with 0.1 wt.% phenolphthalein and stirred at room tem-
perature for 4 h in helium. The mixture was then titrated with
0.01 M benzoic acid in o-xylene until the color changed from color-
less to pink. The concentration of basic sites was calculated based
on the amount of acid consumed. The inert atmosphere was used
here in order to avoid the error introduced by CO2 in air.
[50]. Here, the cystallinity of Nit-NaY was about 52% based on a
comparison of the reflection peak areas of (5 3 3), (6 4 2), and (6 6 0)
between NaY and Nit-NaY. In addition, a significant reduction of
the intensity ratio of the reflections corresponding to (2 2 0) and
(3 1 1) was observed, which is caused by the substitution of nitrogen
into the zeolite framework [22]. There were no further structural
changes in the spent Nit-NaY after reaction at 120 ◦C for 75 h in
a methanol–water solvent. Fig. 2 shows the Ar isotherms of NaY
and Nit-NaY and the corresponding pore size distributions. NaY
The 29Si single-pulse (SP) magic angle spinning (MAS) nuclear
magnetic resonance (NMR) analysis of the fresh Nit-NaY and spent
Nit-NaY was performed on a Varian 360 MHz spectrometer with
4 mm probe using 14 kHz MAS. A 90◦ pulse length of 2.4 ␮s was
used with a recycle delay of 30 s, as established in our previous
experiments [22]. The chemical shifts reported are with respect to
neat liquid tetramethylsilane (TMS) at 0 ppm.
˚
and Nit-NaY present similar isotherms. The pore size of 7.2 A for
NaY was maintained over Nit-NaY. However, the Ar adsorption of
Nit-NaY is less than that of NaY, indicating the reduction of the
porosity. Micropore volumes (Vt-method) of NaY, Nit-NaY and the
spent Nit-NaY are 0.36 mL/g, 0.13 mL/g and 0.17 mL/g (N2 adsorp-
tion), respectively. The BET surface areas of NaY, Nit-NaY and the
spent Nit-NaY are 954 m2/g (CBET = −45), 337 m2/g (CBET = −4.3) and
368 m2/g (CBET = −41, N2 adsorption), respectively. Although the
BET equation is not valid for these microporous materials, this
shows the relative reduction of surface area of ∼2/3, similar to
the micropore volume. The results from the high resolution Ar
adsorption isotherm and XRD are consistent, showing that Nit-
NaY maintains only part of the crystal structure, contributing to the
reduction of the surface area, micropore volume, and crystallinity.
The XRD pattern of the calcined MgO–ZrO2 catalyst (as seen
in the supplementary materials Fig. S1) shows two phases, cor-
responding to cubic MgO and tetragonal ZrO2, respectively. After
reaction, the catalysts underwent a phase transition, where the
spent MgO–ZrO2 contained Mg(OH)2 and ZrO2. ZrO2 retained high
crystallinity without noticeable changes in the spent catalyst, while
the Mg(OH)2 showed broad XRD peaks, indicative of poor crys-
tallinity. The BET surface area of the fresh MgO–ZrO2 is 102 m2/g
and increased to 194 m2/g for the spent MgO–ZrO2, which might be
attributed to a phase transition (MgO to Mg(OH)2) or a roughening
from the fresh catalyst to the spent catalyst.
3. Results
3.1. Catalyst characterization
3.1.1. Powder XRD and high resolution adsorption isotherm
Fig. 1 shows the XRD patterns of NaY, Nit-NaY and the spent Nit-
NaY. Nit-NaY retained all the reflection peaks of NaY, indicating the
FAU structure was maintained after treatment in ammonia flow at
850 ◦C for 24 h. However, the intensities of the reflections over Nit-
NaY were significantly reduced compared to that of NaY, indicating
reduced crystallinity, consistent with our previous reports [21,22].
The percentage of crystalline material remaining in Nit-NaY can
be estimated according to the method introduced by Solache et al.
3.1.2. Characterization of basic sites by CO2-TPD and titration
Fig. 3 shows the CO2 TPD-MS profiles over the fresh and spent
MgO–ZrO2 catalysts. The spent MgO–ZrO2 was collected after two
runs of furfural condensation with acetone at 120 ◦C for 91 h and
washed with excess methanol. It was seen that the spent MgO–ZrO2
after regeneration has almost the same amount of CO2 uptake
at 190 ◦C compared with the fresh MgO–ZrO2 catalyst. The spent
MgO–ZrO2 showed only a very weak CO2 uptake peak at 120 ◦C,
but a sharp CO2 peak at 400 ◦C related to the decomposition of
adsorbed furfural and its degradation products, and a broad peak
at 800 ◦C probably due to coke. There was a very small CO2 peak at
Fig. 1. The powder XRD patterns of NaY, Nit-NaY and the spent Nit-NaY.

W. Shen et al. / Applied Catalysis A: General 392 (2011) 57–68
61
900
900
800
700
600
500
400
300
200
100
0
550 °C
400 °C
740 °C
800
700
600
500
400
300
200
100
0
6
800 °C
5
6
4
120 °C
5
4
3
800 °C
3
150°C
2
190 °C
740 °C
1
2
1
0
20
40
60
80
Time (min)
0
20
40
60
80
Fig. 4. CO2 TPD-MS profiles over NaY, Nit-NaY and spent Nit-NaY. Key: 1. Fresh Nit-
NaY, background control experiment; 2. Fresh NaY, background control experiment;
3. Fresh Nit-NaY; 4. Fresh NaY; 5. Spent Nit-NaY; 6. Spent Nit-NaY, background
control experiment.
Time (min)
Fig. 3. CO2 TPD-MS profiles over the fresh and spent MgO–ZrO2 catalyst. Key: 1.
Fresh MgO–ZrO2, background control experiment; 2. Fresh MgO–ZrO2; 3. Spent
MgO–ZrO2 calcined at 600 ◦C for 1 h; 4. Spent MgO–ZrO2; 5. Spent MgO–ZrO2, sig-
nal at low temperature; 6. Spent MgO–ZrO2, background control experiment. The
signals in Spectra 1, 2, 3 and 5 were multiplied by ten.
amount of adsorbed species over the spent Nit-NaY was much less
than that adsorbed by the spent MgO–ZrO2.
The results from CO2-TPD–TGA (Figs. S2 and S3) are consistent
with those from CO2-TPD-MS experiments. In addition, the fresh
Nit-NaY showed good stability toward heat with the weight loss
similar to NaY during the background control experiment.
740 ◦C (in Thermograms 1–3), which might be due to the decom-
position of magnesium or zirconium carbonate compounds formed
after exposing the catalyst to air.
The CO2 uptakes calculated from the TGA weight changes dur-
ing CO2 chemisorption and TPD and the CO2 uptakes from CO2-MS
peaks over the fresh and spent catalyst are listed in Table 1, respec-
tively. The CO2 uptake over the fresh MgO–ZrO2 and the spent
MgO–ZrO2 after calcination at 600 ◦C for 1 h are very close, indicat-
ing that the catalyst is regenerable. The spent MgO–ZrO2 without
calcination shows only about 10% CO2 uptake. The CO2 uptake
capacities based on both the adsorption and desorption from TGA
results are very close for all three zeolite samples, probably due to
the ability of the zeolite cage to capture CO2. The amount of basic
sites over Nit-NaY was also calculated based on the CO2 TPD-MS,
which was about one order of magnitude smaller than that of the
fresh MgO–ZrO2. However, due to the desorption of the captured
CO2 over a broad temperature range, the baseline of TPD-MS over
the zeolite samples shifted dramatically, thus making the method
less accurate for quantification of basic sites of Nit-NaY.
Fig. 4 shows the CO2 TPD-MS profiles of NaY, Nit-NaY and
the spent Nit-NaY. The spent Nit-NaY catalyst was collected after
two runs of furfural condensation with acetone at 120 ◦C for 75 h
and washed with excess methanol. The fresh Nit-NaY (in Thermo-
gram Curve 3) shows two types of CO2 desorption sites: one at
low temperature near 150 ◦C and the other at high temperature
above 800 ◦C. Since the high temperature CO2 desorption was also
observed on the fresh NaY, it might be related to the decomposition
of carbonate compounds, for instance, Na2CO3, formed during CO2
chemisorption. The low temperature CO2 peaks at 150 ◦C over fresh
NaY and spent Nit-NaY were very broad or absent, indicating the
lack of the same kind of basic sites. In addition to the CO2 desorp-
tion peaks, the CO2 partial pressure increased continuously (almost
linearly) from 200 ◦C to 800 ◦C during CO2 TPD. The spent Nit-NaY
catalyst showed two CO2 uptake peaks at 550 ◦C and 740 ◦C. Again,
these two peaks are associated with adsorbed furfurals, reaction
intermediates, and/or the degradation products, respectively. The
The concentrations of basic sites calculated by titration are also
shown in Table 1. The values are comparable to the results from
Table 1
The concentration of basic sites for fresh and spent catalysts from (1) CO2 adsorption calculated from the TGA weight change, (2) C2 desorption calculated by TGA weight
change and TPD-MS area and (3) benzoic acid titration.
Catalyst
CO2 adsorption
Temperature
CO2 uptake (basic sites concentration) × 104 (mmol/mg catalyst)
Time
Adsorption
Desorption
Titration
By weight change
By weight change
By TPD-MS
MgO–ZrO2
Fresh
50 ◦
50 ◦C
50 ◦
C
3 h
3 h
3 h
5.92
0.68
6.47
5.27
0.43
5.76
6.31
0.64
6.05
4.33
0.65
–
Spent
Spenta
Nit-NaY
Fresh
C
50 ◦C
50 ◦C
3 h
3 h
3.22
3.62
3.26
3.61
0.45
–
0.42
–
Spent
NaY
Fresh
50 ◦C
3 h
3.64
3.82
–
–
a
The spent MgO-ZrO2 catalyst calcined in air at 600 ◦C for 1 h before CO2 adsorption.

62
W. Shen et al. / Applied Catalysis A: General 392 (2011) 57–68
100
90
80
70
60
50
40
30
20
10
0
MgO-ZrO2
NaY
Nit-NaY
Catalysts
Furfural Disappearance
Monomer
Selecꢀvity to Aldol Product
Dimer
Fig. 5. 29Si MAS NMR over the fresh and spent Nit-NaY catalyst. The spent Nit-NaY
after two runs of furfural condensation with acetone at 120 ◦C for 75 h lost 90% of
nitrogen in zeolite framework.
Fig. 6. Furfural disappearance, selectivity to aldol products, and product distribution
for aldol condensation of furfural with acetone over MgO–ZrO2, NaY and Nit-NaY.
Reaction conditions: temperature: 120 ◦C; pressure: 750 psig of helium; reaction
time: 24 h; total organic in solvent: 5 wt.%; furfural:acetone = 1 by moles; catalyst:
15 mg; solvent: water/methanol = 1 by volume.
either TGA or TPD-MS. However, the method could not be applied
to quantify the basic sites of NaY or the spent Nit-NaY either, since
the pH indicator was insufficient to show the color change.
gen occurred after two runs of furfural condensation with acetone
in a methanol–water solution.
3.1.3. 29Si MAS NMR
3.2. Reactions
Fig. 5 shows the 29Si SP MAS NMR spectra of the fresh and
the spent Nit-NaY catalyst. The TO4 units in the zeolite become
TO4−xNx after nitrogen substitution (T: Si or Al). The symbol x
refers the number of nitrogen atoms bonded to a central silicon
(or aluminum) atom. The detailed quantum chemical calculations
of chemical shifts for all possible Si–N environments for NaY and
HY are given in our previous publication [22]. The resonances with
chemical shifts at −73 and −82 ppm could be assigned to 1N envi-
ronments (i.e., x = 1) and the resonance with chemical shifts at −58
and −64 ppm could be assigned to the 2N environments (i.e., x = 2).
No 3N environments were observed for the fresh Nit-NaY sample.
After reaction, the resonance assigned to the 2N environments dis-
appeared and the intensity of the 1N resonance was decreased.
The spectra were deconvoluted so as to extract the intensities
of the different resonances, and thus the concentrations of the
different local environments of TO4−xNx. The deconvolution was
performed by considering the calculated and experimental chem-
ical shift values for nitrogen substituted TO4 units reported. The N
content of the sample could then be estimated from these con-
centrations, as discussed in our earlier paper [22]. The analysis
revealed that the nitrogen content decreased from 10% to 1% for
the spent Nit-NaY. Therefore, a significant loss of substituted nitro-
3.2.1. Aldol condensation reactions
MgO–ZrO2, Nit-NaY and NaY were evaluated for the condensa-
tions of furfural and HMF with acetone or propanal. The reactions
were carried out in the liquid phase at 120 ◦C in 750 psig of He for
16 h or 24 h under vigorous stirring in a high throughput batch reac-
tor. The furfural (or HMF) and acetone (or propanal) molar ratio
was kept at 1. The total organic content of the feed was 5 wt.%.
The solvent was a 1:1 volumetric mixture of water and methanol.
The amount of catalyst used was about 1/6 of the total organics by
weight (about 15 mg).
Fig. 6 shows the catalytic performance of the aldol condensa-
tion of furfural with acetone over MgO–ZrO2, NaY and Nit-NaY.
The results of furfural/propanal, HMF/acetone, and HMF/propanal
condensation reactions over these catalysts are listed in Table 2.
The catalytic performance was evaluated by the disappearance
of furfural, the selectivity to aldol products, and monomer or
dimer distribution. In all experiments, there was no acetone self
condensation product observed. There was also no furfural (or
HMF)/propanal dimer observed as products. MgO–ZrO2 had the
highest activity and selectivity to the aldol products. NaY exhibited
weak basicity with the lowest conversion. Nit-NaY showed dra-
Table 2
The performance of aldol condensation reactions of furfural (or HMF) with acetone (or propanal) over MgO–ZrO2, NaY and Nit-NaY. Reaction conditions: temperature: 120 ◦C;
pressure: 750 psig of helium; total organics in solvent: 5 wt.%; furfural:acetone = 1 by moles; catalyst: 15 mg; solvent: water/methanol = 1 by volume.
Reaction
Reactant (1:1 by molar)
Catalyst
Reaction time
(h)
Disappearance of
furfurals (%)
Selectivity to aldol
products (%)
Distribution (%)
Monomer
Dimer
1
2
3
4
5
6
7
8
9
HMF/acetone
HMF/acetone
HMF/acetone
Furfural/propanal
Furfural/propanal
Furfural/propanal
HMF/propanal
HMF/propanal
HMF/propanal
MgO–ZrO2
NaY
Nit-NaY
MgO–ZrO2
NaY
Nit-NaY
MgO–ZrO2
NaY
16
16
16
16
16
16
16
16
16
51.4
14.1
40.1
60.4
6.7
55.5
86.8
26.4
86.4
91.9
89.7
80.9
90.0
38.6
104.7
57.0
39.6
53.7
56.5
85.9
72.7
100
100
100
100
100
100
43.5
14.1
27.3
–
–
–
–
–
–
Nit-NaY
Note: The performance of furfural condensation with acetone is shown in Fig. 9. The experiment error is about 5%.

W. Shen et al. / Applied Catalysis A: General 392 (2011) 57–68
63
MgO-ZrO
matically increased activity and selectivity to the aldol products
2
100
90
80
70
60
50
40
30
20
10
0
compared to NaY. In the condensation of furfural with acetone as
shown in Fig. 6, the furfural disappearances over MgO–ZrO2, NaY
and Nit-NaY were 59.9%, 7.2% and 35.1%, respectively. Among the
consumed furfural, there were about 100%, 58.3% and 84.3% reacted
to form aldol products. The unbalanced furfural was probably due
to undetected intermediates [46], the degradation products from
either furfural or monomer and dimer, or the chemical species that
adsorbed on the catalyst surface including coke. It was seen that
NaY had the highest ratio of monomer to dimer of about 9. How-
ever, the total yield of monomer is very low (3.5%) due to the low
conversion and aldol product selectivity. The ratio of monomer to
dimer over Nit-NaY was lower than that observed for NaY at about
2.2. However, it was significantly higher than that observed for
MgO–ZrO2 at 0.8.
The catalytic performance of HMF and acetone condensation
over MgO–ZrO2, NaY and Nit-NaY was similar to that of furfural
and acetone condensation as seen in Table 2. Again, MgO–ZrO2 had
the highest catalytic activity and aldol product selectivity. Ninety
two percent of the consumed HMF formed the aldol products with
56.5% of monomer and 43.5% of dimer. NaY showed the lowest
activity with 14.1% of HMF disappearance, but this catalyst selec-
tively produced monomer over dimer. The activity of Nit-NaY was
comparable to MgO–ZrO2 with about 40.1% of HMF conversion and
80.9% aldol product selectivity. The ratio of monomer to dimer was
2.7, higher than that of MgO–ZrO2 which was 1.3. In the aldol con-
densation of furfural/HMF with propanal, the catalytic behavior of
MgO–ZrO2 and Nit-NaY was similar, showing much greater activity
and selectivity towards aldol product than NaY.
100%
75%
67%
50%
33%
25%
Nit-NaY
100
90
80
70
60
50
40
30
20
10
0
100%
75%
67%
50%
33%
25%
Water Concentraꢀon vol.%
Furfural Disappearance
Monomer
Selecꢀvity to Aldol Product
Dimer
In summary, incorporation of nitrogen into the framework of
NaY increased the catalytic activity of Nit-NaY dramatically. The
catalytic activities over Nit-NaY for the aldol condensation reac-
tions of furfural/HMF with acetone/propanal were comparable
to MgO–ZrO2, but more selective towards the furfurals acetone
monomer. The related catalytic activity over the studied catalysts
follows the order: MgO–ZrO2 > Nit-NaY > NaY.
Fig. 7. Solvent effect on furfural disappearance and the product distribution over
MgO–ZrO2 and Nit-NaY. Reaction conditions: temperature: 120 ◦C; reaction time:
24 h; total organic in solvent: 5 wt.%, furfural/acetone = 1 by molar; catalyst: 1/6
total organics by weight (15 mg); solvent: variation of methanol to water volumetric
ratios.
3.2.2. Effect of water
tions carried out over Nit-NaY. Water changes not only the catalytic
activity, but also the product distribution.
Fig. 7 shows the solvent effect on the aldol condensations of
furfural with acetone over MgO–ZrO2 and Nit-NaY with water con-
tent varying from 100 vol.% to 25 vol.% in methanol–water solvents.
With decreasing water content from 100 vol.% to 67 vol.%, the fur-
fural disappearance decreased from 91% to 50% over MgO–ZrO2
catalyst. However, the overall aldol product selectivity increased
and the ratio of monomer to dimer slightly increased from 0.55 to
∼0.8. Further decreasing the water content to less than 67 vol.% pro-
duced no obvious change of either furfural disappearance (≈50%)
or the aldol product selectivity (100%), while, the selectivity of
monomer to dimer varied slightly from 0.7 to 1.1. The average
monomer and dimer yields were approximately 15% and 20%,
respectively. However, the catalytic performance over Nit-NaY
varied significantly with water content in the solvent. The fur-
fural disappearance decreased dramatically from 75% to 20% by
decreasing the water content from 100 vol.% to 25 vol.%. The ratio
of monomer to dimer varied from 0.65 in pure water to 6.5 in
25 vol.% water. After decreasing water content to less than 50 vol.%,
the monomer became the predominant product. Over the entire
water to methanol ranges studied, the aldol product selectivity over
Nit-NaY was maintained at 90%.
3.2.3. Effect of feed concentration
Fig. 8 shows the effect of feed concentration on the catalytic
performance of furfural and acetone condensation over MgO–ZrO2
and Nit-NaY. The reactions were conducted at a constant fur-
fural/acetone molar ratio of 1, total organics/catalyst weight ratio
of 6, and methanol/water volumetric ratio of 1. Increasing the
feed concentration increased the furfural disappearance and the
tendency to generate more dimer product over MgO–ZrO2. The
furfural disappearance increased from 23% to 40% to 57% to 65%
with feed concentrations of 1, 5, 8, and 15 wt.%, respectively. Mean-
while, the ratio of monomer to dimer decreased from 5.7 to 0.96 to
0.67 to 0.35, respectively. Increasing the feed concentration from
1 wt.% to 15 wt.% also gradually increased the furfural disappear-
ance over Nit-NaY from 14% to 56%. However, the selectivity of
monomer to dimer exhibited quite a different trend from that over
MgO–ZrO2. At a very low feed concentration of 1 wt.% over Nit-NaY,
the monomer was the predominant aldol product and the ratio of
monomer to dimer in the aldol products was as high as 24. At feed
concentrations above 5 wt.%, the change in the ratio of monomer to
dimer decreased, varying from 3 to 2.5 to 2 at concentrations of 5, 8,
and 15 wt.%, respectively. Therefore, increasing the feed concentra-
tion increased the disappearance of furfural and the selectivity to
dimer over both catalysts. However, Nit-NaY maintained a higher
selectivity to monomer product at all feed concentrations through
shape selectivity.
The effect of water on aldol condensations of HMF with ace-
tone was also studied over MgO–ZrO2 and Nit-NaY. The results
were summarized in Table 3. Similarly, there was no preference to
HMF acetone monomer or dimer over MgO–ZrO2, but, decreasing
water content to 67 vol.% generated more HMF acetone monomer
than dimer over Nit-NaY. Therefore, water plays a significant role in
the liquid phase aldol condensation reactions, especially the reac-

64
W. Shen et al. / Applied Catalysis A: General 392 (2011) 57–68
Table 3
Summary of solvent effect and catalyst to feed ratio effect over MgO–ZrO2 and Nit-NaY. Reaction conditions: temperature: 120 ◦C; pressure: 750 psig of helium; total organics
in solvent: 5 wt.%; furfural:acetone = 1 by moles.
Run#
Catalyst
Org/Cat
(wt.)
Solvent (vol.)
Reaction time
Reaction
Furfural (or HMF)
disappearance (%)
Selectivity to
aldol product (%)
Product distribution (%)
temperature (◦C)
Monomer
Dimer
Reactant: HMF/acetone
1
2
3
4
5
6
7
8
MgO–ZrO2
MgO–ZrO2
MgO–ZrO2
MgO–ZrO2
Nit-NaY
Nit-NaY
Nit-NaY
Nit-NaY
6
6
6
6
6
6
6
6
Water
16 h
16 h
16 h
16 h
16 h
16 h
16 h
16 h
120
120
120
120
120
120
120
120
93.3
68.5
41.2
38.6
92.0
51.4
34.3
34.1
55.9
84.5
82.2
81.1
74.8
82.5
75.0
75.1
50.4
48.6
55.9
67.8
48.2
63.6
72.0
77.2
49.6
51.4
44.1
32.2
51.8
36.4
28.0
22.8
Water/methanol = 2:1
Water/methanol = 1:1
Water/methanol = 1:2
Water
Water/methanol = 2:1
Water/methanol = 1:1
Water/methanol = 1:2
Reactant: furfural/acetone
9
10
11
12
13
14
15
16
17
18
19
20
21
MgO–ZrO2
MgO–ZrO2
MgO–ZrO2
MgO–ZrO2
MgO–ZrO2
MgO–ZrO2
MgO–ZrO2
Nit-NaY
Nit-NaY
Nit-NaY
Nit-NaY
Nit-NaY
60
30
18
12
6
3
2
12
6
4
Water/methanol = 1:1
Water/methanol = 1:1
Water/methanol = 1:1
Water/methanol = 1:1
Water/methanol = 1:1
Water/methanol = 1:1
Water/methanol = 1:1
Water/methanol = 1:1
Water/methanol = 1:1
Water/methanol = 1:1
Water/methanol = 1:1
Water/methanol = 1:1
Water/methanol = 1:1
24 h
24 h
24 h
24 h
24 h
24 h
24 h
24 h
24 h
24 h
24 h
24 h
24 h
120
120
120
120
120
120
120
120
120
120
120
120
120
32.9
36.4
47.8
51.8
53.9
56.7
66.0
33.5
48.9
59.8
70.6
82.8
97.5
106.1
102.3
100.5
102.3
103.5
95.4
75.7
57.5
55.7
68.7
54.1
50.7
45.0
41.9
44.8
45.6
42.4
82.1
69.9
65.1
55.6
47.1
58.4
45.9
49.3
55.0
58.1
55.2
54.4
57.9
17.9
30.1
34.9
44.4
52.9
41.6
3
2
1
72.1
75.7
59.6
Nit-NaY
3.2.4. Effect of catalyst to feed ratio
The results of aldol condensation of HMF with acetone by varia-
tion of catalyst to feed ratio over MgO–ZrO2 and Nit-NaY are shown
in Fig. 9. In addition, tabulated data on furfural acetone condensa-
MgO-ZrO2
100
90
80
70
60
50
40
30
20
10
0
MgO-ZrO2
100
90
80
70
60
50
40
30
20
10
0
1%
5%
Nit-NaY
8%
15%
1/60
1/30
1/18
1/12
1/6
100
90
80
70
60
50
40
30
20
10
0
Nit-NaY
100
90
80
70
60
50
40
30
20
10
0
1%
5%
8%
15%
Feed Concentraꢀon
1/60
1/30
1/12
1/6
1/4
Catalyst to feed raꢀo
Furfural Disappearance
Monomer
Selecꢀvity to Aldol Product
Dimer
HMF Disappearance
Monomer
Selecꢀvity to Aldol Product
Dimer
Fig. 8. Effect of feed concentration on furfural disappearance, aldol product selectiv-
ity and the product distribution over MgO–ZrO2 and Nit-NaY. Reaction conditions:
temperature: 120 ◦C; reaction time: 24 h; total organics/catalyst ratio of 6, fur-
fural/acetone = 1 by moles; solvent: water/methanol = 1 by volume.
Fig. 9. Effect of catalyst to feed ratio by weight on the HMF disappearance, the
selectivity to aldol products and the product distribution over MgO–ZrO2 and Nit-
NaY. Reaction conditions: temperature: 120 ◦C; reaction time: 16 h; total organics in
solvent: 5 wt.%, HMF/acetone = 1 by moles; solvent: methanol/water = 1 by volume.

W. Shen et al. / Applied Catalysis A: General 392 (2011) 57–68
65
100
100
90
80
70
60
50
40
30
20
10
0
tion over these two catalysts is listed in Table 3 for comparison. The
90
80
70
60
50
40
30
20
10
0
catalytic performance over MgO–ZrO2 was independent of the cat-
alyst to feed ratio (catalyst/total organics ranging from 1/60 to 1/6
by weight). The HMF disappearance slightly increased from 31% to
41% and the total aldol product selectivity was relatively unaffected
at the level of 80% to 90%. The monomer to dimer ratio was slightly
decreased with increase in catalyst amount. However, the HMF dis-
appearance increased dramatically from 9% to 38% by changing the
catalyst to feed ratio from 1/60 to 1/4 over Nit-NaY. The selectivity
of monomer increased at the lower conversion on both catalysts.
Nit-NaY displayed more selectivity to the monomer product over
the entire range than that of the MgO–ZrO2 catalyst. For furfural
condensation with acetone as shown in Table 3, further increasing
the catalyst to feed ratio to 1 resulted in almost 97.5% furfural disap-
pearance over Nit-NaY. But, the higher the furfural disappearance,
the lower the aldol selectivity (some of the degradation products
could not be dissolved in methanol and hence partially contribute
to the unbalanced furfural). There was no obvious preference for the
monomer product at the increased catalyst to feed ratio of above
1/4 over Nit-NaY. The differences observed in the activity and selec-
tivity between the two catalysts due to catalyst to feed ratio might
indicate the different reaction mechanisms over these catalysts. It
should also be noted that the ratio of monomer to dimer for HMF
acetone condensation was higher than that for furfural acetone con-
densation over MgO–ZrO2. This may be due to the relatively short
reaction time for HMF acetone condensation reaction (16 h), which
is insufficient to reach the reaction equilibrium as observed during
the stability testing (Section 3.2.5).
MgO-ZrO2
0
10
20
30
40
50
Reacꢀon Time (h)
100
90
80
70
60
50
40
30
20
10
0
100
90
80
70
60
50
40
30
20
10
0
0
10
20
30
40
50
Reacꢀon Time (h)
Run 1-Furfural Disappearance
Run 1- Dimer Selecꢀvity
Run 2- Monomer Selecꢀvity
Run 1- Monomer Selecꢀvity
Run 2- Furfural Disappearance
Run 2- Dimer Selecꢀvity
Fig. 10. TOS furfural disappearance and the selectivities to monomer and dimer
over MgO–ZrO2 and Nit-NaY. The second run was conducted by using the catalyst
recycled from the first run. Reaction conditions: reaction volume: 75 mL; tempera-
ture: 120 ◦C; pressure: 750 psig of helium; total organics 5 wt.%, furfural/acetone = 1
by moles; solvent: methanol/water = 2; catalysts: 250 mg, about 1/12 total organics
by weight.
3.2.5. Catalyst stability
Fig. 10 shows the time-on-steam (TOS) catalytic performance
of MgO–ZrO2 and Nit-NaY and their stability towards recycling for
furfural/acetone condensation. The reactions were conducted in a
Parr batch reactor as described in Section 2.3. The furfural disap-
pearance increased from 12% at 1 h to 80% after about two days of
reaction over the MgO–ZrO2 catalyst. The selectivity to monomer
increased initially from 13% at 1 h gradually to a maximum of 40%
at 19 h, and then, slowly decreased to 24% at 44 h. The selectivity
to dimer increased continuously from 4% at 1 h to 76% at 44 h. The
selectivity to the aldol products and the product distribution dur-
ing the time-on-stream (TOS) reaction are summarized and listed
in Table 4. It was seen that the initial aldol products selectivity was
as low as 17.3% and increased to about 100% with TOS at 19 h. The
unbalanced furfural at the beginning of the reaction was presum-
ably caused by undetected reaction intermediates. After catalyst
recycling, the MgO–ZrO2 catalyst retained most of the activity with
only a slight decrease in dimer selectivity.
of furfural onto the catalyst surface and/or the base strength of the
spent Nit-NaY catalyst which was insufficient to initiate the reac-
tion intermediates towards the aldol product (reversible reactions
from furfural to the reaction intermediates) at the beginning of the
reaction. The loss of the catalytic activity over the spent Nit-NaY in
the long run was presumably due to the loss of basic sites (nitro-
gen in the framework). This is consistent with the characterization
results from CO2 TPD, benzoic acid titration and 29Si SP MAS NMR
in Sections 3.1.2 and 3.1.3.
4. Discussion
In contrast to MgO–ZrO2, Nit-NaY displayed a different catalytic
performance and stability for the aldol condensation reaction. In
the first run, furfural disappearance increased with TOS from 37%
at 0.5 h to73% at 47 h. The selectivity to monomer increased from
9% at 0.5 h to around 40% at 10.5 h and remained constant until the
end of the reaction. The selectivity to dimer increased at a rela-
tively slow rate, starting at 2% at 0.5 h, increasing to 50% at 20.5 h,
and gradually increasing to 60% after 47 h of reaction. The total
selectivity to aldol products increased with TOS from 10% to about
100% in the period of 30 h reaction at 120 ◦C. In the second run, the
furfural disappearance first decreased from 33% to 12% in the first
5 h of reaction then slowly increased from 11% to 21% after 30 h. In
addition, the selectivity to dimer was relatively low, less than 2% in
the entire TOS range, while, the selectivity to monomer increased
from about 1% to 14%. During the first 5 h of reaction, there were
trace amounts of monomer and dimer products observed despite
the fact that the initial furfural disappearance was close to that of
the first run. This abnormal behavior was reproducible and the rea-
son is not understood at present. It is likely caused by adsorption
4.1. Shape selectivity
Shape selective zeolite catalysis was first reported by Weisz and
co-workers in 1960 [51] in the petroleum industry and was later
applied for the production of high-grade gasoline from biomass
compounds [52,53]. Shape selective catalysis is defined as the
transformation of reactants into products depending on how the
processed molecules fit into the pore of a catalyst [54]. Here, we
estimated the kinetic diameters and critical diameters of various
molecules as listed in Table 5 in order to determine if shape selec-
tivity is expected to occur within different sizes of zeolite pores.
The methods of calculation of kinetic and critical diameter can be
found in the supplementary materials.
The 12-ring widow size of zeolite Y is 7.4 A [6]. The high resolu-
tion Ar adsorption isotherms show that the pore sizes of NaY and
Nit-NaY are about 7.2 A, intermediate to the sizes of furaldehyde
acetone monomer and dimer. The size at the faujasite pore open-
ing is sufficient to accommodate the monomer product but is too
˚
˚

66
W. Shen et al. / Applied Catalysis A: General 392 (2011) 57–68
Table 4
Summary of TOS furfural disappearance and aldol product distribution over MgO–ZrO2 and Nit-NaY. Reaction was conducted in a Parr batch reactor at 120 ◦C in 750 psig of
helium. Organics: furfural/acetone = 1, 5 wt.%; solvent: water/methanol = 2 by volume; catalysts: 1/12 of total organics by weight; reaction volume: 75 mL. The second run
was conducted by using catalyst recycled from the first run.
Reaction time (h) Furfural
disappearance (%)
Selectivity to aldol Distribution (%)
products (%)
Reaction
time (h)
Furfural
disappearance (%)
Selectivity to aldol Distribution (%)
products (%)
Monomer
Dimer
Monomer Dimer
MgO–ZrO2: first run
MgO–ZrO2: second run
1
2
4
6
10
19
26
34
12.2
17.3
36.1
44.5
56.7
68.5
97.6
105
99.9
100.4
86.7
85.9
81.8
76.7
69.6
58.5
50.2
44.4
38.9
13.3
14.1
18.2
23.4
30.1
41.5
49.8
55.6
61.1
1
23.1
29.1
27.8
34.8
47.6
59.9
62.8
71.6
79.6
6.0
10.6
34.4
49.5
73.7
84.0
99.0
94.7
98.1
84.1
87.8
82.4
76.0
61.1
51.2
48.3
42.8
42.7
15.9
12.2
17.6
24.0
38.9
48.8
51.7
57.2
57.3
11.4
20.4
24.9
30.9
40.7
52.9
66.9
77.6
3.5
7.5
11.5
18.5
28
34.5
44
47
44
Nit-NaY: first run
Nit-NaY: second run
0.5
3
5
10.5
13
20.5
30
36.6
40.7
39.5
44.8
51.1
56.8
65.3
73.3
10.1
35.3
47.8
69.9
71.5
92.3
102.0
100.4
91.9
83.3
81.0
72.0
69.0
62.5
62.3
57.3
8.1
16.7
19.0
28
31.0
37.5
37.7
42.7
1
3
5
32.9
24.6
11.5
10.7
14.0
18.2
21.4
1.3
1.7
3.4
6.7
7.5
89.6
60.0
69.0
88.6
92.0
94.6
93.7
10.4
40.0
31.0
11.4
8.0
8
11
21
28
12.7
16.0
5.4
6.3
47
12
10
8
Fig. 11 summarizes the furfural disappearance as a function of
the molar ratio of monomer over dimer at the methanol/water sol-
vent volumetric ratio of 1 and 2. It is clearly seen that the Nit-NaY
was more selective towards monomer in the entire range of fur-
fural disappearance than that of MgO–ZrO2, thus evidence that
the shape selectivity did occur over Nit-NaY. Further experiments
showed that the shape selectivity to monomer over Nit-NaY was
not affected by feed concentration, catalyst to feed ratio, and reac-
tion time. However, the water content in the solvent has a dramatic
effect on the selectivity over Nit-NaY. Since the Nit-NaY catalyst
lost a significant degree of its FAU-type crystallinity and microp-
orosity after nitrogen substitution, the significant amount of dimer
we speculate is generated by the external surface reaction. In the
future, the optimization of nitridation procedure should be per-
formed in order to minimize the aldol dimer formation due to
external surface reaction.
6
4
Nit-NaY
MgO-ZrO
2
2
0
0
20
40
60
80
100
Furfural Disappearance (%)
Fig. 11. Furfural disappearance as a function of molar ratio of furfural acetone
monomer over dimer. The data are compared at methanol/water solvent volumetric
ratio of 1 (solid symbols) and methanol/water volumetric ratio of 2 (open symbols).
4.2. Effect of water
Water content in the solvent plays a significant role on the
performance of both catalysts, especially for Nit-NaY. Increasing
water content leads to a higher conversion and increased dimer
selectivity on both MgO–ZrO2 and Nit-NaY. On the one hand, the
furfural/acetone dimer is not able to dissolve in the solvent with
high water content, thus, the precipitation of dimer drives the reac-
tion towards the generation of more dimer. On the other hand, the
pores of the zeolite might be completely filled with water when
using solvents with high water content in methanol. Therefore, the
reaction takes place on the external zeolite surface, allowing dimer
to be formed more easily.
small to allow the rapid diffusion of the dimer product, thus leading
to a high selectivity towards the monomer product. In contrast, the
˚
fresh MgO–ZrO2 shows a broad pore size distribution from 5 A to
˚
1000 A, thus having no preference to monomer due to the catalyst
texture. Here, NaY was specifically chosen as the starting material
because Na+ might lead to stronger basic strength after nitrogen
substitution than HY [55].
Table 5
List of the kinetic diameters and critical sizes of the molecules of interest.
˚
˚
˚
˚
Critical diameter (A)
Molecule
Kinetic diameter (A)
Critical size (A)
Molecule
Kinetic diameter (A)
Width
Length
Width
Length
Furfural
FA monomer
FA dimer
FP monomer
FP dimer
5.7
6.3
7.4
6.3
6.9
4.56
4.87
6.03
5.24
5.54
5.99
9.89
13.75
8.64
HMF
6.2
6.8
8.0
6.8
7.2
5.42
4.97
6.66
5.22
6.71
8.64
12.15
17.95
10.90
12.91
HA monomer
HA dimer
HP monomer
HP dimer
10.81
Note: “FA” means furfural acetone; “FP” means furfural propanal; “HA” means HMF acetone; “HP” means HMF propanal.

W. Shen et al. / Applied Catalysis A: General 392 (2011) 57–68
67
4.3. Stability
culation of TPD-MS peak). Since these sites were all occupied by
CO2 after exposure to air, they are unlikely to be the catalytically
active sites. The Brønsted base sites were generated due to the het-
erolytic dissociation of water at the surface of MgO. These sites were
indistinguishable on the fresh and the spent catalyst after calcina-
tion, but very noticeable on the spent catalyst as seen in Fig. 3,
indicating that these sites were formed during reaction. The XRD
pattern of the spent MgO–ZrO2 shows broad reflections assigned to
Mg(OH)2, providing further evidence for the formation of surface
hydroxyl groups over MgO–ZrO2. Even though the base strength
of OH groups is considered to be lower than that of the O2− ions
on the MgO surface [64], they were reported to be active for aldol
condensation of acetone [65].
In the stability test, MgO–ZrO2 maintained its catalytic perfor-
mance in the two reaction cycles. An additional experiment (as
shown in the supplementary materials Fig. S5) shows that the pre-
dominate reaction occurred through heterogeneous catalysis over
MgO–ZrO2 despite the fact that a small amount of catalyst leached
out into the hot solution under the reaction conditions. The leached
catalyst underwent re-precipitation after cooling down to room
temperature and could be easily recycled. A similar experiment was
conducted by Barrett et al. [44], in which the MgO–ZrO2 was recy-
cled three times without any loss of catalytic activity, consistent
with our experimental observation.
The stability of Nit-NaY was poor since more than two thirds
of the activity was lost after a single catalytic reaction during the
stability test. Silicon-29 SP MAS NMR further confirmed that the
framework nitrogen in Nit-NaY was leaching out-about 90% loss
occurred after two runs of furfural condensation with acetone at
120 ◦C for a total reaction time of 75 h. The leaching of nitrogen
may be caused by hydrolysis of the nitrogen with water at elevated
temperature. Agarwal et al. [56] conducted DFT calculations on
the kinetic stability of HY, leading to the prediction that nitrogen-
substituted HY remains stable at temperatures below 400 ◦C in sat-
urated water loading in the HY unit cell (200 molecules of water per
unit cell). However, the current experiment reveals that Nit-NaY is
not stable in water–methanol solvent at 120 ◦C. This might be due to
the extra-framework sodium cations or the competitive adsorption
of water, methanol and reactants at the sites, driving the hydroly-
sis reaction to the lower temperature. This discrepancy could also
arise from the fact that the sites that leach may not be precisely the
same kind as those modeled by Agarwal et al. [56,57], prompting
new DFT calculations on nitrogen-substituted NaY. The influence
of water on the activity of nitridated NaX and NaY was studied
previously by Ernst et al. [58]. It was reported that the uptake of
water during storage and subsequent hydrolysis were responsible
for the deactivation of the nitrogen substituted zeolites. The detri-
mental effect of water on the stability of nitrogen-substituted AlPO
was also reported by Benítez et al. [59,60]. These observations of
hydrolysis of the framework nitrogen occurred at room tempera-
ture. However, there was no evidence of the loss of catalytic activity
over Nit-NaY in this study during the approximate six month stor-
age period in desiccators without the protection of nitrogen.
The base strength of NaY is due to the negative charge densi-
ties of the framework oxygen atoms. Zhu et al. [24] reported that
the pKa value of NaY was less than 7.2. Here, NaY had very low
activity in the furfural/acetone condensation reactions which is
consistent with these previous studies and showed the weak base
strength. The nitrided nanoporous materials have been reported
to catalyze the reactions requiring mild to mediate base strength,
including condensations of benzaldehyde with compounds con-
taining a methylenic group such as malononitrile [17,40], ethyl
cyanoacetate [38], and phenylsulfonyl-acetonitrile [66]. Replacing
the oxygen with the less electronegative nitrogen increases the
electron-donating tendency (Lewis base strength) of the lone pair
on the bridging atom, which apparently increases the base strength
from a pKa value of less than 7.2 on NaY to a strong base on Nit-NaY
[30]. The CO2 TPD-MS of the fresh Nit-NaY showed a desorption
peak at 150 ◦C, indicating that the base strength of the framework
nitrogen on Nit-NaY is weaker than that of Mg2+–O2− pairs (190 ◦C),
but stronger than that of Mg(OH)2 (120 ◦C).
Climent et al. [67] conducted the aldol condensation reaction of
heptanal with benzaldehyde over the nitrogen-substituted amor-
phous aluminophosphate (ALPO) and found the cooperative effect
between week acid and base sites. The bi-functional acid-base
mechanism was proposed, which involved the activation of ben-
zaldehyde by protonation-polarization of the carbonyl group on
the acid sites and the nucleophilic addition of the enolate heptanal
intermediate generated on the basic sites. Similarly, we propose
here that the driving force for the formation of enolate is the com-
bination of the free electron pair and the condensed electronegative
charge on the framework nitrogen, and the positively charged
sodium cations (hard Lewis acid [68]) and/or the surface hydroxyl
groups functioned as the site to activate furfural. The existence of
surface silanol groups and Al(OH)2 was confirmed by an initial FT-IR
experiment as seen in the supplementary materials Fig. S4.
4.4. The base strength and the nature of basic sites
Kelly et al. [61] summarized the base strength of the most stud-
ied solid base catalysts and the minimal base strength required to
remove the proton from a R1–CH2–R2 reactant molecule for the
condensation reaction. It was reported that a strong base with pKa
of about 20 was required to deprotonate acetone and propanal.
MgO is a strong solid base catalyst [61]. The basic sites on the MgO
catalyst have been well studied [49,62,63]. The strongest sites are
attributed to Lewis base sites of the low coordination (LC) surface
oxygen; the second strongest basic sites are formed at acid-base
pairs (Mg2+–O2−); the weakest base sites are the surface hydroxyl
groups [62]. The CO2 TPD-MS of MgO–ZrO2 showed three desorp-
tion peaks at temperatures of 740 ◦C, 190 ◦C and 120 ◦C, which
may be assigned to LC Lewis base sites, Mg2+–O2− pair and the
surface hydroxyl groups, respectively. There was no observation
of the extra desorption of CO2 due to ZrO2, which might be due
to the low content of ZrO2 in MgO–ZrO2 at 1/18 by moles. The
CO2 desorption at 740 ◦C was due to the adsorption of CO2 after
exposing the LC Lewis base sites to air, forming carbonate species.
Therefore, this peak was also observed on the fresh catalyst in the
background control experiment. However, there are few of these
strongest basic sites (0.36 × 10−4 mmol/mg catalyst from the cal-
5. Conclusions
Aldol condensations of furaldehydes with acetone and propanal
was studied on solid base catalysts MgO–ZrO2, NaY and Nit-NaY.
Nit-NaY showed catalytic activity comparable to MgO–ZrO2 and
much higher than that of untreated NaY. This indicates that replac-
ing the bridging oxygen by less electronegative nitrogen increases
the base strength of the zeolite framework dramatically. Nit-NaY
showed more selectivity to monomer than MgO–ZrO2 due to the
small cage size in the FAU structure, indicating a shape selective
base catalyst. The size at the faujasite pore opening is sufficient
to accommodate the monomer product but is too small to allow
the rapid diffusion of the dimer product. Increasing water concen-
tration of the feed solution leads to increased catalyst activity and
dimer selectivity on Nit-NaY. On MgO–ZrO2, increasing the water
content had no obvious effect on either furfural disappearance or
dimer selectivity at low water concentration (less than 67 vol.%).
Upon further increasing water concentration to above 75 vol.%, the
catalytic activity and dimer selectivity increased. Increasing the

68
W. Shen et al. / Applied Catalysis A: General 392 (2011) 57–68
feed concentration increased the disappearance of furfural and the
selectivity to dimer on both catalysts. Increasing the catalyst to
feed ratio from 1/60 to 1/6 has no obvious effect over MgO–ZrO2,
but significantly increased the disappearance of furfurals and low-
ered the monomer to dimer ratio over Nit-NaY. Nit-NaY was more
sensitive towards water and catalyst to feed ratio for the studied
reactions compared to MgO–ZrO2. Nit-NaY showed higher selec-
tivity towards monomer than MgO–ZrO2 in the entire range of
furfural disappearance. Nit-NaY was found to be unstable in the
water–methanol solvent with the loss in catalytic activity due to
leaching of the framework nitrogen, while MgO–ZrO2 showed good
recyclability.
[16] A.-J. Han, H.-Y. He, J. Guo, H. Yu, Y.-F. Huang, Y.-C. Long, Micropor. Mesopor.
Mater. 79 (2005) 177–184.
[17] S. Ernst, M. Hartmann, S. Sauerbeck, T. Bongers, Appl. Catal. A: Gen. 200 (2000)
117–123.
[18] K. Narasimharao, M. Hartmann, H.H. Thiel, S. Ernst, Micropor. Mesopor. Mater.
90 (2006) 377–383.
[19] J. Guo, A.-J. Han, H. Yu, J.-P. Dong, H. He, Y.-C. Long, Micropor. Mesopor. Mater.
94 (2006) 166–172.
[20] A.-J. Han, J. Guo, H. Yu, Y. Zeng, Y.-F. Huang, H.-Y. He, Y.-C. Long, ChemPhysChem
7 (2006) 607–613.
[21] K.D. Hammond, M. Gharibeh, G.A. Tompsett, F. Dogan, A.V. Brown, C.P. Grey,
S.M. Auerbach, W.C. Conner Jr., Chem. Mater. 22 (2009) 130–142.
[22] F. Dogan, K.D. Hammond, G.A. Tompsett, H. Huo, W.C. Conner Jr., S.M. Auerbach,
C.P. Grey, J. Am. Chem. Soc. 131 (2009) 11062–11079.
[23] K.D. Hammond, S.M. Auerbach, in: I. Halasz (Ed.), Silica and Silicates in Modern
Catalysis, Research Signpost, Kerala, India, 2010.
CO2 TPD-MS and XRD distinguish three types of basic sites on
MgO–ZrO2, attributed to LC Lewis basic sites, Mg2+–O2− pairs and
surface hydroxyl groups. The latter two sites are likely the catalyt-
ically active sites for liquid phase aldol condensation of furfurals
with acetone/propanal. Due to the strong adsorption of CO2 on NaY,
CO2 TPD-MS and CO2 TPD–TGA are not deemed suitable to quantify
the basic sites over Nit-NaY. Silicon 29 MAS NMR of the fresh and
spent Nit-NaY disclosed that the substituted nitrogens are located
in 1N or 2N environments, where one or two of oxygen atoms
in TO4 units were replaced by nitrogen. The base strength of Nit-
NaY is in between of Mg2+–O2− pairs and Mg(OH)2. The significant
leaching of nitrogen from Nit-NaY during the aldol condensation in
water–methanol solvent was observed.
These results provide the first clear evidence of shape selectiv-
ity in nitrogen-substituted zeolites. This proof-of-principle is an
important step forward for providing a new tool in the arsenal of
heterogeneous catalysis, especially as this field pursues solutions
to problems in biomass refining and biofuel production. Further
research is required to improve the stability and the selectivity of
basic zeolites, either by modifying reaction conditions, solvents,
nitridation methods, or the basic sites themselves.
[24] J.H. Zhu, Y. Chun, Q.-h. Xu, Y. Qin, Mater. Lett. 33 (1998) 331–335.
[25] Y. Xia, R. Mokaya, J. Phys. Chem. C 112 (2008) 1455–1462.
[26] C.W. Jones, K. Tsuji, M.E. Davis, Nature 393 (1998) 52–54.
[27] D. Lesthaeghe, V. Van Speybroeck, G.B. Marin, M. Waroquier, J. Phys. Chem. B
109 (2005) 7952–7960.
[28] A. Cauvel, D. Brunel, F. Di Renzo, P. Moreau, F. Fajula, in: H.K. Beyer, H.G. Karge,
I. Kiricsi, J.B. Nagy (Eds.), Studies in Surface Science and Catalysis, Elsevier, 94,
1995, pp. 286–293.
[29] A. Corma, P. Viruela, L. Fernández, J. Mol. Catal. A: Chem. 133 (1998) 241–250.
[30] R. Astala, S.M. Auerbach, J. Am. Chem. Soc. 126 (2004) 1843–1848.
[31] J. EI Haskouri, S. Cabrera, F. Sapin˜a, J. Latorre, C. Guillem, A. Beltrán-Porter, D.
Beltrán-Porter, M.D. Marcos, P. Amorós, Adv. Mater. 13 (2001) 192–195.
[32] Y. Xia, R. Mokaya, Angew. Chem. Int. Ed. 42 (2003) 2639–2644.
[33] C. Zhang, Q. Liu, Z. Xu, J. Non-Cryst. Solids 351 (2005) 1377–1382.
[34] N. Chino, T. Okubo, Micropor. Mesopor. Mater. 87 (2005) 15–22.
[35] J. Wang, Q. Liu, Micropor. Mesopor. Mater. 83 (2005) 225–232.
[36] M.J. Climent, A. Corma, V. Fornés, A. Frau, R. Guil-López, S. Iborra, J. Primo, J.
Catal. 163 (1996) 392–398.
[37] X. Guan, F. Zhang, G. Wu, N. Guan, Mater. Lett. 60 (2006) 3141–3144.
[38] J. Xiong, Y. Ding, H. Zhu, L. Yan, X. Liu, L. Lin, J. Phys. Chem. B 107 (2003)
1366–1369.
[39] K.D. Hammond, F. Dogan, G.A. Tompsett, V. Agarwal, W.C. Conner, C.P. Grey,
S.M. Auerbach, J. Am. Chem. Soc. 130 (2008) 14912–14913.
[40] C. Zhang, Z. Xu, K. Wan, Q. Liu, Appl. Catal. A: Gen. 258 (2004) 55–61.
[41] G.W. Huber, J.A. Dumesic, Catal. Today 111 (2006) 119–132.
[42] R. Xing, A.V. Subrahmanyam, H. Olcay, W. Qi, G.P. van Walsum, H. Pendse, G.W.
Huber, Green Chem. 12 (2010) 1933–1946.
[43] J.N. Chheda, J.A. Dumesic, Catal. Today 123 (2007) 59–70.
[44] C.J. Barrett, J.N. Chheda, G.W. Huber, J.A. Dumesic, Appl. Catal. B: Environ. 66
(2006) 111–118.
Acknowledgements
[45] G.W. Huber, J.N. Chheda, C.J. Barrett, J.A. Dumesic, Science 308 (2005)
1446–1450.
[46] N. Fakhfakh, P. Cognet, M. Cabassud, Y. Lucchese, M.D. de Los Ríos, Chem. Eng.
Process.: Process Intensification 47 (2008) 349–362.
[47] M.A. Aramendía, V. Borau, C. Jiménez, A. Marinas, J.M. Marinas, J.R. Ruiz, F.J.
Urbano, J. Mol. Catal. A: Chem. 218 (2004) 81–90.
[48] M.A. Aramendía, V. Boráu, C. Jiménez, A. Marinas, J.M. Marinas, J.A. Navío, J.R.
Ruiz, F.J. Urbano, Colloids Surf. A 234 (2004) 17–25.
The author would like to thank for the financial support by the
United States Department of Energy (DOE) (DE-FG02-07ER15918
and DE-FG02-96ER14681), and an NSF-MRI grant (0722802). We
also thank Vishal Agarwal and Dr. A. V. Subrahmanyam for helpful
discussions.
[49] R. Vidruk, M.V. Landau, M. Herskowitz, M. Talianker, N. Frage, V. Ezersky, N.
Froumin, J. Catal. 263 (2009) 196–204.
[50] M. Solache, I. García, P. Bosch, S. Bulbulian, A. Blumenfeld, J. Fripiat, Micropor.
Mesopor. Mater. 21 (1998) 19–25.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.apcata.2010.10.023.
[51] P.B. Weisz, V.J. Frilette, J. Phys. Chem. 64 (1960) 382–1382.
[52] P.B. Weisz, W.O. Haag, P.G. Rodewald, Science 206 (1979) 57–58.
[53] T.R. Carlson, T.P. Vispute, G.W. Huber, ChemSusChem 1 (2008) 397–400.
[54] B. Smit, T.L.M. Maesen, Nature 451 (2008) 671–678.
[55] V. Agarwal, G.W. Huber, W.C. Conner Jr., S.M. Auerbach, J. Phys. Chem. C, sub-
mitted for publication.
References
[56] V. Agarwal, G.W. Huber, W.C. Conner Jr., S.M. Auerbach, J. Catal. 270 (2010)
249–255.
[57] V. Agarwal, G.W. Huber, W.C. Conner Jr., S.M. Auerbach, J. Catal. 269 (2010)
53–63.
[58] S. Ernst, M. Hartmann, T. Hecht, P. Cremades Jaén, S. Sauerbeck, G.G.R. Aiello,
F. Testa, Stud. Surf. Sci. Catal. 142 (2002) 549–556.
[59] J.J. Benítez, P. Malet, I. Carrizosa, J.A. Odriozola, R. Conanec, R. Marchand, Y.
Laurent, J. Eur. Ceram. Soc. 17 (1997) 1979–1982.
[60] J.J. Benítez, A. Díaz, Y. Laurent, J.A. Odriozola, Appl. Catal. A: Gen. 176 (1999)
177–187.
[61] F. King, G.J. Kelly, Catal. Today 73 (2002) 75–81.
[62] V.K. Díez, C.R. Apesteguía, J.I. Di Cosimo, J. Catal. 240 (2006) 235–244.
[63] E. Knözinger, K.-H. Jacob, S. Singh, P. Hofmann, Surf. Sci. 290 (1993) 388–
402.
[64] H. Kabashima, H. Hattori, Catal. Today 44 (1998) 277–283.
[65] G. Zhang, H. Hattori, K. Tanabe, Appl. Catal. 36 (1988) 189–197.
[66] M.J. Climent, A. Corma, R. Guil-Lopez, S. Iborra, J. Primo, Catal. Lett. 59 (1999)
33–38.
[1] Y. Ono, T. Baba, Catal. Today 38 (1997) 321–337.
[2] Y. Ono, J. Catal. 216 (2003) 406–415.
[3] H. Hattori, Appl. Catal. A: Gen. 222 (2001) 247–259.
[4] A. Corma, S. Iborra, C.G. Bruce, K. Helmut, Adv. Catal. 49 (2006) 239–302.
[5] R.J. Davis, J. Catal. 216 (2003) 396–405.
[6] S.M. Auerbach, K.A. Carrado, P.K. Dutta (Eds.), Handbook of Zeolite Sicence and
Technology, Marcel Dekker, New York, 2004.
[7] M. Huang, A. Adnot, S. Kaliaguine, J. Am. Chem. Soc. 114 (2002) 10005–10010.
[8] J. Zhu, Y. Chun, Y. Wang, Q. Xu, Catal. Today 51 (1999) 103–111.
[9] L.R.M. Martens, P.J. Grobet, W.J.M. Vermeiren, P.A. Jacobs, Stud. Surf. Sci. Catal.
28 (1986) 935–941.
[10] P.E. Hathaway, M.E. Davis, J. Catal. 116 (1989) 263–278.
[11] G.J. Suppes, M.A. Dasari, E.J. Doskocil, P.J. Mankidy, M.J. Goff, Appl. Catal. A: Gen.
257 (2004) 213–223.
[12] I. Rodriguez, S. Iborra, F. Rey, A. Corma, Appl. Catal. A: Gen. 194–195 (2000)
241–252.
[13] B.M. Choudary, M.L. Kantam, P. Sreekanth, T. Bandopadhyay, F. Figueras, A. Tuel,
J. Mol. Catal. A: Chem. 142 (1999) 361–365.
[14] X. Zhang, E.S. Man Lai, R. Martin-Aranda, K.L. Yeung, Appl. Catal. A: Gen. 261
(2004) 109–118.
[67] M.J. Climent, A. Corma, V. Fornés, R. Guil-Lopez, S. Iborra, Adv. Syn. Catal. 344
(2002) 1090–1096.
[68] R.G. Pearson, J. Am. Chem. Soc. 85 (1963) 3533–3539.
[15] K.-i. Shimizu, E. Hayashi, T. Inokuchi, T. Kodama, H. Hagiwara, Y. Kitayama,
Tetrahedron Lett. 43 (2002) 9073–9075.